Journal of Applied Bioinformatics & Computational BiologyISSN: 2329-9533

Reach Us +1 850 754 6199
All submissions of the EM system will be redirected to Online Manuscript Submission System. Authors are requested to submit articles directly to Online Manuscript Submission System of respective journal.

Research Article, J Appl Bioinforma Comput Biol Vol: 7 Issue: 3

Profiling of Differential Expression of Genes in Mice Carrying Both Mutant Presenilin 1 and Amyloid Precursor Protein Transgenes with or without Knockout of Β2 Adrenergic Receptor Gene

Yuan Zhou1,2,3,#, Lintao Chen1,4,#, Xi Zhou1,4, Yechun Pei1,4, Shuangshuang Wei1,4, Anum Mehmood1,4, Yang K Xiang2,5* and Dayong Wang1,2,4*

1Laboratory of Biotechnology and Molecular Pharmacology, Hainan Key Laboratory of Sustainable Utilization of Tropical Bioresources, Hainan University, Haikou, Hainan 570208, China

2Department of Molecular and Integrative Physiology, University of Illinois at Urbana-Champaign, Urbana, Illinois 61801, USA

3Amber Glen Alzheimer’s Association, 1704 Amber Ln, Urbana, IL 61802, USA

4Collage of Biology, Institute of Tropical Agriculture and Forestry, Hainan University, Haikou, Hainan 570208, China

5Department of Pharmacology, University of California, Davis, CA95616, USA

#The first two authors contributed equally to the study

*Corresponding Authors : Dayong Wang
Professor of Biochemistry and Molecular Biology, Laboratory of Biotechnology and Molecular Pharmacology, Hainan Key Laboratory of Sustainable Utilization of Tropical Resource
Hainan University, 817 Nong-Ke Lou, 58 People’s Road, Meilan District, Haikou, Hainan 570208, China
Tel:
+86-18789556728/+1-217-721-9757
E-mail: [email protected]

Yang K. Xiang
Department of Pharmacology, University of California, Davis, CA 95616, USA
Tel: 530-752-6895
E-mail: [email protected]

Received: June 05, 2018 Accepted: September 11, 2018 Published: September 18, 2018

Citation: Zhou Y, Chen L, Zhou X, Pei Y, Wei S, et al. (2018) Profiling of Differential Expression of Genes in Mice Carrying Both Mutant Presenilin 1 and Amyloid Precursor Protein Transgenes with or without Knockout of Β2 Adrenergic Receptor Gene. J Appl Bioinforma Comput Biol 7:3. doi: 10.4172/2329-9533.1000155

Abstract

Alzheimer’s disease (AD) is a lifelong progressive neurodegenerative disease related with accumulation of amyloid β peptide (Aβ) produced by processing of amyloid precursor protein (APP) in the brain. In spite of several-decades effort on AD, there is still no medicine used to intervene with its pathological processes. Our previous studies made in transgenic animal models harboring familial AD genes of mutant presenilin 1 and amyloid precursor protein (APP) showed that β2AR gene knock-out (β2AR-KO) is beneficial in senile AD animals. Consistently, an epidemiological study lasted for two decades showed that the sole usage of β blockers as antihypertensive medicines is associated with fewer brain lesions and less brain shrinkage seen in senile AD patients. In order to understand why senile β2AR-KO AD mice had better learning and memory, genomic effects of β2AR-KO in the double transgenic AD mice were investigated. In the analysis, major genomic significance of β2AR-KO was directed to influence proteinprocessing and presentation involving membrane structure and MHC class I and II protein complex, and lysosome and hydrolase activity for protein degradation, which are critical for accumulation of amyloid β peptide, the hallmark of AD.

Keywords: β2 adrenergic receptor; Alzheimer’s disease; Genome; Differential expression; Protein processing and presentation; Lysosome

Introduction

Alzheimer’s disease (AD) is a progressive neurodegenerative disease characterized by accumulation of plaques composed of amyloid β peptides (Aβ) in the brain related with abnormal processing and presentation of Aβ by cells [1]. Studies have shown that the number of β2 adrenergic receptor (β2AR) is increased in the prefrontal cortex and the hippocampus of AD subjects, while it is not correlated with age [2]. Since the cortex and the hippocampus are major brain regions that are highly degenerated in AD and the cortex is the earliest to form Aβ plaques, it is supposed that β2AR may play an important role in AD [3]. A study has shown that chronic treatment with β2AR agonist enhances γ-secretase activity, and related Aβ production and formation of Aβ plaques [4]. Using AD transgenic mouse model harboring human familial AD genes of mutant presenilin 1 (PS1) and amyloid precursor protein (APP), we have found that β2AR gene knock-out (KO) improved learning and memory in 1-year-old senile AD mice, although there was a weak decreasing tendency of the performance in 6-month-old young wild type (WT) mice [5]. In addition, removing the gene encoding β2AR increases the survival rate in P301S mutant tau-transgenic mice [6]. An epidemiological study showed that variations in β2AR gene are involved in the pathogenesis of sporadic late onset Alzheimer’s disease (LOAD) [2]. In an epidemiological study made in 2,197 participants conducted from 1991 to 2010, it was found that clinical use of β-blockers was associated with a lower risk of cognitive impairment, and the association was more obvious in senile men who were more than 75-years old [7]. An autopsy study made on 774 brains of male AD patients after death showed that the patients who took β-blockers had less brain lesions and shrinkage than those who took other medications for blood hypertension and those untreated, and their brains showed that they suffered less microinfarcts [8]. A parallel study showed that patients who took β-blockers experienced less cognitive decline as they aged compared to control groups [9].

