Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Low-grade peripheral inflammation affects brain pathology in the AppNL-G-Fmouse model of Alzheimer’s disease

Low-grade peripheral inflammation affects brain pathology in the AppNL-G-Fmouse model of... Alzheimer’s disease (AD) is a chronic neurodegenerative disease characterized by the accumulation of amyloid β (Aβ) and neurofibrillary tangles. The last decade, it became increasingly clear that neuroinflammation plays a key role in both the initiation and progression of AD. Moreover, also the presence of peripheral inflammation has been exten- sively documented. However, it is still ambiguous whether this observed inflammation is cause or consequence of AD pathogenesis. Recently, this has been studied using amyloid precursor protein (APP) overexpression mouse models of AD. However, the findings might be confounded by APP-overexpression artifacts. Here, we investigated the effect NL-G-F of low-grade peripheral inflammation in the APP knock-in (App ) mouse model. This revealed that low-grade peripheral inflammation affects (1) microglia characteristics, (2) blood-cerebrospinal fluid barrier integrity, (3) periph- eral immune cell infiltration and (4) Aβ deposition in the brain. Next, we identified mechanisms that might cause this effect on AD pathology, more precisely Aβ efflux, persistent microglial activation and insufficient Aβ clearance, neu- ronal dysfunction and promotion of Aβ aggregation. Our results further strengthen the believe that even low-grade peripheral inflammation has detrimental effects on AD progression and may further reinforce the idea to modulate peripheral inflammation as a therapeutic strategy for AD. Keywords: Low-grade peripheral inflammation, Brain barriers, Choroid plexus, Blood-CSF barrier, Alzheimer’s disease Introduction [1]. Worldwide, nearly 50 million people have AD or Alzheimer’s disease (AD) is a devastating age-related related dementia, and this number will multiply in the neurodegenerative disorder that is characterized by the next decades [1]. The speed of disease progression is sub - progressive and disabling deficits in cognitive functions jective to individual variability, but patients are estimated including reasoning, attention, judgment, comprehen- to live from a few up to 20 years after their diagnosis [1]. sion, memory and language. AD is the most common Next to dramatically affecting the life quality and expec - form of dementia and may contribute to 60–70% of cases tancy of patients, the disease also takes its toll on our healthcare system and is becoming one of the most eco- nomically taxing diseases in developed countries. Unfor- *Correspondence: Roosmarijn.Vandenbroucke@irc.VIB-UGent.be tunately, only symptomatic medication that is effective Lien Van Hoecke and Roosmarijn E. Vandenbroucke share senior authorship for some AD patients is available, but no cure nor treat- VIB-UGent Center for Inflammation Research, ment to reverse or even halt disease progression exists. Technologiepark-Zwijnaarde 71, 9052 Ghent, Belgium Full list of author information is available at the end of the article © The Author(s) 2021. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Xie et al. acta neuropathol commun (2021) 9:163 Page 2 of 23 It is already well-known for decades that the depo- the microglial cells. Microglia might also be primed in sition of amyloid-beta (Aβ) protein in senile plaques response to peripheral immune reaction [64]. outside neurons and the formation of neurofibrillary While numerous studies have looked into the pres- tangles (NFT) composed of hyperphosphorylated Tau ence of inflammation in AD using mouse models and (p-Tau) protein inside neurons result in the loss of syn- patient samples [46], only a limited amount of studies apses and neurodegeneration which ultimately leads to have looked into the direct impact of peripheral inflam - symptoms associated with AD [12]. The steady progress mation on AD pathology [29, 31, 68]. Among them is the in the understanding of the etiopathogenesis has led recent publication of Tejera et al. showing that peripheral to the evaluation of therapies aiming to reduce patho- inflammation alters Aβ pathology by negatively regulat - logical aggregates of either Aβ or p-Tau. Unfortunately, ing microglial clearance capacity [64]. Importantly, all none of these strategies has led to clinical success [10]. research performed so far on the effect of peripheral As a consequence, a more in-depth and more compre- inflammation on AD pathology is performed in mouse hensive understanding of the AD pathology is crucial models that overexpress the Aβ precursor protein (APP) for the development of novel effective therapies. in combinations with different familial AD (FAD) asso - Recently, emerging evidence suggests that innate ciated mutations in APP or presenilin 1 (PS1), such as immune activation plays a crucial role in the pathogen- Tg2576, APP/PS1, 5xFAD and 3xTg-AD [53]. These so esis and progression of AD [14, 24, 35]. For example, called first-generation transgenic mouse models exhibit genome-wide association studies (GWAS) have dem- AD pathology, but the overexpression may cause addi- onstrated that genes for immune receptors including tional phenotypes unrelated to AD. In addition, the CR1, CLU, CD33, and TREM2 are associated with AD mouse models use the neuron-specific Thy1-promotor to development[26]. More recently, Sierksma et  al. also overexpress the N-terminal truncated Aβ species, which identified SYK, GRN, SLC2A5, PYDC1, HEXB, and makes APP processing even less similar to the human BLNK as risk genes[59]. All these identified risk genes situation. Contrary, second-generation mouse models are involved in the regulation of the immune response utilize an APP knock-in strategy that closer represents within the central nervous system (CNS) but remark- the physiological accumulation of Aβ without pheno- ably also outside the CNS. Moreover, epidemiologi- types related to overexpression [52]. In this study, we are cal and translational research suggests that peripheral the first to report on the effects of low-grade peripheral inflammation may promote AD pathology [67]. All inflammation in the more representative second-genera - NL-G-F these findings support a substantial involvement of tion mouse models for AD, namely the App mouse both peripheral and central immune function in AD model. Our data reveal that, in agreement with the study pathogenesis. Consequently, understanding the con- of Tejera et  al. [64], microglial activity is also affected nections between the immune system and AD develop- using this AD mouse model. Besides, our results reveal ment might be key in our search for therapies against that upon low-grade peripheral inflammation there is an AD. influx of myeloid cells into the brain and a disruption of The principal resident immune cell of the CNS is the the blood-cerebrospinal fluid (CSF) barrier. Moreover, microglia. These phagocytotic cells are ubiquitously we demonstrate that not only microglial Aβ clearance distributed in the brain and patrol their assigned brain is affected by low-grade peripheral inflammation but regions for the presence of pathogens and cell debris also Aβ transport across the brain barriers and neuronal [23]. Moreover, microglial cells provide factors that functioning. support overall tissue maintenance and plasticity of neuronal circuits [32]. However, when homeostasis is Results disrupted, e.g. in response to inflammation, microglia Low‑grade peripheral inflammation induces adopt an activated state which is characterized by an neuroinflammation in second‑generation Alzheimer’s amoeba-like structure, an increase in proinflammatory disease mouse model. cytokine expression and a higher phagocytic activity During the last years it became increasingly clear that the [64]. If such an imbalance in homeostasis persists, the induction of peripheral inflammation causes an immune microglia cells trigger an exaggerated inflammatory response in the CNS [48, 54, 55], while its impact on e.g. response leading to a sustained exposure of neurons to Alzheimer-like pathology is less well studied. Here, we pro-inflammatory mediators, with neuronal dysfunc - used two i.p. injections (day 0 and day 7) of a low LPS tion and cell death as a consequence [25]. During aging dose (1.0  mg/kg body weight) to study how low-grade and neurodegeneration, microglia show enhanced sen- peripheral inflammation may affect AD pathology in sitivity to inflammatory stimuli, so called priming of NL-G-F 20–23 weeks old App mice 24 h (day 8) and 2 weeks (day 21) after the last LPS injection (Fig. 1a). Xie  et al. acta neuropathol commun (2021) 9:163 Page 3 of 23 Fig. 1 LPS induces transient peripheral inflammation and neuroinflammation. a Schematic representation of the experimental design. b Protein levels of IL-1β, TNF and IL-6 in plasma and hippocampus (n = 5–10). c Expression of the pro-inflammatory genes Il1β, Tnf and Il6 in hippocampus (n = 5–10). d Analysis of relative TLR4 activation by plasma and brain lysate (n = 9–21). Mean ± SEM, two-way ANOVA Bonferroni’s post hoc test for