Interleukin-17 Promotes the Infiltration of CD8+ T Cells into the Brain in a Mouse Model for Alzheimer’s Disease

ABSTRACT Background Interleukin-17 (IL-17) family cytokines play critical roles in inflammation and pathogen resistance. Inflammation in the central nervous system, denoted as neuroinflammation, promotes the onset and progression of Alzheimer’s disease (AD). Previous studies showed that IL-17A neutralizing antibody treatment alleviated Amyloid β (Aβ) burden in rodent models of AD, while overexpression of IL-17A in mouse lateral ventricles rescued part of the AD pathology. However, the involvement of IL-17 in AD and its mechanism of action remain largely unknown. Methods To investigate the role of IL-17 in AD, we crossed mice lacking the common receptor of IL-17 signaling (IL-17RA knockout mice) to the APP/PS1 mouse model of AD. We then analyzed the composition of immune cells and cytokines/chemokines during different phases of AD pathology, and interrogated the underlying mechanism by which IL-17 may regulate immune cell infiltration into AD brains. Results Ablation of IL-17RA in APP/PS1 mice decreased infiltration of CD8+ T cells and myeloid cells to mouse brain. IL-17 was able to promote the production of myeloid- and T cell-attracting chemokines CXCL1 and CXCL9/10 in primary glial cells. We also observed that IL-17 is upregulated in the late stage of AD development, and ectopic expression of IL-17 via adenoviral infection to the cortex trended towards worsened cognition in APP/PS1 mice, suggesting a pathogenic role of excessive IL-17 in AD. Conclusion Our data show that IL-17 signaling promotes neuroinflammation in AD by accelerating the infiltration of CD8+ T lymphocytes and Gr1+ CD11b+ myeloid cells.