β2AR which consists of seven transmembrane α-helices belongs to G protein-coupled receptor superfamily and transduces signal via Gαs and also Gβγ proteins [10]. β2AR is distinct from β1AR in that β2AR internalizes after binding to isoproterenol or Aβ [11]. In response to β2AR endogenous ligands norepinephrine and epinephrine, Gαs dissociates from Gβγ to stimulate adenylyl cyclase to produce cAMP, which activates protein kinase A (PKA) and the exchange protein activated by cAMP (Epac) [10,12]. Gβγ dimer interacts with many different proteins, and different combinations of Gβ and Gγ subtypes transduct signals diversly with or without the association of Gα subunit to inhibit or activate various downstream signaling components, including ion channels, G protein-coupled receptor kinases (GRKs), phosphoinositide 3-kinase (PI3K), mitogenactivated protein kinases (MAPK) signaling mediated by small GTPase Ras, phospholipase A and C, and etc [13-16]. The signal transductions mediated by Gαs and Gβγ regulate a few of transcription factors, such as cAMP response element binding protein (CREB), extracellular signal-regulated kinases (ERKs) and etc., therefore, β2AR activation regulates expression of genes, however integrative influence of β2AR on genomic expression is still unknown. β2ARs are expressed throughout the brain, abundantly in the cortex and the hippocampus, which are the two brain regions essential for higher cognitive functions [6]. Despite being involved in fundamental biological processes, β2AR geneknockout (KO) mice are viable and fertile [6].

In the study, a genetic approach was adopted to remove β2AR gene from a double transgenic mouse model of AD that overexpresses mutant human PS1 gene harboring M146L and L286V familial AD mutations, and APP gene (695) harboring Swedish (K670N, M671L), Florida (I716V) and London (V717I) mutations, and the effects of β2AR-KO on genomic expression profiles in PS1/APP transgenic AD mice was investigated.

Materials and Methods

Animals

PS1/APP double transgenic and β2AR-KO/PS1/APP mice in B6 background were described previously [11,17,18], and 1-year old mice were used for analysis of whole genome gene expression. The PS1/APP mice, which are transgenic animal models of Alzheimer’s disease, were purchased from Jackson laboratory (stock number: 006554), which overexpress human PS1 gene harboring two familial AD mutations, M146L and L286V and APP (695) gene with Swedish (K670N, M671L), Florida (I716V) and London (V717I) familial AD mutations [18]. PS1/APP mice were crossbred with β2AR-KO mice to produce β2AR-KO/PS1/APP mice. All animal experimental procedures were approved by the Animal Care and Use Committee of University of Illinois and Hainan University.

Whole-genome expression analysis

One-year-old PS1/APP and β2AR-KO/PS1/APP mice were sacrificed, six mice in each group. The cerebrums were dissected out, and one side of the cerebral hemispheres of each mouse was used for the study. The cerebral hemispheres were homogenized separately on ice, and total RNA was extracted using RNeasy® Lipid Tissue Kit (Qiagen). Whole-genome expression profiles of PS1/ APP Alzheimer’s disease mice and β2AR-KO/PS1/APP mice were tested using the Mouse WG-6 v2.0 BeadChips and HiScan System (Illumina). The Mouse WG-6 v2.0 BeadChips have probes for the NCBI Mouse Reference Sequence (RefSeq) Release 22, including 26,766 manually annotated and reviewed coding transcripts designated as known mRNAs which begin with NM, 6,856 predicted coding transcripts (XM), and 56 annotation-well-established noncoding transcripts (XR). The BeadChips also have probes for 5,659 transcripts described in the Japan RIKEN Functional Annotation of Mouse (FANTOM) 24-6 database, 3,573 additional sequences listed in the RefSeq Release 5 (Build 33.1), and 2,371 sequences listed in the Mouse Exonic Evidence Based Oligonucleotide (MEEBO) set. In the single factor experimental design, PS1/APP transgenic mice were used as background controls for the gene expression profile of the Alzheimer’s disease model, and β2AR-KO was the only factor of treatment.

Results

Principal components analysis for all biological samples

Principal component analysis (PCA) is a multivariate statistical method to analyze internal correlation or variation among original variables by constructing linear combinations of the variables, which were gene expression levels in the study. During the process the dimensionality of original variables is reduced and the newly constructed variables designed to preserve much original information are independent to each other. The linear combination with biggest variance is treated as the first or the primary principal component (PC), which was designated as PC1 in the study, and the second one was PC2. The PCs which were orthogonal to each other effectively show variation of gene expression levels among thousands of noise in each subject. In the study, PC1 showed that β2AR-KO-induced significant change in genomic expression, in that the two experimental groups which are PS1/APP double transgenic AD mice (PA) and β2AR-KO/PS1/APP mice (B2P) were distributed in two separated regions along the PC1 axis; PC1s for PA samples fell into the region from 19 to 39, however, B2P samples were in the region from -40 to -12 (data not shown). In other words, β2AR-KO in PS1/APP mice significantly changed the distribution of PC1 values, indicating a prominent consequence resulted from β2AR-KO.

Numeric statistics of differentially expressed genes (DEGs) related with β2AR-KO in transgenic animal model of AD

Raw data of the expression levels of the 45,281 transcripts were normalized to systematic differences between different BeadChips. Differential expression analysis on the normalized data was made by DEseq. Those genes whose normalized expression levels have absolute values of log2 ratio that are not less than one and a false discovery rate less than 0.05 (q<0.05) were taken as DEGs. Knock-out of β2AR gene significantly changed the expression levels of more than 50% of total genes manifested as the total transcripts probed, with down-regulated transcripts being more than up-regulated transcripts (Figures 1a, 1b).

