Introduction
Aging robustly affects the bidirectional communication between the brain and immune system [
14,
35]. This essential communication involves microglia and astrocytes, which interpret inflammatory signals from the periphery and propagate them within the central nervous system (CNS) [
17,
29,
44,
46]. Moreover, central inflammatory signaling is critical for normal physiological and behavioral responses to infection [
16]. With aging, this altered neuro-immune communication results in heightened risk of mortality and co-morbidity of depression or dementia [
34,
36,
51,
53,
58]. For instance, acute bacterial infection in elderly patients often presents as acute cognitive impairment and altered mood [
2,
19]. Moreover, these individuals are at an increased risk for progressive dementia and cognitive impairment even after the infection resolves [
33]. These data are consistent with rodent studies showing acute immune challenge triggers prolonged neuroinflammatory responses, altering affective behavior and cognition [
17,
28]. For instance, immune challenge by lipopolysaccharide (LPS) or
Escherichia coli in aged rodents induces elevated neuroinflammation, prolonged sickness behavior, and acute cognitive impairment, which are attributed to activation of microglia and astrocytes [
5,
6,
26,
27,
73]. In humans, these infection-related neurological and psychiatric complications reduce both quality of life and life expectancy [
51,
52,
62,
67,
69]. Therefore, understanding how aging impacts glial interactions in the brain and thereby leads to cognitive impairment is of paramount importance.
There is evidence that microglia and astrocytes develop a more pro-inflammatory or “primed” profile as a result of normal aging [
47]. For instance, microglia in the aged brain have increased expression of several inflammatory markers, including major histocompatibility complex (MHC) class II proteins, and adopt a de-ramified morphology with thicker processes [
13,
24,
29,
32,
56,
61,
63,
64,
68]. Additionally, astrocytes in the aged brain have increased baseline levels of glial fibrillary acidic protein (GFAP) and vimentin, both of which indicate increased reactivity [
15,
26,
42]. While the presence of these primed glia is insufficient to induce cognitive dysfunction, primed glia mediate exaggerated and prolonged neuroinflammatory responses to peripheral immune challenge. This hyper-inflammatory response in the CNS is not mirrored by the peripheral innate immune response, which is intact in aged animals [
4,
12,
14,
26,
29,
73]. Indeed, when the CNS is stimulated directly with intracerebroventricular (i.c.v.) LPS or gp120, aged mice still exhibit an exaggerated and prolonged sickness responses [
1,
31]. Thus, aged glia adopt a primed profile with age, leaving the elderly susceptible to hyper-inflammatory CNS reactions to acute peripheral stimuli.
Recent studies show that microglia can be depleted from the rodent CNS through colony-stimulating factor 1 receptor (CSF1R) antagonism without significant complications [
20,
55]. Moreover, cessation of this antagonism results in rapid microglial repopulation. Rice et al. (2017) used this approach to promote microglial turnover following inducible hippocampal neuron death and found microglial depletion and repopulation following hippocampal lesion ameliorated chronic microgliosis, leukocyte infiltration, and inflammatory gene expression [
55]. Furthermore, this was associated with improved cognitive and behavioral recovery. Recently, Elmore et al. (2018) found that depletion and repopulation of microglia in aged mice restored age-associated changes in microglial morphology [
21]. This was associated with a reversal of age-associated hippocampal dendritic spine loss and cognitive decline. Thus, depletion and repopulation of microglia may present a therapeutic strategy for redirecting chronic microglia-mediated inflammation.
The purpose of this study was to determine the degree to which CSF1R antagonist-mediated depletion of microglia in the aged brain would result in repopulation with new and unprimed microglia. Here, we provide novel evidence that promoting forced turnover of aged microglia reduced intracellular accumulation of lipofuscin and restored lysosome size to adult levels. While repopulated microglia in the aged brain had an intermediate RNA signature compared to aged controls, they remained primed to peripheral immune challenge and were hyper-inflammatory when activated. Moreover, age-associated reactive astrogliosis persisted independent of microglial turnover and ex vivo data shows the aged CNS microenvironment promotes microglial priming in neonatal microglia.