Introduction
Alzheimer's disease (AD) is the most common type of dementia among the elderly. The pathology of AD is complex, encompassing thousands of genetic risk factors and multiple environmental risks (Blalock et al. 2005;Eikelenboom et al. 2012;Ertekin-Taner 2007;Masters et al. 2015;Seshadri et al. 2010). However, the disease can be broadly characterized by two major pathological hallmarks: the formation of extracellular amyloid β (Aβ) plaques and intracellular neurofibrillary tangles (NFTs). Accumulation of Aβ, in both fibrillar and oligomeric forms, plays a key role in initiating AD. Studies in recent decades have shown that both brain injury and infection are related to the pathogenesis of AD, especially to Aβ deposition in the brain (Olsen and Singhrao 2015;Roberts et al. 1994;Victoria et al. 2010). Gram-negative bacteria, LPS and herpes simplex virus have been found colocalized with Aβ deposition in the brain, in close vicinity to neurons and blood vessels (Civitelli et al. 2015;De Chiara et al. 2010;Maheshwari and Eslick 2015;Piacentini et al. 2011). Infection causes rapid Aβ seeding and accumulation in cultured neurons and brains of both WT and 5x Familial AD (5X FAD) model mice (Kumar et al. 2016). Aβ aggregates trap bacterial pathogens and hyper-activate microglia and astrocytes, leading to neuroinflammation (Zheng et al. 2016). Aβ is cleaned up via diffusion to the blood, active uptake by macrophages and microglia, and degradation by endoproteases from astrocytes Masters et al. 2015). In AD, exacerbated Aβ production exceeds the pace of its clearance, resulting in Aβ accumulation in the hippocampus and temporal cortex. Both oligomeric and fibrillar forms of Aβ can alarm brain cells, including neurons, microglial cells, astrocytes, oligodendrocytes, and peripheral immune cells, to trigger neurodegenerative processes Meraz-Rios et al. 2013;Tuppo and Arias 2005).
Neuroinflammation is a central player in AD onset and progression. Multiple inflammatory factors, including various cytokines, have been shown to play crucial roles in AD. These include interleukin-1 (IL-1), IL-6, IL-10, and tumor necrosis factor alpha (TNF-α) (Akiyama et al. 2000;Colangelo et al. 2002;Eikelenboom et al. 2002;Wyss-Coray 2006). The link between inflammation and AD suggests that AD might arise as a consequence of neuron-immune crosstalk. In brains of AD patients, CD3+ T cells were found in the hippocampus, amygdala, entorhinal cortex, and temporal gyrus. These T cells were predominately CD8+ and CD27 -. In addition, most of these T cells were also CD45RA-and negative for Ki67 and PCNA, indicating that they were fully differentiated, nonproliferating T cells (Bryson and Lynch 2016;Sardi et al. 2011;Togo et al. 2002). This is consistent with the notion that naïve T cells do not cross the blood-brain barrier (BBB) (Westermann and Pabst 1996), while activated T cells more frequently cross the BBB into the central nervous system (CNS) (Hickey BLH and Kimura 1991;Togo et al. 2002). Clonally expanded CD8+ effector memory T cells were also found in the peripheral blood and cerebrospinal fluid of AD patients, and the presence of these cells was negatively associated with cognition (Gate et al. 2020). Despite the presence of T lymphocytes in AD brains, the mechanism that controls T cell infiltration to CNS during AD development remains elusive.
Interleukin-17 (IL-17) is a potent pro-inflammatory cytokine that promotes autoimmune inflammation, including in the CNS (Sun et al. 2017;Waisman et al. 2015;Zimmermann et al. 2018), and evidence suggests it may be involved in T cell infiltration in AD. Human studies reported higher levels of IL-17 in PBMC and serum of AD patients compared to healthy individuals (Chen et al. 2014;Swardfager et al. 2010), and mouse studies showed Aβ immunization increased IL-17 production by T helper 17 (Th17) cells, suggesting a direct link between Aβ deposition and Th17 activation (McQuillan et al. 2010;Zhang et al. 2013). IL-17 induces the production of granulocyte colony-stimulating factor (G-CSF) and chemokines, such as CXCL1, CXCL2, and CXCL8, which mediate the recruitment of neutrophils to inflammatory sites (Datta et al. 2010;Kolls and Linden 2004;Park et al. 2005;. In certain cases, IL-17 also activates the CXCL9/ CXCL10/CXCR3 axis and promotes the migration of T cells and macrophages (Chen et al. 2019a;Jin and Dong 2013). Elevated expression of CXCL10 was found in both AD patients and animal models of AD (Cattaneo et al. 2017;Duan et al. 2008;Verite et al. 2018). On the other hand, loss of CXCR3 in mice resulted in significantly reduced Aβ deposition and enhanced Aβ uptake by microglia (Krauthausen et al. 2015;Xia et al. 2000). Junling  in the lateral brain ventricle of mice that harbor AD-related mutations APPswe and PSEN1dE9 (APP/PS1 mice) and observed improved glucose metabolism in brain plasma and reduced cerebral amyloid angiopathy (Yang et al. 2017). However, antibody-mediated neutralization of IL-17 in Aβ-injected mice rescued memory impairment and prevented the increase of pro-inflammatory cytokines and chemokines (Cristiano et al. 2019). Thus far, the function of IL-17, at a physiological level, remains unknown in AD development.
In the current study, we employed the APP/PS1 model of AD to mimic Aβ accumulation in the brain, and crossed this model to the IL-17RA knockout mouse line to interrogate the role of IL-17 in this process. We also stimulated primary neurons, microglia, and astrocytes with recombinant IL-17, and assayed their cytokine and chemokine production. With these tools, we aimed to uncover the link between IL-17-mediated neuroinflammation and Aβ pathogenesis during AD development.

Methods and materials
Mice C57BL/6 and APP/PS1 (APPswe, PSEN1dE9, C57BL/6 background) mice were purchased from the Model Animal Research Center of Nanjing University and the Jackson Laboratory. Il17ra −/− mice were from Amgen . Il17ra −/− mice were crossed to APP/PS1 mice to generate the Il17ra −/− APP/PS1 mouse line, with Il17ra +/− mice with or without APP/PS1 transgenes serving as controls. Both male and female mice were used in this study.
All animal experiments were performed in accordance with animal use guidelines approved by the Committee for the Ethics of Animal Experiment at the Shenzhen-Peking University-The Hong Kong University of Science and Technology Medical Center (SPHMC), and the University of Connecticut Health Center.

q-RT-PCR
For q-RT-PCR, mRNA was reverse transcribed using RevertAid First Strand cDNA synthesis Kit (Thermo-Fisher, #K1622). Five-fold diluted cDNA was added to SsoAdvanced Universal SYBR Green Supermix (BioRad, #1725270) and amplified on a CFX96 Real-Time PCR detection system (BioRad). Levels of house-keeping genes GAPDH and RPL32 were used as internal controls. Relative expression levels of different genes were calculated as 2 (Ct House keeping -Ct indicated gene ), and are presented in arbitrary units. Primer sequences for qPCR are listed in Table S2.