Figure 1: Numeric statistics of differentially expressed genes whose expression was affected by β2AR-KO. a. Count of differentially expressed genes. The statistical analysis was made with DEseq. The criteria for differentially expressed genes were defined as |log2Ratio| ≥ 1 and q<0.05, where Ratio equals to fold change. b. Volcano plot showing expression levels of genes affected by β2AR-KO. PA, PS1/APP double transgenic mice. B2P, β2AR-KO PS1/APPdouble transgenic mice.

Hierarchical cluster analysis of DEGs related with β2AR-KO in transgenic animal model of AD

Hierarchical cluster analysis is a statistical approach to assign normalized DEG values into different clusters, with intra-cluster difference being much smaller than intercluster difference. It is different from classification in that it is an unsupervised mathematical grouping regardless of biological functions related with a normalized value. Phenotype structures of the samples are revealed by a hierarchical clustering figure. The R software was used by us to show two-dimensional hierarchical clustering. One dimension of the construction is the biological samples listed individually in horizontal, the other is DEGs identified in the above statistics. A dendrogram along the vertical axis showing hierarchical clustering of DEGs was achieved according to the Euclidean metrics of the Log2 values of the expression levels of DEGs (Figure 2). Zooming in the hierarchical clustering diagram, the most intuitive changes are that the DEGs with moderate transcriptional levels in PS1/APP mice are obviously decreased by β2AR-KO, and some other genes with relatively lower transcriptional levels are obviously increased.

Figure 2: Hierarchical clustering analysis for differentially expressed genes. The dendrogram on the left shows the hierarchical clustering of the genes which was made according to the Euclidean distance of Log2 values of the expression levels of the genes. The red color indicates relatively high levels of genes expression, and the blue color indicates relatively low gene expression levels. PA, PS1/APP double transgenic mice. B2P, β2AR-KO PS1/APP-double transgenic mice.

Gene ontology (GO) classification of DEGs related with β2AR-KO in transgenic animal model of AD

Currently, GO construction is aimed to pave the way to interpret or predict functions of genes or proteins by extracting and analyzing gene-related knowledges accumulated in large amount of literatures, leading to artificial intelligence. In GO consortium, there are three independent ontology domains, which are biological processes, cellular component and molecular function, using controlled vocabularies which are updated dynamically. In each of the three domains, there are subdomains to form tree structures with levels, nodes and relationships. According to the biological terms at the third level specified in the GO database typically, the expression of genes up- or down-regulated by β2AR-KO are classified and counted (Figure 3). There are four biological processes that are prominently affected by β2AR-KO, which are cellular process, single-organism process, metabolic process and biological regulation, each comprises more than 35% of total DEGs. In the domain of cellular components, gene expression levels that are mostly changed by β2AR-KO are cell part, membrane, membrane part, organelle, organelle part and macromolecular complex. In the domain of molecular function, the most prominent result is that the genes whose biological function is relavent to binding comprises more than 50% of total DEGs. At this step, these results primarily associated β2AR with membrane and its main function which is binding, which are indispensable for protein processing and presentation, which are immunological functions and shown in the following section.

Figure 3: GO classification of differentially expressed genes. Up- and down-regulated genes are assorted under GO terms in the three GO domains, which are biological process, cellular component and molecular function. PA, PS1/APP double transgenic mice. B2P, β2AR-KO PS1/APP-double transgenic mice.

Q-value distribution of GO terms enriched with DEGs related with β2AR-KO in transgenic animal model of AD

In order to control increased probability of type I error within all rejections associated with multiple hypothesis test of differential expression levels, false discovery rate (FDR) was calculated. The q-value is the minimum positive FDR (pFDR) to reject a null hypothesis for a given rejection region Γα of a GO term enriched with genes and an observed statistic T=t, and is defined as inf{Γα:t∈Γα} pFDR(Γα) (19). The distribution charts for GO term-enriched DEGs related with β2AR-KO were drawn according to q-values (Figures 4, 5a, 5b). The biological process (BP) domain comprises terms describing series of events accomplished by organized assemblies of molecular functions in living units. In GOBP, most of the terms enriched with DEGs are directly related with immunological function; there are 88 in total 117 terms are related, and 54 in 68 terms have very high strength which were highlighted in red (Figure 4). The top 3 highest enriched GO terms showed that β2AR primarily affects protein processing and presentation in the transgenic animal model of AD, which is pertinent to abnormal Aβ accumulation in the brain (Figure 4). Besides the top 3, twelve terms showed facets of the protein processing and presentation affected by β2AR-KO, including MHC class I and II mediating presentation of endogenous and exogenous proteins, peptides and even polysaccharides (Figure 4). In addition to the primary effects of β2AR on gene expression, other genes are also affected. There are six enriched GO terms related with protein folding, four related with nitric oxide synthesis, three related with response to stimulus, and ten related with chemical metabolism (Figure 4). Another two terms drawn attention are mRNA splice site selection and sphingolipid biosynthesis (Figure 4). In the GO domain of cellular component (CC), almost all of the terms are membranerelated events (27 out of total 30 terms), including MHC protein complex, plasma membrane, external side of plasma membrane, intrinsic component of membrane, endosome, vesicle, secretory granule and etc. (Figure 5a). The most prominent GOCC results are consistent with GOBP (Figure 5a). In the GO domain of molecular function, there are only two significantly enriched terms, and the most affected one is unfolded protein binding, which is a membrane function related with protein processing and presentation (Figure 5b).

Figure 4: Q-value distribution of GO terms enriched with differentially expressed genes in GO domain of biological process. Statistical significance is denoted by shades of color; a smaller q-value is signified by a darker color, and vice versa. PA, PS1/APP double transgenic mice. B2P, β2AR-KO PS1/ APP-double transgenic mice.