Materials & methods
Mice and PLX5622 administration
All procedures were performed in accordance with the National Institute of Health Guide for the Care and Use of Laboratory Animals and were approved by The Ohio State University Institutional Animal Care and Use Committee. PLX5622 was provided by Plexxikon (Berkeley, CA) and formulated in standard rodent chow by Research Diets (New Brunswick, NJ) at 1200 mg PLX5622/kg chow. Adult (6–8 weeks old) and aged (16–18 months old) male BALB/c mice from Charles River (Wilmington, MA) were provided vehicle or PLX5622 chow ad libitum for 21 d, after which PLX5622 treatment was withdrawn for 21 d to allow microglia to fully repopulate the CNS. Behavioral testing was performed immediately following this repopulation period.
Percoll-enrichment of microglia
Microglia were isolated from brain homogenates using a Percoll density gradient as previously described [
48]. In brief, mice were sacrificed by CO
2 asphyxiation. Brains were collected after decapitation and homogenized in ice-cold phosphate-buffered saline (PBS) using a 10 mL Potter-Elvehjem tissue grinder (Wheaton) and the cell pellet resuspended in 70% isotonic Percoll (GE Healthcare). A discontinuous Percoll density gradient was layered and centrifuged at 2000×
g for 20 min. Enriched microglia were collected from the interface between the 70 and 50% Percoll layers. Of the cells collected from this interface, > 80% of the cells were CD11b
+/CD45
low microglia.
Microglial lipofuscin detection by flow cytometry
Percoll-enriched microglia were incubated with anti-mouse CD11b-PE-Cyanine7 and CD45-PerCP-Cyanine5.5 antibodies (1:50; eBioscience). Expression was determined using a BD FACSCalibur cytometer. Microglia were identified by CD11b+/CD45low expression. Lipofuscin was detected at 488 nm excitation and 515–545 nm emission. Flow data were analyzed using FlowJo software (Tree Star).
NanoString nCounter analysis of mRNA copy number
Mice were sacrificed by CO
2 asphyxiation and the hippocampus was dissected and snap frozen in liquid nitrogen (− 196 °C). Hippocampal RNA was isolated using the Tri-Reagent protocol (Sigma-Aldrich). RNA quality and integrity was determined using the Agilent 2200 TapeStation assay (Agilent Technologies). nCounter analysis (NanoString Technologies) was performed by the OSU Comprehensive Cancer Center (OSUCCC) Genomics Shared Resource facility (The Ohio State University, Columbus, OH) using the Mouse Inflammation v2 Panel for 248 inflammation-related mouse genes, 20 custom genes, and 6 internal reference controls. Copy numbers were normalized using DESeq2 Bioconductor package in R [
41].
Immunohistochemistry for Iba1 and GFAP
Mice were sacrificed by CO2 asphyxiation and transcardially perfused with ice-cold PBS (pH 7.4) and 4% paraformaldehyde (PFA). Brains were collected and post-fixed in 4% PFA for 24 h, cryoprotected in 30% sucrose for 24 h, frozen using dry ice-cooled isopentane (− 78 °C), sectioned coronally at 30 μm on a Leica CM1800 cryostat, and stored in cryoprotectant (30% ethylene glycol, 30% polyethylene glycol, 40% 0.2 M phosphate buffer) for immunolabeling. Next, sections (Bregma - 1.5 mm) were washed in PBS, blocked (5% normal donkey serum, 0.1% Triton X-100, 1% bovine serum albumin in PBS) for 1 h, and incubated with rabbit anti-mouse Iba1 (1:1000; Wako Chemicals) or rabbit anti-mouse GFAP antibody (1:500; Abcam) overnight at 4 °C. Next, sections were washed in PBS and incubated with a fluorochrome-conjugated secondary antibody (Alexa Fluor 594 or Alexa Fluor 488). Sections were mounted on slides and cover-slipped with Fluoromount-G (Invitrogen). Slides were then imaged using a Leica DM5000 B epifluorescent microscope at 20X magnification and captured using a Leica DFC300 FX camera and imaging software. For each animal, 2–4 images were quantified and averaged, and these values were used to calculate the mean and standard error for each experimental group.