Cell viability assay
To evaluate cell viability, CellTiter 96 AQueous (Promega, #G5421) was added to PC12 cells or primary cortical neurons treated with indicated cytokines or solvent control at 1/10 volume of the medium. Cells were incubated at 37ºC for 4 hours. Two hundred microliters of mixed medium was transferred into a 96-well ELISA plate. The absorbance was measured at 490 nm using a microplate reader (BIO-RAD). Cell viability was presented as the percentage of absorbance obtained in control cells.

Brain cell isolation
Mice were sacrificed and the brains were carefully removed from the skull. The brains were chilled in ice-cold normal saline, and associated meninges as well as choroid plexus were removed. Cortex or hippocampus tissues were collected for downstream cell isolation. Adult mouse brain tissues were digested with DNase I (0.5 mg/ml, Sigma, # D5025) and collagenase D (0.2 mg/ml) at 37°C for 30 min. Cells were further dissociated by triturating sequentially with 1 mm, 0.75 mm, and 0.5 mm Pasteur pipettes. Cells were then filtered with a 70 μm cell sieve, and myelin was removed by spinning in 40% Percoll for 25 min at 650 g.
Neonatal mouse brain tissues were digested with 0.25% trypsin, 1 mM EDTA for 13 min, with subsequent addition of 1 mg/ml DNase I and another 7 min digestion. Cells were then purified as described above using 40% Percoll.
Primary glial cells were isolated from the cortex of day 1 C57BL/6J male and female mice. Glial cells were cultured in DMEM with 2 mM L-glutamine, 10% fetal bovine serum, 50 U/ ml penicillin, and 50 μg/ml streptomycin. Cell culture medium was refreshed every 3 days. Primary glial cells were digested with Trypsin-EDTA (0.05%) 14 days after their initial seeding and plated. These cells contain astrocytes (75% GFAP positive cells) and microglia (10% IBA-1 immunopositive cells).

ELISA
Brain tissues were lysed in RIPA buffer with 1 mM PMSF and a protease and phosphatase inhibitor cocktail. Levels of chemokines were analyzed by using MILLIPLEX® MAP Mouse Chemokine Panel. Data were normalized against the amount of total protein.

Immunohistochemistry
Immunohistochemistry was performed according to the methods of Hu et al. (2018). Briefly, Brains were cryo-sectioned into 16 µm slices on a freezing microtome. PFA-fixed brain slices were subjected to antigen retrieval by boiling in 0.01 M citrate-buffered saline (pH 6.0) for 5 min. Brain slices were then stained with anti-APP antibody (6E10 1:1,000, Biolegend, # 803015) or Gr-1 antibody (Biolegend, #108402), followed with secondary antibody staining. Ten slices per mouse were imaged and analyzed with Image J for the percentages of Aβ-positive areas.

Stereotactic injection
After anesthetization with pentobarbital (40 mg/kg), mice were placed in a stereotactic frame (RWD, China) for intraperitoneal injection. A midline incision was made on the scalp, one hole was drilled in the skull according to the following coordinates: anteroposterior (AP) 1.5 mm, lateral (LAT) 0.5 mm, dorsoventral (DV) 1.9 mm (bregma as reference). 1 × 10 7 GFU IL-17-overexpressing adenovirus or GFP-expressing control virus were injected at a rate of 0.2 µl/min with 26 G needles in a 5 µl Hamilton syringe with a compact infusion pump. After the injection, the needle was maintained for another 10 min and then raised to a rate of 1 mm/5 min. The bone hole was sealed by bone wax, and mice were placed at a 37°C heating plate until awake and returned to home cages. Sham operation and saline injections were performed as controls. Mice were tested in a Morris water maze 1 month after injection.

Cognitive behavior test
Morris water maze test was performed in an apparatus that is 122 cm in diameter and 51 cm in height. The tank was filled with water and mixed with skim milk to make the water opaque. Mice were adapted to the behavior test room for three days prior to the test. The Morris water maze test involves 5-day acquisition trials. The circular platform, 12 cm in diameter, was placed at a specific location in the tank, submerged 1 cm below the surface of the water. Random probes (cards with different colors and shapes) were stuck around the tank. A video camera was mounted directly above the tank to record the movement and the trails of mice. Mice were put into the tank from four starting locations (four quadrants) and allowed 90 s to search for the hidden platform. The time and numbers of trials to reach the platform were recorded by Any Maze. Mice standing on the platform for 10 s were considered successful in finding the spot. When mice failed to find the platform within 90 s, they were guided to the platform and allowed to remain on it for 10 s. The platform was kept at the same location and mice were tested for 5 consecutive days.