Figure 5: Q-value distribution of GO terms enriched with differentially expressed genes in GO domains of cellular component and molecular function. Statistical significance is denoted by shades of color; a smaller q-value is signified by a darker color, and vice versa. PA, PS1/APP double transgenic mice. B2P, β2AR-KO PS1/APP-double transgenic mice.

Hierarchical relationships of GO terms highly enriched with DEGs related with β2AR-KO in transgenic animal model of AD

Hierarchical relationships among GO terms highly enriched with DEGs were shown by directed acyclic graphs (DAG) in detail (Figures 6-8). DAG was drawn based on automated analysis in each of the three independent GO ontologies which are biological process, cellular component and molecular function. In each of the three GO ontologies, top five DEGs-enriched GO terms are major nodes in DAG. In GOBP, all six highly enriched nodes are related to immune, with q-values ranging from 3.05 e-14 to 1.57 e-9, and 4 out of the 6 nodes showed that β2AR-KO fundamentally changed expression levels of genes regulating antigen processing (Figure 6). Top genes annotated in GO to be related to antigen processing and presentation were listed in the sequence of the fold of change (Table 1). In GOCC, the GO terms enriched with DEGs whose expression was affected by β2AR-KO are directed toward two directions. One group of GOCC nodes are directed toward MHC protein complex which was highly enriched with a q value of 1.3 e-9 through 3 paths, two of which are significantly enriched and related with membrane (Figure 7). The other group of GOCC nodes are directed toward a one-way path consisted of three consecutive nodes starting from vacuole (q=1.93 e-10), which in animals are typically founded in the protein-processing steps of endocytosis and exocytosis, and ending with lysosome (q=6.32 e-11) which is a component of protein processing and presentation (Figure 7). Top genes annotated in GO to be related to lysosome (Tables 2a and 2b) and MHC protein complex (Tables 3a and 3b) were listed in the sequence of the fold of change. In GOMF, there are 3 unrelated paths. One is the hydrolase activity path, which can be found in lysosomes and related to protein processing and presentation. The hierarchical node sequence in the hydrolase activity path is catalytic activity (GO:0003824, q=7.064 × 10-3, enrichment rate (ER):1181/6588), hydrolase activity (GO:0016787, q=2.29 × 10-4, ER: 521/2686), hydrolase activity acting on glycosyl bonds (GO:0016798, q=3.23 × 10-4, ER: 37/125) and hydrolase activity hydrolyzing O-glycosyl compounds (GO: 0004553, q=2.11 e-5, ER: 34/99) (Figure 8). Top genes annotated in GO to be related to hydrolase activity were listed in the sequence of the fold of change (Tables 4a and 4b). Another path is directed toward unfolded protein binding, which is a step of protein processing and presentation and may be related with AD pathology. The node sequence of the binding path is binding (GO: 0005488, q=2.0986 × 10-2, ER: 2376/13682), protein binding (GO: 0005515, q=4.253 × 10-3, ER: 1383/7736) and unfolded protein binding (GO: 0051082, q=2.05 e-05, ER: 36/107) (Figure 8). Top genes annotated in GO to be related to unfolded protein binding were listed in the sequence of the fold of change (Table 5). The third path is from the node of enzyme regulatory activity (GO: 0030234, q=1.6829 × 10-2, ER: 175/889), which is reserved for cases when the regulator directly interacts with the enzyme, to nitric-oxide synthase regulator activity (GO: 0030235, q=1.41 × 10-4, ER: 5/5) (Figure 8). Besides, by further reviewing of DEG data, it was found that the expression of all 5 genes comprised in GO nodes of nitric-oxide synthase regulator activity was decreased by β2AR-KO in the transgenic AD animal model. In addition to the above GO analysis, the genes that are annotated to be related to AD in the KEGG’s pathways were listed in Table 6, including Apolipoprotein E, which is the strongest genetic risk factor for both early- and late-onset AD found in one third of the cases of AD.

Items Gene ID Fold Chg. pvalue qvalue
Histocompatibility 2, class II antigen A, beta 1 14961 8.871354 1.91E-06 0.000375
CD74 antigen (invariant polypeptide of major histocompatibility complex, class II antigen-associated) 16149 6.778798 2.57E-05 0.002088
Cathepsin E 13034 5.167045 2.08E-07 9.51E-05
Histocompatibility 2, D region locus 1 14964 3.86412 6.03E-07 0.000192
Histocompatibility 2, K1, K region 14972 3.259361 3.27E-05 0.002469
Proteasome (prosome, macropain) subunit, beta type 9 (large multifunctional peptidase 2) 16912 2.908455 3.20E-05 0.00244
Histocompatibility 2, class II, locus Mb1 14999 2.430133 5.14E-07 0.000173
Histocompatibility 2, class II, locus Mb2 15000 2.132515 3.91E-08 2.70E-05
Solute carrier family 11 (proton-coupled divalent metal ion transporters), member 1 18173 2.006802 1.74E-05 0.001625
Histocompatibility 2, O region alpha locus 15001 1.993441 0.000614 0.01491
Proteasome (prosome, macropain) subunit, beta type 9 (large multifunctional peptidase 2) 16913 1.861686 0.000539 0.013966
Fc receptor, IgE, high affinity I, gamma polypeptide 14127 1.769161 8.53E-05 0.004534
Beta-2 microglobulin 12010 1.684328 4.52E-05 0.003045
Transporter 2, ATP-binding cassette, sub-family B (MDR/TAP) 21355 1.628326 0.003592 0.042868
Fc receptor, IgG, low affinity Iib 14130 1.62805 0.000151 0.006517
Histocompatibility 2, M region locus 3 14991 1.626856 1.77E-05 0.001647
Unc-93 homolog B1 (C. elegans) 54445 1.538428 0.000216 0.00799
Nucleotide-binding oligomerization domain containing 1 107607 1.503754 0.000728 0.016624
Fc receptor, IgG, low affinity III 14131 1.411944 0.001029 0.021051
Fc receptor, IgG, alpha chain transporter 14132 1.33175 0.003783 0.044161

Table 1: Profiles of differentially expressed genes in β2AR-KO transgenic AD model annotated to be related to antigen processing and presentation in Gene Ontology.