Immunohistochemistry for lipofuscin and CD68
To visualize microglia and neurons, sections (Bregma - 1.5 mm) were incubated with rabbit anti-Iba1 (1:1000; Wako) or rabbit anti-NeuN (1:1000; Millipore) primary antibody, respectively, followed by Alexa Fluor 647 donkey anti-rabbit IgG secondary antibody (1:500, Invitrogen). In order to visualize microglial lysosomes, sections were incubated with rabbit anti-Iba1 (1:1000; Wako) and rat anti-CD68 (1:500; Abcam) primary antibodies, Alexa Fluor 594 donkey anti-rabbit IgG and Alexa Fluor 647 donkey anti-rat IgG secondary antibodies (1:500, Invitrogen), and 0.1% Sudan Black B (Sigma-Aldrich) solution in 70% ethanol for 2 min prior to cover-slipping. Slides were then imaged using a Leica SP8 upright confocal microscope at 63X magnification and sequential optical sections captured using the Leica Application Suite X imaging software. Lipofuscin was imaged at 488 nm excitation and 495–545 nm emission. Sequential optical sections were analyzed using ImageJ software (NIH). For each animal, 2–4 images were quantified and averaged, and these values were used to calculate the mean and standard error for each experimental group.
RNA sequencing of sorted microglia and coronal brain section
Percoll-enriched microglia were sorted using a Becton-Dickinson FACSAria III cell sorter at the OSUCCC Analytical Cytometry facility. Microglia were identified by CD11b+/CD45low expression. Cells were pelleted and lysed in Arcturus PicoPure Extraction Buffer immediately after sorting. The Arcturus PicoPure RNA Isolation Kit (Applied Biosystems) was used to purify and concentrate total RNA from sorted microglia. During RNA isolation, samples were treated with on-column DNase digestion for 15 min at 23 °C to eliminate contaminating genomic DNA. A 1-mm coronal brain section was also collected from each brain and snap frozen in liquid nitrogen. RNA was isolated using the Tri-Reagent protocol (Sigma-Aldrich). RNA quality and integrity was determined using the Agilent 2200 TapeStation assay (Agilent Technologies). RNA-Seq was performed on sorted microglia and brain section (Bregma - 1.5 mm) RNA at the Hussman Institute for Human Genomics Sequencing Core Facility (University of Miami, Miami, FL). Briefly, RNA-Seq libraries were prepared using the Ovation SoLo RNA-Seq System with AnyDeplete rRNA to remove rRNA and other abundant transcripts according to the manufacturer’s recommendation (Nugen). RNA-Seq libraries were run on an Illumina NextSeq 500 sequencing instrument according to the protocols described by the manufacturer.
Differential gene expression and pathway analysis of microglial and whole-brain RNA sequencing
FASTQ files were aligned to the mouse mm10 genome using STAR Aligner [
18]. Raw counts were normalized and differentially expressed genes were determined using the DESeq2 Bioconductor package in R [
41]. Factors of unwanted variation were determined with RUVseq and included in the model for differential expression [
57].
P-values were adjusted for false discovery rate using the Benjamini-Hochberg procedure [
7]. To determine reversal of age effects by microglial repopulation, differentially expressed genes between Aged Control and Adult Control microglia (
Padj < 0.05; absolute fold change > 1.5) and brain section (
Padj < 0.05) samples were determined. Next, these genes were determined to be exacerbated (Aged Repopulation vs. Aged Control:
P < 0.05; conserved directionality), partially reversed (Aged Repopulation vs. Aged Control:
P < 0.05; Aged Repopulation vs. Adult Control:
P < 0.05), or reversed (Aged Repopulation vs. Aged Control:
P < 0.05; Aged Repopulation vs. Adult Control:
P ≥ 0.05) by microglial repopulation. Pathway analysis was performed using Ingenuity Pathway Analysis (IPA; QIAGEN) [
37] and Protein Analysis through Evolutionary Relationships (PANTHER) gene list analysis [
66] on genes with baseMean > 10,
Padj < 0.05, and absolute fold change > 1.5. Heat maps were generated and normalized by gene using ‘pheatmap’ in R. The normalized RNA sequencing data is supplied (Additional file
1).
Immune challenge with lipopolysaccharide
Adult and aged BALB/c mice received a single intraperitoneal (i.p.) injection of saline or
Escherichia coli LPS (0.33 mg/kg; serotype 0127:B8; Sigma-Aldrich). This LPS dosage was selected because it elicits a pro-inflammatory cytokine response in the brain resulting in a transient sickness response in adult BALB/c mice without mortality in aged mice [
8,
26,
72].