Statistical analysis
Both male and female mice were used in the animal experiments. Data are presented as the mean ± SEM and were analyzed using MS Excel 2018 (Microsoft). Statistical differences between two groups were determined by Student's t test. F-test was used to determine the types of t-test. We applied two-sample equal variance (homoscedastic) when the p-value of f-test ≥0.5 and two-sample unequal variance (heteroscedastic) was applied when the p-value of f-test <0.5. The p-value of t-test and the n-number for each individual experiment are described in the figure legends. Differences with p < .05 were considered statistically significant. In experiments with more than one factor (such as APP/PS1 combined with IL-17RA KO), we used two-way ANOVA and Tukey's multiple comparisons test to analyze the data.

IL-17 is upregulated during late-stage AD development in APP/PS1 mice
We used the APP/PS1 mouse model to analyze how cytokines regulate innate and adaptive immunity in the CNS (Jankowsky et al. 2004). Amyloid plaque deposition can be found in the brains of APP/PS1 mice at 6 months of age, followed by cognitive loss. To examine the involvement of inflammatory cytokines in AD development, we analyzed their mRNA levels by q-RT-PCR. APP/PS1 mice showed increased expression of IL-17A and F in the cortex and hippocampus, compared with controls ( Figure 1). The expression levels of IL-17 receptor C (IL-17RC) and IL-17RE were also increased in 15-month-old APP/PS1 mice ( Figure S1a). IL-17RC and RE are two co-receptors that pair with IL-17RA and mediate signaling from IL-17A/F and IL-17C, respectively. The expression level of IL-17RA during AD development was similar to non-AD controls ( Figure S1a). In addition, expression of several Toll-like receptor (TLR) signaling molecules were also increased in 9-monthold APP/PS1 mice ( Figure S1b). TLRs, including TLR2, TLR4, and TLR9, generally promote innate immune responses, and their activation promotes IL-17-mediated inflammation (Derkow et al. 2015). These results thus suggested that AD pathogenesis is associated with broad-spectrum inflammation, including elevated IL-17 responses in the brain. Upregulation of IL-17 family cytokines was observed at and after 9 months of age, later than the accumulation of Aβ plaques in APP/PS1 mice. These results suggest that Aβ accumulation may be the cause of IL-17 overexpression in the CNS. The increase in IL-17 levels in APP/PS1 mice was more profound in the hippocampus than the cortex (Figure 1b). On the other hand, upregulation of TNF-α was higher in the cortex, compared with the hippocampus, of APP/PS1 mouse brains (Figure 1b).

IL-17 increases expression of CXCL9/CXCL10 in glial cells
To interrogate the targets of IL-17 signaling in the brain, we measured the expression of IL-17 receptors in different brain cells. Microglia, astrocytes, and neurons were purified by flow cytometry from day 2 postnatal C57BL/6 mice and analyzed via q-RT-PCR. IL-17RA mRNA was expressed in all three cell types, while IL-17RC was restricted to astrocytes and microglia ( Figure S2a). IL-17RC and RE are indispensable co-receptors for IL-17A, C, and F, and the lack of IL-17RC and RE expression on neonatal neurons indicates that IL-17 does not signal directly to these cells. The expression of IL-17RC on astrocytes and microglial cells points to the possibility that IL-17 signals to these cells during neuroinflammation. Indeed, stimulation of primary mouse glial cells and BV2 microglial cells with recombinant IL-17 increased expression of the chemokines CCL2, CXCL1, and CXCL9/10 ( Figure 2). It is plausible that IL-17 participates in the pathogenesis of AD by regulating chemokine-mediated immune cell infiltration.

IL-17 induces CXCL9/10 chemokines in APP/PS1 mouse brain
Consistent with the notion that IL-17 promotes CXCL9/10 expression (Figure 2), ELISA analysis in the cortex and hippocampus showed increased CXCL10 in APP/PS1 mice starting from 9 months of age (Figure 3a), when IL-17 upregulation was also evident (Figure 1). To establish a causal link between IL-17 and CXCL9 family chemokines in AD, we crossed IL-17RA knockout mice to APP/PS1 mice. Loss of IL-17 signaling resulted in a trend of decreased levels of CXCL9/10 in the hippocampus (Figure 3b). Ablation of IL-17RA in APP/PS1 mice did not impact mouse survival (Table S1), nor did it affect total Aβ deposition in the cortex or hippocampus (Figure 3c). These data support that the increase in IL-17 expression is a consequence of Aβ deposition in APP/PS1 mouse brains, and that at a pathological level, IL-17 does not impact Aβ accumulation. To further  Percentages of 6E10-positive areas in the cortex and hippocampus were quantified from 3 pairs of animals (1 male and 2 females). Data are represented as mean ± SEM and were analyzed with Student's t test (*p < .05, **p < .01). m=month. assess IL-17 function in AD progression, we injected IL-17A-overexpressing adenovirus into the cortex of 7-month-old APP/PS1 and WT mice. Prior to viral injection, we confirmed consistent cognitive loss in APP/PS1 mice as measured by the Morris water maze (data not shown). We tested the impact of IL-17 overexpression on memory loss 1 month after adenovirus injection. Ectopic expression of IL-17 in the brain slightly increased the time required for APP/PSI mice to find the hidden platform, indicating worsened cognitive loss ( Figure S3).

IL-17 promotes the infiltration of CD8+ T cells and Gr-1+ CD11b+ myeloid cells to the hippocampus of aged APP/PS1 mice
CXCL9 and 10 bind to CXCR3 and attract the infiltration of T cells, neutrophils, NK cells, and monocytes (Muller et al. 2010;Tokunaga et al. 2018). Given the role of IL-17 in activating CXCL9/10 production in APP/PS1 mouse brains, we proposed that IL-17 promotes the infiltration of T cells and monocytes in response to Aβ accumulation. To this end, we performed flow cytometry analysis from APP/PS1 and WT mice that harbor intact or defective IL-17 signaling circuitry. The gating strategies to identify lymphocytes, invading macrophages, and microglia are shown in Figure S4a. Increased frequencies of CD11b-CD45 hi lymphocytes, CD8+ T lymphocytes, and Gr1+CD11b int CD45 int myeloid cells in the cortex and hippocampus were found in 15-and 24-month-old APP/PS1 mice when compared with age-matched wild type controls (Figure 4a). Loss of IL-17RA in APP/PS1 mice resulted in markedly reduced numbers of these cell populations (Figure 4a,c). Loss of IL-17 signaling in APP/PS1 mice did not result in gross abnormality of different blood cell populations in the blood or spleen ( Figure S4b). The reduction of CD8+ T cells and Gr1+ CD11b int CD45 int monocytes in the brain of IL-17RA knockout mice may thus be due to reduced migration from the periphery. Previous studies have shown that T cells can be attracted to and activated at the choroid plexus (CP), then cross the CP epithelium into the cerebral spinal fluid (CSF) (Kaur et al. 2016). It is possible that IL-17 promotes migration of T cells through the CP-CSF route.

Discussion
IL-17 family cytokines are important players in infection and autoimmunity (Isailovic et al. 2015;Moseley et al. 2003;Onishi and Gaffen 2010). In healthy brains, physiological levels of IL-17 inhibit intrinsic neuronal excitability, and negatively regulate the expression of proneuronal genes in neuronal progenitor cells in the adult dentate gyrus of the hippocampus under non-inflammatory conditions (Liu et al. 2014a). The role of IL-17 in AD remains largely unknown, although a polymorphism within the IL-23 locus, whose protein product activates the production of IL-17 by Th17 and innate immune cells, is associated with AD development (Liu et al. 2014b). Ablation of IL-12p40 (which is shared between IL-12 and IL-23) results in decreased cerebral amyloid load (Tan et al. 2014;Vom Berg et al. 2012). In addition, elevated expression of IL-17 has been reported in serum or CSF in both AD patients and animal models (Chen et al. 2014;Swardfager et al. 2010). IL-17 neutralizing antibody has also been found to rescue the Aβ burden and cognitive deficits in an AD model based on intracerebroventricular injection of Aβ (Cristiano et al. 2019). Furthermore, IL-17 has been shown to disrupt the BBB in AD that results in increased . IL-17 promotes the infiltration of CD11b-CD45 hi lymphocytes, CD8+ T cells, and Gr1+ CD11b int CD45 int monocytes into the brain of APP/PS1 mice. (a) Cortex (CX) and hippocampal (HP) cells from 15and 24-month-old of APP/PS1 or WT mice were isolated and subjected to flow cytometry analysis. Frequencies of CD45 hi cells, CD11b-CD45 hi lymphocytes, and Gr1+ CD11b int CD45 int monocytes are shown as percentages of all living cells (Ghost dye negative). Frequencies of CD8+ CD3+ T cells were immune cell infiltration into the brain (Katayama 2020;Kebir et al. 2007;Yang et al. 2017). These reports suggest a possible pathogenic role of IL-17 in AD, although the exact function and underlying mechanism remains unclear.
IL-17 is known to promote neutrophil infiltration to sites of inflammation, which may be related to its role in AD pathology. Other studies suggest that FasL-positive Th17 cells interact with Fas-expressing neurons to induce neuronal death in AD (Zhang et al. 2013). Derkow et al. found that IL-17+ γδ T cells are neurotoxic by co-culturing these cells with neurons, but this effect is not mediated by IL-17 (Derkow et al. 2015). To decipher the mechanism by which IL-17 regulates AD pathology, we generated Il17ra-null APP/PS1 mice. Loss of IL-17 signaling resulted in reductions in both lymphocytes and myeloid cells in the brain of APP/PS1 mice. IL-17 stimulation in glial cells induced the production of CCL2 and CXCL1/2/9/10 chemokines, which may contribute to the recruitment of lymphocytes and myeloid cells. Therefore, the role of IL-17 in AD is mainly mediated through its activation of immune cell recruitment.
In our study, we found that neonatal primary neurons express IL-17RA, but not IL-17RC or RE. Given that the latter two IL-17 co-receptors are required for the sensing of IL-17A, C, and F cytokines, it is unlikely that neonatal neurons respond to IL-17 stimulation. However, one cannot rule out the possibility that as animals age, neurons may start to express IL-17 co-receptors. This may be further complicated by the pathological progression of AD. Thus, during late-stage AD development, such as in the case of aged APP/PS1 mice where we found increased IL-17 expression, it is possible that IL-17 signals to neuronal cells and modulate their function. Indeed, a recent publication showed that IL-17A signals to cortical glutamatergic neurons and promotes anxiety-like behavior (Alves de Lima et al. 2020). It will be important to also test the possibility of direct IL-17-neuron circuitry in aged mice during AD progression.
Sex has been shown to be a variable of AD development in humans and experimental animals. In the APP/PS1 model that we used in this study, it is known that female mice exhibit the loss of cognitive capability earlier than males (Mifflin et al. 2021). Male mice, however, show earlier impairment of glucose and insulin tolerance (Li et al. 2016). In a different model of AD, IL-17 has been shown to accumulate in the brains of female, but not male animals (Brigas et al. 2021). These findings point to a possible gender-based bias on IL-17-mediated neuroinflammation during AD development. Nonetheless, in the APP/ PS1 model, we did observe upregulated IL-17A in male mice in the hippocampus and cortex starting from 9 months of age, compared to age-and gender-matched WT controls (Figure 1). Our limited animal numbers did not allow us to interrogate the differences between male and female APP/PS1 mice in IL-17 production and immune cell infiltration. Further investigation in this area may contribute to additional insights on the genderspecific neuro-inflammation that promotes AD development.
IL-17-producing Th17 cells infiltrate into the AD brain and localize at Aβ plaques (Zhang et al. 2013). Lymphocytes, including Th17, γδ T cells, ILC3 cells, cytotoxic T cells, and macrophages can all produce IL-17 in response to infection or injury. Aβ may also activate the production of IL-17 by peripheral T cells, including cells in the gut (Diaz et al. 2012;Pistollato et al. 2016;Yin et al. 2009). The infiltration of immune cells to AD brain may be associated with deteriorated BBB integrity during AD development. Defective BBB function has been reported in both AD patients and animal models. Neuroimaging and immunostaining of human AD brain specimens showed loss of cerebrovascular integrity, impaired BBB structure, abnormal peripheral deposits, and macrophage infiltration (Hultman et al. 2013;Montagne et al. 2017;Sweeney et al. 2018;van de Haar et al. 2016). In animal models of AD based on disruptions in APP, PSEN1, and Tau proteins, BBB degeneration was demonstrated by the detection of leaking capillary contents and intravenously administrated Evans's blue dye (Montagne et al. 2017;Nelson et al. 2016). In APP/ PS1 mice, increased BBB permeability, including microhemorrhages and fibrin perivascular deposits in the cortex, hippocampus, and periventricular zone, were reported at 12 months of age (Cao et al. 2019;Montagne et al. 2017). It is plausible that activated T cells migrate across the defective BBB to initiate and/or exacerbate neuroinflammation (Farkas et al. 2003;Mietelska-Porowska and Wojda 2017). In our study, we show that IL-17 facilitates the accumulation of myeloid cells and T lymphocytes in the brain of APP/PS1 mice. It remains to be addressed whether IL-17 signaling has any impact on the integrity of BBB, and if so, how this process regulates immune infiltration into AD brains.
In AD pathology, it is now known that microbiota and injury induce the production of Aβ, which consequently leads to Aβ accumulation in the brain followed by cognitive loss. Oligomeric and fibral forms of Aβ stimulate brain cells, including neurons, microglia, astrocytes, oligodendrocytes, and ependymocytes. Activated microglial cells treated by Aβ produce a collection of pro-inflammatory cytokines, including IL-17 (Chen et al. 2019b). These observations suggest that Aβ-induced brain pathology leads to gradual upregulation of IL-17, which may in turn accelerate the pace of AD development. The APP/PS1 mouse model of AD typically shows Aβ accumulation in the brain by 6 months of age (Jankowsky et al. 2004), and cognitive loss by 7 months (Serneels et al. 2009). The increase in IL-17 expression occurs at later stages of AD development, first observed at 9 months. Therefore, the increase in IL-17 level in APP/PS1 mouse brains is plausibly a consequence of Aβ deposition, which further exacerbates neuroinflammation. This timeline also explains the lack of change in Aβ density after ablation of the Il17ra gene in APP/PS1 mice. It is likely that, during AD development in the APP/PS1 model, IL-17 plays no role in Aβ deposition and early phase neuronal injury and memory loss. The impact of IL-17 is more prominent in the recruitment of myeloid and T cells to the brain. In our study, we did not find changes in Aβ deposition in IL-17RA Knockout APP/ PS1 mice while reduced CD8+ T cells and myeloid cells were observed. Activated CD8+ T cells induce neuron damage, microglia activation, T cell infiltration and other broad proinflammation responses in the AD brain (Ní Chasaide and Lynch 2020). There is also a strong correlation between CD8+ T cells and tau pathology in AD patients (Laurent et al. 2017;Merlini et al. 2018) . However, the association between CD8+ T cells and Aβ deposition in AD progress is still not clear. Myeloid cells regulate the function of microglia, but their role in engulfing β-amyloid is debatable (Ní Chasaide and Lynch 2020). No change in amyloid deposition were found when microglia were replaced with myeloid cells. Instead, macrophages in the brain promote microglia activation and reduce synaptic plasticity (Barrett et al. 2015;Blau et al. 2012;Prokop et al. 2015;Varvel et al. 2015). Therefore, our observed reduced infiltration of CD8+ T cells and myeloid cells might not change the Aβ deposition, but alter the neuroinflammation in the brain. Our data also suggest that IL-17 promotes the infiltration of lymphocytes and myeloid cells in the late stage of AD in APP/PS1 mice by signaling to glial cells and inducing the production of CXCL1/2/9/10. Previous studies have found that various immune cells, mainly Th cells, aggregate around Aβ plaques together with activated microglial cells Van Eldik et al. 2016). Future studies are needed to address the impact of the IL-17-T cell axis in late-stage AD development. On the other hand, IL-17 may also be upregulated when the CNS is stressed, such as during brain trauma or infections that are linked to AD (Kumar et al. 2016;Smith et al. 1998;Victoria et al. 2010). IL-17 induced via these mechanisms, preceding Aβ deposition, may play a role in early-stage AD.
Taken together, we show that IL-17 signals to glial cells to induce the production of T cell and myeloid cell-attracting chemokines CXCL9 and CXCL10. IL-17 promotes memory loss in APP/PS1 mice, and accelerates infiltration of CD8+ T cells and Gr1+ CD11b+ CD45+ myeloid cells to late-stage APP/PS1 hippocampus. Blocking IL-17-mediated inflammation may prove beneficial in limiting memory loss in human AD patients, a notion that should be tested in subsequent studies.