Items Gene ID Fold Chg. pvalue qvalue
CD74 antigen (invariant polypeptide of major 16149 6.778798 2.57E-05 0.002088
histocompatibility complex, class II antigen-associated)        
Cathepsin C 13032 5.514072 0.002342 0.033643
Interleukin 4 induced 1 100328588 2.760841 0.000134 0.006
Solute carrier family 15, member 3 65221 2.480609 1.68E-07 8.88E-05
CD68 antigen 12514 2.203099 6.29E-06 0.000837
Heparanase 15442 2.188486 6.48E-07 0.000198
RIKEN cDNA 5430435G22 gene 226421 2.148704 4.72E-05 0.003128
Histocompatibility 2, class II, locus Mb2 15000 2.132515 3.91E-08 2.70E-05
Alpha-N-acetylglucosaminidase (Sanfilippo disease IIIB) 27419 2.13038 6.49E-06 0.000851
Solute carrier family 11 (proton-coupled divalent metal ion transporters), member 1 18173 2.006802 1.74E-05 0.001625
Prolylcarboxypeptidase (angiotensinase C) 72461 1.972664 5.00E-05 0.003226
N-sulfoglucosamine sulfohydrolase (sulfamidase) 27029 1.965037 1.81E-05 0.001656
Glucuronidase, beta 110006 1.954243 1.93E-05 0.001715
Cathepsin Z 64138 1.944783 4.78E-05 0.003156
Lysosomal-associated protein transmembrane 5 16792 1.943501 2.60E-06 0.000457
Mannosidase 2, alpha B1 17159 1.755669 6.14E-05 0.003691
Arylsulfatase G 74008 1.669844 0.00139 0.024861
LPS-induced TN factor 56722 1.644251 1.01E-05 0.001124
Lysosomal-associated membrane protein 2 16784 1.628241 2.08E-06 0.000394
Niemann Pick type C1 18145 1.623278 6.90E-06 0.000884
CD63 antigen 12512 1.619914 0.000166 0.006884
Ceroid lipofuscinosis, neuronal 3, juvenile (Batten, Spielmeyer-Vogt disease) 12752 1.617276 5.01E-05 0.003226
Galactosidase, beta 1 12091 1.616097 3.79E-05 0.00271
Niemann Pick type C2 67963 1.614508 7.80E-06 0.00094
Deoxyribonuclease II alpha 13423 1.595048 0.00179 0.029031
Unc-93 homolog B1 (C. elegans) 54445 1.538428 0.000216 0.00799

Table 2a: Profiles of differentially expressed genes in β2AR-KO transgenic AD model annotated to be related lysozyme in Gene Ontology (Part I).

Items Gene ID Fold Chg. p value q value
Cathepsin K 13038 7.580496 0.000158 0.006652
Cathepsin S 13040 10.41212 0.001679 0.027926
Sialidase 4 241159 5.623464 0.002891 0.037488
Peroxiredoxin 6 11758 5.989792 0.002557 0.035172
Syntaxin 8 55943 7.56421 0.001079 0.021551
Neuraminidase 1 18010 10.8563 0.000602 0.014729
Solute carrier family 29 (nucleoside transporters), member 3 71279 8.107022 0.003121 0.039352
RIKEN cDNA 0610031J06 gene 56700 9.661221 0.000288 0.009493
Iduronidase, alpha-L- 15932 7.167215 0.003211 0.039979
Immunity-related GTPase family M member 1 15944 7.71878 0.003203 0.039973
ATP-binding cassette, sub-family B (MDR/TAP), member 9 56325 9.114246 0.001726 0.028518
Corticotropin releasing hormone binding protein 12919 11.26217 0.000188 0.007354
Glucosidase, alpha, acid 14387 11.3279 0.00054 0.013971
Glucosidase, beta, acid 14466 9.797483 0.002908 0.037603
Phospholipase A2, group XV 192654 10.78939 0.002593 0.035401
Mannosidase, beta A, lysosomal 110173 8.529202 0.000965 0.020127
Ceroid-lipofuscinosis, neuronal 5 211286 9.777018 0.001512 0.026133
Vacuolar protein sorting 11 (yeast) 71732 9.435465 0.000712 0.016323
Lysosomal-associated membrane protein 1 16783 12.1733 0.001168 0.022445
Arylsulfatase B 11881 11.80603 0.002461 0.034612
VMA21 vacuolar H+-ATPase homolog (S. cerevisiae) 67048 10.57272 0.003456 0.041699
Transmembrane protein 55A 72519 11.26718 0.002119 0.031692
Sphingosine kinase 2 56632 4.96189 0.000586 0.014558
Iduronate 2-sulfatase 15931 6.259901 0.00208 0.031336
Transmembrane protein 74 239408 5.943605 2.08E-06 0.000394

Table 2b: Profiles of differentially expressed genes in β2AR-KO transgenic AD model annotated to be related to lysozyme in Gene Ontology (Part II)