Social exploratory behavior
Social exploration was determined as a measure of sickness behavior as previously described [
26]. In brief, mice were injected (
t = 0) intraperitoneally with saline or 0.33 mg/kg LPS. At
t = 4 and 24 h, a novel male juvenile C57BL/6 mouse was introduced into the test subject’s home cage for 5 min. Behavior was recorded and the total duration of time the experimental subject engaged in social investigation of the juvenile (e.g. anogenital sniffing, trailing) was determined.
Gene expression by qPCR
RNA was isolated from a 1-mm coronal brain section (Bregma - 1.5 mm) using the Tri-Reagent protocol (Sigma-Aldrich). Reverse transcription was performed using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems) to produce cDNA. Quantitative real-time (q)-PCR was performed using the TaqMan Gene Expression Assay (Applied Biosystems). In brief, experimental cDNA was amplified using qPCR such that a target gene and reference gene (Gapdh) were amplified simultaneously using an oligonucleotide probe with a 5′ fluorescent reporter dye (FAM) and 3′ non-fluorescent quencher (NFQ). When Taq DNA polymerase synthesizes a new strand and reaches the TaqMan probe, the FAM is cleaved from the NFQ and increases the fluorescent intensity proportional to the amount of amplicon synthesized. Fluorescence was determined using a QuantStudio 3 or 5 Real-Time PCR System (Applied Biosystems). Data were analyzed using the comparative threshold cycle (ΔΔCT) method and results are expressed as fold change from a control group.
CNS-conditioned neonatal microglia culture
Adult (8–10 weeks old) and aged (20 months old) male BALB/c mice were sacrificed by CO
2 asphyxiation and three 1-mm coronal brain sections (Bregma - 1, 2, and 3 mm) were collected from each mouse. These sections were placed three-per-well onto 0.4-μm nylon mesh inserts in 1.5 mL DMEM/F12 (Millipore) supplemented with 25% horse serum and incubated at 37 °C and 5% CO
2. Media was refreshed after 3 h and conditioned media (CM) collected after another 24 h. CM was then supplemented with fresh DMEM/F12 containing 10% fetal bovine serum (FBS) at a 1:3 ratio. Control media received identical treatment without contact with brain tissue. Primary microglia cultures were established from the brain of neonatal mice as previously described [
22]. In brief, whole brains were isolated from neonatal (post-natal day 1–3) BALB/c mice and digested in 0.25% trypsin EDTA (Gibco) for 15 min at 37 °C. Samples were then pelleted and resuspended in growth medium (20% FBS, 100 U/mL penicillin, 100 U/mL streptomycin, 0.25 μg/mL amphotericin B, and 50 μg/mL gentamicin in DMEM/F12). Pellets were mechanically dissociated by pipetting and filtered through a 70-μm cell strainer. Mixed glia cultures were maintained at 37 °C and 5% CO
2, and media was replenished every 3–4 days until confluency. Mixed glia cultures were then shaken at 150 rpm on a Forma Orbital Shaker (Thermo) at 37 °C for 3 h. Microglia were harvested from the confluent layer and plated at a density of 1 × 10
4 cells per well on a 96-well plate. Purity was assessed at > 98% using immunocytochemistry (Iba1
+/DAPI
+). After 48 h, microglia were washed twice with DMEM/F12, incubated in adult or aged CM for 24 h, and then stimulated with 1 μg/mL LPS. After 4 h, media was aspirated, and microglial RNA extracted using the RNeasy Plus Mini kit (QIAGEN). Parallel cultures were processed using the CellTiter 96 AQ
ueous One Solution Cell Proliferation Assay (Promega) to determine cell viability in adult or aged CM.
Statistical analysis
Statistical analysis was performed using GraphPad Prism software (San Diego, CA) unless otherwise specified. To determine significant main effects and interactions between variables, two-way analysis of variance (ANOVA) was performed. Comparison between groups was performed using Student’s t-test and Bonferroni’s correction for multiple comparisons. A value of P < 0.05 was considered statistically significant. All data are presented as mean ± standard error of the mean (SEM).
