TREM2 and CD163 Ameliorate Microglia-Mediated Inflammatory Environment in the Aging Brain

Aging decreases cognitive functions, especially learning and memory. Neuroinflammation is mediated by microglia and occurs in age-related neurodegenerative diseases. The expression profiles in a dataset of cognitively normal controls (GSE11882) were obtained from the Gene Expression Omnibus (GEO) database. Microarray data were used to explore the expression of age-related genes in the human hippocampus. A total of 120 differentially expressed genes (DEGs) were identified and subjected to Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses. A protein–protein interaction (PPI) network was constructed. A total of 18 key genes were identified by the plugin cytoHubba in Cytoscape software. Two genes with a positive impact on cognition during aging were teased out: triggering receptor expressed on myeloid cells 2 (TREM2) and a scavenger receptor (CD163). Finally, the results of reverse transcription–quantitative polymerase chain reaction (RT-qPCR) and western blotting (WB) verified that the mRNA expression of these two genes was significantly upregulated in aged mice. Moreover, the levels of the inflammatory factors IL-1β and IL-6 were significantly increased. TREM2 and CD163 may be upregulated to alleviate the inflammatory environment resulting from microglial activation in the aging brain, thereby delaying cognitive decline.


Introduction
Aging leads to physiological neurodegeneration in the brain. The features of physiological neurodegeneration, such as the accumulation of phospho-Tau and amyloid peptides, the release of proinflammatory mediators and the presence of oxidative stress (Mecca et al. 2018;Teissier et al. 2020), can also be observed in the early stage of pathological neurodegeneration, such as that occurring in Alzheimer's disease (AD). Studies have found that the incidence of AD increases sharply with age (Schrijvers et al. 2012;Rocca et al. 2011). Therefore, efficient tools should be developed to diagnose pathological aging in the brain.
The human cerebral cortex undergoes significant changes between ages 60 and 70, a critical period for brain degeneration (Berchtold et al. 2008). In the case of cognitive dysfunction, age-related genes are dysregulated in the frontal and temporal regions. Previous studies have focused on those genes in the prefrontal/frontal cortex (Irizarry et al. 2003). However, the expression profiles of these genes in other regions of the human brain remain unclear. The hippocampus is responsible for cognitive functions, such as spatial learning and memory (Joaqu¨ªn et al. 2017). Therefore, characterizing the expression of age-related genes in the hippocampus may provide new information about neurodegenerative diseases.
Bioinformatics approaches have been widely used to analyze the pathogenesis of diseases. Berchtold NC et al. used microarray analysis to obtain RNA expression profiles of genes in different brain regions in postmortem samples from young and aged humans with normal brain function (Berchtold et al. 2008). In the present study, we used microarray data to assess the expression of age-related genes in the hippocampal region in humans.

Microarray Data
The expression profiles in the GSE11882 dataset were obtained from the Gene Expression Omnibus (GEO) database (https:// www. ncbi. nlm. gov/ geo/). The microarray consisted of postmortem brain samples from young and aged individuals in the Alzheimer's Disease Research Center (ADRC) brain banks. We assessed gene expression profiles in the hippocampus in 43 cognitively normal individuals aged 20-99 years. The Young group (20-59 years) contained 18 samples, and the Aged group (60-99 years) contained 25 samples.

Data Preprocessing
The R package Affy was used to read the raw data, and a robust multichip average (RMA) algorithm was used to normalize the data between samples (Irizarry et al. 2003). Then, the annotation package in R was used to convert probe names to gene names. When a gene corresponded to multiple probes, the median expression value of the multiple probe IDs was calculated.

Identification of DEGs
The limma (Ritchie et al. 2015) software package in R was utilized to screen the differentially expressed genes (DEGs) between the two groups. The DEGs were defined as genes with p < 0.01, q < 0.05, and log 2 |fold change| > 0.9. Then, a volcano plot and heatmap were generated in R (http:// cran.r-proje ct. org/ packa ge).

Enrichment Analysis of DEGs
Gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed with the Database for Annotation, Visualization and Integrated Discovery (DAVID; version 6.8) (https:// david. ncifc rf. gov/) tool. p < 0.05 and enrichment count >2 were set as the significance thresholds, and the enrichment results are displayed in tables.

Analysis of the PPI Network
PPI network analysis of the DEGs was performed with the Search Tool for the Retrieval of Interacting Genes (STRING; version 11.0) database (https:// string-db. org/) (Szklarczyk et al. 2017). Then, Cytoscape software (Shannon et al. 2003) was used to visualize the network. The key genes were identified with the plugin cytoHubba. The top 30 genes were ranked by intersecting the results of four algorithms: maximal clique centrality (MCC), maximum neighborhood component (MNC), density of maximum neighborhood component (DMNC), and degree (Lin et al. 2008;Rajendran and Paolicelli 2018). Through the intersection of four algorithms, 18 genes were obtained. Then, of which the key genes were identified to have compensatory effects on neuroinflammation.

Animal Experiments
Male C57BL/6J (weight, 20-25 g and 40-45 g; n = 8) mice were assigned to two groups-the Young (3 months) group and the Aged (18 months) group, and housed under constant environmental conditions (a temperature of 22 ± 2 °C and a humidity of 50-70%) on a light/dark cycle of 12 h, with 2 weeks of adaptation and free access to food and water. All animal experiments were approved by the Experimental Animals Welfare and Ethical Inspection of Nanjing Drum Tower Hospital (Nanjing, China) (No. 2019AE01077).

RNA Extraction
The mice in the two groups were decapitated, and the brain hemispheres were dissected along the midline for collection of hippocampal tissue. Total RNA was extracted from the hippocampus using an RNA extraction kit (BioTeke, Beijing, China). A NanoDrop UV spectrophotometer (NanoDrop Technologies; Thermo Fisher Scientific, Inc., Waltham, MA, USA) was used to determine the RNA concentration.

Statistical Analysis
All experimental data are expressed as the mean ± standard deviation values. SPSS software version 24.0 (IBM Corporation, New York, USA) was used to perform t tests. The number of experimental animals is indicated by "n." Differences were considered statistically significant at p < 0.05, and GraphPad Prism 7.0 was used to generate the graphs.

Identification of DEGs
With the criterial of a p value < 0.01, adjusted p (q) value < 0.05 and log 2 | fold change | > 0.9, a total of 120 DEGs between the two groups were identified, among which 116 were upregulated and 4 were downregulated (NECAB1, THEMIS, TRIM23, CRH), as shown in the volcano plot ( Fig. 1a). A heatmap was generated for cluster analysis (Fig. 1b).

GO and KEGG Pathway Analyses
Gene Ontology (GO) ( Table 1) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses (Table 2) were performed. The DEGs were mainly enriched in the terms plasma membrane (GO:0005886) and extracellular space (GO:0005615) in the cellular component (CC) category. In the biological process (BP) category, the DEGs were mainly enriched in the terms regulation of inflammatory response (GO:0006954), phagocytosis (GO:0006909), and immune response (GO:0006955). KEGG pathway enrichment analysis showed that the DEGs were enriched in the phagosome (hsa04145), complement and coagulation cascades (hsa04610) and cell adhesion molecules (CAMs) (hsa04514) pathways.

PPI Network of DEGs
In our study, the PPI network was constructed based on the STRING database. After deleting unconnected nodes in the network, clusters containing a total of 97 genes were visualized with the MCODE plugin in Cytoscape (Fig. 2).

Identification of Key Genes
In Cytoscape, the 30 most enriched genes were ranked with four algorithms (MCC, DMNC, MNC, and degree), and an intersection was considered to indicate a key gene (Table 3 and Fig. 3a). Finally, 18 genes were obtained. As Table 4 indicates, LY86, C3AR1, and CD14 were found to be related to the inflammatory response, and HLA-DRA, HLA-DPA1, C1QC, and CD74 were related to the immune response (Table 4 and Fig. 3b). Interestingly, two of the key genes, TREM2 (triggering receptor expressed on myeloid cells 2) and CD163 (a scavenger receptor), can ameliorate inflammation in the normal aging brain.

RT-qPCR and WB Validation
RT-qPCR and WB were used to verify the reliability and accuracy of the bioinformatics analysis results. The expression of TREM2, CD163 and the inflammatory cytokines IL-1β and IL-6 in the Aged group was significantly upregulated compared with that in the Young group (p < 0.01, p < 0.05) (Figs. 4 and 5).

Discussion
Cognitive function declines as the brain ages, especially in the hippocampus. In age-related diseases, neurogenesis is inhibited in the hippocampal dentate gyrus, and microglia are activated (Morel et al. 2015). Furthermore, aging exerts negative effects on neurite growth and neuronal function by dysregulating the expression of genes in the hippocampus (Yan et al. 2015). In the present study, we explored the expression of and functional changes in age-dependent genes in the hippocampus. Bioinformatics analysis indicated that TREM2 and CD163 were expressed in an age-dependent manner. The results of RT-qPCR and WB were consistent with those of the bioinformatics analysis.
We found a wide distribution of gene expression levels as calculated by the binary logarithm of the fold change (log 2 |FC|). Based on the microarray data, we identified 120 DEGs between the two groups, with 4  downregulated and 116 upregulated, as shown in the volcano plot (Fig. 1a) and heatmap (Fig. 1b). Gene ontology (GO) enrichment analysis indicated that these DEGs were mainly enriched in the terms plasma membrane and cell surface in the cellular component (CC) category, as well as the terms immune response and inflammatory response in the biological process (BP) category (Table 1), consistent with the results of Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis (Table 2). A PPI network was constructed using the STRING database and visualized with Cytoscape software (Fig. 2). Among the 18 hub genes obtained by intersecting the results of the four algorithms in the Cytoscape plugin MCODE (Table 3 and Fig. 3), most of the upregulated genes were related to immune and inflammatory responses, a result that was further validated by the high levels of IL-1β and IL-6 in the aging brain (Fig. 5a, d, e). TREM2 and CD163 upregulation in microglia can promote phagocytosis and the release of anti-inflammatory mediators from microglia. Therefore, we hypothesized that TREM2 and CD163 may counter microglia-mediated neuroinflammation and thus mitigate age-related cognitive decline. Our RT-qPCR and western blot experiments also confirmed the upregulation of TREM2 and CD163 and the high levels of IL-1β and IL-6 in the hippocampus in aged mice (Figs. 4 and 5).
The aged brain exhibits accumulation of endogenous factors (such as fibrillar Aβ and proinflammatory cytokines) and upregulation of complement components, inflammasome components and Toll-like receptors (David et al. 2012). Microglia in aging mice and mice with neurodegeneration show loss of the homeostatic phenotype, which protects against neurodegenerative disorders by phagocytosing debris, removing protein aggregates, and assisting neural repair (Puigdell¨ªvol et al. 2020). Microglial activation is a hallmark of the aging brain; however, the neuroprotective effects of the homeostatic phenotype are impaired (Soto et al. 2015;Michelle et al. 2007). Microglia show a moderately activated phenotype in the normal human brain. In addition, reactive microglia are indispensable for neuroinflammation in AD, posing a high risk of cognitive decline (Azam et al. 2021;Hui-Ming et al. 2003;Mandrekar-Colucci and Landreth 2010).
TREM2, an innate immune receptor in the immunoglobulin superfamily, is mainly expressed in myeloid cells and uniquely expressed by microglia in the central nervous system (Frank et al. 2008). The cytoplasmic portion of TREM2 cooperates with DNAX-activating protein of 12 kDa (DAP12, also called TYROBP), which is a type I transmembrane adaptor protein, forming a molecular complex for its signaling and functions (Daws et al. 2001;Fig. 3 Venn diagrams for the identification of key genes. (a) Eighteen key genes were identified by intersecting the results of four algorithms. (b) Intersection of key genes and immune-inflammatory response BP terms to identify immuneinflammation-related genes  Humphrey et al. 2006). Studies have found that activated TREM2 can regulate the functions of microglia, including stimulating phagocytosis and suppressing cytokine production (Takahashi et al. 2005). Recent evidence has also proven the crucial role of TREM2 in maintaining the expansion and survival of microglia (Kawabori et al. 2015;Zhong et al. 2015). Previous genome-wide association studies have shown that a rare mutation (R47H) in TREM2 with lossof-function effects is a known risk factor for AD (Colonna and Wang 2016;Guerreiro et al. 2013;Tanzi 2015); in addition, silencing of TREM2 in microglia impairs their capacity to phagocytize cell membrane debris and increases the production of proinflammatory cytokines (Hsieh et al. 2009;Kleinberger et al. 2014). Previous research has found that the expression levels of TREM2 in the brains of AD patients and AD mice are positively correlated with age and are higher than those in the brains of normal controls of the same age (Zhao 2019;Piccio et al. 2016).
In the present study, we found that the expression of TREM2 was markedly upregulated in the hippocampus and positively correlated with age ( Figs. 4a and 5a, b). Animal studies have shown that soluble TREM2 levels in the brain increase in parallel with amyloidosis and microglial activation during aging (Brendel et al. 2017). Previous studies Fig. 4 The key genes are expressed in both young and aged mice. The mRNA expression of TREM2 and CD163 was detected by reverse transcription-quantitative polymerase chain reaction. TREM2: Comparison between the Young group and Aged group, (a) p = 0.000; CD163: Comparison between the Young group and Aged group, (b) p = 0.000. *p < 0.05, **p < 0.01, ***p < 0.001 versus the Young group. The results are presented as the mean ± SD values Fig. 5 (a) The expression of the key genes TREM2 and CD163 and the inflammatory factors IL-1β and IL-6 were assessed through western blot analysis and were significantly upregulated in the Aged group compared with the Young group. (b) p = 0.008, (c) p = 0.000), (d) p = 0.002, (e) p = 0.000. *p < 0.05, **p < 0.01, ***p < 0.001 versus the Young group. The results are presented as the mean ± SD values have demonstrated that microglia-induced neuroinflammation is predominant in the white matter of aging mice and humans as well as in the early-onset AD (EOAD) brain (Raj et al. 2017). TREM2 has emerged as a critical factor in regulating the activity of microglia and the polarization from the proinflammatory M1 toward the anti-inflammatory M2 microglial phenotype (Garcia-Revilla et al. 2019). In addition, TREM2 is involved in neuroprotection mediated by the disease-associated microglia (DAM) phenotype (Deczkowska et al. 2018), which mediates the clearance of misfolded and aggregated proteins in neurodegenerative diseases and during general aging (Keren-Shaul et al. 2017). Defective TREM2 function affects the microglial response to Aβ plaques, exacerbating tissue damage, whereas TREM2 overexpression attenuates this pathology in AD mouse models of amyloid β (Aβ) accumulation (Wang et al. 2020). The current consensus suggests that low-grade chronic neuroinflammation in the aging brain causes the risk of AD. Here, we confirmed the presence of increased expression of TREM2 in the aging brain and revealed differential expression of inflammatory factors between the Aged and Young groups. Our results indicated that TREM2 may be involved in delaying cognitive function decline and progression to AD by alleviating microglia-mediated inflammatory responses in the aging brain.
However, some studies have arrived at unexpected conclusions. One study suggested that knocking down astrocytic TREM2 may have beneficial effects on learning and memory abilities in elderly mice (Wei et al. 2021). A possible interpretation is that TREM2 expressed on microglia and astrocytes may play different regulatory roles. Another study revealed that loss of TREM2 affects neuronal structure and confers resilience to age-related synaptic and cognitive impairment during nonpathogenic aging. It is possible that dysregulation of microglial phagocytosis during aging contributes to age-related loss of synapses and that TREM2 deficiency exerts protective effects by ameliorating microglial phagocytosis to reduce synaptic pruning (Qu and Li 2021). Clearly, future studies using different conditional TREM2 models may provide deeper insight into the temporal role of TREM2 during aging.
CD163, a member of the group B scavenger receptor cysteine-rich (SRCR) family, is expressed in macrophages, monocytes and microglia (Fabriek et al. 2005;Moestrup and Moller 2004). CD163 has homeostatic capacity, and soluble CD163 generated by shedding of the outer domain can exert anti-inflammatory effects (Pey et al. 2014). Endogenous proinflammatory cytokines, such as IFN-γ and tumor necrosis factor-alpha (TNF-α), that regulate classical activation of macrophages/microglia, can decrease the expression of CD163, while glucocorticoids, IL-6 and IL-10 can increase CD163 expression (Christa et al. 2000). In recent years, CD163 has been considered a specific marker with strong anti-inflammatory properties in microglia (Gorp et al. 2010), monocytes and macrophages. CD163 is the hemoglobin-haptoglobin (HbHp) complex receptor, whereas its metabolites, such as bilirubin, free iron, and CO released by HbHp degradation upon the enzymatic activity of HO-1, have strong antioxidative and anti-inflammatory effects (Otterbein et al. 2003;Soares Miguel and Bach Fritz 2009). CD163 not only acts as an anti-inflammatory marker but also binds to its ligands to transduce signals, leading to the release of anti-inflammatory mediators, such as interleukin-10 (IL-10). IL-10 further upregulates the expression of CD163 and HO-1 in an autocrine or paracrine manner, forming a positive feedback loop involved in the clearance of hemoglobin (Hb) and in preventing extracellular Hb from triggering inflammatory responses. In addition, the increase in CD163-immunoreactive microglia is a specific immune response to AD neuropathology. Upregulation of CD163 and its positive feedback loop with IL-10 are prominent in aged people, suggesting that the protective effect of CD163 against Hb-induced inflammation is associated not only with the scavenging of HbHp complexes from the extracellular milieu but also with the active release of IL-10 and heme metabolites. Our research data showed that the expression of CD163 in the hippocampus increased in an age-dependent manner (Figs. 4b and 5a, c), while microglial polarization is a key regulator of aging-induced chronic inflammation in the brain. As a marker of M2 microglia, upregulation of CD163 in the aged brain helps promote the polarization of microglia into an anti-inflammatory phenotype.
Consistent with the changes in inflammatory factors, the expression of TREM2 and CD163 on microglia increased in an age-dependent manner, suggesting that they may ameliorate the inflammatory environment resulting from activation of hippocampal microglia in the aged brain through regulating microglial phagocytosis and anti-inflammatory abilities, thereby delaying age-related cognitive decline. The molecular mechanisms of TREM2 and CD163 in regulating the microglia-mediated inflammatory milieu should be further explored to ensure that appropriate interventions can be designed to counter age-related neuroinflammation and reverse age-related cognitive degeneration.