Items Gene ID Fold Chg. p value q value
Histocompatibility 2, class II antigen A, beta 1 14961 8.871354 1.91E-06 0.000375
CD74 antigen (invariant polypeptide of major histocompatibility complex, class II antigen-associated) 16149 6.436063 9.77E-06 0.001104
cathepsin E 13034 5.167045 2.08E-07 9.51E-05
Histocompatibility 2, D region locus 1 14964 3.86412 6.03E-07 0.000192
Histocompatibility 2, Q region locus 8 15019 3.623889 1.40E-05 0.001408
Histocompatibility 2, K1, K region 14972 3.259361 3.27E-05 0.002469
Predicted gene 8909 667977 3.114538 8.43E-07 0.000229
Proteasome (prosome, macropain) subunit, beta type 9 (large multifunctional peptidase 2) 16912 2.908455 3.20E-05 0.00244
Histocompatibility 2, class II, locus Mb1 14999 2.430133 5.14E-07 0.000173
Eukaryotic translation initiation factor 2-alpha kinase 2 19106 2.220892 4.71E-07 0.00016
Histocompatibility 2, class II, locus Mb2 15000 2.132515 3.91E-08 2.70E-05
Lymphocyte-activation gene 3 16768 2.047944 0.000108 0.005212
Solute carrier family 11 (proton-coupled divalent metal ion transporters), member 1 18173 2.006802 1.74E-05 0.001625
Histocompatibility 2, O region alpha locus 15001 1.993441 0.000614 0.01491
Proteasome (prosome, macropain) subunit, beta type 8 (large multifunctional peptidase 7) 16913 1.861686 0.000539 0.013966
Fc receptor, IgE, high affinity I, gamma polypeptide 14127 1.769161 8.53E-05 0.004534
Histocompatibility 2, K1, K region 14972 1.763206 2.74E-06 0.000472
DnaJ (Hsp40) homolog, subfamily C, member 16 214063 1.713747 1.49E-05 0.001453
beta-2 microglobulin 12010 1.684328 4.52E-05 0.003045
Transporter 2, ATP-binding cassette, sub-family B (MDR/TAP) 21355 1.628326 0.003592 0.042868
Fc receptor, IgG, low affinity Iib 14130 1.62805 0.000151 0.006517
Histocompatibility 2, M region locus 3 14991 1.626856 1.77E-05 0.001647
Protein tyrosine phosphatase, non-receptor type 6 15170 1.62081 0.000859 0.018684
Unc-93 homolog B1 (C. elegans) 54445 1.538428 0.000216 0.00799
Nucleotide-binding oligomerization domain containing 1 107607 1.503754 0.000728 0.016624

Table 3a: Profiles of differentially expressed genes in β2AR-KO transgenic AD model annotated to be related to MHC protein complex in Gene Ontology (Part I).

Items Gene ID Fold Chg. p value q value
Cyclin D1 12443 1.489153 0.000724 0.016556
DnaJ (Hsp40) homolog, subfamily C, member 18 76594 1.461304 0.000678 0.015844
CD3 antigen, epsilon polypeptide 12501 1.42577 0.003388 0.04123
Fc receptor, IgG, low affinity III 14131 1.411944 0.001029 0.021051
Fc receptor, IgG, alpha chain transporter 14132 1.33175 0.003783 0.044161
ATP-binding cassette, sub-family B (MDR/TAP), member 9 56325 1.305657 0.001726 0.028518
Wolfram syndrome 1 homolog (human) 22393 1.289007 0.003984 0.045336
SEC63-like (S. cerevisiae) 140740 1.193609 0.001819 0.029331
HERPUD family member 2 80517 -1.22882 0.003121 0.039352
Heat shock protein 90, alpha (cytosolic), class A member 1 15519 -1.23512 0.001109 0.021868
Heat shock 105kDa/110kDa protein 1 15505 -1.24793 0.001371 0.024635
DnaJ (Hsp40) homolog, subfamily B, member 11 67838 -1.29418 0.000546 0.014039
Chaperonin containing Tcp1, subunit 3 (gamma) 12462 -1.35609 0.000138 0.006112
Heat shock protein 90 alpha (cytosolic), class B member 1 15516 -1.4133 0.000366 0.011053
Heat shock protein 90, beta (Grp94), member 1 22027 -1.41384 0.002119 0.031692
T-complex protein 1 21454 -1.41447 0.002478 0.034708
Heat shock protein 5 14828 -1.43609 0.001112 0.021872
DnaJ (Hsp40) homolog, subfamily A, member 4 58233 -1.46868 5.87E-05 0.00359
DnaJ (Hsp40) homolog, subfamily C, member 3 1.00E+08 -1.50583 0.000205 0.007744
Serine/arginine-rich splicing factor 10 14105 -1.54974 0.002451 0.034597
Heat shock protein 1 (chaperonin) 15510 -1.76402 0.000238 0.008435
Heat shock protein 8 15481 -1.8216 0.000412 0.011816
Calreticulin 12317 -1.86027 0.001086 0.021609
p53 and DNA damage regulated 1 68559 -15.3898 1.55E-12 5.96E-09

Table 3b: Profiles of differentially expressed genes in β2AR-KO transgenic AD model annotated to be related to MHC protein complex in Gene Ontology (Part II).