Discussion
Our primary objective was to test if age-associated microglial priming and hyperactivity to immune challenge is reversed by promoting microglial turnover by CSF1R antagonism. We depleted microglia from both adult and aged BALB/c mice and subsequently allowed new microglia to fully repopulate the brain. We provide novel evidence that forcing microglial turnover in aged mice reversed age-associated increases in intracellular lipofuscin accumulation and CD68+ lysosome size. Furthermore, microglia-specific RNA sequencing revealed a 511-gene signature differentiating aged microglia from adult microglia. Of these, expression of 127 genes was reversed by microglial repopulation. Despite this partial reversal of the aged microglial transcriptome, LPS challenge still elicited exaggerated sickness behavior and CNS inflammation in aged mice regardless of microglial repopulation. Thus, we provide novel evidence that the aged brain microenvironment may cause microglial priming. For instance, the whole-brain transcriptional signature of aging was unaffected by repopulation of microglia. Furthermore, soluble factors from the aged microenvironment prime neonatal microglia to LPS ex vivo. While there may be microglia-intrinsic aspects of aging that can be reversed by forced microglial turnover, the brain microenvironment re-establishes the pro-inflammatory and primed profile characteristic of aging.
An important finding is that microglia present after forced turnover in aged mice were different from those in aged controls. “New” microglia had reduced intracellular lipofuscin and smaller CD68
+ lysosomes. A previous report showed aged microglia had increased lipofuscin burden and larger CD68
+ lysosomes due to homeostatic clearance of myelin debris throughout the lifespan [
60]. Furthermore, previous reports indicate that microglial repopulation reverses the proportion of CD68
+ microglia in aged mice [
21]. We extend these findings to show that, while all microglia contain CD68
+ lysosomes, confocal microscopy revealed a restoration of age-associated CD68
+ lysosome enlargement. Additionally, our findings indicate that the lipid-laden phenotype observed in aged microglia was reversed by microglial depletion and subsequent proliferation. It is important to highlight that these changes were observed with only 80–85% depletion of microglia. Collectively, these findings suggest that promoting microglial turnover restores age-related dysfunction in lysosomal structure and intracellular lipofuscin accumulation.
In addition to structural alterations in repopulated aged microglia, repopulated microglia had an intermediate RNA signature. For instance, of the 511 genes differentially expressed by age, 127 were reversed by repopulation. These include various scavenging receptors implicated in Alzheimer’s disease pathogenesis that were increased with age and reversed by microglial repopulation (e.g., A2m, Apoe, Olr1, Sorl1). TGF-β family ligands, Bmp6 and Tgfb, were increased with age and restored by microglial repopulation. Microglial repopulation also reversed age-associated decreases in expression in mRNA encoding suppressor of cytokine signaling-3 (Socs3), which inhibits inflammatory cytokine and chemokine production and may reflect a mechanism for elevated cytokine production in aged microglia. Age-associated differential expression of 58 genes (53 increased, 5 decreased) was partially reversed by microglial repopulation, such that expression moved towards adult control levels but remained significantly different. For instance, age-associated alterations in genes related to innate-immune sensing (Lyz2, Tlr5), chemokines (Ccr1), and cytokine signaling (Tgfbr3) were partially reversed. Collectively, these findings suggest that forced turnover of aged microglia was sufficient to restore lipofuscin accumulation and lysosome dysfunction, and to partially reverse the aged microglial transcriptional signature.
Despite the intermediate RNA profile described above, 307 of 511 differentially regulated genes with age were unaffected by microglial repopulation. Unaffected genes included the pro-inflammatory cytokine
Il1b and the MHC class II component
H2-Eb1, which are implicated in age-associated microglial priming [
24,
29,
68]. Notably, genes associated with a conserved neurodegenerative microglial phenotype were altered with age and either exacerbated (
Axl) or unaffected by microglial repopulation (
Clec7a,
Itgax) [
38]. Despite divergent aging signatures in human and mouse microglia, we found conserved genes (
Cxcr4,
Tnfaip2) were unaffected by repopulation [
25]. Despite partial or full reversal of 185 genes, pathway analysis showed little benefit of repopulation. For example, differential gene expression was consistent with age-associated increases in NF-κB signaling, production of NO/ROS by macrophages, and elevated IFNγ, IFNα, IL-1β and TNF signaling in microglia. This is consistent with previous reports of elevated baseline inflammation and ROS production in aged microglia [
70]. Furthermore, elevated STAT3 and increased cell motility/movement were previously described in mice and humans, respectively [
25,
30]. Neither these pathways nor overall ontological classifications of genes regulated by age were influenced by microglial repopulation. These findings are consistent with PCA, suggesting there is an overall effect of age on transcription that is not fully reversed by repopulation. Collectively, microglial repopulation reverses age-associated increases in intracellular lipid accumulation, but incompletely restores the aged microglial transcriptome.
Intermediate restoration of the microglial mRNA profile in aged mice was insufficient to prevent age-associated exacerbation of sickness behavior or amplified neuroinflammation following peripheral LPS challenge. Aged mice, regardless of microglial repopulation, had prolonged and exaggerated sickness behavior following LPS challenge compared to adult controls. Analysis of RNA copy number in the hippocampus confirmed and extended previous findings that aged mice have exaggerated neuroinflammatory responses to LPS [
29,
46]. Here, we extend the previous literature to differentiate the aged LPS response into two categories: genes increased by LPS in adult mice but
exacerbated by aging and genes
uniquely increased by LPS in aged mice but not in adults. Neither genes exacerbated nor uniquely regulated in aged mice following peripheral LPS were dramatically influenced by microglial repopulation. This is consistent with recent findings reported by Elmore et al. in which whole-brain inflammatory gene expression remained exaggerated in aged mice 6 h after peripheral LPS administration regardless of microglial repopulation. Notably, we found that numerous complement components (
C1qa, C1qb, C3ar1,
Cfb) and the inflammatory cytokine
Il1b were exacerbated by age and unaffected by repopulation. Furthermore, the aged LPS response was more comprehensive than adult mice, and included genes related to extracellular remodeling (
Mmp9), pathogen recognition (
Tlr2, Tlr7), and interferon responsiveness (
Ifit5). Of the 43 genes comprising the LPS signature in aged mice, only 5 were reversed by microglial repopulation. Thus, microglial repopulation was insufficient to reverse age-associated microglial priming to peripheral immune challenge.
Another relevant point of discussion is that the aged microenvironment was unaffected by microglial renewal and likely influences repopulating microglia. It is important to note that the inflammatory signature of the aged brain is conserved throughout the brain and across species [
40]. Thus, the persistence of an inflammatory/damaged microenvironment in the aged brain may explain why microglial repopulation in mice was insufficient to reverse age-associated exacerbation of sickness behavior and neuroinflammation following LPS challenge. In support of this, microglial depletion and repopulation reduced the level of lipofuscin in aged microglia, but not in the aged neurons. Thus, neurons remained lipid-laden in the aged brain, which is associated with elevated oxidative stress [
65]. Moreover, the aging mRNA signature in a coronal brain section (10% microglia) was unaffected by microglial depletion and repopulation. Genes related to astrocyte reactivity (
Gfap,
S100b,
Vim), neurotrophic/growth factors (
Negr1,
Nrep,
Ntrk3), cell death (
Anxa4), neurotransmitter signaling (
Grin3a,
Glra2), and myelin (
Mbp,
Mobp) were all dysregulated with age and unaffected by microglial repopulation. It is important to highlight that these findings differ from a recent report that microglial repopulation restores age-associated dysregulation in genes associated with neuronal health [
21]. Elmore et al. found age-associated increases in whole-brain inflammatory gene expression was unaffected by microglial repopulation. This was in contrast to genes related to cytoskeletal rearrangement and synaptogenesis, which were restored with microglial renewal in aged C57BL/6 mice. Of the 820 transcripts regulated by age in whole-brain RNA between Elmore et al. and our dataset, only 29 are shared in both analyses. This may explain the discrepancies between our conclusions regarding the overall benefit of microglial repopulation with age. Nonetheless, we both observed that microglial repopulation was insufficient to prevent immune and inflammatory priming to peripheral LPS. We further characterized the microglia-specific transcriptome and found an intermediate expression profile, with restoration of some, but not all, inflammatory genes. We interpret these findings to suggest that some microglia-intrinsic aspects of microglial aging can be reversed by repopulation, but overall they become primed as they repopulate in the aged brain.
Microglial repopulation did not reverse evidence of age-induced astrogliosis, which may play a role in priming repopulating microglia. Elevated
Gfap,
S100b, and
Vim expression is associated with reactive astrogliosis in the aged brain [
11,
48,
50,
74], all of which were increased with age in whole-brain RNA regardless of mciroglial repopulation. Consistent with these results, we detected higher GFAP mRNA and protein expression in the hippocampus. Others report astrocytes in the aged brain have an mRNA profile associated with dysfunction (i.e., less supportive of growth, repair, and regulation) [
59]. This is relevant because recent evidence shows that microglia-astrocyte communication helps to resolve microglial activation after peripheral immune challenge [
46,
48]. Furthermore, astrocytes dynamically respond to environmental cytokines, including IL-1, TNFα, and C1q, and this neurotoxic/inflammatory phenotype may persist independent of microglial repopulation [
39]. Here, we found neonatal microglia cultured with conditioned media from aged coronal brain sections had increased responsiveness to direct LPS stimulation and had higher levels of
Il1b,
Il6 and
Tnf compared to those cultured with adult conditioned media, a response consistent with microglial priming [
45,
47]. Thus, either the presence or absence of soluble factors from the aged brain causes microglia to be primed to LPS challenge. We do not, however, know which factor or factors are critical for this re-direction of the microglial LPS response. Collectively, these data indicate age-associated microglial priming is not intrinsic to microglia; rather, microglia develop a primed phenotype in response to elevated inflammation, oxidation, or damage present in the aged brain [
40].
Despite persistence of immune priming, microglial turnover may have some benefits. Elmore et al. report microglial repopulation restored age-associated cognitive decline and synapse loss [
21]. Furthermore, the benefit of microglial repopulation may be more profound in contexts with spatially or temporally restricted injury, in contrast to advanced age where CNS damage is altered progressively throughout the brain. For example, microglial turnover following inducible hippocampal lesion and neuron death ameliorated chronic microgliosis, leukocyte infiltration, and inflammatory gene expression [
55]. These experiments were completed using adult and otherwise healthy mice. We also have data that forced microglial turnover in a model of stress (e.g., repeated social defeat) reversed microglial priming to LPS challenge ex vivo and in vivo [
71]. These data support the conclusion that whether forced turnover of microglia is sufficient to alter responses to subsequent stimuli is context-dependent. In the context of advanced age, our results indicate that the aged CNS microenvironment plays a major role in the development of the pro-inflammatory profile of microglia. This may have significant ramifications for therapeutic approaches looking to replace microglia in the aged, injured, or diseased brain. Based on our current findings, microenvironmental consequences of aging or disease are likely to influence the development of repopulating microglia towards their original compromised phenotype.
There may be specific functional benefits to microglial physiology with depletion and repopulation in the aged brain. As discussed, microglial repopulation reduced CD68 expression, cleared lipofuscin, and partially restored the microglial RNA signature. Because microglia have myriad proposed functions beyond host-defense, including support of neurodegeneration and dynamic phagocytosis of synapses, it is plausible there is a direct benefit of repopulation. We, however, did not systemically test each proposed function of microglia here. Rather, our goal was to determine if microglial priming and exacerbated immune reactivity to LPS challenge in aged mice could be reversed. Priming and immune reactivity in the microglia of aged mice was not reversed by forcing microglial turnover. Nonetheless, we provide significant findings that indicate the local microenvironment of microglia contributes strongly to their phenotype in the brain.
In summary, we provide original and compelling evidence that microglia can be removed and repopulated in the aged brain. As such, the lipid-laden aged phenotype was reversed by forced microglial turnover. This turnover also resulted in partial reversal of the aged microglia RNA signature. Nonetheless, priming and immune-reactive RNA signatures were still detected after repopulation of aged microglia. As a result, LPS challenge still induced an exaggerated microglial inflammatory response in the aged brain. To explain why these new microglia remain primed, we provide RNA sequencing of whole-brain tissue with clear evidence of “inflammaging” that was not restored by microglial turnover. Indeed, conditioned media generated from the brain of aged mice was sufficient to recreate the primed response in developing neonatal microglia ex vivo. Thus, we conclude that the forced renewal of microglia in the aged brain cannot overcome the environmental cues of the inflamed brain as they repopulate.