multiple comparisons. *p < 0.05, **p < 0.01, ***p < 0.001 NL-G-F Both WT and App mice show a drop in body Next, we looked into the effect of low-grade periph - weight and an increase in plasma IL-6 levels in eral inflammation on neuropathological changes in both NL-G-F response to both LPS injections (Additional file  4: WT and App mice. We investigated the effect on (1) Fig.  1). The IL-6 response is slightly more pronounced microglia characteristics, (2) influx in the CNS of periph - NL-G-F in the App compared to the wild type mice, and eral myeloid cells and (3) integrity of the blood-CSF in all case we observed a less strong inflammatory barrier. response to the second LPS injection. Figure  1b shows elevated levels of the pro-inflammatory cytokines Low‑grade peripheral inflammation affects microglia IL-1β, TNF and IL-6 in the plasma 24 h after LPS injec- characteristics. tion (Fig.  1b). In contrast, all cytokines were back at Microglia, the brain-resident immune cells, are the key baseline levels 2  weeks later. Also, neuroinflammation players in regulating central inflammation. Tejera et  al. was observed by an increase in protein and mRNA lev- recently described that microglial cells of WT mice show els of IL-1β and TNF, but not IL-6, in the hippocampus morphological signs of activation 24  h after LPS injec- (Fig.  1b, c). Similar to the peripheral cytokine levels, tion [64]. Here, we elaborated further on the effect of also the increase in brain cytokine levels was transient low-grade peripheral inflammation on brain inflamma - NL-G-F as this increase was detectable  24  h after LPS injec- tion in WT mice and App mice by studying micro- tion while the levels were normalized both in WT mice glia proliferation and activation. Moreover, we compared NL-G-F and App mice 2 weeks later. The increase in brain microglia characteristics when peripheral inflammation inflammation at 24 h is not due to the presence of LPS is still ongoing (24 h after the last low dose LPS injection) in the brain as we could not detect TLR4 activation in to when peripheral inflammation is resolved (2  weeks brain lysates, despite the fact that LPS is still present in after the last LPS injection). Our analysis revealed more NL-G-F the blood at that timepoint (Fig. 1d). microglia and increased proliferation in App mice Xie et al. acta neuropathol commun (2021) 9:163 Page 4 of 23 Low‑grade peripheral inflammation induces leukocyte compared to WT mice at baseline. Moreover, LPS chal- trafficking to the brain lenge further increases both of these parameters. In addi- NL-G-F Next to brain resident microglial cells, also infiltrating tion, the relative effect of LPS on WT and App mice NL-G-F leukocytes can contribute to neuroinflammation. To is similar, but LPS stimulated App mice display with investigate the infiltration of peripheral macrophages the highest number of IBA1 + microglia cells, the high- in the brain after LPS injection, brain sections were est rate of proliferation, the shortest dendrite length, the immunostained for IBA1 and TMEM119. TMEM119 smallest number of branch points and the smallest vol- is specifically expressed on microglial cells, but not on ume. All of this is consistent with the increasing number IBA1 infiltrating macrophages [16, 22]. As shown in of microglia and their amoeboid stage (Fig.  2a–d). The Fig.  3a, b, no macrophage infiltration was observed in significant increase of Ki67 microglial cells was only NL-G-F the cortex or hippocampus 24  h after LPS stimulation, observed in the cortex of App mice 2  weeks after NL-G-F neither in WT nor App mice. In contrast, 2 weeks peripheral inflammation and in the hippocampus of WT later, a clear increase in infiltrating macrophages mice 24  h after the last LPS injection (Fig.  2a–d). Espe- was visible in LPS stimulated mice compared to con- cially microglial cells that are not located in the area of trol mice. This was again observed in both WT and Aβ plaques showed a high proliferation upon peripheral NL-G-F App mice. Although we observed an increased inflammation (Fig. 2a–d). infiltration of peripheral immune cells into the brain Also, the microgliosis marker Aif1 and the microglial upon LPS challenges, the infiltration is still very mod - activation marker CD69 were significantly upregulated in est. In addition, perivascular macrophages (PVMs) are the hippocampus 24  h after LPS treatment both in WT + − NL-G-F also IBA1 and TMEM119 and these cells have been and in App mice (Fig. 2e). This increase is transient shown to increase in neurodegenerative disease models as no differences were observed 2  weeks after the last [20]. Therefore, other techniques which can accurately LPS treatment. distinguish invading peripheral monocytes from brain Despite the fact that peripheral inflammation and neu - endogenous PVMs and microglia, for example using roinflammation are not detectable  2  weeks after LPS genetic labeling of different myeloid populations [49], injection based on cytokine levels (Fig.  1b), the quanti- should be used to further validate this result. tative morphometric three-dimensional (3D) measure- Furthermore, we checked the expression of leuko- ments of IBA1 microglial cells revealed a significant cyte trafficking molecules in the brain during periph - decrease in the number of segments, branch and termi- NL-G-F eral inflammation and we show that exposure to LPS nal points in hippocampus of WT and App mice significantly increases the expression levels of integrin 2 weeks after peripheral inflammation compared to their ligand (Icam1) and chemokines (Ccl2 and Cxcl10) in respective controls (Fig. 2f, g). However, the shorter pro- the hippocampus and/or choroid plexus (CP) of WT cesses and smaller volumes were only observed in WT NL-G-F and App mice 24  h after the second LPS injec- mice injected with LPS compared to WT mice injected NL-G-F tion. However, only Ccl2 expression in the CP remained with PBS. Interestingly, the App mice also showed increased 2  weeks after LPS stimulation. In addition, more activated microglia compared with WT mice in the gene expression of Icam1, Ccl2 and Cxcl10 was basal conditions (Fig.  2g). Additionally, we also inves- more significantly upregulated in response to low- tigated the morphological changes of microglia in the grade peripheral inflammation in the CP compared cortex (Additional file  4: Fig. 2). WT mice showed more to the hippocampus (Fig.  3c, d). Taken together, these pronounced changes in all the examined parameters, findings indicate that low-grade peripheral inflamma - namely dendrite length, number of segments, branch tion induces immune cell infiltration into the brain. and terminal points and cell volume 2 weeks after the last LPS injections compared to the control condition, while NL-G-F this was not observed in App mice. (See figure on next page.) Fig. 2 Low-grade peripheral inflammation affects microglia proliferation and activation. a Representative images of IBA1, 6E10 and Ki67 staining in hippocampus. Scale bar: 100 μm and 20 μm (insert). b Quantification of microglial proliferation in hippocampus (n = 4–6). c Representative images of IBA1, 6E10 and Ki67 staining in cortex. Scale bar: 20 μm. d Quantification of microglial proliferation in cortex (n = 4–6). d Gene expression of microgliosis marker Aif1 and Cd69 in hippocampus (n = 5–10). f Representative 3D reconstruction images of IBA1 microglia from hippocampus 2 weeks after LPS stimulation. Scale bars: 20 µm. g Imaris-based quantification of cell morphology of IBA1 microglia in hippocampus. Each symbol represents one mouse, each mouse was randomly selected with 3–5 cells outside the Aβ for analysis (n = 4–5). Mean ± SEM, two-way ANOVA Bonferroni’s post hoc test for multiple comparisons. *p < 0.05, **p < 0.01 Xie  et al. acta neuropathol commun (2021) 9:163 Page 5 of 23 Fig. 2 (See legend on previous page.) Xie et al. acta neuropathol commun (2021) 9:163 Page 6 of 23 Loss of blood‑CSF barrier integrity in response and OCLN (Additional file  4: Fig. 4a-c). Also, a fragmented to low‑grade peripheral inflammation border staining of ZO-1 and OCLN was observed upon We have previously shown that high dose LPS has detri- IL-1β treatment (Additional file  4: Fig.  4b). These results mental effects on blood-CSF barrier integrity, while the suggest that the proinflammatory cytokine IL-1β is suffi - effect on blood–brain barrier (BBB) was less pronounced cient to induce loss of blood-CSF barrier integrity, but only [2, 66]. To investigate whether low-grade peripheral induces limited differences in the expression of TJ proteins inflammation alters the tight junction (TJ) complex at the and genes in primary CP epithelial cells. CP epithelial cells, the cellular localization of the TJ pro- teins was evaluated by immunocytochemistry and confo- Low‑grade peripheral inflammation results in a higher NL‑G‑F cal microscopy. In control conditions, E-cadherin (CDH1), amyloid deposition over time in App mice Occludin (OCLN), Claudin-1 (CLDN1) and Claudin-5 It is already well-known for decades that the deposition (CLDN5) immunoreactivity appeared as a near continu- of Aβ protein