Items Gene ID Fold Chg. pvalue qvalue
Cathepsin C 13032 5.514072 0.002342 0.033643
Kallikrein related-peptidase 6 19144 5.377842 2.08E-07 9.51E-05
Cathepsin E 13034 5.167045 2.08E-07 9.51E-05
Proteasome (prosome, macropain) subunit, beta type 9 (large multifunctional peptidase 2) 16912 2.908455 3.20E-05 0.00244
Plasminogen activator, urokinase 18792 2.825956 1.57E-09 2.64E-06
Lysozyme 1 17110 2.51725 1.58E-06 0.000345
Chymotrypsin-like elastase family, member 1 109901 2.386791 1.55E-08 1.38E-05
Ectonucleotide pyrophosphatase/phosphodiesterase 6 320981 2.239908 1.35E-07 7.72E-05
CTD (carboxy-terminal domain, RNA polymerase II, polypeptide A) phosphatase, subunit 1 67655 2.216111 6.35E-06 0.000841
Peptidyl arginine deiminase, type II 18600 2.204623 9.15E-06 0.001061
Heparanase 15442 2.188486 6.48E-07 0.000198
Dual specificity phosphatase 1 19252 2.12332 0.002496 0.034772
Caspase 1 12362 2.095731 6.53E-05 0.00386
Phosphatidic acid phosphatase type 2C 50784 2.080684 1.81E-07 9.01E-05
Ribonuclease, RNase A family 4 58809 2.073185 2.71E-07 0.000106
Phospholipase D family, member 4 104759 2.053021 1.40E-06 0.000322
Lysozyme 2 17105 2.031321 6.58E-06 0.000859
Agmatine ureohydrolase (agmatinase) 75986 2.030783 0.000198 0.007654
ATP-binding cassette, sub-family B (MDR/TAP), member 1B 18669 2.025545 3.16E-05 0.002415
Cathepsin Z 64138 2.001336 0.000111 0.0053
Prolylcarboxypeptidase (angiotensinase C) 72461 1.972664 5.00E-05 0.003226
N-sulfoglucosamine sulfohydrolase (sulfamidase) 27029 1.965037 1.81E-05 0.001656
glucuronidase, beta 110006 1.954243 1.93E-05 0.001715
Suppression of tumorigenicity 14 (colon carcinoma) 19143 1.915971 0.000214 0.007947
Proprotein convertase subtilisin/kexin type 4 18551 1.914996 0.000861 0.0187
2',3'-cyclic nucleotide 3' phosphodiesterase 12799 1.91379 8.14E-06 0.000968
Proteasome (prosome, macropain) subunit, beta type 8 (large multifunctional peptidase 7) 16913 1.861686 0.000539 0.013966
paraoxonase 3 269823 1.86039 1.43E-05 0.001416
Phospholipase C, gamma 2 234779 1.819061 9.23E-06 0.001066
Mannosidase 2, alpha B1 17159 1.755669 6.14E-05 0.003691

Table 4a: Profiles of differentially expressed genes in β2AR-KO transgenic AD model annotated to be related to hydrolase activity in Gene Ontology (Top 30 increase).

Items Gene ID Fold Chg. p value q value
DEAD (Asp-Glu-Ala-Asp) box polypeptide 3, Y-linked 26900 -2.10674 0.002199 0.032358
Phosphodiesterase 4A, cAMP specific 18577 -1.98027 7.42E-05 0.004177
Protein tyrosine phosphatase, receptor type, O 19277 -1.97666 0.00269 0.035996
Proteasome (prosome, macropain) subunit, beta type 5 19173 -1.96916 1.71E-05 0.001616
Glucosamine-6-phosphate deaminase 1 26384 -1.95682 4.02E-05 0.0028
N-acyl phosphatidylethanolamine phospholipase D 242864 -1.93054 0.000285 0.009404
acyl-CoA thioesterase 1 26897 -1.86093 0.000536 0.013913
Spastin 50850 -1.75118 9.36E-06 0.001069
DIS3 mitotic control homolog (S. cerevisiae) 72662 -1.74871 0.000286 0.009435
Ectonucleoside triphosphate diphosphohydrolase 6 12497 -1.74754 5.98E-05 0.003631
Rhomboid, veinlet-like 3 (Drosophila) 246104 -1.7434 2.35E-05 0.001954
Acyl-CoA thioesterase 7 70025 -1.73511 4.17E-05 0.002877
DEAD (Asp-Glu-Ala-Asp) box polypeptide 5 13207 -1.70454 0.000419 0.011946
Spastin 50850 -1.7034 0.001148 0.022269
Protease, serine, 53 330657 -1.69693 0.001317 0.024066
Esterase D/formylglutathione hydrolase 13885 -1.6703 1.03E-05 0.001143
Mannan-binding lectin serine peptidase 2 17175 -1.65176 0.00124 0.023313
Haloacid dehalogenase-like hydrolase domain containing 2 76987 -1.64692 9.52E-05 0.004837
Bisphosphate 3'-nucleotidase 1 23827 -1.64313 0.000246 0.008591
OTU domain, ubiquitin aldehyde binding 2 68149 -1.63286 0.000229 0.008222
Endo/exonuclease (5'-3'), endonuclease G-like 208194 -1.58646 0.001819 0.029331
Histocompatibility 13 14950 -1.57331 0.000107 0.005193
Iduronate 2-sulfatase 15931 -1.573 0.00208 0.031336
YME1-like 1 (S. cerevisiae) 27377 -1.57208 0.000551 0.014096
Adenosine deaminase-like 75894 -1.57155 0.002845 0.037159
Angiogenin, ribonuclease, RNase A family, 5 11727 -1.5621 0.000464 0.012644
Inositol polyphosphate-4-phosphatase, type II 234515 -1.54742 0.004207 0.046872
Expressed sequence AI464131 329828 -1.54346 0.000783 0.017454
Myotubularin related protein 2 77116 -1.5103 0.003353 0.041002
Alkaline ceramidase 2 230379 -1.50696 8.64E-05 0.004576

Table 4b: Profiles of differentially expressed genes in β2AR-KO transgenic AD model annotated to be related to hydrolase activity in Gene Ontology (Top 30 decrease).

Items Gene ID Fold Chg. p value q value
Eukaryotic translation initiation factor 2-alpha kinase 2 19106 2.220892 4.71E-07 0.00016
DnaJ (Hsp40) homolog, subfamily C, member 16 214063 1.713747 1.49E-05 0.001453
Cyclin D1 12443 1.489153 0.000724 0.016556
DnaJ (Hsp40) homolog, subfamily C, member 18 76594 1.461304 0.000678 0.015844
Wolfram syndrome 1 homolog (human) 22393 1.289007 0.003984 0.045336
SEC63-like (S. cerevisiae) 140740 1.193609 0.001819 0.029331
HERPUD family member 2 80517 -1.22882 0.003121 0.039352
Heat shock protein 90, alpha (cytosolic), Class A member 1 15519 -1.23512 0.001109 0.021868
Heat shock 105kDa/110kDa protein 1 15505 -1.24793 0.001371 0.024635
DnaJ (Hsp40) homolog, subfamily B, member 11 67838 -1.29418 0.000546 0.014039
Chaperonin containing Tcp1, subunit 3 (gamma) 12462 -1.35609 0.000138 0.006112
Heat shock protein 90 alpha (cytosolic), class B 15516 -1.4133 0.000366 0.011053
Heat shock protein 90, beta (Grp94), member 1 22027 -1.41384 0.002119 0.031692
T-complex protein 1 21454 -1.41447 0.002478 0.034708
Heat shock protein 5 14828 -1.43609 0.001112 0.021872
DnaJ (Hsp40) homolog, subfamily A, member 4 58233 -1.46868 5.87E-05 0.00359
DnaJ (Hsp40) homolog, subfamily C, member 3 1.00E+08 -1.50583 0.000205 0.007744
Serine/arginine-rich splicing factor 10 14105 -1.54974 0.002451 0.034597
Heat shock protein 1 (chaperonin) 15510 -1.76402 0.000238 0.008435
Heat shock protein 8 15481 -1.8216 0.000412 0.011816
Calreticulin 12317 -1.86027 0.001086 0.021609
p53 and DNA damage regulated 1 68559 -15.3898 1.55E-12 5.96E-09

Table 5: Profiles of differentially expressed genes in β2AR-KO transgenic AD model annotated to be related to unfolded protein binding in Gene Ontology.

Items Gene ID Fold Chg. p value q value
Cytochrome c oxidase, subunit VI a, polypeptide 2 12862 2.523729 1.44E-05 0.001417
ATPase, Ca++ transporting, cardiac muscle, fast twitch 1 11937 1.512032 0.000439 0.012244
Beta-site APP-cleaving enzyme 2 56175 1.506286 8.44E-05 0.004494
Inositol 1,4,5-triphosphate receptor 2 16439 1.451202 0.000228 0.008213
Lipoprotein lipase 16956 1.413552 0.000202 0.007714
Apolipoprotein E 11816 1.35843 0.000376 0.01129
Caspase 8 12370 1.336963 0.003373 0.041103
Calpain 1 12333 1.311871 0.001259 0.023432
NADH dehydrogenase (ubiquinone) flavoprotein 3 78330 1.231451 0.004652 0.049888
ATP synthase, H+ transporting, mitochondrial F0 complex, subunit B1 11950 -1.26575 0.004255 0.047107
Guanine nucleotide binding protein, alpha q polypeptide 14682 -1.26842 0.003271 0.040466
Phospholipase C, beta 4 18798 -1.29642 0.000965 0.020124
Anterior pharynx defective 1c homolog (C.elegans) 68318 -1.56577 0.000409 0.011797
RIKEN cDNA 1500003O03 gene 56398 -1.58374 2.07E-05 0.001806
Cytochrome c oxidase subunit VIIa polypeptide 2-like 20463 -10.3201 8.36E-14 1.12E-09

Table 6: Profiles of differentially expressed genes in β2AR-KO transgenic AD model annotated to be related to Alzheimer’s disease in KEGG’s pathways.

Figure 6: DAG showing hierarchical relationships of GO terms highly enriched with differentially expressed genes in GO domain of biological process. Five top-enriched GO terms denoted with rectangles are major nodes of the DAG, and relationships of interrelated GO terms were shown in the diagram. Each node in the DAG was shown with a unique GO access number, title, q-value and the proportion of enriched genes in the node. Statistical significance is denoted by shades of color; a smaller q-value is signified by a darker color, and vice versa.

Figure 7: DAG showing hierarchical relationships of GO terms highly enriched with differentially expressed genes in GO domain of cellular component. Five top-enriched GO terms denoted with rectangles are major nodes of the DAG, and relationships of interrelated GO terms were shown in the diagram. Each node in the DAG was shown with a unique GO access number, title, q-value and the proportion of enriched genes in the node. Statistical significance is denoted by shades of color; a smaller q-value is signified by a darker color, and vice versa.

Figure 8: DAG showing hierarchical relationships of GO terms highly enriched with differentially expressed genes in GO domain of molecular function. Five top-enriched GO terms denoted with rectangles are major nodes of the DAG, and relationships of interrelated GO terms were shown in the diagram. Each node in the DAG was shown with a unique GO access number, title, q-value and the proportion of enriched genes in the node. Statistical significance is denoted by shades of color; a smaller q-value is signified by a darker color, and vice versa.

Discussion

AD has baffled scientists ever since German psychiatrist Alois Alzheimer reported the first case of AD in 1906, the cause of which has been poorly understood until now [19,20], and there is neither treatment to intervene its development, nor particular measure to be effective in preventing or delaying the onset of AD [21].

Traditional amyloid hypothesis proposes that abnormal accumulation and aggregation of Aβ in the brain is the central event triggering neuronal degeneration in AD, which may be related to abnormal protein processing and presentation. Since the early days, the aggregated Aβ has been believed to be the causal factor of AD disrupting homeostasis of calcium ions in neurons and inducing apoptosis [