Alterations in T-Cell Transcription Factors and Cytokine Gene Expression in Late-Onset Alzheimer’s Disease

Abstract

Background:

Late-onset Alzheimer’s disease (LOAD) is associated with many environmental and genetic factors. The effect of systemic inflammation on the pathogenesis of neurodegenerative diseases such as AD has been strongly suggested. T helper cells (Th) are one of the important components of the immune system and can easily infiltrate the brain in pathological conditions. The development of each Th-subset depends on the production of unique cytokines and their main regulator.

Objective:

This study aimed to compare the mRNA levels of Th-related genes derived from peripheral blood mononuclear cells of LOAD patients with control. Also, the identification of the most important Th1/Th2 genes and downstream pathways that may be involved in the pathogenesis of AD was followed by computational approaches.

Methods:

This study involved 30 patients with LOAD and 30 non-demented controls. The relative expression of T-cell cytokines (IFN-γ, TNF-α, IL-4, and IL-5) and transcription factors (T-bet and GATA-3) were assessed using Real-time PCR. Additionally, protein-protein interaction (PPI) was investigated by gene network construction.

Results:

A significant decrease at T-bet, IFN-γ, TNF-α, and GATA-3 mRNA levels was detected in the LOAD group, compared to the controls. However, there was no significant difference in IL-4 or IL-5 mRNA levels. Network analysis revealed a list of the highly connected protein (hubs) related to mitogen-activated protein kinase (MAPK) signaling and Th17 cell differentiation pathways.

Conclusion:

The findings point to a molecular dysregulation in Th-related genes, which can promising in the early diagnosis or targeted interventions of AD. Furthermore, the PPI analysis showed that upstream off-target stimulation may involve MAPK cascade activation and Th17 axis induction.

Keywords

Alzheimer’s disease cytokine gene expression LOAD mitogen-activated protein kinase protein-protein interaction T helper cell Th17 cell transcription factor

INTRODUCTION

Alzheimer’s disease (AD) is the most common cause of dementia and has a progressive and neurodegenerative nature. Since most cases occur after the age of 65, the majority of patients fall into the late-onset AD (LOAD) category. It is estimated that 11.3% of the US population aged 65 and older have AD dementia [1]. AD is generally defined by classic symptoms such as a progressive loss of memory, and cognitive changes. Biologically, AD is distinguished by the presence of two specific neurological markers, extracellular deposition of amyloid-β (Aβ) and intraneuronal neurofibrillary tangles [2, 3]. In AD, underlying dementia progressively worsens and leads to a disability that generally results in death within 5 to 12 years of onset of symptoms [4]. While AD has a high economic and health-care burden, disease-modifying therapies that hinder or slow the progression of the disease are not currently available [5, 6].

LOAD is considered a complex disease and numerous environmental and genetic factors are involved in its pathogenesis [7, 8]. Some studies on LOAD have emphasized a greater share of genetic abnormalities than environmental factors [9, 10]. A lot of evidence shows that inefficiency in regulating and controlling inflammatory responses in the brain may play a major causal role in AD [11 –16]. More detailed genetic studies have supported the theory that AD is dependent on immune-related mechanisms [17 –22]. Neuroinflammation can be triggered by dying neurons, invading viruses or bacteria, harmful chemicals, or endogenous protein aggregates such as seen with Aβ peptides [13 , 24]. Inflammation is usually considered a protective response, but in the case of chronic inflammation, it contributes to the disease process, as is seen in many neurodegenerative diseases [25]. In the brain of AD patients, reactive astrogliosis and microgliosis are recognized as prominent pathological features [6, 26]. The inflammatory response is predominantly mediated by activated microglia [27]. Active microglial cells have been observed in both AD patients and animal models of the disease [28]. Based on the signals they receive, resting microglia may turn into one of two distinct phenotypes. The classical M1 mode release pro-inflammatory cytokines and cytotoxic substances that cause further nerve damage. The alternative M2 mode, in turn, plays a neuroprotective role in the CNS by producing trophic factors and anti-inflammatory cytokines [29].

Immune mediators, such as cytokines, play an important role in regulating AD pathology. Cytokines can be produced by resident glial cells in the brain parenchyma or secreted by peripheral leukocytes that can then indirectly influence the central nervous system (CNS) [30 –32]. Some reports have shown an increase in the level of pro-inflammatory cytokines, such as tumor necrosis factor-alpha (TNF-α), interleukin (IL)-1α/β, IL-6 or IL-8 in the brain, blood, and cerebrospinal fluid (CSF) of patients with AD [33, 34]. T helper (Th) cells are considered essential for optimal homeostasis of immune cells and host defense and therefore play an important role in the pathology of autoimmune and inflammatory diseases [35]. Th cells are identified and classified into distinct subsets based on different cytokine profiles that cause different functional characteristics. Among the subsets of Th, the Th1 cells are identified by the secretion of interferon-gamma (IFN-γ) and TNF-α, which are responsible for cell-mediated immune responses, and their excessive activity leads to tissue damage. Other major subsets of Th are known as Th2 cells. This subset is characterized by IL-4, IL-5, and IL-13 secretion, and they are responsible for humoral-mediated immunity. Their activation is associated with atopy/hypersensitivity [36, 37]. The normal function of both subsets of Th cells is to provide adequate immunity to pathogens. However, dysregulation of both subsets has been observed in disease states.

Brain tissue is easily affected by systemic inflammation, and this peripheral effect can be potent [38 –40]. Under pathological conditions, peripheral immune cells such as monocytes, T, and B lymphocytes, can easily penetrate the brain [39]. Although AD was commonly considered a strict brain disorder, it has now become clear that it also has a systemic component [41 –44]. Using genome-wide gene expression microarrays in AD whole blood, Naughton et al. showed a strong blood-brain correlations in AD and identified a profound inflammatory condition in patients with AD that included almost all components of the immune system [45]. It is important to evaluate comparative changes in gene expression in AD patients and healthy individuals for a variety of purposes. These include early diagnosis, identifying potential pathological pathways, and developing targeted interventions. Although access to brain tissue samples, or CSF, is a challenge for genetic studies, peripheral blood is considered a viable alternative [46, 47]. It is worth noting that several studies have shown that in 80% of cases, blood and brain-derived cells have similar transcriptome and in fact, it may be a reflection of molecular events occurring during AD pathogenesis [45 , 49]. In this regard, blood cells and several brain tissues have a remarkable similarities in gene expression [47 , 51].

Moreover, LOAD is a complex disorder in which several proteins and molecular pathways are involved. Previously, multiple genes and proteins were investigated experimentally to be associated with AD and provide a precious resource for computational approaches which can substantially increase the yield of downstream studies at the prioritization of genes in human diseases such as AD [52, 53]. Text mining (or text data mining) is a powerful approach for targeted scanning of potential biomarkers among massive biological published data and provides a cross-cut to clarify the complex disorder modality. Moreover, integrative strategy (e.g., Gene Ontology, KEGG) and Network constructive software accelerate the assessment of the molecular pathways involved in high altitude diseases. Also, functional analysis by combined literature mining data from various sources and computational methods, enable to empower the discovery of new potential biological hypotheses in complex disorders including LOAD [53]. Therefore, as a major part of interactomes, protein-protein interaction (PPI) networks are powerful tools for decoding the complexity of the biological process. The identification of highly connected protein (hubs) in PPI networks underscores their involvement in essential complex biological modules. Network analysis, using topological characteristics (including degree and betweenness), has provided an important tool that can help identify biomarkers and probable therapeutic targets in complex diseases such as LOAD [22 , 55].

In the present study, we examined the mRNA levels of T-cell transcription factors (T-bet and GATA-3) and T-cell cytokines (IFN-γ, TNF-α, IL-4, and IL-5) derived from peripheral blood using real-time PCR. The selected genes in this study purposefully follow the expression of the main genes that differentiate Th cells into two subsets of “Th1” and “Th2” cells, and thus changes in the homeostasis of the immune cells at the molecular level can be assessed. It was hypothesized that there would be a significant difference between the mRNA levels of selected genes and, consequently, the homeostasis of the immune cells in favor of “cellular immunity” or “humoral immunity” in the peripheral blood of control individuals and LOAD patients. To further understand, these genes can be considered as seeds in the text mining approaches to extract the most related gene in the recently published experimental data on AD. Then, by using the final gene list and constructing a PPI network to investigate downstream pathways that play a pivotal role in the pathogenesis of AD, a set of potential hub genes and important pathways for future studies can be identified.

MATERIALS AND METHODS

Research sample profile

Due to the multifunctional nature of LOAD, the study population was confined to a city with no significant environmental pollutants. All participants had the same ethnicity and were from a similar geographical area. In addition, due to the constant presence of the participants in a residential facility and the same impact of social, cultural and nutritional factors on them, the study of changes in genetic factors has become more reliable. The participants were males and females aged 65 years and older selected from a residential aged care facility and two day-care centers situated in Mehriz County, Yazd, Iran. The inclusion criteria for the patient group were 1) aged 65 or above; 2) having at least 6 years of formal education; 3) suffering from moderate to severe AD based on the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (DSM-5) criteria [56]. The exclusion criteria were: 1) having early-onset/familial AD, non-AD dementia (secondary to other CNS or systemic conditions), delirious state, major mood disorders, or psychotic disorders; 2) having systemic inflammation-associated syndromes (e.g., sepsis); 3) using psychotropic or non-psychotropic medications that could substantially interfere with cognitive performance.

Initially, the participants were tested by Mini-Mental State Examination (MMSE). Individuals who scored≤21 on MMSE (according to the optimal cut-off score of the Persian version of MMSE [57], were classified as having a cognitive impairment and were further evaluated by a geriatric psychiatrist to confirm the diagnosis of Major Neurocognitive Disorder due to Alzheimer’s disease based on the DSM-5 criteria. Information from the personal and family history, general physical and neurological examination, laboratory tests and neuroimaging were used by the geriatric psychiatrist to make a more accurate diagnosis of AD. Participants who scored > 21 on MMSE were classified as the non-demented control group.

The local Ethics Committee of the National Institute of Genetic Engineering and Biotechnology Research (NIGEB) approved the study protocol. All participants or their primary caregivers delivered a written informed consent before the enrollment into the study.

Mononuclear cell separation, RNA isolation, and reverse transcription

Complete blood samples were collected in 5 ml volume from all participants by phlebotomy from the antecubital vein. The samples were injected into tubes containing 0.5 ml of ethylenediaminetetraacetic acid (EDTA) with anticoagulant activity. Peripheral blood mononuclear cells (PBMCs) were separated from the rest of the blood components via a gradient density centrifuge technique using Ficoll-Hypaque (Pharmacia, Uppsala, Sweden). RNA extraction from PBMCs was conducted using the High Pure RNA Isolation Kit (Roche, Germany) in accordance with the manufacturer’s instructions. The quantity and quality of the extracted RNA samples were verified by optical density ratio of 260 nm to 280 nm using NanoDrop 2000 instrument (Wilmington, USA) and agarose gel analysis, respectively. For reverse transcription-polymerase chain reaction (RT-PCR), 400 ng of total RNA was reverse transcribed into First Strand complementary DNA (cDNA) by cDNA synthesis kit (Fermentase, Germany) according to the manufacturer’s instruction. Finally, the cDNA samples were kept at –70°C before they were used for PCR.

PCR and real-time PCR analysis

Beta-actin was used as a housekeeping gene to normalize the expression of target genes. The primers were designed using Oligo7 software (version 7.56). To omit the amplification of genomic DNA and pseudogenes, primers were verified using Primer-BLAST online software (Table 1). In all samples, a PCR reaction was carried out to confirm transcription of target genes and to determine the contamination of the DNA using master mix PCR (Cinnagen, Tehran, Iran), specific gene primers and cDNA. Real-time PCR was performed for relative quantification of the mRNA expression of all targeted genes using SYBR Green Real-time PCR Master Mix (Light Cycler Fast Start DNA Master Plus SYBR Green I, Roche, Germany), previously Designed primers, and cDNA. A Thermocycler Rotor-Gene 6000 instrument (Corbett Research/Australia) was employed to detect fluorescent intensity in each sample. Melting curve analysis represented only one peak per reaction, as was later confirmed by observing a band by electrophoresis of PCR products.

Table 1

Primers used in Real-time PCR for cytokine, transcription, and internal control genes

Genes	Genbank Accession Number	Forward primer (5^′ ⟶ 3′)	Reverse primer (5^′ ⟶ 3′)	Amplicon size (bp)
T-bet	NM_013351	CCAAGTTTAATCAGCACCAGACAG	CAAGACCACGTCCACAAACATCC	131
IFN- γ	NM_000619	CGGTAACTGACTTGAATGTCCAAC	ACCATTACTGGGATGCTCTTCGAC	141
TNF- α	NM_000594	CTCCTTCAGACACCCTCAACC	CGAAGTGGTGGTCTTGTTGC	101
GATA-3	NM_001002295	CCCTCATTAAGCCCAAGCGAAG	CAGGGGTCTGTTAATATTGTGAAG	167
IL-4	NM_172348	CTTCCCCCTCTGTTCTTCCTG	CCTTCTCAGTTGTGTTCTTCTGCTC	130
IL-5	NM_000879	TGCTGATAGCCAATGAGACTCTG	GACTCTCCAGTGTGCCTATTCC	102
ACTB	NM_001101	AGACGCAGGATGGCATGGG	GAGACCTTCAACACCCCAGCC	161

T-bet, T-box transcription factor 21; IFN-γ, interferon-gamma; TNFα, Tumor necrosis factor alpha; GATA3, GATA binding protein 3; IL-4, interleukin-4; IL-5, interleukin-5; ACTB, β-actin.

Data analysis

PCR efficiency (E) and crossing point deviation (ΔCP) were calculated using the LinRegPCR Software (Version 11.0) according to Real-time PCR data of each sample. Next, statistical analyses and the calculation of relative gene expression were carried out using the Relative Expression Software Tool (REST © 2009, Version 2.0.13). The normal distribution of variables was examined via GraphPad Prism Software (version 8.0.2.263; San Diego, CA) using the Kolmogorov-Smirnov test. Mann-Whitney test was used to compare the means of the groups and Spearman’s rank correlation coefficient was carried out to determine the relationship between variables. p < 0.05 was considered statistically significant.

Protein-protein interaction (PPI)

Immune system and AD-associated genes (data set collection)

The associated genes, including T-cell transcription factors (T-bet and GATA-3) and T-cell cytokines (IFN-γ, TNF-α, IL-4 and IL-5) were selected as the seed proteins to construct the PPI network associated with AD. We identified the susceptibility genes associated with AD based on data mining through previous protein-protein interaction studies, Gene Expression Omnibus (GEO) studies, EVEX database, and a review of recent PubMed registered publications. The literature search was performed by using keywords: Protein-protein interaction, AD Gene Network, and Seed proteins.

Network construction

We used data mining to identify genes associated with AD. These were subsequently expanded to 224 which GeneMANIA (version 3.5.2) identifies based on the most related genes to a query using a guilt-by-association approach. Ten PPI networks for AD were visualized using the Cytoscape software (version3.7.2) [58] and the intersection network between them was extracted. In a given intersection network, we detected highly interconnected regions (clusters) using Molecular Complex Detection (MCODE) (https://baderlab.org/Software/MCODE). This method finds clusters based on vertex weighting by local neighborhood density and outward traversal from a locally dense seed protein to isolate the dense proteins. MCODE clustered the whole network by network cut-off set at 2, node score cut-off of 0.2 (K-Core:2), and a maximum depth of 100 [59].

Network analysis

Biological network analysis was performed with CentiScaPe (Version 2.2) to screen for hub proteins [60]. The topological characteristics selected were Degree, Betweenness, Closeness, Eigenvector, Bridging, Centroid Value, and Eccentricity [61]. The degree refers to the number of edges connecting a node. Betweenness relates to communication flow. Closeness is an estimate of the rate of flow of information from a given node to other nodes. The eigenvector measures the relevance of a node in a network. The Bridging is based on information flow and topological locality in the network. For instance, the centroid value of a node in a biological network can be interpreted as the “probability” of a protein being functionally capable of organizing discrete protein clusters or modules. The eccentricity can be interpreted as the accessibility of a protein to be functionally reached by all other proteins in the network [62]. This description is according to a network analysis of protein interaction data: an introduction published on EMBL-EBI Train online (https://www.ebi.ac.uk/training/online).

The values of (degree and betweenness), (degree and Bridging), (EigenVector and Betweenness), and (EigenVector and Degree) for all nodes in the module were plotted in a scatter plot to identify nodes with high values in both the centrality parameters using the Minitab1 17.3.1.

Enrichment analysis

To determine the biological meaning behind the intersection network, we performed enrichment analysis using DAVID (Database for Annotation, Visualization and Integrated Discovery), the functional annotation tool and retrieved Gene Ontology (GO) terms (for more details, see Supplementary Table 1). This database allowed us to identify the most appropriate biological processes for a gene/protein set.

Using charts for molecular function (GOTERM_MF_FAT), biological process (GOTERM_BP_FAT), and cellular component (GOTERM_CC_FAT), three lists of GO terms and their p values were generated independently (for more details, see Supplementary Table 2). For summarizing and visualizing GO categories, Reduce + Visualizes Gene Ontology (REVIGO) (http://revigo.irb.hr/) was applied, with the following parameters: “Small (0.5)” for the allowed similarity and “SimRel” for semantic similarity measures [63]. To gain insight into the most enriched biological pathways of the intersection network and its clusters, the Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis was also performed.

Second extended PPI network

All the genes were ranked according to seven topological characters and the total score for each one was assimilated. Afterwards, this gene list was sorted by total score and the lowest score in the seed group was selected as the cut-off. Based on this cutoff method, 20 genes were generated for the foundation. They were key to building the PPI network and constituted the backbone of the network. With these proteins as input, we used STRING (STRING v10.5) to construct the 2nd extended PPI network for a further comprehensive review. STRING can search for neighbor interactors and the proteins that have direct interactions with the inputted proteins. Furthermore, STRING generated a PPI network consisting of all selected proteins and the interactions between them. Gephi (version 0.9.2), a program for large network analysis, was used to analyze the structure of the PPI networks.

RESULTS

Comparison of demographic characteristics between LOAD and control groups

Means and standard deviations for age, MMSE total score, age of disease onset, and duration of illness of LOAD group and control group are presented in Table 2. The LOAD group had a significant difference in the mean MMSE score of 9.1 (SD±3.1) compared to the control group which had a mean score of 24.8 (SD±3.2). Based on these results, the majority of patients in the LOAD group were in the moderate to severe cognitive impairment range while the majority of the control group had normal cognitive status. In the LOAD group, the mean age of onset was 78.9 years (SD±8.8) and the average disease duration was 9.8 years (SD±6.5).

Table 2

Demographic and clinical characteristics of participants in the study groups

	LOAD patients (n = 30)	Controls (n = 30)
Male/female, n	15/15	15/15
Age (y), mean±SD	88.7±6.4	82.5±5.1
MMSE total score, mean±SD	9.1±3.1	24.8±3.2
Age of disease onset, mean±SD	78.9±8.8	–
Duration of illness (y), mean±SD	9.8±6.5	–

SD, standard deviation; LOAD, Late-onset Alzheimer’s disease; MMSE, Mini-Mental State Examination. Interpretation of the scores on the MMSE (out of 30 points) includes the categories of severe (9 or less), moderate (10–18 points), mild (19–23 points), and normal cognition (24 or more points).

LOAD decreased the mRNA expression of Th1-related transcription factor

The LOAD group had a significantly (Z = –6.638, p < 0.0001) lower mRNA expression of T-bet transcription factor in comparison to the control group (Table 3). The relative levels of T-bet mRNA expression based on the log2-transformed value are shown in Fig. 1.

Table 3

Statistical analysis of relative expression of the selected genes in study groups

	T-bet		INF- γ		TNF- α		GATA3		IL-4		IL-5
	Control	LOAD	Control	LOAD	Control	LOAD	Control	LOAD	Control	LOAD	Control	LOAD
Number of values	30	30	30	30	30	30	30	30	30	30	30	30
Mann-Whitney U	1.000		12.000		0.000		0.000		438.000		409.000
Wilcoxon W	466.000		477.000		465.000		465.000		903.000		874.000
Z	–6.638		–6.475		–6.652		–6.652		–0.177		–0.606
Asymp. Sig. (2-tailed)	0.000		0.000		0.000		0.000		0.859		0.544
(p < 0.05)?	yes		yes		yes		yes		No		No

Fig. 1

T-bet mRNA expression in PBMCs samples among control individuals and LOAD patients. The fold change ratio in LOAD patients was 0.007 compared to the control group. The relative levels of T-bet mRNA expression were significantly decreased in LOAD patients when compared to control individuals (p = 0.000). Error bars represent standard error of mean. The number of subjects in each of the groups = 30. Statistical significance is denoted by an asterisk: ^****p < 0.0001. PBMCs, peripheral blood mononuclear cells; LOAD, late-onset Alzheimer’s disease.

LOAD decreased the mRNA expression of Th1-related cytokines

Significantly lower relative levels of IFN-γ (Z = –6.475, p < 0.0001) and TNF-α (Z = –6.652, p < 0.0001) mRNA expression was shown in PBMCs derived from the LOAD group compared to control (Fig. 2).

Fig. 2

mRNA expression of Th1-related cytokines (IFN-γ and TNF-α) in PBMCs samples among control individuals and LOAD patients. The fold change ratio of IFN-γ and TNF-α in LOAD patients was 0.032 and 0.003 compared to the control group, respectively. The relative levels of Th1-related cytokines (IFN-γ and TNF-α) mRNA expression were significantly decreased in LOAD patients when compared to control individuals (p = 0.000 and p = 0.000, respectively). Error bars represent standard error of mean. The number of subjects in each of the groups = 30. Statistical significance is denoted by an asterisk: ^****p < 0.0001. PBMCs, peripheral blood mononuclear cells; LOAD, late-onset Alzheimer’s disease.

LOAD decreased the mRNA expression of Th2-related transcription factor

The LOAD group had significantly (p < 0.0001) lower mRNA expression of GATA-3 transcription factor in comparison to the control. The relative levels of GATA-3 mRNA expression based on the log2-transformed value are shown in Fig. 3.

Fig. 3

GATA-3 mRNA expression in PBMCs samples among control individuals and LOAD patients. The fold change ratio in LOAD patients was 0.006 compared to the control group. The relative levels of GATA-3 mRNA expression were significantly decreased in LOAD patients when compared to control individuals (p = 0.000). Error bars represent standard error of mean. The number of subjects in each of the groups = 30. Statistical significance is denoted by an asterisk: ^****p < 0.0001. PBMCs, peripheral blood mononuclear cells; LOAD, late-onset Alzheimer’s disease.

The mRNA expression of Th2-related cytokines were unchanged in LOAD

Expression of IL-4 and IL-5 did not significantly differ between the LOAD group and control. The relative levels of IL-4 and IL-5 mRNA expression based on the log2-transformed value are shown in Fig. 4.

Fig. 4

mRNA expression of Th2-related cytokines (IL-4 and IL-5) in PBMCs samples among control individuals and LOAD patients. The fold change ratio of IL-4 and IL-5 in LOAD patients was 0.911 and 0.753 compared to the control group, respectively. The relative levels of Th2-related cytokines (IL-4 and IL-5) mRNA expression in LOAD patients were not significantly different in comparison to control individuals (p > 0.05). Error bars represent standard error of mean. The number of subjects in each of the groups = 30. PBMCs, peripheral blood mononuclear cells; LOAD, late-onset Alzheimer’s disease.

Network studies related to immune system involvement in AD

Network analysis

We performed data mining to evaluate key factors involved in AD pathogenesis. Gene selection and raw gene sets were expanded during gene ontology enrichment analysis by GeneMANIA (version 3.5.2). There were 244 nodes and 6668 edges discovered in the AD network. (For more details about the Gene list, see Supplementary Table 3). Highly connected regions of the intersection network were identified using the MCODE plugin implemented in the Cytoscape platform. A total of ten clusters were detected. The number of nodes in clusters 1, 2, 3, 4,5, 6, 7, 8, 9, and 10 was 36, 15, 30, 24, 19, 6, 20, 3, 3, and 10, respectively (score: cluster 1 : 19.371, cluster 2 : 12.143, cluster 3 : 10.207, cluster 4 : 7.217, cluster 5 : 5.778, cluster 6 : 4.000, cluster 7 : 3.474, cluster 8 : 3.000, cluster 9 : 3.000, and cluster 10 : 2.889). Table 4 details these results. We utilized MCODE to recognize networks clustering using the gene set extracted from the data mining. MCODE: Table 5 shows the numbers of nodes and edges derived from data mining for AD genes. MCODE plugins identified the most important modules in the network by the score of each cluster. The most important module is related to cluster 1 with a score of 19.371, which consists of T-cell Cytokines and AD seeds (Fig. 5).

Table 4

PPI data derived from data mining for AD genes associated; MCODE clusters

Clusters	Node	Edges	Score
1	36	461	19.371
2	15	123	12.143
3	30	195	10.207
4	24	109	7.217
5	19	59	5.778
6	6	10	4.000
7	20	38	3.474
8	3	3	3.000
9	3	4	3.000
10	10	16	2.889

Table 5

MCODE clusters nodes

MCODE	Node IDs
Clusters
1	AREG,EREG,GZMB,IL15,IFNG,CXCL10,CTSS,PPP1R15A,NR4A1,CSF3,PSME1,CXCL2,IRF7,PSMB10,ZFP36,CD4,TNF,PSMB8,STAT1,CXCL8,TRIM22,EGR1,FOSL1,JUNB,IL1B,CXCL3,CCL4,ICAM1,CCL5,CASP1,TNFAIP3,IL4R,HLA-A,HLA-DRB5,BCL3,TIMP1
2	TNFAIP2,NAMPT,HLA-G,CCL2,CXCL1,IL6,IRF1,FOS,NFKBIA,GBP1,TAP1,HBEGF,CCL3,JAK2,STAT3
3	IL11,IFNA1,IL12A,IL2RA,IRF4,PRDM1,CYBB,S100A12,IL12B,MX1,FOSB,JUND,CTF1,APOE,CYBA,CD68,DAPK1,IL18,STAT5A,JAK1,SYK,RELA,CXCL9,NFKB1,IL19,IL21,NCF1,A2M,IL5,IFNA17
4	IL4,ICOS,CD80,IL17A,IL7,PIK3CD,CD28,MAPK13,SP1,MMP2,NFATC3,CD200,PTK2B,GATA-3,RAC1,KDR,TP53,IL33,CR1,TGFB1,CD55,PPP3CA,PLD1,IFNAR1
5	FASLG,RIPK2,PF4,STAT5B,SDC4,CIITA,INPP5D,IL16,FCGRT,S100A11,LGALS3,MS4A1,PRKCD,PDGFB,CYP27A1,PNP,PIK3CA,JUN,TBX21
6	SETBP1,SATB1,ARNT,CD34,PPARGC1A,EPHA1
7	TNFRSF11B,IL1RL1,MUC1,CMA1,SORL1,MAPK9,CFTR,RARRES2,FERMT2,HMGB1,HSP90AB1,ANGPTL4,TNFRSF10A,SUMO1,GJA1,CELF1,EGF,GAB2,CDKN1B,MAPK10
8	POMC,CEL,IL9
9	HSPD1,GAPDH,TRAP1
10	CRP,AMBP,WWC1,MEF2C,SERPINA3,FOXO3,FOXP3,TG,WISP1,BDKRB2

Fig. 5

Ten subnetworks constructed of significant (p < 0.01) KEGG enrichment pathways of intersection network identified by MCODE plugin. Nodes and links represent human genes and gene interactions, respectively. a) Cluster 1, b) Cluster 2, c) Cluster 3, d) Cluster 4, e) Cluster 5, f) Cluster 6, g) Cluster 7, h) Cluster 8, i) Cluster 9, j) Cluster 10. Red circles represent the Hub genes with a higher score in each cluster.

Using the DAVID annotation system, we identified the most significant KEGG pathways within each cluster. The highest scored cluster (Cluster 1) contains nodes that associate with immune system involvement and most clusters contain nodes that are associated with Th17 cell differentiation signaling pathway (hsa04659), Th1 and Th2 cell differentiation (hsa04658), and MAPK signaling pathway (hsa04010) (Table 6).

Table 6

KEGG pathways of intersection network and each MCODE cluster by DAVID

Clusters	Number of Nodes	KEGG Pathway	Genes	p
Cluster 1	12	Cytokine-cytokine receptor interaction	CCL4,CCL5,CXCL10,CXCL2,CXCL3,CXCL8,CSF3,IFNG,IL1B,IL15,IL4R,TNF	1.2E-8
Cluster 2	7	Chemokine signaling pathway	CCL2,FOS,JAK2,NFKBIA,IL6,HLA-G,TAP1	9.3E-6
Cluster 3	12	Jak-STAT signaling pathway	JAK1,CTF1,IFNA1,IFNA17,IL11,IL12A,IL12B,IL19,IL2RA,IL21,IL5,STAT5A	1.5E-12
Cluster 4	6	MAPK signaling pathway	MAPK13,NFATC3,PPP3CA,RAC1,TGFB1,TP53	8.1E-4
Cluster 5	3	MAPK signaling pathway	FASLG,JUN,PDGFB	9.2E-2
Cluster 7	3	MAPK signaling pathway	EGF,MAPK10,MAPK9	8.0E-2

The second enrichment analysis, using selected hub genes with the highest score (total score cut-off < 300), further revealed the involvement of Th1 and Th2 cell differentiation (hsa04658), Th17 cell differentiation (hsa04659), and KEGG pathways in AD (Table 7).

Table 7

The proteins in the second extended PPI network involved in the KEGG pathways of AD

Pathways	description	count in gene set	false discovery rate
hsa04659	Th17 cell differentiation	12 of 102	1.13e-20
hsa04658	Th1 and Th2 cell differentiation	9 of 88	7.35e-15
hsa04010	MAPK signaling pathway	3 of 293	0.0055

Enrichment analysis

The results of the enrichment analysis were applied to construct three GO networks. The GO terms provided by the DAVID annotation system with a smaller EASE Score (a modified Fisher Exact p-value) were further enriched and associated with a ranked gene list in the intersection. According to the p values, the most enriched GO terms were those involved in cytokine production (GO:0001816) and immune response (GO:0006955) (Fig. 6).

Fig. 6

GO enrichment analysis visualized by REVIGO. Significantly enriched GO terms related to; a) scatterplot: biological processes, b) TreeMap: Cell components, and c) interactive graph: Molecular functions. the most important paths are marked with stars. Clustering of circles (representing GO terms) was based on semantic similarities to other GO terms in the gene ontology (larger circles represent more general terms whereas adjoin circles depict close relationships). The size and color of the circle represent the GO term frequency and the log10 P-value for the enrichment derived from the AgriGO analysis (red higher, yellow lower) respectively for part a and (red higher, blue lower) for part b. The most pivotal pathway marked with stars in part c.

Identification of hubs

The Cytoscape plugin, jActiveModules, permits the scoring of the connected regions of the merged network. The total number of nodes, in the most expressed active subnetwork (Score: 19.371), consisted of 36 nodes and 461 edges. These were examined for centrality parameters. In a scatter plot of two topological parameters, hub nodes with the highest degree and betweenness, Badging and degree, Betweenness and Eigenvector, and Eigenvector and degree value were identified. Figure 7 and Table 8 show these results, while Supplementary Table 3 provides more detail.

Fig. 7

Scatter plots of the centralities parameters. All the most important hubs genes with a high score are determined by name in the plot which contribute to important pathways associated with the AD. Scatter-plot of the betweenness centrality versus degree, Bridging versus degree, centroid versus degree, Eigenvector versus betweenness centrality, Eigenvector versus Bridging and Eigenvector versus Degree, respectively. Joint hub genes extracted from different scatter plots provide the final gene list for 2nd extended PPI network construction.

Table 8

Hub nodes

Node Name	Description	KEGG pathway	Betweenness	Degree	Bridging	Eigen Vector
NFKB1	Translocates into the nucleus and stimulates the expression of genes involved in a wide variety of biological functions	MAPK signaling pathway/ Th1 and Th2 cell differentiation/ Th17 cell differentiation	706.010	164	1.778	0.135
STAT3	Transcription activators	Th17 cell differentiation	591.774	161	1.564	0.121
STAT1	Transcription activators	Th1 and Th2 cell differentiation/ Th17 cell differentiation	516.150	143	1.730	0.120
GATA-3	Regulator of T-cell development	Th1 and Th2 cell differentiation/ Th17 cell differentiation	611.302	87	4.033	0.086
IRF1	Activator of genes involved in both innate and acquired immune responses	Th1 and Th2 cell differentiation	517.546	127	2.153	0.136
JUN	Increased steroidogenic gene expression upon cAMP signaling pathway stimulation	MAPK signaling pathway	932.485	178	1.861	0.128

It is noteworthy that eight out of ten hubs related to Th1 and Th2 cell differentiation and three out of ten hubs corresponded to MAPK signaling pathway.

The second extended PPI network

With the use of the backbone nodes as input, we implemented STRING to construct the second extended PPI network associated with AD. This network (without co-expression interactions) demonstrated 9 nodes for the KEGG pathways of Th1 and Th2 cell differentiation (hsa04658) (JUN, NFKB1, NFKB1A, JAK2, STAT1, GATA-3, IFNG, IL4R, CD4), 12 nodes of Th17 cell differentiation (hsa04659) (JUN, NFKB1, NFKB1A, IL6, STAT3, GATA-3, STAT1, JAK2, IL1B, IFNG, IL4R, CD4), and 3 nodes of MAPK signaling pathway (hsa04010) (JUN, NFKB1, IL1B) (Table 8). Moreover, the second extended PPI network, with co-expression interactions, is displayed in Fig. 8. This network demonstrated that most nodes related to the KEGG pathway are related to the immune system and further strengthens the involvement of potential aberrant immune response in AD.

Fig. 8

2nd extended PPI network by STRING represented three potential pathways associated with AD. The nodes of the 2nd extended PPI network involved in the KEGG pathways of Th1 and Th2 cell differentiation, Th17 cell differentiation and MAPK signaling pathway were colored in blue, green and red respectively.

DISCUSSION

The aim of this study was to compare the mRNA level of Th-related genes in PBMCs of LOAD patients and controls. A significant decrease at the mRNA levels of T-bet, IFN-γ, TNF-α, and GATA-3 was observed in PBMCs collected from LOAD patients compared to controls. However, there was no significant difference in IL-4 and IL-5 mRNA levels between the two groups. These changes in the expression of immune-related genes can be interpreted from two perspectives: component functions and biological pathways.

The main subsets of peripheral CD4 + T cells are Th1 and Th2. Each of these subpopulations is classically controlled by a master regulator: Th1 is controlled by T-bet (Tbx21) and Th2 by GATA-3. The production of cytokines in each of these two main subgroups has also been identified. For example, Th1 cells are known to produce IFN-γ, TNFα and IL-2 which makes them pro-inflammatory and pathogenic, while Th2 cells produce IL-4, IL-5 and IL-13, giving them the ability to neutralize the Th1 response [64, 65].

Transcription factors control the growth of specific immune cells lineages by activating or suppressing genes that are vital to cellular phenotype [66]. The transcription factor T-bet, encoded by the Tbx21 gene, is the Th1 cell-defining transcription factor [67, 68]. T-bet expression in different cells of the innate and adaptive immune system indicates that it is essential for the maintenance of immunity and the effective clearance of pathogens [66, 68]. This is the first paper analyzing T-bet expression in PBMCs of patients with LOAD in comparison to control individuals. The reduction observed in relative T-bet expression in LOAD may reflect a peripheral pathway that contributes to the pathophysiology of AD. This is reinforced by the finding that the pattern of peripheral gene expression is reasonably related to CNS gene expression [50]. Simultaneous studies of T-bet expression in peripheral blood and CNS tissue samples would help validate this observation and determine its significance in AD patients. Another possibility underlying the reduced T-bet expression in PBMCs of LOAD may relate to a protective role. T-bet silencing can protect and treat autoimmunity due to its regulatory effects on the pro-inflammatory immune response [69]. Thus, the downregulation of T-bet in our study may have occurred as a result of a protective mechanism in LOAD patients.

IFN-γ is induced by T-bet expression and plays a critical role in Th1 cell biology [70, 71]. IFN-γ is expressed by infiltrating Th1 cells, resident microglia, and neurons and is involved in the development of AD and systemic autoimmunity. IFN-γ signaling is known to worsen AD pathology and cognitive function [72, 73]. Activation of microglia by INF-γ also inhibits Aβ clearance in the brain [74, 75]. In the present study, we expected that in the LOAD group a relative decrease in T-bet expression would lead to a relative decrease in IFN-γ expression. Although no change in IFN-γ levels in serum or CSF of AD patients has been reported [76, 77], an association between low IFN-γ-producing genotype and rapid disease progression has been shown [78]. Therefore, the downregulation of IFN-γ in the present study may have occurred as a result of a protective mechanism in LOAD patients.

TNF-α is a pro-inflammatory cytokine that acts as a pivotal molecule in inflammation [79]. Like other cytokines, TNF-α can cross the healthy blood-brain barrier and affect cognitive function [80 –82]. Important roles for TNF-α have been suggested in many neurological disorders, including AD [83 –86]. TNF-α is involved in the regulation of several physiological processes in the CNS, including immune surveillance, immune reactions to fight microbial infections, and induction of cell death [87 –89]. In animal models of AD, an increase in TNF-α is shown to be a key element in the inflammatory cascade [90]. In humans, short-term anti-TNF-α treatment has been shown to improve cognition in AD patients [76, 91]. Although, in general, the level of cytokines detected in AD patients has been reported to be inconsistent among different research groups [76], the TNF-α cytokines have been reported to be predominantly elevated in CSF and plasma of AD patients [92]. Consistent with the downregulation of T-bet, we observed a decrease in TNF-α expression in the AD group compared to the control group. Other groups have not detected any differences in the serum TNF-α levels in AD patients compared to healthy individuals [93]. Furthermore, a meta-analysis study showed that there was a significantly higher level of TNF-α in peripheral blood samples from AD patients compared to controls [77]. Since the present study is based on gene expression in PBMCs and TNF-α is produced by different types of cells including lymphoid cells, mast cells, endothelial cells, cardiac myocytes, adipose tissue, fibroblasts, and neurons [94], additional studies are recommended that compare the gene expression profile in specific cells to the actual protein content of cytokines in the plasma.

Induction of the master transcription factor GATA-3 occurs when CD4 + T cells are exposed to IL-4 [95]. These Th2-derived subtypes can penetrate the CNS parenchyma and have an important regulatory effect on microglia and peripheral macrophages. Th2-derived cytokines not only attenuate M1 and M1-like phenotypes in macrophages/microglia but also induce M2 and M2-like phenotypes respectively in these cells and are required for neurogenesis [96 –98]. Network analysis of transcription factor gene expression datasets in mild cognitive impairment (MCI) and AD clearly showed that GATA-3 is one of the associated transcription factors involved in both MCI and AD [99]. In our study, a decrease in GATA-3 transcription factor expression was observed in the LOAD group compared with the control group. This data is consistent with the findings of another study which showed that AD and MCI were both directly related to low CD4 + GATA-3 cell percentage [100]. These findings contradict the report of increased activity and protein levels of GATA-3 transcription factor in familial AD patients [101], which may be of a different nature to sporadic forms of the disease. GATA-3 was significantly increased in MCI subjects compared to either AD or age-matched healthy individuals [102]. Therefore, GATA-3 expression changes in different studies may reflect different stages of AD.

IL-4 is an anti-inflammatory cytokine produced by mature Th2, mast cells or basophils. This interleukin is involved in T cell proliferation and higher normal brain functions such as memory and learning [103]. The protective role of IL-4 in AD has been linked to the regulation of microglia, Aβ phagocytosis, and the secretion of anti-inflammatory and neurotrophic cytokines [104 –106]. Some studies have reported a possible association between IL-4 polymorphism and increased susceptibility to AD in certain populations [107, 108]. IL-4 levels in MCI subjects were significantly higher than in the control and AD groups [109], and it appears that these changes may also have been influenced by disease stages. Meta-analysis showed no significant difference between IL-4 blood levels in AD patients and controls [77]. Therefore, the results of comparing the relative expression of IL-4 gene between LOAD and control in the present study are consistent with some previous studies.

IL-5 is often expressed in Th2 cells and is regulated by several transcription factors, including GATA-3 [110]. IL-5 is mainly associated with eosinophil involvement in allergen-induced late-stage airway inflammation [111]. There is some evidence that IL-5 is also involved in the pathogenesis of AD. IL-5 blocked apoptosis and tau hyperphosphorylation [112]. A negative association between CSF IL-5 levels and disease progression has been observed in patients with AD [16]. The present study did not show a significant difference in the relative expression of IL-5 gene between the AD group and the control group. This finding differed from the results of a study in which the mean serum and CSF IL-5 concentrations were significantly increased in the MCI/AD groups compared to the control group [113]. These varied results may be due to the sample collection or patient profile analyzed.

From a biological pathways perspective, the genes selected in this study followed exactly the main pathways of differentiation into two subtypes of T-helper cells, “Th1” and “Th2” cells. In the late 1980s, the “Th1/Th2” theory was established, linking the development of each CD4 + T cell subset to the production of unique cytokines and their main regulator [36]. In line with this hypothesis, genome-wide analyses showed that T-bet directly induces many genes associated with Th1 but indirectly suppresses the genes associated with Th2 [114].

In the present study, the pathway of Th1 differentiation was completely downregulated from the expression of Th1 cell transcription factor (T-bet) to related cytokines (IFN-γ and TNF-α), which may be evidence of impaired “cellular immunity” in LOAD patients. However, based on the homeostasis hypothesis between “Th1 and Th2 activity”, our previous expectation of enhancing “humoral immunity” by increasing the expression of Th2 cell transcription factor GATA-3 and related cytokines (IL-4 and IL-5) was not observed. To better clarify this imbalance of “Th1 / Th2” in LOAD patients additional studies are warranted. The classic Th1 / Th2 paradigm of CD4 + T helper cells have changed in recent years. Some scientists have suggested new Th subsets as members of the Th1 and Th2 families. These subsets form a broad and delicate network of the CD4 + T cell family. Epigenetic regulation may help determine the fate of T cells [71] since activation and differentiation of Th cells is a complex process in which a significant number of genes are involved [115]. The pathogenesis of AD is linked to the modulation of several immune events and the serotonergic pathway [116, 117]. Therefore, it seems that to better understand how the immune system changes in AD, the interactions of neuronal systems on the immune system should be investigated.

This study focused mainly on peripheral changes of inflammatory mediators in LOAD patients. But whether these changes in the peripheral immune system play a mechanical, compensatory, or perhaps merely associative role in the pathogenesis of AD is unclear and needs further investigation. On the other hand, the local production of mediators by microglia is also very important. Microglia as immune cells residing in the brain are constantly monitoring the microenvironment under physiological conditions. Microglia play a delicate and complex role in the development of AD. Active microglia may play a potentially harmful role by enhancing the expression of pro-inflammatory cytokines such as IL-1β, IL-6, and TNF-α, which affect surrounding brain tissue [118, 119]. To better understand the crosstalk of peripheral cytokines with CNS-secreted cytokines, a number of studies have been performed. Pro-inflammatory cytokines have been shown to create a pro-inflammatory environment in the central nervous system by crossing the blood-brain barrier, signaling through endothelial cells or organs around the ventricles, and stimulating the vagus nerve. Through each of these pathways, systemic inflammation is believed to induce pro-inflammatory microglia and astrocytic phenotypes that can increase tau hyperphosphorylation, Aβ oligomerization, complement activation, and the breakdown of neurotransmitters into potentially harmful bioactive metabolites [120, 121]. Therefore, in the field of AD, T helper cells can modulate the microglial by contact or release of cytokines.

To complement our investigations, we conducted a network analysis using a vast gene-set related to AD to gain a better understanding of the underlying molecular mechanisms involved in the observed Th1 / Th2 imbalance. We provided a gene pool by data mining of recent biomedical literature and expression database according to our Th-related genes as seed. Afterward, we constructed a PPI network using a gene pool list to yield a more comprehensive perspective of biological processes employing pathway-based and network-based analyses in AD. Ultimately the crosstalk between several involved biological pathways revealed MAPK signaling pathway as the most relevant process to explore probable explaining the Th1/Th2 imbalance observed in this study. Previous reports indicated enhanced Th2 and down-regulated Th1 immunity following immune challenge with Aβ in AD [122]. The present study demonstrated controversial results, that contradict the Th1/Th2 differentiation theory. Two mechanisms including STAT6-dependent and -independent were involved in the process of Th2 polarization, including Notch, mTORC2, IL-2/STAT5, and Wnt [123]. All these mechanisms promote the expression of the Th2 master regulator GATA-3, thereby the downstream expression of Th2 cytokines (IL-4 and IL-5). GATA-3 expression is drastically reduced and the IL-4 and IL-5 expression levels were not significant in the LOAD group compared to the control. To explore the underlying molecular mechanisms in the Th1/Th2 imbalanced observed, network analysis has been designed, which proposed that MAPK signaling pathways can exhibit off-target effects. Down-expression of IL-4 and IL-5 is expected in presence of T-bet and GATA-3 reduction [124], which is contradicted by observations in the present study. One another hypothesis is that, in the GATA-3 reduction state, IL-4 and IL-5 expression induction could regulate through alternate P38/JUN via MAPK signaling pathway, which represents a GATA-3-independent pathway. MAPK signaling over-activation has been observed in AD [125]. Other suppressors, such as RUNX3 could maintain the overexpression of these cytokines in a normal range.

MAPK signaling has a crucial role in anti-aging and anti-inflammatory processes [126]. Inflammation and oxidative stress are present in the brain of AD mice, modulated via the hippocampal MAPK signaling pathways which respond strongly to diverse environmental stimuli [125, 127]. MAPK signaling pathway plays pivotal roles in Aβ deposition, tau phosphorylation, and synaptic dysfunction. Microglia and astrocyte-mediated neuroinflammation and impairment have been mediated by MAPK [125]. Recently a study reports that the phosphorylation level of p38, JNK, and ERK are increased in the hippocampus of AD mice. MAPK signaling inhibition in glial cells may offer a therapeutic strategy in AD [128 –131]. Network analysis results determined JUN as the most pivotal hub related to AD. In the support of our result, the activation of NF-κB, c-Jun/AP1, and STAT3 signaling promoted neuroinflammation in AD [132].

A computational study, using microarray data for both gene expression and DNA methylation for ageing-related genes in the brain and their relationship with AD, revealed the maximum positive correlation r = 0.85 in the upstream CpG island of RUNX3 gene [133]. RUNX3 diminished the GATA-3 activity through direct interaction (T-bet independent manner) [134]. We assumed that despite the reduction in T-bet, GATA-3 could be suppressed by RUNX3 mediated pathway in a T-bet-independent manner [135]. Besides, network analysis suggested the role of Th17 cell differentiation pathway (hsa04659) in AD. This result is supported by findings that an imbalance of circulating Th17 contributes to the neuropathological changes in AD [102, 136]. T-bet and GATA-3 act as repressors of the Th17 cytokine, therefore a decline in Th1 and Th2 polarization could promote Th17 pathway [123].

The present study has several limitations. The focus of these studies was on the mRNA analysis and did not include an evaluation of the protein derived from each gene. We considered and analyzed only a limited number of cytokines. Furthermore, the sample size for each group may not be large enough to be representative of the overall disease profile. Nevertheless, this study does provide a comprehensive analysis of the Th polarization pathway and its potential role in AD pathogenesis. This information can be used to design more extensive studies on the role of T cell subsets in neurodegenerative disorders.

Conclusion

The present study compared the relative mRNA expression of genes related to the Th differentiation pathway in the peripheral blood of LOAD and control groups. The results showed that there may be a molecular dysregulation in the peripheral immune homeostasis of AD patients. These changes in the expression of Th-related genes may be of interest in early diagnosis or targeted interventions of LOAD. In addition, PPI analysis indicated that off-target upstream stimulation may involve activation of the MAPK cascade and induction of the Th17 axis. These pathways may play a role in Th1/Th2 imbalance in LOAD and may even be considered as a new LOAD treatment strategy in the future, which needs further study. Overall, our findings may highlight the role of the peripheral immune system in the pathogenesis of LOAD, but further research on larger groups of patients are needed to confirm the findings of this study.

Footnotes

ACKNOWLEDGMENTS

The authors thank the participation of the medical staff, nurses and patients in the Neshat residential aged-care facility, Mehriz County, Yazd, Iran.

Authors’ disclosures available online ().

The supplementary material is available in the electronic version of this article: .

References

Rajan

, Weuve

, Barnes

, McAninch

, Wilson

, Evans

(2021) Population estimate of people with clinical Alzheimer’s disease and mild cognitive impairment in the United States (2020–2060). Alzheimers Dement, doi: 10.1002/alz.12362.

Jack

Jr , Wiste

, Schwarz

, Lowe

, Senjem

, Vemuri

, Weigand

, Therneau

, Knopman

, Gunter

(2018) Longitudinal tau PET in ageing and Alzheimer’s disease. Brain 141, 1517–1528.

Duyckaerts

, Delatour

, Potier

M-C

(2009) Classification and basic pathology of Alzheimer disease. Acta Neuropathol (Berl) 118, 5–36.

Vermunt

, Sikkes

, van den Hout

, Handels

, Bos

, van der Flier

, Kern

, Ousset

P-J

, Maruff

, Skoog

(2019) Duration of preclinical, prodromal, and dementia stages of Alzheimer’s disease in relation to age, sex, and APOE genotype. Alzheimers Dement 15, 888–898.

Alzheimer’s Association (2019) 2019 Alzheimer’s disease facts and figures. Alzheimers Dement 15, 321–387.

Long

, Holtzman

(2019) Alzheimer disease: An update on pathobiology and treatment strategies. Cell 179, 312–339.

Kanatsu

, Tomita

(2017) Molecular mechanisms of the genetic risk factors in pathogenesis of Alzheimer disease. Front Biosci (Landmark Ed) 22, 180–192.

Sanabria-Castro

, Alvarado-Echeverría

, Monge-Bonilla

(2017) Molecular pathogenesis of Alzheimer’s disease: An update. Ann Neurosci 24, 46–54.

Gatz

, Reynolds

, Fratiglioni

, Johansson

, Mortimer

, Berg

, Fiske

, Pedersen

(2006) Role of genes and environments for explaining Alzheimer disease. Arch Gen Psychiatry 63, 168–174.

10.

Dorszewska

, Prendecki

, Oczkowska

, Dezor

, Kozubski

(2016) Molecular basis of familial and sporadic Alzheimer’s disease. Curr Alzheimer Res 13, 952–963.

11.

McGeer

, Itagaki

, Boyes

, McGeer

(1988) Reactive microglia are positive for HLA-DR in the substantia nigra of Parkinson’s and Alzheimer’s disease brains. Neurology 38, 1285–1285.

12.

Rogers

, Webster

, Lue

L-F

, Brachova

, Civin

, Emmerling

, Shivers

, Walker

, McGeer

(1996) Inflammation and Alzheimer’s disease pathogenesis. Neurobiol Aging 17, 681–686.

13.

Frank-Cannon

, Alto

, McAlpine

, Tansey

(2009) Does neuroinflammation fan the flame in neurodegenerative diseases? Mol Neurodegener 4, 47.

14.

Amor

, Puentes

, Baker

, Van Der Valk

(2010) Inflammation in neurodegenerative diseases. Immunology 129, 154–169.

15.

Heneka

, Kummer

, Latz

(2014) Innate immune activation in neurodegenerative disease. Nat Rev Immunol 14, 463–477.

16.

Taipa

, das Neves

, Sousa

, Fernandes

, Pinto

, Correia

, Santos

, Pinto

, Carneiro

, Costa

(2019) Proinflammatory and anti-inflammatory cytokines in the CSF of patients with Alzheimer’s disease and their correlation with cognitive decline. Neurobiol Aging 76, 125–132.

17.

Heppner

, Ransohoff

, Becher

(2015) Immune attack: The role of inflammation in Alzheimer disease. Nat Rev Neurosci 16, 358–372.

18.

Consortium IGoAsD, Jones

, Lambert

, Wang

, Choi

, Harold

, Vedernikov

, Escott-Price

, Stone

, Richards

(2015) Convergent genetic and expression data implicate immunity in Alzheimer’s disease. Alzheimers Dement 11, 658–671.

19.

Karch

, Goate

(2015) Alzheimer’s disease risk genes and mechanisms of disease pathogenesis. Biol Psychiatry 77, 43–51.

20.

Sims

, Van Der Lee

, Naj

, Bellenguez

, Badarinarayan

, Jakobsdottir

, Kunkle

, Boland

, Raybould

, Bis

(2017) Rare coding variants in PLCG2, ABI3, and TREM2 implicate microglial-mediated innate immunity in Alzheimer’s disease. Nat Genet 49, 1373–1384.

21.

Jansen

, Savage

, Watanabe

, Bryois

, Williams

, Steinberg

, Sealock

, Karlsson

, Hagg

, Athanasiu

(2018) Genetic meta-analysis identifies 9 novel loci and functional pathways for Alzheimers disease risk. Biorxiv, 258533.

22.

Malamon

, Kriete

(2020) Erosion of gene co-expression networks reveal deregulation of immune system processes in late-onset Alzheimer’s disease. Front Neurosci 14, 228.

23.

Wyss-Coray

, Rogers

(2011) Inflammation in Alzheimer disease–a brief review of the basic science and clinical literature. Cold Spring Harb Perspect Med 2, a006346.

24.

Calsolaro

, Edison

(2016) Neuroinflammation in Alzheimer’s disease: Current evidence and future directions. Alzheimers Dement 12, 719–732.

25.

Reddy

, Seth

(2019) Role of tumor necrosis factor in neurodegeneration. EC Endocrinol Metab Res 4, 162–171.

26.

Efthymiou

, Goate

(2017) Late onset Alzheimer’s disease genetics implicates microglial pathways in disease risk. Mol Neurodegener 12, 43.

27.

Ransohoff

(2016) How neuroinflammation contributes to neurodegeneration. Science 353, 777–783.

28.

Cagnin

, Brooks

, Kennedy

, Gunn

, Myers

, Turkheimer

, Jones

, Banati

(2001) In-vivo measurement of activated microglia in dementia. Lancet 358, 461–467.

29.

Boche

, Perry

, Nicoll

(2013) Activation patterns of microglia and their identification in the human brain. Neuropathol Appl Neurobiol 39, 3–18.

30.

Heneka

, Carson

, Khoury

, Landreth

, Brosseron

, Feinstein

, Jacobs

, Wyss-Coray

, Vitorica

, Ransohoff

, Herrup

, Frautschy

, Finsen

, Brown

, Verkhratsky

, Yamanaka

, Koistinaho

, Latz

, Halle

, Petzold

, Town

, Morgan

, Shinohara

, Perry

, Holmes

, Bazan

, Brooks

, Hunot

, Joseph

, Deigendesch

, Garaschuk

, Boddeke

, Dinarello

, Breitner

, Cole

, Golenbock

, Kummer

(2015) Neuroinflammation in Alzheimer’s disease. Lancet Neurol 14, 388–405.

31.

Hong

, Kim

, Im

H-I

(2016) Pathophysiological role of neuroinflammation in neurodegenerative diseases and psychiatric disorders. Int Neurourol J 20, S2.

32.

Leszek

, E Barreto

, Gasiorowski

, Koutsouraki

, Aliev

(2016) Inflammatory mechanisms and oxidative stress as key factors responsible for progression of neurodegeneration: Role of brain innate immune system. CNS Neurol Disord Drug Targets 15, 329–336.

33.

Rubio-Perez

, Morillas-Ruiz

(2012) A review: Inflammatory process in Alzheimer’s disease, role of cytokines. ScientificWorldJournal 2012, 756357.

34.

Smith

, Das

, Ray

, Banik

(2012) Role of pro-inflammatory cytokines released from microglia in neurodegenerative diseases. Brain Res Bull 87, 10–20.

35.

Raphael

, Nalawade

, Eagar

, Forsthuber

(2015) T cell subsets and their signature cytokines in autoimmune and inflammatory diseases. Cytokine 74, 5–17.

36.

Mosmann

, Cherwinski

, Bond

, Giedlin

, Coffman

(1986) Two types of murine helper T cell clone. I. Definition according to profiles of lymphokine activities and secreted proteins. J Immunol 136, 2348–2357.

37.

Berger

(2000) Th1 and Th2 responses: What are they? BMJ 321, 424.

38.

Van Gool

, Van de Beek

, Eikelenboom

(2010) Systemic infection and delirium: When cytokines and acetylcholine collide. Lancet 375, 773–775.

39.

Wang

, Song

, Chen

, Leng

(2018) Connection between systemic inflammation and neuroinflammation underlies neuroprotective mechanism of several phytochemicals in neurodegenerative diseases. Oxid Med Cell Longev 2018, 1972714.

40.

Magalhães

, Weiler

, Teixeira

, Hayata

, Moraes

, Boldrini

, Dos Santos

, de Campos

, de Rezende

, Joaquim

(2018) Systemic inflammation and multimodal biomarkers in amnestic mild cognitive impairment and Alzheimer’s disease. Mol Neurobiol 55, 5689–5697.

41.

Walker

, Gottesman

, Wu

, Knopman

, Gross

, Mosley

Jr. , Selvin

, Windham

(2019) Systemic inflammation during midlife and cognitive change over 20 years: The ARIC Study. Neurology 92, e1256–e1267.

42.

Zhang

, Miller

, Madison

, Jin

, Honrada

, Harris

, Katz

, Forshew

, McGrath

(2013) Systemic immune system alterations in early stages of Alzheimer’s disease. J Neuroimmunol 256, 38–42.

43.

Holmes

, Cunningham

, Zotova

, Woolford

, Dean

, Kerr

, Culliford

, Perry

(2009) Systemic inflammation and disease progression in Alzheimer disease. Neurology 73, 768–774.

44.

Goldeck

, M. Witkowski

, Fülop

, Pawelec

(2016) Peripheral immune signatures in Alzheimer disease. Curr Alzheimer Res 13, 739–749.

45.

Naughton

, Duncan

, Murrey

, Meadows

, Newsom

, Stoicea

, White

, Scharre

, McCarty

, Fu

(2014) Blood genome-wide transcriptional profiles reflect broad molecular impairments and strong blood-brain links in Alzheimer’s disease. J Alzheimers Dis 43, 93–108.

46.

Lunnon

, Sattlecker

, Furney

, Coppola

, Simmons

, Proitsi

, Lupton

, Lourdusamy

, Johnston

, Soininen

(2013) A blood gene expression marker of early Alzheimer’s disease. J Alzheimers Dis 33, 737–753.

47.

Maes

, Schipper

, Chertkow

, Wang

(2009) Methodology for discovery of Alzheimer’s disease blood-based biomarkers. J Med Biomed Sci 64, 636–645.

48.

Liew

C-C

, Ma

, Tang

H-C

, Zheng

, Dempsey

(2006) The peripheral blood transcriptome dynamically reflects system wide biology: A potential diagnostic tool. J Lab Clin Med 147, 126–132.

49.

Achiron

, Gurevich

(2006) Peripheral blood gene expression signature mirrors central nervous system disease: The model of multiple sclerosis. Autoimmun Rev 5, 517–522.

50.

Sullivan

, Fan

, Perou

(2006) Evaluating the comparability of gene expression in blood and brain. Am J Med Genet B 141, 261–268.

51.

Maes

, Xu

, Yu

, Chertkow

, Wang

, Schipper

(2007) Transcriptional profiling of Alzheimer blood mononuclear cells by microarray. Neurobiol Aging 28, 1795–1809.

52.

Moreau

, Tranchevent

L-C

(2012) Computational tools for prioritizing candidate genes: Boosting disease gene discovery. Nat Rev Genet 13, 523–536.

53.

Battaglia

, Venturin

, Sojic

, Jesuthasan

, Orro

, Spinelli

, Musicco

, De Bellis

, Adorni

(2019) Candidate genes and MiRNAs linked to the inverse relationship between cancer and alzheimer’s disease: Insights from data mining and enrichment analysis. Front Genet 10, 846.

54.

Meng

, Mei

(2019) Transcriptional dysregulation study reveals a core network involving the progression of Alzheimer’s disease. Front Aging Neurosci 11, 101.

55.

Safari-Alighiarloo

, Taghizadeh

, Rezaei-Tavirani

, Goliaei

, Peyvandi

(2014) Protein-protein interaction networks (PPI) and complex diseases. Gastroenterol Hepatol Bed Bench 7, 17.

56.

American Psychiatric Association (2013) Diagnostic and statistical manual of mental disorders: DSM-5. American Psychiatric Association, Washington, DC.

57.

Foroughan

, Jafari

, Shirin Bayan

, Ghaem Magham Farahani

, Rahgozar

(2008) Validation of Mini-Mental State Examination (MMSE) in the elderly population of Tehran. Adv Cogn Psychol 10, 29–37.

58.

Shannon

, Markiel

, Ozier

, Baliga

, Wang

, Ramage

, Amin

, Schwikowski

, Ideker

(2003) Cytoscape: A software environment for integrated models of biomolecular interaction networks. Genome Res 13, 2498–2504.

59.

Bader

, Hogue

(2003) An automated method for finding molecular complexes in large protein interaction networks. BMC Bioinformatics 4, 2.

60.

Scardoni

, Tosadori

, Faizan

, Spoto

, Fabbri

, Laudanna

(2014) Biological network analysis with CentiScaPe: Centralities and experimental dataset integration. F1000Research 3, 139.

61.

Scardoni

, Petterlini

, Laudanna

(2009) Analyzing biological network parameters with CentiScaPe. Bioinformatics 25, 2857–2859.

62.

Hevey

(2018) Network analysis: A brief overview and tutorial. Health Psychol Behav Med 6, 301–328.

63.

Supek

, Bošnjak

, Škunca

, Šmuc

(2011) REVIGO summarizes and visualizes long lists of gene ontology terms. PLoS One 6, e21800.

64.

Allen

, Wynn

(2011) Evolution of Th2 immunity: A rapid repair response to tissue destructive pathogens. PLoS Pathog 7, e1002003.

65.

Kubick

, Flournoy

PCH

, Enciu

A-M

, Manda

, Mickael

M-E

(2020) Drugs modulating CD4+T cells blood–brain barrier interaction in Alzheimer’s disease. Pharmaceutics 12, 880.

66.

Kallies

, Good-Jacobson

(2017) Transcription factor T-bet orchestrates lineage development and function in the immune system. Trends Immunol 38, 287–297.

67.

Szabo

, Kim

, Costa

, Zhang

, Fathman

, Glimcher

(2000) A novel transcription factor, T-bet, directs Th1 lineage commitment. Cell 100, 655–669.

68.

Lazarevic

, Glimcher

, Lord

(2013) T-bet: A bridge between innate and adaptive immunity. Nat Rev Immunol 13, 777–789.

69.

Gocke

, Cravens

, Ben

L-H

, Hussain

, Northrop

, Racke

, Lovett-Racke

(2007) T-bet regulates the fate of Th1 and Th17 lymphocytes in autoimmunity. J Immunol 178, 1341–1348.

70.

Jenner

, Townsend

, Jackson

, Sun

, Bouwman

, Young

, Glimcher

, Lord

(2009) The transcription factors T-bet and GATA-3 control alternative pathways of T-cell differentiation through a shared set of target genes. Proc Natl Acad Sci U S A 106, 17876–17881.

71.

Zhang

, Zhang

, Gu

, Sun

(2014) TH1/TH2 cell differentiation and molecular signals In T helper cell differentiation and their function. Adv Exp Med Biol 841, 15–44.

72.

Mastrangelo

, Sudol

, Narrow

, Bowers

(2009) Interferon-γ differentially affects Alzheimer’s disease pathologies and induces neurogenesis in triple transgenic-AD mice. Am J Pathol 175, 2076–2088.

73.

Monteiro

, Ferreira

, Pinto

, Roque

, Morais

, de Sá-Calçada

, Mota

, Correia-Neves

, Cerqueira

(2016) Absence of IFNγ promotes hippocampal plasticity and enhances cognitive performance. Transl Psychiatry 6, e707.

74.

Bate

, Kempster

, Last

, Williams

(2006) Interferon-γ increases neuronal death in response to amyloid-β 1-42. J Neuroinflammation 3, 7.

75.

Browne

, McQuillan

, McManus

, O’Reilly

J-A

, Mills

, Lynch

(2013) IFN-γ production by amyloid β–specific Th1 cells promotes microglial activation and increases plaque burden in a mouse model of Alzheimer’s disease. J Immunol 190, 2241–2251.

76.

Zheng

, Zhou

X-W

, Wang

J-Z

(2016) The dual roles of cytokines in Alzheimer’s disease: Update on interleukins, TNF-α, TGF-β and IFN-γ. Transl Neurodegener 5, 7.

77.

Swardfager

, Lanctôt

, Rothenburg

, Wong

, Cappell

, Herrmann

(2010) A meta-analysis of cytokines in Alzheimer’s disease. Biol Psychiatry 68, 930–941.

78.

Asselineau

, Benlhassan

, Arosio

, Mari

, Ferri

, Casati

, Gussago

, Tedone

, Annoni

, Mazzola

(2015) Interleukin-10 production in response to amyloid-β differs between slow and fast decliners in patients with Alzheimer’s disease. J Alzheimers Dis 46, 837–842.

79.

Decourt

, K Lahiri

, N Sabbagh

(2017) Targeting tumor necrosis factor alpha for Alzheimer’s disease. Curr Alzheimer Res 14, 412–425.

80.

Pan

, Kastin

(1999) Penetration of neurotrophins and cytokines across the blood–brain/blood–spinal cord barrier. Adv Drug Del Rev 36, 291–298.

81.

Banks

, Kastin

, Broadwell

(1995) Passage of cytokines across the blood-brain barrier. Neuroimmunomodulation 2, 241–248.

82.

Banks

, Farr

, Morley

(2002) Entry of blood-borne cytokines into the central nervous system: Effects on cognitive processes. Neuroimmunomodulation 10, 319–327.

83.

Michaud

, Balardy

, Moulis

, Gaudin

, Peyrot

, Vellas

, Cesari

, Nourhashemi

(2013) Proinflammatory cytokines, aging, and age-related diseases. J Am Med Dir Assoc 14, 877–882.

84.

Tweedie

, Sambamurti

, Greig

(2007) TNF-α inhibition as a treatment strategy for neurodegenerative disorders: New drug candidates and targets. Curr Alzheimer Res 4, 378–385.

85.

Montgomery

, Bowers

(2012) Tumor necrosis factor-alpha and the roles it plays in homeostatic and degenerative processes within the central nervous system. J Neuroimmune Pharmacol 7, 42–59.

86.

Fischer

, Maier

(2015) Interrelation of oxidative stress and inflammation in neurodegenerative disease: Role of TNF. Oxid Med Cell Longev 2015, 610813.

87.

Clark

, Alleva

, Vissel

(2010) The roles of TNF in brain dysfunction and disease. Pharmacol Ther 128, 519–548.

88.

Locksley

, Killeen

, Lenardo

(2001) The TNF and TNF receptor superfamilies: Integrating mammalian biology. Cell 104, 487–501.

89.

Bradley

(2008) TNF-mediated inflammatory disease. J Pathol 214, 149–160.

90.

Janelsins

, Mastrangelo

, Park

, Sudol

, Narrow

, Oddo

, LaFerla

, Callahan

, Federoff

, Bowers

(2008) Chronic neuron-specific tumor necrosis factor-alpha expression enhances the local inflammatory environment ultimately leading to neuronal death in 3xTg-AD mice. Am J Pathol 173, 1768–1782.

91.

Tobinick

, Gross

, Weinberger

, Cohen

(2006) TNF-alpha modulation for treatment of Alzheimer’s disease: A 6-month pilot study. Medscape J Med 8, 25.

92.

Brosseron

, Krauthausen

, Kummer

, Heneka

(2014) Body fluid cytokine levels in mild cognitive impairment and Alzheimer’s disease: A comparative overview. Mol Neurobiol 50, 534–544.

93.

Kim

, Lee

, Kim

(2017) Serum tumour necrosis factor-α and interleukin-6 levels in Alzheimer’s disease and mild cognitive impairment. Psychogeriatrics 17, 224–230.

94.

Carito

, Ciafrh

, Tarani

, Ceccanti

, Natella

, Iannitelli

, Tirassa

, Chaldakov

, Ceccanti

, Boccardo

(2015) TNF-α and IL-10 modulation induced by polyphenols extracted by olive pomace in a mouse model of paw inflammation. Ann Ist Super Sanita 51, 382–386.

95.

Dardalhon

, Korn

, Kuchroo

, Anderson

(2008) Role of Th1 and Th17 cells in organ-specific autoimmunity. J Autoimmun 31, 252–256.

96.

Appel

(2009) CD4+T cells mediate cytotoxicity in neurodegenerative diseases. J Clin Investig 119, 13–15.

97.

Derecki

, Cardani

, Yang

, Quinnies

, Crihfield

, Lynch

, Kipnis

(2010) Regulation of learning and memory by meningeal immunity: A key role for IL-4. J Exp Med 207, 1067–1080.

98.

Shechter

, Miller

, Yovel

, Rosenzweig

, London

, Ruckh

, Kim

K-W

, Klein

, Kalchenko

, Bendel

(2013) Recruitment of beneficial M2 macrophages to injured spinal cord is orchestrated by remote brain choroid plexus. Immunity 38, 555–569.

99.

Bottero

, Potashkin

(2019) Meta-analysis of gene expression changes in the blood of patients with mild cognitive impairment and Alzheimer’s disease dementia. Int J Mol Sci 20, 5403.

100.

Gironi

, Borgiani

, Farina

, Mariani

, Cursano

, Alberoni

, Nemni

, Comi

, Buscema

, Furlan

, Grossi

(2014) A global immune deficit in Alzheimer’s disease and mild cognitive impairment disclosed by a novel data mining process. J Alzheimers Dis 43, 1199–1213.

101.

Qin

, Zhou

, Zuo

, Wang

, Cheng

, Jia

(2011) The study on the role of transcription factor GATA binding protein 3 in familial Alzheimer’s disease pathogenesis. Chin J Neurol 44, 159–162.

102.

Saresella

, Calabrese

, Marventano

, Piancone

, Gatti

, Alberoni

, Nemni

, Clerici

(2011) Increased activity of Th-17 and Th-9 lymphocytes and a skewing of the post-thymic differentiation pathway are seen in Alzheimer’s disease. Brain Behav Immun 25, 539–547.

103.

Sachin

, Gadani

, Cronk

, Norris

, Kipnis

(2012) Interleukin-4: A cytokine to remember. J Immunol 189, 4213–4421.

104.

Gordon

, Martinez

(2010) Alternative activation of macrophages: Mechanism and functions. Immunity 32, 593–604.

105.

Gadani

, Cronk

, Norris

, Kipnis

(2012) IL-4 in the brain: A cytokine to remember. J Immunol 189, 4213–4219.

106.

Liu

H-Y

, Tang

R-H

, Qi

R-Q

(2019) Interleukin-4 affects microglial autophagic flux. Neural Regen Res 14, 1594.

107.

Ribizzi

, Fiordoro

, Barocci

, Ferrari

, Megna

(2010) Cytokine polymorphisms and Alzheimer disease: Possible associations. Neurol Sci 31, 321–325.

108.

, Qian

, Teng

, Ding

, Zhang

(2014) Association of interleukin-4 genetic polymorphisms with sporadic Alzheimer’s disease in Chinese Han population. Neurosci Lett 563, 17–21.

109.

King

, O’Brien

, Donaghy

, Morris

, Barnett

, Olsen

, Martin-Ruiz

, Taylor

J-P

, Thomas

(2018) Peripheral inflammation in prodromal Alzheimer’s and Lewy body dementias. J Neurol Neurosurg Psychiatry 89, 339–345.

110.

Kaminuma

, Mori

, Kitamura

, Hashimoto

, Kitamura

, Inokuma

, Miyatake

(2005) Role of GATA-3 in IL-5 gene transcription by CD4+T cells of asthmatic patients. Int Arch Allergy Immunol 137, 55–59.

111.

Hoppenot

, Malakauskas

, Lavinskiene

, Sakalauskas

(2015) p-STAT6, PU. 1, and NF-κB are involved in allergen-induced late-phase airway inflammation in asthma patients. BMC Pulm Med 15, 122.

112.

Zhou

, Li

, Zheng

, Wang

(2017) IL-5 blocks apoptosis and tau hyperphosphorylation induced by Aβ 25–35 peptide in PC12 cells. J Physiol Biochem 73, 259–266.

113.

de la Monte

, Daiello

, Hapel

, Tong

, Ott

(2017) Altered serum and cerebrospinal fluid inflammatory cascades in mild cognitive impairment and Alzheimer’s disease. J Neuroinflamm Neurodegener Dis 1, 1–24.

114.

Zhu

, Jankovic

, Oler

, Wei

, Sharma

, Hu

, Guo

, Yagi

, Yamane

, Punkosdy

(2012) The transcription factor T-bet is induced by multiple pathways and prevents an endogenous Th2 cell program during Th1 cell responses. Immunity 37, 660–673.

115.

Kanduri

, Tripathi

, Larjo

, Mannerström

, Ullah

, Lund

, Hawkins

, Ren

, Lähdesmäki

, Lahesmaa

(2015) Identification of global regulators of T-helper cell lineage specification. Genome Med 7, 122–122.

116.

Arreola

, Becerril-Villanueva

, Cruz-Fuentes

, Velasco-Velázquez

, Garcés-Alvarez

, Hurtado-Alvarado

, Quintero-Fabian

, Pavón

(2015) Immunomodulatory effects mediated by serotonin. J Immunol Res 2015, 1–21.

117.

Neshan

, Campbell

, Malakouti

, Zareii

, Ahangari

(2020) Gene expression of serotonergic markers in peripheral blood mononuclear cells of patients with late-onset Alzheimer’s disease. Heliyon 6, e04716.

118.

Sarlus

, Heneka

(2017) Microglia in Alzheimer’s disease. J Clin Investig 127, 3240–3249.

119.

Wang

W-Y

, Tan

M-S

, Yu

J-T

, Tan

(2015) Role of pro-inflammatory cytokines released from microglia in Alzheimer’s disease. Ann Transl Med 3, 136.

120.

Walker

, Ficek

, Westbrook

(2019) Understanding the role of systemic inflammation in Alzheimer’s disease. ACS Chem Neurosci 10, 3340–3342.

121.

Cunningham

, Hennessy

(2015) Co-morbidity and systemic inflammation as drivers of cognitive decline: New experimental models adopting a broader paradigm in dementia research. Alzheimers Res Ther 7, 33.

122.

Town

, Vendrame

, Patel

, Poetter

, DelleDonne

, Mori

, Smeed

, Crawford

, Klein

, Tan

(2002) Reduced Th1 and enhanced Th2 immunity after immunization with Alzheimer’s β-amyloid1–42. J Neuroimmunol 132, 49–59.

123.

Maier

, Duschl

, Horejs-Hoeck

(2012) STAT6-dependent and-independent mechanisms in T h2 polarization. Eur J Immunol 42, 2827–2833.

124.

Zhu

, Min

, Hu-Li

, Watson

, Grinberg

, Wang

, Killeen

, Urban

, Guo

, Paul

(2004) Conditional deletion of Gata3 shows its essential function in TH 1-TH 2 responses. Nat Immunol 5, 1157–1165.

125.

Lee

, Kim

N-J

(2017) Recent advances in the inhibition of p38 MAPK as a potential strategy for the treatment of Alzheimer’s disease. Molecules 22, 1287.

126.

Shal

, Ding

, Ali

, Kim

, Khan

(2018) Anti-neuroinflammatory potential of natural products in attenuation of Alzheimer’s disease. Front Pharmacol 9, 548.

127.

Franco

, Martínez-Pinilla

, Navarro

, Zamarbide

(2017) Potential of GPCRs to modulate MAPK and mTOR pathways in Alzheimer’s disease. Prog Neurobiol 149, 21–38.

128.

Mehan

, Meena

, Sharma

, Sankhla

(2011) JNK: A stress-activated protein kinase therapeutic strategies and involvement in Alzheimer’s and various neurodegenerative abnormalities. J Mol Neurosci 43, 376–390.

129.

Yamamoto

, Fujii

, Kasahara

, Tanida

, Ohora

, Ono

, Suzuki

, Sobue

(2016) Simvastatin and atorvastatin facilitates amyloid β-protein degradation in extracellular spaces by increasing neprilysin secretion from astrocytes through activation of MAPK/E rk1/2 pathways. Glia 64, 952–962.

130.

Yang

, Liu

, Zhou

, Bu

, Zhang

(2017) Andrographolide attenuates microglia-mediated Aβ neurotoxicity partially through inhibiting NF-κB and JNK MAPK signaling pathway. Immunopharmacol Immunotoxicol 39, 276–284.

131.

Wang

, Wu

, Ma

, Xie

, Liu

, Yao

, Zhang

, Yang

, Guo

(2020) Resveratrol preincubation enhances the therapeutic efficacy of hUC-MSCs by improving cell migration and modulating neuroinflammation mediated by MAPK signaling in a mouse model of Alzheimer’s disease. Front Cell Neurosci 14, 62.

132.

Cui

, Wu

, Mou

, Xie

, Wang

, Fang

, Xu

, Dong

, Liu

(2018) Exosomes derived from hypoxia-preconditioned mesenchymal stromal cells ameliorate cognitive decline by rescuing synaptic dysfunction and regulating inflammatory responses in APP/PS1 mice. FASEB J 32, 654–668.

133.

Meng

, Zhong

, Mei

(2016) A systematic investigation into aging related genes in brain and their relationship with alzheimer’s disease. PLoS One 11, e0150624.

134.

Kohu

, Ohmori

, Wong

, Onda

, Wakoh

, Kon

, Yamashita

, Nakayama

, Kubo

, Satake

(2009) The Runx3 transcription factor augments Th1 and down-modulates Th2 phenotypes by interacting with and attenuating GATA3. J Immunol 183, 7817–7824.

135.

Yagi

, Junttila

, Wei

, Urban

Jr , Zhao

, Paul

, Zhu

(2010) The transcription factor GATA3 actively represses RUNX3 protein-regulated production of interferon-γ. Immunity 32, 507–517.

136.

Oberstein

, Taha

, Spitzer

, Hellstern

, Herrmann

, Kornhuber

, Maler

(2018) Imbalance of circulating Th17 and regulatory T cells in Alzheimer’s disease: A case control study. Front Immunol 9, 1213.