in senile plaques outside neurons results in ous staining at the apical cell border (Fig.  4a). Upon low- the loss of synapses and neurodegeneration which ulti- grade peripheral inflammation, a loss of these TJ proteins mately leads to symptoms associated with AD [12]. Next immunostaining was observed, leading to a fragmented to this, peripheral inflammation has been associated with border and diffuse distribution staining, although the AD [28]. Here, we analyzed whether low-grade inflam - NL-G-F NL-G-F effect on App mice was more pronounced (Fig.  4a). mation has an impact on Aβ pathology in App mice. NL-G-F In addition, immunofluorescence quantitative analysis In unchallenged App mice, we observed an confirmed lower expression of these TJ proteins in the CP increased Aβ deposition at both examined timepoints. NL-G-F of LPS injected WT and App mice compared with Interestingly, 6E10 staining of Aβ revealed that the the corresponding PBS injected mice. However, the lower amount of Aβ plaques is increased in both hippocam- expression of TJ proteins only showed a significant down - pus and cortex upon low-grade peripheral inflammation NL- regulation of CLDN5 2 weeks after LPS injection in App (Fig. 5a, b) and this increase is significant 2 weeks after the G-F NL- mice (Fig.  4b). Consistently, the gene expression levels second LPS injection compared to PBS injected App G-F showed downregulation trends after LPS challenge and mice. In addition to the amount of Aβ aggregation, NL-G-F the App mice seemed more susceptible to disruption also the degree of plaque compactness and the surface by systemic LPS than the WT mice (Fig. 4c). These results area reflect AD pathology [72]. Morphometric analysis show that peripheral inflammation has more pronounced of Aβ stained brain sections revealed an increase in small effects on redistribution of TJ proteins and less on their plaques (< 10 µm ) in the cortex, while in the hippocam- gene expression. Additionally, we also investigated the TJ pus all different sizes of plaques were increased (< 10 µm , 2 2 protein ZO-1 but no differences were visible either on pro - 10–20 µm and > 20 µm ) (Fig.  5c). In agreement with tein level nor on mRNA level, at least not at the examined the Aβ plaque analysis, we also observed a significant time points (Additional file  4: Fig. 3a-c). Collectively, these increase in both soluble and insoluble Aβ in the cortex 1-40 NL-G-F NL-G-F results suggest that the blood-CSF barrier in App of App mice 2 weeks after LPS stimulation (Fig. 5d). mice is more vulnerable to low-grade peripheral inflam - For Aβ only the soluble fraction was increased 2 weeks 1-42 mation compared to their WT counterparts. after LPS stimulation (Fig.  5d). However, the levels of Aβ IL-1β is a major pro-inflammatory cytokine released peptides are not significantly changed 24 h after LPS stim - from activated microglia and have been demonstrated that ulation (Fig. 5d). Taken together, these results suggest that IL-1β treatment increases BBB permeability in  vitro [69]. low-grade peripheral inflammation affects Aβ deposition NL-G-F Here, we hypothesize that IL-1β may also directly affect in App mice, which may be the result of aberrant Aβ blood-CSF barrier permeability. To this end, we studied clearance from the brain. the effect of IL-1β on blood-CSF barrier integrity using primary CP epithelial cells followed by transepithelial elec- Low‑grade peripheral inflammation disturbs Aβ transport trical resistance (TEER) measurements and TJ proteins across the choroid plexus (CP) epithelial cells staining. This revealed that IL-1β treatment significantly Increased levels of Aβ deposition in the brain may be the reduced TEER and TJ proteins expression including ZO-1 result of impaired Aβ clearance that on its term can be the (See figure on next page.) Fig. 3 Low-grade peripheral inflammation induces leukocytes trafficking to the brain. a Representative images of IBA1 and TMEM119 staining in + - hippocampus and cortex. Scale bar: 20 μm. b Quantification of IBA1 TMEM119 macrophages in hippocampus (left graph) and cortex (right graph). Each symbol represents one mouse. Expression of the gene Icam1, Ccl2 and Cxcl10 in hippocampus c and in CP d. Mean of 5 ± SEM, two-way ANOVA Bonferroni’s post hoc test for multiple comparisons. *p < 0.05, **p < 0.01, ****p < 0.0001 Xie  et al. acta neuropathol commun (2021) 9:163 Page 7 of 23 Fig. 3 (See legend on previous page.) Xie et al. acta neuropathol commun (2021) 9:163 Page 8 of 23 result of a disturbed balance of transport across the brain To address whether the reduced Aβ transport can be barriers and/or defective degradation of Aβ aggregates. linked to a reduction in LRP2 transport of the Aβ, we first First, we investigated the effect of low-grade peripheral investigated the expression of LRP2 upon IL-1β stimula- inflammation on Aβ transcytosis across the blood-CSF bar - tion. We showed that there is a reduction in LRP2 expres- rier. As shown in Fig. 6a, Aβ CSF levels were increased sion by CP epithelial cells upon stimulation with IL-1β 1-42 NL-G-F in App mice 24  h after LPS stimulation. Addition- (Fig.  6g, h). The LRP2-dependency of this reduced trans - ally, the level of Aβ in plasma was significantly lower in port was further studied using a blocking anti-LRP2 anti- 1-42 NL-G-F App mice at 24 h, but not at 2 weeks post LPS chal- body. Interestingly, sequential treatment of CP epithelial lenge compared to PBS injected mice (Fig.  6a). To ascer- cells with IL-1β and anti-LRP2 showed no additive effect tain whether this effect correlates with Aβ efflux from the on Aβ transport (Fig.  0.6f ). Importantly, neither the anti- CSF via the blood-CSF barrier, we analyzed the expression LRP2 antibody nor the IgG control had detrimental effects of genes responsible for Aβ influx and efflux in the CP. In on blood-CSF barrier integrity (Additional file  4: Fig. 4a, b). line with the changes of Aβ in CSF and plasma, the expres- Altogether, these data indicate that the pro-inflammatory sion of the Aβ efflux transporters P-gp and Lrp2 showed cytokine IL-1β blocks Aβ transport at least partially by increased trends 2  weeks post LPS injection compared inhibiting LRP2. NL-G-F to the 24  h timepoint in App mice, and only Lrp2 showed significant change (Fig.  6b). However, no signifi - Low‑grade peripheral inflammation affects Aβ cant differences were observed compared to PBS injected phagocytosis by microglial cells NL-G-F App mice. Moreover, protein levels of LRP2 shows Next to Aβ efflux, also defective phagocytotic clearance by a similar trend as the mRNA expression profile (Fig.  6c). microglial cells can explain the increased Aβ deposition Additionally, it has been reported that also LRP1 plays an in the brain [50]. As shown in Fig.  2, low-grade periph- important role in Aβ transport across CP epithelial cells eral inflammation significantly increases microglia prolif - [15, 47]. Therefore, we looked into the expression of LRP1 eration, mainly of the non-plaque-associated microglia. in the CP but no obvious changes were seen 24 h after LPS Moreover, the number of microglia around Aβ plaques NL-G-F stimulation either at protein nor at mRNA-level in the is reduced 24  h after LPS stimulation in App mice NL-G-F App mice (Additional file 4 : Fig. 5a-c). Taken together, (Fig. 7a, b). We also found more microglia migration to the these data suggest that low-grade peripheral inflammation vessel at that timepoint (Additional file  4: Fig. 7a, b). Based disturbs Aβ efflux from the CSF into the blood mainly. on these findings, the question arises if the reduced num - Next, we examined whether the observed IL-1β increase ber of microglial cells around Aβ plaques has an effect on in hippocampus during low-grade peripheral inflammation Aβ engulfment and clearance. To study this, we quantified (Fig.  1b) can explain the effect seen on Aβ transport. To the amount of internalized Aβ and observed a decrease in address this, we established an in vitro primary CP epithe- Aβ engulfment by microglial cells 24  h after LPS stimula- NL-G-F lial transport system using fluorescently labelled Aβ [34, tion in App mice (Fig. 7a, c). Furthermore, we exam- 1-42 62]. The addition of fluorescently labelled Aβ monomers ined CD68 phagocytic microglial cells by performing a 1-42 to the apical side of the epithelial cells allows the cells to co-staining of CD68, Aβ and IBA1 on hippocampus and transport the Aβ to the basolateral side. The rate of Aβ cortex (Fig. 7d). A transient increase in CD68 expression is 1-42 transport can be calculated by measuring the transported observed 24 h after LPS challenge in the hippocamps and NL-G-F Aβ in the basolateral compartment. To eliminate the the cortex of App mice. Interestingly, we observed 1-42 effects of paracellular diffusion of Aβ , scrambled Aβ that less CD68 microglial cells are recruited to the Aβ 1-42 1-42 NL-G-F was added to the cell cultures together with Aβ and the plaques in App mice after LPS challenge, especially at 1-42 transcytosis quotient (TQ) of Aβ was normalized to the the 24-h timepoint (Fig. 7d, e). 1-42 diffusion rate of scrambled Aβ . Confirming our previ - Next, we investigated whether the changes in Aβ engulf- 1-42 ous findings, TQ of Aβ was reduced upon IL-1β stimu- ment can be explained by the activation status of the 1-42 lation of CP epithelial cells (Fig. 6f ). IL-1β did not alter the microglial cells in response to low-grade peripheral inflam - viability of primary CP epithelial cells, even at concentra- mation. To study this, we examined the correlation of the tions ten times greater than that which caused changes in activation state of the microglia and the distance to Aβ the levels of Aβ transport (Additional file 4 : Fig. 6). plaques during low-grade peripheral inflammation. In the (See figure on next page.) Fig. 4 Characterization blood-CSF barrier integrity during peripheral immune challenge. a, b Representative images of CP and quantification of the percentage red staining of stained for CDH1, OCLN, CLDN1 and CLDN5. The dotted line indicates the ependymal cells that line the ventricle (n = 4–6). Scale bar: 20 μm. c Expression of the genes Cdh1, Ocln, Cldn1 and Cldn5 in hippocampus (n = 5). Mean ± SEM, two-way ANOVA Bonferroni’s post hoc test for multiple comparisons. *p < 0.05 Xie  et al. acta neuropathol commun (2021) 9:163 Page 9 of 23 Fig. 4 (See legend on previous page.) Xie et al. acta neuropathol commun (2021) 9:163 Page 10 of 23 NL-G-F Fig. 5 Low-grade peripheral inflammation affects Aβ deposition in App mice. a Representative images of 6E10 staining in hippocampus and NL-G-F cortex of App mice. Scale bar: 100 μm. b Quantification of Aβ plaque area and number (n = 5–6). Hippocampus (left two graphs); cortex (right two graphs). c Quantification of Aβ plaque size distribution. Hippocampus (left); cortex (right); (n = 5–6). d Soluble and insoluble Aβ and Aβ 1-40 1-42 levels in prefrontal cortex tissues (n = 5). Mean ± SEM, two-way ANOVA Bonferroni’s post hoc test for multiple comparisons. *p < 0.05, **p < 0.01 hippocampus, our analysis shows that if the microglial of segments, branch points and terminal points (Fig.  7f, cell is in close proximity to Aβ deposits, the LPS stimula- g). In contrast, peripheral LPS challenge induces a more tion does not lead to further morphological changes of the pronounced microglial activation with increasing distance microglia in terms of dendrite length, cell volume, numbers of the microglia to Aβ plaques compared to PBS injected Xie  et al. acta neuropathol commun (2021) 9:163 Page 11 of 23 NL-G-F App mice (Fig.  7f, g). These morphological changes Correspondingly, coimmunostaining of IBA1 and NeuN are reflected by a reduction in dendrite length, cell volume, showed a higher percentage of microglia and neurons number of segments, branch points and terminal points. that are in close contact with each other in response to The same morphological changes of microglia in the cortex low-grade peripheral inflammation (Additional file  4: were observed upon low-grade peripheral inflammation Fig. 10a, b). (Additional file 4 : Fig. 8). Next, we studied whether peripheral inflammation Altogether, these results show that peripheral inflam - can affect Aβ aggregation by affecting neuron func - mation affects microglial activation, migration and motil - tion. Firstly, we performed live cell imaging to moni- ity and reduces microglial Aβ phagocytosis. tor the morphologic changes and Aβ biodistribution in primary neuron cultures after IL-1β treatment. Com- Low‑grade peripheral inflammation induces neuronal pared to neuronal cell cultures without IL-1β treatment dysfunction and Aβ (Additional file  1: movie 1), the addition of Aβ The phagocytotic ability of a microglial cell is not only affects neuronal function that promotes Aβ aggregation important to clear Aβ aggregates, but is also beneficial close to the soma (Additional file  2: movie 2). Pretreat- for the clearance of apoptotic neurons [23, 36]. However, ment with IL-1β leads to neuronal dysfunction with a activation of microglia upon peripheral inflammation has reduction in dendrites and a similar Aβ as in the condi- shown to cause microglial phagocytosis of healthy neu- tion where only Aβ was added (without the pretreatment rons and synapses that lead to neuronal loss and dysfunc- with IL-1β) (Additional file  3: movie 3). However, it can’t tion [41, 58]. With this in mind, we tested whether an be excluded that Aβ aggregation induces fluorescence enhanced microglial activation and impaired microglial quenching of HF488-Labeled Aβ , reflected by a reduc - 1-42 phagocytosis induced by low-grade peripheral inflamma - tion in fluorescence intensity [13]. To circumvent this NL-G-F tion causes neuronal and synaptic loss in App mice. phenomenon and study the impact of IL-1β on Aβ aggre- In line with a previous study [71], we observed increased gation in neurons, we used fluorescence lifetime imaging trends in cell and neuronal death in the hippocampus (FLIM) of HF488-Labeled Aβ to detect Aβ aggrega- 1-42 and cortex after LPS challenge. Yet, neuronal death only tion as described previously [13]. The fluorescence life - reaches statistical significance in WT mice 24  h after time depends on conformational changes associated LPS challenge (Additional file  4: Fig.  9a-c). Interestingly, with Aβ aggregation and decreases with the increase of NL-G-F the App mice showed a significant higher percent - Aβ aggregation. As shown in Fig. 9a–c, we observed less age of neuronal death in the hippocampus compared to and slower Aβ aggregation in the condition without cells WT mice 2  weeks after PBS injection (Additional file  4: compared to Aβ in the presence of neuronal cells. Inter- Fig. 9a-c). estingly, IL-1β treatment further significantly enhances In the cortex, immunostainings of the presynap- the rate of aggregation in the latter condition. Taken tic marker synaptophysin (SYP) revealed a decrease in together, these data indicate that neuronal dysfunction expression 2 weeks after low-grade peripheral inflamma - affected by peripheral inflammation may play a role in NL-G-F tion in App mice, but no significant differences at accelerating or exacerbating AD pathology. NL-G-F the 24  h timepoint compared to PBS injected App mice. In WT mice, no LPS-induced obvious differences Discussion in SYP expression were observed (Fig.  8a, b). Next, we Already more than 30 years ago, Aβ was for the first time used western blotting to detect the differences of SYP isolated and proposed to play an important role in AD expression in the hippocampus 2  weeks after LPS stim- pathogenesis [18, 19]. Initial efforts in the development ulation. As displayed in Fig.  8c, d, this confirmed the of a curative AD treatment mainly focused on strate- decrease in SYP expression in the hippocampus upon gies to lower Aβ levels and decrease toxic Aβ aggregates. NL-G-F low-grade peripheral inflammation in App mice. Unfortunately, these attempts are without any clinical (See figure on next page.) Fig. 6 Low-grade peripheral inflammation affects Aβ transport across the blood-CSF barrier in vivo and in vitro. a Quantification of Aβ and 1-40 Aβ in CSF and plasma (n = 5–8). b Expression of the genes P-gp and Lrp2 in CP (n = 5). c Representative images of LRP2 immunostaining of CP. 1-42 Scale bar: 20 μm. d Quantification of the relative red staining of LRP2 in CP (n = 5 per group). e Schematic diagram of Aβ transcytosis analysis in an 5-FAM TAMRA CP epithelial transport system using fluorescently labeled Aβ peptides. f movement of scrambled Aβ (left) and Aβ (middle) from 1-42 1-42 the basolateral to the apical chamber, and normalized Aβ transcytosis quotient of Aβ from the basolateral to the apical chamber (right) (n = 3). 1-42 g Representative images of LRP2 staining in primary CP epithelial cells. h Quantification of the relative red staining of LRP2 in primary CP epithelial cells (n = 3). Mean ± SEM, two-way ANOVA Bonferroni’s post hoc test for multiple comparisons (a, b, d), one-way ANOVA Bonferroni’s post hoc test for multiple comparisons f Nonparametric Mann–Whitney U test h. *p < 0.05, **p < 0.01, ***p < 0.001 Xie et al. acta neuropathol commun (2021) 9:163 Page 12 of 23 Fig. 6 (See legend on previous page.) Xie  et al. acta neuropathol commun (2021) 9:163 Page 13 of 23 success so far, indicating that there is an urgent need for peripheral immune stimulation in contrast to microglia a more in-depth understanding of the onset and pro- located at a greater distance from the plaques; both in the NL-G-F gression of AD pathology. Over the last years, it became hippocampus and cortex of App mice. Indeed, dis- increasingly clear that the innate immune system plays tant microglia showed sustained changes in morphology, a crucial role in the pathogenesis of AD, even in the reflected by a reduction in dendrite length, cell volume, early clinically silent period [17, 24, 64]. Considerable number of segments, branch points and terminal points. efforts have been devoted to understand the interactions These changes were rather moderate 24  h after the last between the peripheral immune system and the CNS [37, LPS challenge but further increased after the second 51, 63] in APP overexpression mouse models. However, LPS challenge. The fact that this is not the case for the the underlying mechanisms are not yet fully understood. microglia in direct contact with the Aβ plaques might With this manuscript, we are the first to report the effects be explained by the fact that these microglia are already of LPS-induced low-grade peripheral inflammation on activated and may not further react to an additional trig- AD pathology in a second-generation mouse model for ger. Next, we showed that the microgliosis marker Aif1 NL-G-F AD, namely the App mouse model. The use of this and microglial activation marker Cd69 are transiently more representative mouse model may help to give a bet- increased in the hippocampus after LPS challenge. This ter understanding of the deleterious impact of peripheral is however in contrast to a previous study that found that immune activation on AD development. Aif1 is significantly downregulated in microglial cells Tejera et al. recently reported that peripheral and CNS upon LPS treatment but is unchanged in the whole brain inflammation related symptoms return to basal condi - [63]. Another study reported an increase in Cd69 expres- tions 10  days after a low dose LPS challenge [64]. Based sion after exposure of glial cells to LPS or IL-4, but the -/- on this study, we used two low-dose LPS injections microglia isolated from CD69 mice still exhibited an separated by 7  days to mimic two discrete infectious enhanced production of proinflammatory mediators [6]. events associated with low-grade peripheral inflamma - Al these results together suggest that an increase in cell tion. Importantly, IL-6 plasma level analysis revealed a number may be responsible for the maintenance or slight NL-G-F slightly stronger inflammatory response in the App increase of Aif1 mRNA levels and CD69 reported in our mice compared to wild type controls and showed in both study. Moreover, CD69 exerts some negative regulation genotypes a reduced impact of the second compared to of inflammation upon exposure to stimulus [6]. the first LPS injection, suggesting partial desensitiza - The expression levels of the pro-inflammatory cytokines tion. Next, we investigated both the short and long-term IL-1β and TNF in the brain increase transiently after LPS effects by analyzing the response 24 h and 2 weeks after challenge. These results are consistent with the results the last LPS injection, respectively. of previous studies in first-generation AD mouse mod - Microglial cells are the most prominent immune cells els. In those studies, LPS treatment of APP overexpress- of the CNS and they continually survey and maintain ing mice induces an increased expression of cytokines homeostasis in the brain [23]. Any changes in the CNS [37, 64]. However, the extent of the LPS effects are dif - lead to activation, proliferation, and morphological ferent from our results. For example, a recent study in 5 changes of these cells [60]. On the one hand, activated and 15 months old APP/PS1 mice combined with a single microglia can clear neuronal damage by phagocytosis injection of LPS shows high pro-inflammatory cytokine while on the other hand, activated microglia will also levels 2 days after LPS injection which returns to baseline release molecules that can initiate neuroinflammation levels 8 days later [64]. The differences between our results [23]. In concordance with previous findings [64], we and this study can be explained by a different experimen - observed that microglia located at the site of Aβ depo- tal setup, including LPS type (ultrapure LPS from S. typh- sition don’t show striking morphological changes upon imurium vs. LPS from S. abortus equi, number of LPS (See figure on next page.) NL-G-F Fig. 7 Low-grade peripheral inflammation impairs microglial phagocytosis of Aβ in App mice. a Representative images of IBA1 and 6E10 staining in hippocampus and cortex. The dotted circle shows the border of Aβ plaques. Scale bar: 20 μm. b Quantification of the IBA1 microglia within the Aβ plaque surface. Hippocampus (left); cortex (right) (n = 5–6; each symbol represents the average from 4–6 plaques in one mouse). c Quantification of the percentage of internalized Aβ in the hippocampus (left graph) and cortex (right) (n = 5–6). d Representative images of IBA1, 6E10 and CD68 staining in hippocampus and cortex. Arrow points to the CD68 microglia. Scale bar in hippocampus: 100 μm and 20 μm (insert); Scale bar in cortex: 20 μm. e Quantification of CD68 microglia (n = 5) in hippocampus (left two graphs) and cortex (right two graphs). f Representative images of relationship between microglial morphology and its distance to Aβ deposition. Scale bars: 20 µm. g Imaris-based quantification of cell morphology of IBA1 microglia in hippocampus. Co-staining of IBA1, CD68 and 6E10 was performed on 5 µm paraffin sections. Each symbol represents the average in one cell and 10–15 cells are analyzed per mouse (n = 4–5). Mean ± SEM, two-way ANOVA Bonferroni’s post hoc test for multiple comparisons (b, c, e), nonparametric Mann–Whitney U test g. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 Xie et al. acta neuropathol commun (2021) 9:163 Page 14 of 23 Fig. 7 (See legend on previous page.) Xie  et al. acta neuropathol commun (2021) 9:163 Page 15 of 23 Fig. 8 Low-grade peripheral inflammation and Aβ induces neuronal dysfunction. a Representative images of SYP staining in the cortex. Scale bar: 10 μm. b Quantification of relative SYP area in cortex (n = 5). c Western blot of SYP protein levels in hippocampus 2 weeks after LPS challenge. d Densitometric analysis of relative protein levels of SYP normalized to β-actin (n = 5). Scale bar: 50 μm. Mean ± SEM, two-way ANOVA Bonferroni’s post hoc test for multiple comparisons. *p < 0.05, **p < 0.01 injections (single vs. twice, once a week). The different the brain. So far, different mechanisms are proposed types of LPS may induce varying degrees of inflammatory among which: transmission by the vagal afferents, periph - response and repeated LPS challenge causes less inflam - eral mediators crossing/transporting via circumventricu- matory response than the first LPS injection. In addition, lar organs (CVOs) or the BBB, or by signaling through IL-1β is produced predominantly by microglia and plays the endothelial cells of the BBB [30]. More recently, the an important role in microglia activation and proliferation CP epithelial cells that form the blood-CSF barrier are [36]. Under chronic peripheral inflammation, microglia identified as a novel player in this regard [2, 38]. Here, may reach an effector phenotype that express low levels we show that in WT mice, low-grade peripheral inflam - of cytokines [51]. In our study, we performed two LPS mation has only limited effects on the blood-CSF barrier injections to mimic long-term chronic peripheral inflam - integrity, both at short and long-term. This is in contrast mation and the second stimulus may lead to an exagger- to our previous study in which we showed that a single ated microglial response. This may explain the results of high dose LPS challenge causes severe blood-CSF bar- sustained microglial activation but transient changes in rier disturbance [66]. However, long-term peripheral inflammation level in our study. inflammation did induce significant changes in TJ pro - A pressing question is how pathological insults that teins expression and therefore in blood-CSF barrier NL-G-F reside in circulation and peripheral organs can get into integrity in App mice, especially 2  weeks after LPS (See figure on next page.) Fig. 9 IL-1β stimulation increases Aβ aggregation in primary neuron culture. a Schematic diagram of kinetics of Aβ aggregation in primary neuron HF488 after stimulation with IL-1β. b Representative images of different conditions at 0, 12 and 24 h after adding PBS/Aβ oligomers. The images 1-42 show an overlay of the color-coded FLIM image and the transmitted image. Scale bar: 20 μm. c Quantification of HF488 fluorescence lifetime of HF488 Aβ oligomers (n = 4). Mean ± SEM, nonparametric Mann–Whitney U test. *p < 0.05 1-42 Xie et al. acta neuropathol commun (2021) 9:163 Page 16 of 23 Fig. 9 (See legend on previous page.) Xie  et al. acta neuropathol commun (2021) 9:163 Page 17 of 23 stimulation. While we previously showed that intracer- peripheral immune stimulation affects Aβ pathology. ebroventricular injection of Aβ oligomers contributes This revealed that low-grade peripheral inflammation 1-42 NL-G-F to loss of blood-CSF barrier integrity [7, 61], App significantly increases Aβ aggregation and in particu - mice showed only a trend in increased blood-CSF barrier lar the formation of small size plaques. Previous stud- leakiness compared to WT mice. This might be explained ies demonstrated that low dose LPS injection(s) lead to by the low Aβ levels (around 0.15  ng/ml) in the CSF increased Aβ deposition in APP/PS1 mice [40, 64] while 1-42 NL-G-F of App mice which is not sufficient to significantly other studies failed to show this LPS effect in first-gen - disrupt blood-CSF barrier integrity compared to the dose eration AD mouse models [31]. The varied outcomes of Aβ oligomers (125 ng/ml) injected in our previous could be due to the experimental setup, including mouse 1-42 studies. In addition, in  vitro experiments revealed that genetic background, age-at-onset of plaque formation, IL-1β exposure of primary CP epithelial cells resulted in age at LPS stimulation, number of injections, LPS dose, a significant barrier disruption as previously observed route of administration, and the time between injec- upon TNF stimulation [5]. These results suggest that tion and sacrifice. In our study, we found that sustained there are mechanisms in the brain to protect the blood- inflammation stimulates microglia to further transform CSF barrier from peripheral inflammation. into a phagocytic phenotype associated with proliferation Loss of brain barrier integrity may lead to peripheral and morphological changes. This is in line with a previ - immune cell infiltration in the brain via paracellular path - ous study which indicated that peripheral inflammation ways. In addition, these cells can also cross the brain bar- reduces the microglial Aβ clearance [64]. riers by the expression of integrin ligands and chemokines In addition, analysis of microglia biodistribution in the by epithelial cells [5]. We identified an increased expres - brain suggests that, shortly after induction of periph- sion of Icam1, Ccl2 and Cxcl10, which are crucial for leu- eral inflammation, microglia migrate to blood vessels, a kocyte trafficking [5 ], in the CP and hippocampus of WT phenomenon that has been observed before [22]. Con- NL-G-F and App mice 24 h after LPS stimulation. These find - sequently, this migration leads to fewer microglia avail- ing are consistent with previous reports [39, 65]. It is there- able for Aβ phagocytosis. Two weeks after induction of fore important to further understand whether the repeated peripheral inflammation, however, more microglia were exposure to a peripheral stimulus ultimately results in the again observed close to the Aβ plaques. Nonetheless, Aβ entry of immune cells into the brain. Based on immuno- aggregation further continued in response to the chronic histochemical stainings, we observed an increase in leu- low-grade peripheral inflammation, which might be the NL-G-F kocyte infiltration in the brain of both WT and App result of the chronically activated microglial cells which mice 2 weeks after the last LPS stimulation. Although we are no longer able to process Aβ [3, 27]. Altogether, our observed an increased infiltration of peripheral immune data indicate that peripheral inflammation induces dys - cells into the brain upon LPS challenges, the infiltration is functional microglia that may account for the increased still very modest. In addition, perivascular macrophages Aβ deposition which we observed in the brains of LPS + − NL-G-F (PVMs) are also IBA1 and TMEM119 and these cells challenged App mice. have been shown to increase in neurodegenerative disease Aβ clearance from the brain is also mediated by a com- models [20]. Therefore, other techniques which can accu - bination of transcellular transport mechanisms across rately distinguish invading peripheral monocytes from the blood–brain and blood-CSF barriers [8, 62]. LRP2 is brain endogenous PVMs and microglia, for example using an efflux transporter expressed at the blood-CSF barrier genetic labeling of different myeloid populations [49], and is involved in the elimination of Aβ across CP epithe- should be used to further validate this result. Our results lial cells [8]. According to our results low-grade chronic are in agreement with a previous study that showed that peripheral inflammation induces a short-term effect on a single low dose LPS challenge is sufficient to induce leu - Aβ transporter expression levels associated with changes kocyte infiltration in the brain of aged APP/PS1 mice [64]. in levels of peripheral and central inflammation. In Moreover, a sustained exposure triggers a constant influx agreement with this, Ott and colleagues previously sug- of leukocytes [65] and this can be caused by brain barrier gested that changes in blood-CSF barrier transport are disruption. On the other hand, one recent study indicates related to the expression of inflammatory cytokines and that leukocytes can also pass directly through epithelial chemokines, such as higher IL-1β and TNF in serum and cells [42]. Our findings suggest that leukocytes can adapt CSF of mild cognitive impairment patients are associated their behavior to different circumstances as there is no with low efficiency of transport small and much larger obvious barrier damage in WT mice but still immune cell molecules in the blood-CSF barrier [44]. IL-1β is pro- infiltration. duced predominantly by microglia and plays an impor- As microglial activation is associated with Aβ deposi- tant role in microglia activation and proliferation [36]. To tion [11], we tested whether in our set-up the low-grade specifically examine the role of IL-1β in the regulation of Xie et al. acta neuropathol commun (2021) 9:163 Page 18 of 23 Aβ efflux, we tested Aβ transcytosis in  vitro after IL-1β [53], which overcome intrinsic drawbacks of the APP stimulation. We show that IL-1β not only disrupts the overexpression mouse models by utilizing an App knock- barrier integrity but also affects Aβ transcytosis by inhib - in strategy were generated to overproduce pathogenic iting, in part, the LRP2 transporter. Indeed, we observed Aβ such as Aβ without overexpressing APP. In agree- 1-42 that inhibition of LPR2 in CP epithelial cells reduces Aβ ment with previous studies in APP overexpression mouse NL-G-F efflux from CSF to blood and sequential treatment of models [31, 40, 57, 64], App mice showed higher anti-LRP2 and IL-1β did not have additive effect on Aβ Aβ deposition during low-grade peripheral inflam - transport. Altogether, our findings indicate that low- mation. Importantly, we identified several pathways grade peripheral inflammation affects Aβ pathology by which are activated upon peripheral inflammation and multiple pathways including microglial phagocytosis and subsequently contribute to aggravation of AD pathol- Aβ transcytosis. ogy, including sustained microglial activation, neuronal Previous evidence suggests that Aβ plays a vital role dysfunction and Aβ efflux from the brain. Our results in the induction of neuronal dysfunction from synapses strengthen previous findings in APP overexpressing mice toward neural networks [45] and Aβ is considered to and emphasize the potential vital role of preventing and/ 1-42 be the most neurotoxic form of Aβ [9]. In concordance or intervening with peripheral inflammation to impact with this, we found that the soluble Aβ levels were AD development and progression. 1-42 elevated 2  weeks after low-grade peripheral inflamma - tion. Interestingly, we identified synapse loss at the same Methods timepoint and showed that in vitro treatment of primary Mice neurons with Aβ oligomers also induces synapse loss. NL-G-F 1-42 The generation of App mice carrying Arctic, Swed- These results suggest that Aβ may mediate synaptic 1-42 ish, and Beyreuther/Iberian mutations was described activity and lead to synapse loss. previously [52]. The colony of these mice was main - Here, we also show that low-grade peripheral inflam - tained in our animal house at VIB Center for Inflamma - mation transient induces total cell and neuron death in tion Research. C57BL/6  J mice were used as a control. NL-G-F NL-G-F WT and App mice. In addition, the App mice Mice were kept in individually ventilated cages under a show a higher percentage of neuronal cell death com- 12-h dark/12-h light cycle in specific pathogen-free ani - pared to WT mice, which might be due to Aβ pathology mal facility and received food and water ad libitum. Both induced cell death as demonstrated in previous stud- male and female mice were used and all mice were aged ies [33, 43]. Microglial cells are important players in the between 20–23  weeks at the start of the experiment. maintenance and plasticity of neuronal circuits, con- All animal studies were conducted in compliance with tributing to the protection and remodeling of synapses. governmental and EU guidelines for the care and use The ability of microglial phagocytosis is essential for the of laboratory animals and were approved by the ethical clearance of apoptotic neurons and regulation of neu- committee of the Faculty of Sciences, Ghent University, ronal activity [23, 36]. In our study, we observed a higher Belgium. percentage of microglia that are in contact with neurons in short term and long term low-grade peripheral inflam - Reagents mation, which might lead to the removal of pre-synaptic LPS from S. abortus equi (L-5886) was obtained from input from neurons as described by Sharma et  al. [56]. Sigma-Aldrich. Recombinant mouse IL-1β was pro- The accumulation of Aβ around synapses has also been duced in E. coli and purified by VIB protein core. IL-1β shown to induce microglial complement activation in had specific activity of 6.12 × 10   IU/mg and no detect- neuronal synapses and subsequent microglial engulfment able endotoxin contamination. Synthetic Aβ , HiLyte [4]. Recent study report that TNF induces extracellular 1-42 Fluor 488-labeled (AS-60479), TAMRA-labeled (AS- Aβ production and aggregation in neuronal cell cultures 60476) and scrambled Aβ , 5-FAM labeled (AS-60892) [70]. Therefore, we hypothesize that pro-inflammatory 1-42 at the N-terminus were purchased from Anaspec. Aβ cytokine IL-1β may be also involved in Aβ aggregation 1-42 monomers and oligomers were prepared according to the by affecting neuronal functions. In line with this assump - manufacturer’s instructions. tion, we found that IL-1β induces accelerated Aβ aggre- gation in neurons. These findings indicate that peripheral inflammation influences neuronal functioning which Induction of low‑grade inflammation subsequently plays an important role in AD pathogenesis. LPS injections (1.0  mg/kg body weight, i.p.) were per- In conclusion, our study sheds light on how low-grade formed on day 0 and 7 as displayed in Fig.  1a. Control peripheral inflammation affects AD pathogenesis in a mice received i.p. PBS injections. Depending on the NL-G-F new mouse model of AD, namely the App model experiment, the mice were sacrificed 24  h or 2  weeks Xie  et al. acta neuropathol commun (2021) 9:163 Page 19 of 23 after the second LPS/PBS injection. Body weight, tem- are given as relative expression values normalized to the perature loss, and sickness behavior were checked the geometric mean of the housekeeping genes. Primers used first 3 days after LPS injection. for qPCR are depicted in Additional file 4: Table 1. TLR4 activation analysis Measurement of Aβ by ELISA The level of TLR4 activation in plasma and brain was The Aβ was extracted and measured using a standard measured using the HEK-Blue mTLR4 assay (InvivoGen) protocol as described previously [61]. Briefly, cortex according to the manufacturer’s instructions. HEK-Blue samples were homogenized in Tissue Protein Extraction mTLR4 cells were seeded at 25,000 cells per well in a Buffer containing complete protease inhibitor (Thermo 96-well plate in detection medium. The brain sample was Scientific) and phosphatase inhibitor tablets (Sigma- homogenized in PBS supplemented with 1% penicillin/ Aldrich) using a Precellys (Bertin Technologies) and streptomycin (Gibco), and brain supernatant was col- subsequently centrifuged at 5000  g for 5  min at 4  °C. lected by centrifugation at 20,000g in a microcentrifuge Supernatant was collected and centrifuged at 4 °C for 1 h at 4 °C for 15 min. The next day, cells were incubated for at 100,000 g (TLA-100 Rotor; Beckman Coulter). Super- 12  h with the different samples (20  μl plasma with 1/5 natant containing soluble Aβ was removed and stored at dilutions and 20  μl brain supernatant corresponding to − 80 °C. The pellet was further processed in GuHCl solu - 50  μg total protein), followed by absorption measure- tion containing complete protease inhibitor, sonicated, ment of the culture medium at 655  nm (iMark Micro- vortexed, incubated for 60  min at 25  °C and centrifuged plate Absorbance Reader, Bio-Rad) and relative TLR4 at 70,000  g for 20  min at 4  °C. Supernatant containing activation was calculated. insoluble Aβ was 12 times diluted with GuHCl diluent and immediately frozen at -80  °C. The levels of Aβ 1-42 Cytokine measurements and Aβ in plasma, CSF, and cortex lysates were deter- 1-40 The hippocampus was dissected and immediately snap- mined using a sandwich ELISA assay. Briefly, the sample- frozen in liquid nitrogen and stored at −  80  °C until detection mixtures were added to the ELISA plate coated use. After homogenizing with lysis buffer (0.5% CHAPS with anti-Aβ (1.5 μg/ml; JRF/cAb42/26) or anti-Aβ 1-42 1-40 in PBS supplemented with complete protease inhibitor antibody (1.5 μg/ml; JRF/cAb40/28) and incubated ON at (Thermo Scientific)), supernatant was collected by cen - 4 °C with slow shaking. Absorption at 450 nm was meas- trifugation at 20,000g in a microcentrifuge at 4  °C for ured after adding substrate solution (BD Biosciences 15 min. Protein concentration of the samples was meas- OptEIA ) followed by stopping buffer (1 M H SO ). The 2 4 ured using the Pierce BCA Protein Assay Kit (Thermo amount of Aβ was determined with GraphPad Prism 8.0 Scientific). Cytokines in plasma and hippocampus lysates using a nonlinear regression model. were measured using V-PLEX Custom Mouse Cytokine Kit for IL-1β, TNF and IL-6 (Meso Scale Discovery) Primary mouse CP epithelial cell isolation according to the manufacturer’s instructions. Signals Primary mouse CP epithelial cells were isolated from were measured on a MESO QuickPlex SQ 120 reader P2-P4 WT pups as previously described [2]. CP tissue (Meso Scale Discovery). was isolated from the lateral and fourth ventricles, pooled and digested with pronase (53,702, Sigma-Aldrich) RNA extraction and RT‑qPCR analysis for 7  min. For monolayer cultures, cells were plated in Samples were snap-frozen in liquid nitrogen and stored 24-well Transwell polyester inserts (pore size, 0.4  μm; at -80 °C until further use. After homogenizing the tissue surface area, 33.6 mm ; Corning) coated with laminin with TRIzol (Invitrogen) in a tube containing zirconium (L2020, Sigma-Aldrich). Cells were grown in DMEM/ oxide beads on a TissueLyser (QIAGEN), chloroform F12 supplemented with 10% FBS, 2  mM L-Glutamine was added and the homogenate was separated into 3 (Gibco), 1% penicillin/streptomycin at 37  °C and 5% phases by centrifugation at 20,000g in a microcentri- CO for 1 to 2  weeks until their TEER values reached a fuge at 4  °C for 15  min. Total RNA was extracted from plateau. the upper aqueous phase using Aurum total RNA kit (Bio-Rad) according to the manufacturer’s instructions. In vitro transcytosis of Aβ 1‑42 The concentration of total RNA was determined by the The Aβ transcytosis across the epithelial monolayers was Nanodrop-1000 (Thermo Scientific) and total RNA was measured in triplicate based on the protocol previously reverse-transcribed into cDNA with SensiFAST cDNA described in [21, 34, 62]. In brief, mouse CP epithelial Synthesis Kit (Bioline). qPCR was performed on the cells (5 × 10 cells per well) were grown to confluence Roche LightCycler 480 System (Applied Biosystems) and treated with PBS or IL-1β (2.5  ng/ml) in the basal ™ ® using SensiFAST SYBR No-ROX Kit (Bioline). Results side of Transwell, followed by 1  h blockage with 15  µg/ Xie et al. acta neuropathol commun (2021) 9:163 Page 20 of 23 ml anti-LRP1 or 15  µg/ml control IgG (10500C, Invitro- the FIFO image mode using 256 time channels (Becker gen). After washing the wells with fresh medium, 1  μM & Hickl GmbH, Berlin). All acquisitions were done TAMRA 5−FAM Aβ monomers and 1 μM scrambled Aβ with the 820  nm laser line of the MaiTai DeepSee mul- 1-42 1-42 monomers in assay medium (DMEM, 5% FBS, 25  mM tiphoton laser with a repetition rate of 80 MHz. A Plan- HEPES (Gibco), 2  mM L-Glutamine) were added to the Apochromat 40x/1.4 Oil DIC objective lens was used and apical chamber of each well. Subsequently, the cells were the scanning resolution was set at 512 × 512 pixels. The incubated at 37  °C. Samples were taken from the baso- selected positions were followed for a total of 24  h and lateral chamber and replaced with fresh medium every every 2 h a FLIM image was captured. All TCSPC images 20  min for 2  h and then transferred to 96-well plates were processed initially using SPCImage version 5.2 (Nunc). The fluorescence was determined using FLU - (Becker & Hickl) and fitted as mono-exponential decays. Ostar Omega reader (BMG LABTECH) at an excitation Further image analysis was carried out using NIH Image J wavelength of 485  nm and an emission wavelength of software. Statistical analysis was performed in GraphPad 520  nm. Relative fluorescence units were converted to Software. The average lifetimes in 4 images with healthy values of ng/ml, using corresponding standard curves, cells were computed and these were then averaged to and were corrected for background fluorescence and obtain the mean lifetime (± SEM) for each particular serial dilutions over the course of the experiment. Trans- time point and treatment. Significant differences between port of Aβ across the monolayer was calculated asTQ treatments were obtained by means of comparison using 1-42 using the following formula: nonparametric Mann–Whitney U test. TQ = Aβ (dQ/dT )/Scrambled Aβ (dQ/dT ). 1−42 1−42 Immunohistochemistry where dQ/dT (μg/s) is the rate of appearance of Aβ on For immunostainings on brain sections, mice were tran- the receiver side after application and is calculated by scardially perfused with ice-cold 4% PFA in PBS. Subse- plotting the cumulative amount (Q) versus time(s). quently, brains were carefully extracted from the skull and split into two halves (mid-sagittal). The right hemi - Primary neuron culture spheres were embedded in Frozen Section Medium Cortical neuronal cells were prepared from embryonic (Thermo Scientific) immediately in cryomolds (Sakura) (E16-E18) mouse pups. To dissociate the cortex tissue, that were frozen on dry ice, and stored at − 80  °C until the tissue was finely chopped and digested in papain further use. The left hemispheres were post-fixed over - (Sigma-Aldrich) for 20  min. Neurons were triturated night 4% PFA in PBS at 4 °C. After dehydration, samples gently with a pipette and passed through the 70  µm cell were embedded in paraffin in cryomolds and stored at strainer (BD Falcon) before plating them onto Poly-d-Ly - RT until further use. The brains were cut into 5 μm slices sine (Thermo Scientific) coated 8-well chamber (iBidi) at paraffin sections (HM 340 E, Thermo Scientific) or 20 μm a density of 1 × 10 cells per well. Cultures were main- cryosections (CryoStar NX70, Thermo Scientific). For tained in Neurobasal medium (Thermo Scientific) con - immunofluorescence staining, sections were permeabi - taining 1% penicillin/streptomycin, 0.5 mM GlutaMAX lized in PBS containing 0.3–0.5% Triton X-100. Following I Supplement (Thermo Scientific) and 2% B27 supple - blocking with goat immunomix (GIM) (5% goat serum, ment (Thermo Scientific) at 37  °C with 5% CO . The 0.1% BSA, 0.3–0.5% Triton X-100 in PBS) at RT for 1  h, absence of astrocytes (< 2%) was confirmed by the virtual sections were incubated with primary Abs in GIM at 4 °C absence of glial fibrillary acidic (GFAP) protein immu - overnight. After washing with PBS, sections were stained nostaining (data not shown). with fluorophore-conjugated secondary Abs in PBS or PBS containing 0.1% Triton X-100 at RT for 1–1.5  h. FLIM Counterstaining was done with Hoechst reagent (Sigma- FLIM experiments were carried out using HF488-labeled Aldrich, 1:1000 in PBS). For immunostainings on pri- Aβ oligomers according to a standard protocol as mary cells, cells were fixed with 2% PFA for 20  min on 1-42 described previously [13]. The images were recorded by ice. Next, cells were washed three times with PBS and time-correlated single photon counting (TCSPC) using permeabilized with 0.1% Triton X-100 for 10 min on ice. Zeiss LSM780 confocal microscope equipped with a Samples were washed with blocking buffer (1% BSA in MaiTai DeepSee multiphoton laser (SpectraPhysics) and PBS) and incubated with primary Abs (diluted in block- a modular FLIM system (Becker & Hickl). The emitted ing buffer) for 2 h at RT/overnight at 4 °C. After washing photons were detected with a hybrid detection module with PBS, cells were stained with fluorophore-conjugated (HPM-100–40). Data was recorded by a Simple-Tau sys- secondary Abs in PBS for 1 h at RT. Next, samples were tem (M1-SPC-150 FLIM module, Becker & Hickl) with counterstained with Hoechst and the sections were the instrument recording software SPCM version 9.52 in Xie  et al. acta neuropathol commun (2021) 9:163 Page 21 of 23 mounted. Confocal laser scanning microscopy was per- proximity of Aβ plaques between 2 groups were ana- formed using Zeiss LSM780 or Zeiss Axioscan Z.1. lyzed using nonparametric Mann–Whitney U test. Com- parisons of percentages of body weight loss between 2 Image analysis groups were analyzed using 2-way repeated measures Quantification was performed using ImageJ soft - ANOVA. Comparison of 2 factors was analyzed using ware (version 1.53c, National Institutes of Health) and 2-way ANOVA with Bonferroni’s post hoc test for mul- reported as area fraction of region of interest. The tiple comparisons for parametric data. GraphPad Prism expression levels of TJ proteins and transporter proteins 8.0 was used for statistical analysis. Differences were con - were calculated as relative protein positive signal inten- sidered significant at P < 0.05. sity divided by protein positive area. Infiltrating immune + − cells were calculated as the number of IBA1 TMEM119 Supplementary Information cells. For inside or outside Aβ plaque-associated micro- The online version contains supplementary material available at https:// doi. org/ 10. 1186/ s40478- 021- 01253-z. glia quantification, IBA1 cells differentiated by whether they had contact with Aβ plaques and were manually Additional file 1. Representative movie of primary neuron culture after counted. For quantification of the Aβ internalization treatment with PBS for 5 h and traced for 24.h + + ratio, the area of Aβ-internalized microglia (Aβ IBA1 ) Additional file 2. Representative movie of primary neuron culture after was normalized to the Aβ-positive area. The expression treatment with PBS for 5 h, followed by theaddition of Aβ HF488 1-42 of synaptic molecule expression was calculated as per- oligomers and traced for 24 h. centage of SYP within the Hoechst area. Additional file 3. Representative movie of primary neuron culture after treatment with IL-1β for 5 h, followed by theaddition of Aβ HF488 1-42 oligomers and traced for 24 h. Microglia 3D reconstruction Additional file 4. Supplementary Figures S1-S12 and Appendix Tables For 3D reconstruction of microglia, 50  μm vibratome S1-S2. brain sections were stained overnight with anti-IBA1 Ab at 4 °C, followed by secondary Abs for 2 h at RT. Z-stack Acknowledgements images were taken with a Zeiss LSM 880 with Fast Airys- We thank M. Mercken (Johnson & Johnson Pharmaceuticals Research and can (Zeiss, Germany), using a Plan-Apochromat 40 × 1.3 Development, Beerse, Belgium) and Bart De Strooper ( VIB-KU Leuven) for providing the antibodies against Aβ (JRF/ABN/24, JRF/cAB40/28, and JRF/ oil DIC UV-IR M27 objective. The 3D reconstructions cAB42/26) for the Aβ ELISA. We thank the VIB BioImaging Core for training, and measurements were done by filament tracing algo - support and access to the instrument park. This work was supported by the rithm from Imaris software (Bitplane). Research Foundation-Flanders (FWO), Chinese Scholarship Council (CSC), The Foundation for Alzheimer’s Research Belgium (SAO-FRA) and the Baillet Latour Fund. The graphical abstract was created with BioRender.com. Western blot Authors’ contributions Hippocampus extracts were prepared in 0.5% CHAPS J.X., N.G., L.V.H., and R.E.V. designed experiments. J.X., N.G., C.V., G.V.I., E.V.W, buffer containing complete protease inhibitor (Thermo C.V.C. and L.V.H performed experiments. J.X., L.V.H., and R.E.V. wrote the manu- Scientific) and centrifuged for 15 min at 20,000g at 4 °C, script. All authors read and approved the final manuscript. and protein concentration was determined using a Pierce BCA Protein Assay Kit (Thermo Scientific). Supernatants Declarations were denatured in 5xLaemmli buffer, separated by SDS- Competing interests PAGE gel electrophoresis, and transferred to nitrocel- The authors have declared that no conflict of interest exists. lulose. Following blocking with Odyssey Blocking buffer Author details (LI-COR Biosciences), the membrane was first incubated VIB-UGent Center for Inflammation Research, Technologiepark -Zwijnaarde 71, with primary Abs, and then with fluorophore-conjugated 9052 Ghent, Belgium. Department of Biomedical Molecular Biology, Ghent secondary Abs. Protein bands were visualized Odyssey Fc University, 9000 Ghent, Belgium. VIB BioImaging Core, VIB, Ghent, Belgium. Imaging System (LI-COR Biosciences) and quantification Received: 26 May 2021 Accepted: 1 September 2021 was done in Image Studio (LI-COR Biosciences). Statistics Data are shown as mean ± SEM. 2-tailed Student’s t test References for parametric data was used to compare 2 groups. To 1. Alzheimer’s disease facts and figures. Alzheimers Dement (2020). https:// doi. org/ 10. 1002/ alz. 12068 compare 3 or more groups, 1-way ANOVA with Bon- 2. Balusu S, Van Wonterghem E, De Rycke R, Raemdonck K, Stremersch S, ferroni’s post hoc test for multiple comparisons for Gevaert K, Brkic M, Demeestere D, Vanhooren V, Hendrix A et al (2016) parametric data was used. Comparisons of the relation- Identification of a novel mechanism of blood-brain communication during peripheral inflammation via choroid plexus-derived extracellular ship between the activation state of microglia and the Xie et al. acta neuropathol commun (2021) 9:163 Page 22 of 23 vesicles. EMBO Mol Med 8:1162–1183. https:// doi. org/ 10. 15252/ emmm. 20. Gomez-Nicola D, Schetters ST, Perry VH (2014) Differential role of CCR2 in 20160 6271 the dynamics of microglia and perivascular macrophages during prion 3. Bard F, Cannon C, Barbour R, Burke RL, Games D, Grajeda H, Guido T, Hu disease. Glia 62:1041–1052. https:// doi. org/ 10. 1002/ glia. 22660 K, Huang J, Johnson-Wood K et al (2000) Peripherally administered anti- 21. Gu H, Zhong Z, Jiang W, Du E, Dodel R, Liu J, Farlow MR, Zheng W, Du Y bodies against amyloid beta-peptide enter the central nervous system (2014) The role of choroid plexus in IVIG-induced beta-amyloid clearance. and reduce pathology in a mouse model of Alzheimer disease. Nat Med Neuroscience 270:168–176. https:// doi. org/ 10. 1016/j. neuro scien ce. 2014. 6:916–919. https:// doi. org/ 10. 1038/ 7868204. 011 4. Bartels T, De Schepper S, Hong S (2020) Microglia modulate neurode- 22. Haruwaka K, Ikegami A, Tachibana Y, Ohno N, Konishi H, Hashimoto A, generation in Alzheimer’s and Parkinson’s diseases. Science 370:66–69. Matsumoto M, Kato D, Ono R, Kiyama H et al (2019) Dual microglia effects https:// doi. org/ 10. 1126/ scien ce. abb85 87 on blood brain barrier permeability induced by systemic inflammation. 5. Baruch K, Kertser A, