Role of Suppressor of Cytokine Signaling 3 (SOCS3) in Altering Activated Microglia Phenotype in APPswe/PS1dE9 Mice

Abstract

In response to changes of the central nervous system environment, microglia are capable of acquiring diverse phenotypes for cytotoxic or immune regulation and resolution of injury. Alzheimer’s disease (AD) pathology also induces several microglial activations, resulting in production of pro-inflammatory cytokines and reactive oxygen species or clearance of amyloid-β (Aβ) through phagocytosis. We previously demonstrated that microglial activation and increase in oxidative stress started from the middle age in APPswe/PS1dE9 mice, and hypothesized that M1 activation occurs in middle-aged AD mice by Aβ stimulation. In the present study, we analyzed in vivo expressions of pro-inflammatory cytokines (M1 microglial markers), M2 microglial markers, and suppressor of cytokine signaling (SOCS) family, and examined the microglial phenotypic profile in APPswe/PS1dE9 mice. Then we compared the in vitro gene expression patterns of Aβ- and lipopolysaccharide (LPS)-stimulated primary-cultured microglia. Microglia in APPswe/PS1dE9 mice exhibited an M1-like phenotype, expressing tumor necrosis factor α (TNFα) but not interleukin 6 (IL6). Aβ-stimulated primary-cultured microglia also expressed TNFα but not IL6, whereas LPS-stimulated primary-cultured microglia expressed both pro-inflammatory cytokines. Furthermore, both microglia in APPswe/PS1dE9 mice and Aβ-stimulated primary-cultured microglia expressed SOCS3. Reduction of SOCS3 expression in Aβ-challenged primary-cultured microglia resulted in upregulation of IL6 expression. Our findings indicate that SOCS3 suppresses complete polarization to M1 phenotype through blocking IL6 production, and Aβ-challenged primary-cultured microglia replicate the in vivo gene expression pattern of microglia in APPswe/PS1dE9 mice. Aβ may induce the M1-like phenotype through blocking of IL6 by SOCS3.

Keywords

Alzheimer’s disease amyloid-β APPswe/PS1dE9 mice IL6 inflammation microglia SOCS3 STAT3

INTRODUCTION

Alzheimer’s disease (AD) is characterized by the presence of senile plaques (SPs) composed of amyloid-β (Aβ) peptides, followed by formation of neurofibrillary tangles and neuronal loss. In addition to these key pathogenic features, activated microglia cluster around SPs [1]. Several microglial PET imaging studies using [¹¹C](R)-PK11195 have shown that activation of microglia starts in the early stage of AD in response to Aβ [2, 3]. Microglia function to maintain normal tissue homeostasis and adopt different activated phenotypes in response to changes in the central nervous system environment. Microglial activation phenotypes are now categorized into M1 (classically activated) with cytotoxic properties and M2 (alternatively activated). M2 microglia are further divided into M2a with a repair and regeneration phenotype and M2c with an acquired-deactivated phenotype, while some researchers have added M2b with an immuno-regulatory phenotype [4, 5]. However, their marker molecules have not been clearly identified, and their roles in AD remaincontroversial.

Lipopolysaccharide (LPS), an activator of toll-like-receptors (TLRs), is a typical stimulator of M1 microglia. LPS induces production of pro-inflammatory cytokines such as tumor necrosis factor α (TNFα) and interleukin 6 (IL6). Additionally, LPS also induces M2b microglia. M2b microglia produce not only pro-inflammatory cytokines but also anti-inflammatory cytokines such as interleukin 10 (IL10) [5]. Pro-inflammatory cytokines are also readily induced from microglia by Aβ stimulation, and the pro-inflammatory cytokines released from M1 microglia are known to accelerate neuronal dysfunction [6]. Therefore, microglia are considered an exacerbation factor of AD. Pro-inflammatory cytokines such as TNFα and IL6 produced by M1 microglia directly induce neuronal loss [7, 8], or act in an autocrine manner to activate glial cells and exacerbate oxidative stress, promoting neuronaldegradation [9].

The suppressor of cytokine signaling (SOCS) family is a negative regulator of the Janus kinase (JAK)/signal transducer and activator of transcription (STAT) pathway. The SOCS family prevents unregulated prolonged activation of the JAK/STAT pathway. SOCS3 is co-upregulated with pro-inflammatory cytokines upon LPS stimulation, and has been reported to be an M1/M2b or M2c marker [5]. SOCS3 is expected to suppress excessive inflammation of M1 microglia through negative feedback of the JAK2/STAT3 pathway [10].

On the other hand, M2a and M2c microglia are induced typically by Th2 cytokines such as interleukin 4 (IL4), tumor growth factor β (TGFβ), and IL10. M2 microglia mainly function to maintain homeostasis and restore abnormal environment to the normal condition. Recent studies have demonstrated neuroprotective effects of M2a and M2c microglia [11 –13], and their important role in the clearance of Aβ [14 –16] or deactivation of inflammation through producing anti-inflammatory cytokines [11].

We previously transplanted primary-cultured microglia into an AD rat model and demonstrated that donor microglia increased Aβ clearance. However, the microglia also enhanced the production of cytotoxic pro-inflammatory cytokines [15]. Aβ is considered to induce several phenotypes of activated microglia, and microglial activation may change with AD progression. Knowledge on the microglial activation phenotypes in AD is necessary for the development of therapeutic strategies targeting microglia in AD brain. In order to observe microglial activation and reactive oxygen species (ROS) production, we previously developed methods to monitor oxidative stress in the living state using in vivo electron paramagnetic resonance imaging. We found that oxidative stress started to appear in the brain from the middle age in APPswe/PS1dE8 mice [17]. Thereafter, CD68-positive microglia started to increase, phagocytosing fibrillar Aβ [18]. From these findings, we hypothesize that M1 activation occurs in the middle age of AD mice through stimulationby Aβ.

To confirm this hypothesis, we analyzed the microglial activation phenotypes in middle-aged (9-month-old) APPswe/PS1dE9 mice using immunostaining and qRT-PCR, focusing especially on SOCS3 expression. Moreover, we compared the phenotypes of Aβ-challenged and LPS-stimulated primary-cultured microglia. Finally, we knocked-down SOCS3 expression in Aβ-challenged primary-cultured microglia, and examined the function of microglial SOCS3 in AD.

METHODS

Animals

All the animal studies were approved by the Animal Care and Use Committee of Sapporo Medical University (approval number: 15-063), and all the procedures were carried out in accordance with the institutional guidelines. Pregnant Slc: ddY mice were purchased from Sankyo Labo Service Corporation, Japan. Hemizygous APPswe/PS1dE9 founder mice expressing chimeric mouse/human amyloid precursor protein (APP) swe (mouse APP695 harboring a human Aβ domain and mutations K594N and M595L linked to Swedish familial AD pedigrees) and human mutated presenilin 1 (PS1) – dE9 (deletion of exon 9) were purchased from Jackson Laboratory, USA. Male and female APPswe/PS1dE9 mice and male and female wild-type littermates were used in this study. All the mice used in this study were bred by mating male APPswe/PS1dE9 mice with female C57BL6/J mice in the animal facilities at Sapporo MedicalUniversity. The mice were maintained at 25°C with a 12 h light/dark cycle and provided food and water ad libitum.

Preparation of Aβ oligomers

Aβ oligomers (oAβ) were prepared as described previously with minor modification [19]. Human synthetic Aβ₄₂ (Ana Spec, San Jose, CA, US) was suspended in 1,1,1,3,3,3 hexafluoro-2-propanol (Sigma Aldrich, St. Louis, MO, US) to 1 mM and lyophilized. To form oAβ, peptide films werere-suspended to 5 mM in DMSO and sonicated for 10 min, then diluted to 100 μM in PBS and vortexed for 30 s. Aggregation was allowed to proceed for 24 h at 4°C.

Microglia separation from mouse brain tissue

Brain tissues from 3-, 6-, 9-, and 12-month-old APPswe/PS1dE9 mice (n = 4 for each time point) and age-matched wild-type littermates (n = 4 for each time point) were dissociated to single-cell suspensions using neural tissue dissociation kits (P) (Miltenyi Biotec, Bergisch, Gladbach, Germany) according to the manufacturer’s protocol. To remove myelin, dissociated brain tissue was re-suspended in 35% Percoll (GE Healthcare, Little Chalfont, UK) and slowly underlaid by 70% Percoll. After centrifuging for 10 min at 300 g, glial cells were collected from the 70–35% interphase. To separate primary microglia from the glial cells, CD11b-positive cells were magnetically labeled by MicroBeads-conjugated anti-CD11b antibody (Miltenyi Biotec), and loaded onto a MACS Column (Miltenyi Biotec), which was placed in the magnetic field of a MACS Separator (Miltenyi Biotec). After removing the column from the magnetic field, microglia were eluted as the CD11b-positive cell fraction.

Primary-cultured mouse microglia

Primary-cultured mouse microglia were prepared as described previously with minor modification [20]. Forebrains from newborn Slc:ddy mice were minced and passed through a 70-μm nylon mesh. The cells were cultured in DMEM with 10% fetal bovine serum at 37°C in humidified 5% CO₂, 95% air. On day 17, floating microglia were isolated from the cell cultures. The purified microglia were transferred to 24-well plates (2.0×10⁵ cells/well) or 96-well plates (0.5×10⁵ cells/well) and allowed to adhere at 37°C overnight. The cells were then treated with 10 ng/ml LPS from E. coli serotype O55:B5 (Enzo Life Science, Tokyo, Japan), 1 μM oAβ, or PBS.

Knock-down of SOCS3 in primary-cultured mouse microglia

Primary-cultured microglia were transfected with 0.4 μM siRNA against SOCS3 (SMARTpool: siGENOME Socs3 siRNA; GE Healthcare) or universal control siRNA (NIPPON GENE, Tokyo, Japan) using the hemagglutinating virus of Japan (HVE-J) envelope vector (GenomONE SI, Ishihara Sangyo, Osaka, Japan). At 24 h after siRNA transfection, the cells were treated with 1 μM oAβ or PBS.

Immunohistochemistry

Three-, 6-, 9-, and 12-month-old APPswe/PS1dE9 mice (n = 3 for each time point) were deeply anesthetized by intraperitoneal injection of xylazine (10 mg/kg) and pentobarbital sodium (50 mg/kg). Then the brain was quickly removed and hemisected in the midsagittal plane. The brain was postfixed for 2 days in 4% paraformaldehyde, and subsequently transferred to 10% sucrose followed by 20% sucrose in 10 mM PBS at 4°C for immunohistochemical and histological analysis. The brains of APPswe/PS1dE9 mice were cut into 20-μm thick slices using a cryostat, and collected in 10 mM PBS containing 0.1% sodium azide at 4°C. The brain sections of APPswe/PS1dE9 mice were incubated with rat anti-CD11b (550282, 1 : 50; BD Pharmingen, San Jose, CA US), mouse anti-GFAP (G3893, 1 : 2,000; Sigma Aldrich, St. Louis, MO, US), or mouse anti-NeuN (ABN78, 1 : 2,000; Millipore, Darmstadt, Germany), and rabbit anti-SOCS3 (ab16030, 1 : 1,000; Abcam, Cambridge, UK), mouse anti-TNFα antibody (AB8348, 1 : 250; Abcam) or rabbit anti-IL6 antibody (ab6672, 1 : 1000; Abcam) for 3 days at 4°C. Then, the sections were probed with anti-rat or anti-mouse IgG antibody conjugated with Alexa Fluor 488 and anti-rabbit or anti-mouse IgG antibody conjugated with Alexa Fluor 594 (each diluted 1 : 1,000) for 2 h at room temperature. The brain sections of APPswe/PS1dE9 mice were further incubated with 1-fluoro-2,5-bis(3-carboxy-4-hydroxystyryl) benzene (FSB, 1 : 10,000; Dojindo Laboratories, Kumamoto, Japan) for 30 min or Hoechst 33342 (Hoechst, 1 : 1000; Dojindo Laboratories) for 5 min. Fluorescence was observed under a laser scanning confocal microscope (LSM510 Meta; Carl Zeiss, Jena, Germany).

RNA isolation and quantitative PCR

RNA was prepared from microglia isolated from brain tissues or primary-cultured microglia using RNeasy mini kit (Qiagen, Hilden, Germany) according to the manufacturer’s instruction. Single-stranded complementary DNA (cDNA) was synthesized from more than 1 μg total RNA in a 50 μl reaction mix for 1 h at 42°C using 1000 U ReverScript(R) I (Wako, Tokyo, Japan) and 5 μM oligo-dT primer. For quantitative PCR (qPCR), the cDNA products were diluted to 400 μl in 1/10 TE buffer. Amplification was performed using Power SYBR Green Master Mix (Thermo Fisher, Walthan, MA, US) with specific primers: GAPDH (forward 5’-GGGCTGGCATTGCTCTCA-3’ and reverse 5’-TGTAGCCGTATTCATTGTCATACCA-3’), TNFα (forward 5’-TTCCGAATTCACTGGAGCCTCGAA-3’ and reverse 5’-TGCACCTCAGGGAAGAATCTGGAA-3’), IL6 (forward 5’-TGGCTAAGGACCAAGACCATCCAA-3’ and reverse 5’-AACGCACTAGGTTTGCCGAGTAGA-3’), arginase1 (forward 5’-GGAAGACAGCAGAGGAGGTG-3’ and reverse 5’-TATGGTTACCCTCCCGTTGA-3’), IL10 (forward 5’-GGCAGAGAACCATGGCCCAGAA-3’ and reverse 5’-AATCGATGACAGCGCCTCAGCC-3’), TGFβ (forward 5’-GACTCTCCACCTGCAAGACC-3’ and reverse 5’-GGACTGGCGAGCCTTAGTTT-3’), SOCS1 (forward 5’-TGGTTGTAGCAGCTTGTGTC-3’ and reverse 5’-CCCCTGGTTTGTGCAAAGATAC-3’), and SOCS3 (forward 5’-AAGGCCGGAGATTTCGCTTC-3’ and reverse 5’-GGAAACTTGCTGTGGGTGAC-3’). The cy-cling condition was 30 cycle of 95°C for 30 s, 60°C for 30 s, and 72°C for 30 s, followed by a single 10-min extension at 72°C. The level of mRNA expression in each sample was standardized against GAPDH.

ELISA

Primary-cultured microglia were plated on 96-well plate and incubated overnight before stimulating with oAβ. The intracellular protein levels were determined using In Cell ELISA Base Kit (R&B Systems, Mckinley, NE, USA) with rabbit anti-SOCS3 (1 : 1,000) and mouse anti-GAPDH (1 : 1000, attached to In Cell ELISA Base Kit). TNFα and IL6 concentrations in the conditioned medium were determined usingTNF-α Mouse ELISA kit (Thermo Fisher) and IL-6 Mouse ELISA kit (Thermo Fisher), according to manufacturer’s instructions.

Statistics

All the in vitro experiments were performed at least three times. Statistic analysis was performed using JMP11 (SAS Institute, Cary, NC, USA). Data are presented as mean±SD. The differences between groups were analyzed by Student’s t-test or one-way ANOVA followed by post hoc comparison with Tukey-Kramer HSD test.

RESULTS

Activated microglia produced TNFα but not IL6 in APPswe/PS1dE9 mice

We first examined microglial TNFα and IL6 expressions in APPswe/PS1dE9 mouse brain by microscopic observation (Fig. 1). We observed immunofluorescence for CD11b; a microglial marker (green; Fig. 1b, d) or GFAP; an astrocytic marker (green; Fig. 1e), and TNFα (red; Fig. 1a, b) or IL6 (red; Fig. 1c-e) in 9-month-old APPswe/PS1dE9 mouse brain sections. Aβ deposition was stained by FSB (blue; Fig. 1b, d, e). Ameboid microglia, morphologically activated microglia, were observed surrounding Aβ deposition and astrocytes located next to microglia. TNFα was expressed in CD11b-positive microglia surrounding Aβ deposition (Fig. 1b). However, IL6 was not detected in CD11b-positive microglia (Fig. 1d), but was observed in GFAP-positive astrocytes (Fig. 1e).

Activated microglia in APPswe/PS1dE9 mice expressed SOCS3

We next examined SOCS3 expression in APPswe/PS1dE9 mouse brain by immunofluorescence staining. To analyze the cellular specificity in brain tissue, we double-stained SOCS3 with CD11b (Fig. 2a, b), GFAP (Fig. 2c, d), or NeuN; a neuronal marker (Fig. 2e, f). Each cell type was labeled green by the respective marker, SOCS3 was labeled red, and nuclei were stained blue. In the cerebral cortex of APPswe/PS1dE9 mice, SOCS3 was detected on all the three cell types. Activated microglia and astrocytes accumulated around Aβ deposition and expressed SOCS3 (Fig. 2a-d), but SOCS3 was rarely observed on ramified “resting” microglia.

We previously reported that the number of microglia and the oxidative stress level in APPsws/PS1dE9 mouse brain increased from 9 months of age [17, 18], and estimated that activation of microglia progressed around 6 to 9 months of age. We analyzed the temporal change in microglial SOCS3 expression in the brains of 3- (Fig. 3a, b), 6- (Fig. 3c, d), 9- (Fig. 3e, f), and 12- (Fig. 3g, h) month-old APPswe/PS1dE9 mice. Aβ deposition (blue) was rarely observed in 3-month-old APPswe/PS1dE9 mouse brain (Fig. 3b) and appeared in 6-month-old mice (Fig. 3d). We previously reported that Aβ deposition appeared around 6 months of age and increased in a time-dependent manner [18]. We analyzed the morphology of microglia that expressed SOCS3 by performing immunofluorescence staining for CD11b (green) and SOCS3 (red). No SOCS3 expression was observed in resting microglia at 3, 6, 9, and 12 months of age, while SOCS3 expression in activated microglia around Aβ increased at 6 to 9 months of age (Fig. 3 c-f).

Gene expressions of M1 markers and SOCS3 were upregulated in 9-month-old APPswe/PS1dE9 mice

To confirm the temporal changes of microglial TNFα, IL6 and SOCS3 expressions, we isolated CD11b-positive microglia (Supplementary Figure 1) from APPswe/PS1dE9 mice. Then we examined the gene expression patterns of M1, M2 markers, and SOCS family to profile their phenotypes (Fig. 4). We measured the mRNA levels of TNFα and IL6 as M1 markers (Fig. 4a, b) [5], arginase1 as an M2a marker (Fig. 4c) [21], IL10 and TGFβ as M2c markers (Fig. 4d, e) [5], and SOCS1 and SOCS3 for the SOCS family (Fig. 4f, g). In microglia of APPswe/PS1dE9 mice, TNFα and IL6 mRNA levels tended to increase from 3 to 9 months of age, and TNFα mRNA level at 9 months of age was significantly higher than that of wild-type mice (p < 0.001) (Fig. 4a), but IL6 mRNA level at the same age was not significantly different from that of wild-type mice (Fig. 4b). Contrary to the M1 markers, arginase1 mRNA level in microglia of APPswe/PS1dE9 mice was the lowest at 9 months of age, and significantly lower than that of wildtype mice (p < 0.001) (Fig. 4c). IL-10 mRNA level in APPswe/PS1dE9 mouse microglia also decreased from 3 to 9 months of age, and the level at 6 months of age was significantly lower than that of wild-type mice (p < 0.05) (Fig. 4d). In spite of the decreases in arginase1 and IL-10 expression, TGF-βmRNA level in APPswe/PS1dE9 mouse microglia peaked at 9 months of age and the level was significantly higher than that of wild-type mice (p < 0.001) (Fig. 4e). In microglia of APPswe/PS1dE9 mice, SOCS1 mRNA level did not change significantly (Fig. 4f), whereas SOCS3 mRNA level increased from 3 to 9 months of age and was significantly higher than that of wild-type mice at 9 and 12 months of age (p < 0.001 and p < 0.01, respectively) (Fig. 4g). The gene expression pattern of TNF-α, IL6 and SOCS3 in microglia of APPswe/PS1dE9 mice was consistent with the pattern observed immunohistochemically (Figs. 1 and 3), and was different from the gene expression pattern of typical activated M1microglia [5].

LPS induced M1/M2b activation in primary-cultured microglia

Both immunohistochemical study and microglial gene expression analysis showed the induction of TNFα and SOCS3 in microglia of APPswe/PS1dE9 mice. Microglial TNFα and SOCS3 expressions were observed around Aβ deposition, and TNFα and SOCS3 mRNA levels were upregulated after Aβ started to accumulate on the cerebral cortex. To examine the mechanism of how Aβ affects microglial activation, we used primary-cultured microglai established from Slc:ddy mice.

Before using Aβ to stimulate cultured microglia, we analyzed microglial activation in response to LPS by incubating microglia for 3 h with 10 ng/ml LPS. The LPS that we used was purified as a specific activator for TLR4, and did not activate other TLRs. LPS strongly induced the mRNA levels of pro-inflammatory cytokines including TNFα and IL6 (p < 0.001 for both) (Fig. 5a, b) and did not match the expression pattern of microglia in APPswe/PS1dE9 mice. LPS also increased IL10 mRNA expression (p < 0.001) (Fig. 5d), but decreased arginase1 mRNA (p < 0.05) (Fig. 5c), and did not change TGFβ mRNA level (Fig. 5e). The mRNA levels of both SOCS1 and SOCS3 were upregulated in response to LPS (p < 0.001, for both) (Fig. 5f, g). Overall, LPS induced typical activation of M1/M2b microgliain vitro.

Aβ-induced pro-inflammatory cytokines and SOCS3 expressions in primary-cultured microglia but failed to increase IL6 and SOCS1 expressions

Contrary to LPS-stimulated primary-cultured microglia, Aβ-challenged primary-cultured microglia showed M1-like gene expression pattern similar to that observed in microglia of 9-month-old APPswe/PS1dE9 mice (Fig. 4). Aβ stimulation induced TNFα and SOCS3 mRNA expressions (p < 0.001 and p < 0.001, respectively) (Fig. 6a, g), and upregulated their protein levels (p < 0.05 and p < 0.001, respectively) (Fig. 6k, m). TNFα mRNA level was upregulated in the early phase (3 h) of Aβ exposure, and thereafter decreased with time (Fig. 6h). On the other hand, SOCS3 mRNA level showed bimodal increase in the early phase (3 h of exposure) and the late phase (48 h) (Fig. 6j). Both IL6 mRNA (Fig. 6b) and protein levels (Fig. 6l) did not increase after 6 h of Aβ exposure in vitro. The mRNA levels of arginase1 and IL10 tended to decrease but with no significant differences (Fig. 6c, d). Contrary to other M2 markers, TGFβ mRNA level was elevated by Aβ stimulation (p < 0.001) (Fig. 6e), and was consistent with the microglial expression pattern of 9-month-old APPswe/PS1dE9 mice (Fig. 4). The mRNA level of SOCS1 did not change at 6 h of Aβ exposure (Fig. 6f), but increased gradually with time (Fig. 6i). SOCS3 is known to be upregulated by IL10, interferon γ (IFNγ), or IL6 in macrophage through JAK2/STAT3 activation [22]. We therefore examined the effect of STAT3. However, STAT3 was sparsely expressed on CD11-positive microglia surrounding SPs, and Stattics, a STAT3 specific inhibitor, did not suppress SOCS3 induction (Supplementary Figure 2).

Reduction of SOCS3 gene induced IL6 expression and secretion but did not increase TNFα expression

We finally performed SOCS3 reduction in primary-cultured microglia to analyze the function of SOCS3 in AD. Microglial SOCS3 was knocked down by RNA interference using siRNA against mouse SOCS3 mRNA in primary-cultured microglia. At 24 h after siRNA treatment, SOCS3 mRNA level decreased by approximately 50% compared with universal control siRNA (p < 0.01) (Fig. 7 g). The protein level of SOCS3 also decreased by approximately 20% (p < 0.05), and Aβ-induced upregulation was abolished (Fig. 7j). More severe SOCS3 reduction decreased microglial viability in our model (data not shown).

After knock-down of SOCS3, we further analyzed the Aβ-stimulated gene expression profiles of microglial activation markers and SOCS family. The reduction of microglial SOCS3 did not change TNFα mRNA level (Fig. 7a) or TNFα concentration in the conditioned medium (Fig. 7h). However, microglial SOCS3 reduction increased IL6 mRNA level with or without Aβ stimulation (p < 0.001 and p < 0.05, respectively) (Fig. 7b), and increased Aβ-stimulated IL6 concentration (p < 0.05) (Fig. 7i). Additionally, SOCS3 reduction decreased Aβ-stimulated IL10 mRNA level (p < 0.01) (Fig. 7d). The mRNA levels of arginase1, TGFβ, and SOCS1 did not change significantly as a result of SOCS3 reduction (Fig. 7c, e, f).

DISCUSSION

We previously reported that microglia accumulated in the brain of APPswe/PS1dE9 mice in the middle age [18], and oxidative stress in the brain started to increase during the same period [17]. With the hypothesis that cytotoxic M1 polarization of microglia occurs in this stage, we initially analyzed the microglial characteristics in the brain of middle-aged APPswe/PS1dE9 mice. In this study, TNFα was detected on activated microglia surrounding SPs (Fig. 1), and TNFα mRNA level increased with AD progression at 9 months of age (Fig. 4). On the other hand, IL6 was not observed on microglia surrounding SPs (Fig. 1), and upregulation of IL6 mRNA level was not significant (Fig. 4). Using another AD model mice (Tg2576), Apelt et al. [23] also observed no IL6 production on microglia and IL6 production only on astrocytes. In short, microglia phenotype was changed to an M1-like profile in AD model mice and IL6 expression was different compared with that of typical M1 or M2b microglia. Moreover, the expressions of M2c markers in microglia of AD model mice were also different from those of typical M1 and M2a microglia. The mRNA level of TGFβ, known as an M2c marker, increased in APPswe/PSdE9 mice at 9 months of age (Fig. 4), but that of IL10, another M2c marker, did not increase (Fig. 4). Additionally, mRNA level of SOCS3 increased but that of SOCS1 did not (Fig. 4). This expression profile in microglia of APPswe/PSdE9 mice was different from that of other inflammation models such as the LPS-stimulated model [24]. Therefore, we suspect that AD pathology is the cause of this unique M1-like phenotype. To the best of our knowledge, this is the first report of microglial expression profiles of SOCS1 and SOCS3 in AD model mice.

To confirm the relationship between AD pathology and microglial activation phenotypes observed in vivo, we compared microglial cultures stimulated by oAβ and by LPS in vitro. LPS is the classical compound used in producing infection-induced acute inflammation model. In this study, we used LPS that specifically activates TLR4 and not other TLRs. This experiment demonstrated different patterns of microglial activation between oAβ and LPS stimulation, in terms of mRNA expressions of IL6, IL10, TGFβ, and SOCS1. LPS at 10 ng/ml strongly induced the expressions of TNFα, IL6, IL10, SOCS1, and SOCS3 (Fig. 5). Chhor et al. [5] also reported the induction of TNFα, IL6, IL10, and SOCS3 expressions by LPS stimulation, and concluded that LPS induced M1 and/or IL10-positive M2b microglia. In contrast, oAβ at 1 μM induced the expression of TNFα, but not IL6 (Fig. 6). Additionally, oAβ induced the expression of TGFβ but not IL10 or SOCS1 (Fig. 6). The gene expression profile of Aβ-challenged primary microglia culture matched the in vivo profile in microglia of middle-aged APPswe/PS1dE9 mice (Fig. 4). These findings suggest the possibility that Aβ stimulation is primarily responsible for the change in gene expression profile in microglia of APPswe/PS1dE9 mice. Moreover, microglial activation in middle-aged APPswe/PS1dE9 mice can be replicated without glia-to-glia or neuron-to-glia interaction.

Microglia recognize Aβ via several amyloid receptors such as complement receptors, FPRs (formyl peptide receptors), TLRs, CD14, CD33, CD36, SRs (scavenger receptors), RAGE (receptor for advanced glycosylation end products), and Fc receptors, inducing cytokine production and Aβ clearance [25]. Microglial TLR4 is the main receptor that activates the internal cascade of Aβ-stimulated production of pro-inflammatory cytokines such as TNFα and IL6 [26 –28]. Aβ also promotes TNFα secretion, and the autocrine action of TNFα upregulates TNFα and IL1β gene expressions in microglial culture [29]. However, oAβ did not increase IL6 production, and induced TNFα to a weaker extent than typical TLR4 stimulation by LPS (Figs. 5 and 6). Several studies have reported that primary-cultured microglia and microglial cell lines produced IL6 in response to Aβ stimulation, but these studies used a higher concentration of oAβ (5 μM) or fibrillar Aβ [30, 31]. Induction of IL6 may depend on the strength of stimulation or the form of Aβ.

Some anti-inflammatory cytokines were also upregulated in middle-aged APPswe/PS1dE9 mice, and microglial TGFβ and IL10 gene expressions were different between Aβ and LPS stimulation in vitro (Figs. 5 and 6). TGFβ mRNA level increased in response to Aβ stimulation (Fig. 6). TGFβ is an anti-inflammatory cytokine that suppresses glial and T cell-mediated neuroinflammation in Aβ injection model rats [32], and TGFβ enhances IL4-induced microglial M2a activation by increasing the expressions of arginase1 and YM1/Chitinase-3-like protein 3 in microglial culture [33]. However, in middle-aged APPswe/PS1dE9 mice, no induction of arginase1 was detected (Fig. 4). Upregulation of arginase1 may require other Th2 cytokines such as IL4 and IL10, or may occur with a time lag. In fact, arginase1 mRNA level increased slightly after middle age (12 months) in APPswe/PS1dE9 mice (Fig. 4). Aβ itself is able to induce TGFβ expression in microglial cell lines [34]. Aβ may induce M2a microglia preferentially rather than M2b and M2c microglia, whereas LPS upregulates IL10 expression to induce M2b and M2c microglia [5].

Furthermore, we found differences in the expression pattern of the SOCS family. Microglial SOCS3 was induced in APPswe/PS1dE9 mice, but SOCS1 expression was not induced (Fig. 4). Walker et al. [35] reported increases in SOCS3 and SOCS2 expressions, but not SOCS1 in the brain of AD patients. The expression pattern of the SOCS family in Aβ-challenged primary-cultured microglia matched the in vivo pattern in microglia of APPswe/PS1dE9 mice (Fig. 6), and was different from that of LPS-stimulated primary-cultured microglia (Fig. 5). Primary-cultured microglia were reported to express several members of the SOCS family, such as SOCS1, SOCS2, SOCS3, and cytokine-inducible SH2 containing protein (CIS), in response to higher concentration of Aβ [35], and they may also depend on the strength of stimulation or the form of Aβ. We focused on the mechanism of SOCS3 induction. The mechanism of Aβ-stimulated SOCS3 induction is not fully understood. LPS induces IL10 secretion, and endogenous IL10 activates STAT3 resulting in SOCS3 production [36]. However, we observed no IL10 induction and sparse STAT3 expression on microglia (Supplementary Figure 2). We also tested the effects of a STAT3 inhibitor, Stattics, but it did not suppress microglial SOCS3 induction (Supplementary Figure 2). These findings suggest that Aβ possibly induces microglial SOCS3 without direct activation of JAK2/STAT3 signaling in AD. TNFα induces microglial SOCS3 through the MKK6/p38^MAPK/MK2 cascade that stabilizes SOCS3 mRNA [37]. Therefore, Aβ probably induces microglial SOCS3 by mechanisms other than JAK2/STAT3 signaling.

Finally, we attempted to analyze the function of SOCS3 in AD using Aβ-challenged primary-cultured microglia. Microglial SOCS3 in AD model mice is expected to suppress pro-inflammatory cytokines production, because macrophage SOCS3 suppresses LPS-stimulated IL1β, IL6, and TNFα production [38], and microglial SOCS3 suppresses peripheral nerve injury-induced IL6 and TNFα production [39]. To examine this hypothesis, we knocked down microglial SOCS3 by RNA interference using siRNA and examined the response to Aβ exposure. Reduction of SOCS3 increased IL6 mRNA level, but did not change TNFα mRNA level (Fig. 7). Additionally, SOCS3 reduction resulted in low expression level of IL10 with Aβ stimulation (Fig. 7). Qui et al. [38] also reported that deletion of SOCS3 in macrophage led to reduction of IL10 expression. These data indicate that reduction of SOCS3 enhances IL6 expression in Aβ-challenged primary-cultured microglia, while inhibiting IL10 expression. The function of microglial SOCS3 in AD has not been fully examined, and our findings show an important novel function of microglial SOCS3 in AD. IL6 is associated with progression of AD. IL6 promotes AβPP production [40] and hyperphosphorylation of tau in neurons [41]. Furthermore, IL6 enhances neuronal damage by interacting with N-methyl-D-aspartate (NMDA) [42]. SOCS3 may function as a suppressor of M1 polarization in AD, and induces the AD-specific M1-like phenotype through blocking IL6 production.

In conclusion, we demonstrated that microglial activation in APPswe/PS1dE9 mice exhibited a unique M1-like phenotype with upregulation of TNFα and SOCS3, but not IL6. Furthermore, Aβ-challenged primary-cultured microglia replicated the in vivo M1-like microglial activation observed in middle-aged APPswe/PS1dE9 mice. Our findings indicate that Aβ directly induces an AD-specific M1-like microglial phenotype different from the typical M1 or M2b phenotype. Aβ-challenged primary-cultured microglial may be useful for researching the role of the M1-like microglia in AD.

Footnotes

ACKNOWLEDGMENTS

This work was supported in part by the Japan Society for the Promotion of Science (JSPS) KAKENHI Grant Number 16K19776 (N.I), 15K19288 (A.M.), 15K09840 (J.K.) and 16H05279 (S.S.) and the Smoking Research Foundation (No. 1503837, S.S.).

Authors’ disclosures available online ().

Appendix

References

Haga

, Akai

, Ishii

(1989) Demonstration of microglial cells in and around senile (neuritic) plaques in the Alzheimer brain. An immunohistochemical study using a novel monoclonal antibody. Acta Neuropathol 77, 569–575.

Okello

, Edison

, Archer

, Turkheimer

, Kennedy

, Bullock

, Walker

, Kennedy

, Fox

, Rossor

, Brooks

(2009) Microglial activation and amyloid deposition in mild cognitive impairment: A PET study. Neurology 72, 56–62.

Yokokura

, Mori

, Yagi

, Yoshikawa

, Kikuchi

, Yoshihara

, Wakuda

, Sugihara

, Takebayashi

, Suda

, Iwata

, Ueki

, Tsuchiya

, Suzuki

, Nakamura

, Ouchi

(2011) In vivo changes in microglial activation and amyloid deposits in brain regions with hypometabolism in Alzheimer’s disease. Eur J Nucl Med Mol Imaging 38, 343–351.

Varnum

, Ikezu

(2012) The classification of microglial activation phenotypes on neurodegeneration and regeneration in Alzheimer’s disease brain. Arch Immunol Ther Exp (Warsz) 60, 251–266.

Chhor

, Le Charpentier

, Lebon

, Ore

, Celador

, Josserand

, Degos

, Jacotot

, Hagberg

, Savman

, Mallard

, Gressens

, Fleiss

(2013) Characterization of phenotype markers and neuronotoxic potential of polarised primary microglia in vitro. Brain Behav Immun 32, 70–85.

Pan

, Chen

, Zhu

, Chen

, Zhang

, Huang

, Ye

, Huang

(2009) Tripchlorolide protects neuronal cells from microglia-mediated beta-amyloid neurotoxicity through inhibiting NF-kappaB and JNK signaling. Glia 57, 1227–1238.

Samavati

, Lee

, Mathes

, Lottspeich

, Huttemann

(2008) Tumor necrosis factor alpha inhibits oxidative phosphorylation through tyrosine phosphorylation at subunit I of cytochrome c oxidase. J Biol Chem 283, 21134–21144.

Behrens

, Ali

, Dugan

(2008) Interleukin-6 mediates the increase in NADPH-oxidase in the ketamine model of schizophrenia. J Neurosci 28, 13957–13966.

Rubio-Perez

, Morillas-Ruiz

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

10.

Baker

, Akhtar

, Benveniste

(2009) SOCS1 and SOCS3 in the control of CNS immunity. Trends Immunol 30, 392–400.

11.

Kiyota

, Ingraham

, Swan

, Jacobsen

, Andrews

, Ikezu

(2012) AAV serotype 2/1-mediated gene delivery of anti-inflammatory interleukin-10 enhances neurogenesis and cognitive function in APP+PS1 mice. Gene Ther 19, 724–733.

12.

Yamamoto

, Kiyota

, Walsh

, Liu

, Kipnis

, Ikezu

(2008) Cytokine-mediated inhibition of fibrillar amyloid-beta peptide degradation by human mononuclear phagocytes. J Immunol 181, 3877–3886.

13.

Shimizu

, Kawahara

, Kajizono

, Sawada

, Nakayama

(2008) IL-4-induced selective clearance of oligomeric beta-amyloid peptide(1-42) by rat primary type 2 microglia. J Immunol 181, 6503–6513.

14.

Kakimura

, Kitamura

, Takata

, Umeki

, Suzuki

, Shibagaki

, Taniguchi

, Nomura

, Gebicke-Haerter

, Smith

, Perry

, Shimohama

(2002) Microglial activation and amyloid-beta clearance induced by exogenous heat-shock proteins. FASEB J 16, 601–603.

15.

Takata

, Kitamura

, Yanagisawa

, Morikawa

, Morita

, Inubushi

, Tsuchiya

, Chishiro

, Saeki

, Taniguchi

, Shimohama

, Tooyama

(2007) Microglial transplantation increases amyloid-beta clearance in Alzheimer model rats. FEBS Lett 581, 475–478.

16.

Takata

, Takada

, Ito

, Asai

, Tawa

, Saito

, Ashihara

, Tomimoto

, Kitamura

, Shimohama

(2012) Microglial Amyloid-beta1-40 phagocytosis dysfunction is caused by high-mobility group box protein-1: Implications for the pathological progression of Alzheimer’s disease. Int J Alzheimers Dis 2012, 685739.

17.

Matsumura

, Emoto

, Suzuki

, Iwahara

, Hisahara

, Kawamata

, Suzuki

, Yamauchi

, Sato-Akaba

, Fujii

, Shimohama

(2015) Evaluation of oxidative stress in the brain of a transgenic mouse model of Alzheimer disease by in vivo electron paramagnetic resonance imaging. Free Radic Biol Med 85, 165–173.

18.

Matsumura

, Suzuki

, Iwahara

, Hisahara

, Kawamata

, Suzuki

, Yamauchi

, Takata

, Kitamura

, Shimohama

(2015) Temporal changes of CD68 and alpha7 nicotinic acetylcholine receptor expression in microglia in Alzheimer’s disease-like mouse models. J Alzheimers Dis 44, 409–423.

19.

Stine

, Jungbauer

, Yu

, LaDu

(2011) Preparing synthetic Abeta in different aggregation states. Methods Mol Biol 670, 13–32.

20.

Takata

, Kitamura

, Saeki

, Terada

, Kagitani

, Kitamura

, Fujikawa

, Maelicke

, Tomimoto

, Taniguchi

, Shimohama

(2010) Galantamine-induced amyloid-{beta} clearance mediated via stimulation of microglial nicotinic acetylcholine receptors. J Biol Chem 285, 40180–40191.

21.

Fenn

, Henry

, Huang

, Dugan

, Godbout

(2012) Lipopolysaccharide-induced interleukin (IL)-4 receptor-alpha expression and corresponding sensitivity to the M2 promoting effects of IL-4 are impaired in microglia of aged mice. Brain Behav Immun 26, 766–777.

22.

Wormald

, Zhang

, Krebs

, Mielke

, Silver

, Alexander

, Speed

, Nicola

, Hilton

(2006) The comparative roles of suppressor of cytokinesignaling-1 and -3 in the inhibition and desensitization of cytokine signaling. J Biol Chem 281, 11135–11143.

23.

Apelt

, Schliebs

(2001) Beta-amyloid-induced glial expression of both pro- and anti-inflammatory cytokines in cerebral cortex of aged transgenic Tg2576 mice with Alzheimer plaque pathology. Brain Res 894, 21–30.

24.

, Ma

, Kulesskaya

, Voikar

, Tian

(2014) Microglia are polarized to M1 type in high-anxiety inbred mice in response to lipopolysaccharide challenge. Brain Behav Immun 38, 237–248.

25.

Doens

, Fernandez

(2014) Microglia receptors and their implications in the response to amyloid beta for Alzheimer’s disease pathogenesis. J Neuroinflammation 11, 48.

26.

Jin

, Kim

, Maxwell

, Li

, Fukuchi

(2008) Toll-like receptor 4-dependent upregulation of cytokines in a transgenic mouse model of Alzheimer’s disease. J Neuroinflammation 5, 23.

27.

Trotta

, Porro

, Calvello

, Panaro

(2014) Biological role of Toll-like receptor-4 in the brain. J Neuroimmunol 268, 1–12.

28.

Zhang

, Wu

, Wang

, Gu

, Zhang

, Xia

(2016) Neural stem cell transplantation decreases neuroinflammation in a transgenic mouse model of Alzheimer’s disease. J Neurochem 136, 815–825.

29.

Jekabsone

, Mander

, Tickler

, Sharpe

, Brown

(2006) Fibrillar beta-amyloid peptide Abeta1-40 activates microglial proliferation via stimulating TNF-alpha release and H2O2 derived from NADPH oxidase: A cell culture study. J Neuroinflammation 3, 24.

30.

Jana

, Palencia

, Pahan

(2008) Fibrillar amyloid-beta peptides activate microglia via TLR2: Implications for Alzheimer’s disease. J Immunol 181, 7254–7262.

31.

Jamalidoust

, Ravanshad

, Namayandeh

, Zare

, Asaei

, Ziyaeyan

(2016) Construction of AAV-rat-IL4 and evaluation of its modulating effect on Abeta (1-42)-induced proinflammatory cytokines in primary microglia and the B92 cell line by quantitative PCR assay. Jundishapur J Microbiol 9, e30444.

32.

Chen

, Ke

, Lu

, Qiu

, Peng

(2015) Protection of TGF-beta1 against neuroinflammation and neurodegeneration in Abeta1-42-induced Alzheimer’s disease model rats. PLoS One 10, e0116549.

33.

Zhou

, Spittau

, Krieglstein

(2012) TGFbeta signalling plays an important role in IL4-induced alternative activation of microglia. J Neuroinflammation 9, 210.

34.

, Chen

, Yang

, Qi

, Lu

, Zhao

(2015) Amyloid-beta activates microglia and regulates protein expression in a manner similar to prions. J Mol Neurosci 56, 509–518.

35.

Walker

, Whetzel

, Lue

(2015) Expression of suppressor of cytokine signaling genes in human elderly and Alzheimer’s disease brains and human microglia. Neuroscience 302, 121–137.

36.

Qin

, Roberts

, Niyongere

, Cong

, Elson

, Benveniste

(2007) Molecular mechanism of lipopolysaccharide-induced SOCS-3 gene expression in macrophages and microglia. J Immunol 179, 5966–5976.

37.

Ehlting

, Lai

, Schaper

, Brenndorfer

, Matthes

, Heinrich

, Ludwig

, Blackshear

, Gaestel

, Haussinger

, Bode

(2007) Regulation of suppressor of cytokine signaling 3 (SOCS3) mRNA stability by TNF-alpha involves activation of the MKK6/p38MAPK/MK2 cascade. J Immunol 178, 2813–2826.

38.

Qin

, Holdbrooks

, Liu

, Reynolds

, Yanagisawa

, Benveniste

(2012) SOCS3 deficiency promotes M1 macrophage polarization and inflammation. J Immunol 189, 3439–3448.

39.

Dominguez

, Mauborgne

, Mallet

, Desclaux

, Pohl

(2010) SOCS3-mediated blockade of JAK/STAT3 signaling pathway reveals its major contribution to spinal cord neuroinflammation and mechanical allodynia after peripheral nerve injury. J Neurosci 30, 5754–5766.

40.

Del Bo

, Angeretti

, Lucca

, De Simoni

, Forloni

(1995) Reciprocal control of inflammatory cytokines, IL-1 and IL-6, and beta-amyloid production in cultures. Neurosci Lett 188, 70–74.

41.

Quintanilla

, Orellana

, Gonzalez-Billault

, Maccioni

(2004) Interleukin-6 induces Alzheimer-type phosphorylation of tau protein by deregulating the cdk5/p35 pathway. Exp Cell Res 295, 245–257.

42.

Qiu

, Gruol

(2003) Interleukin-6, beta-amyloid peptide and NMDA interactions in rat cortical neurons. J Neuroimmunol 139, 51–57.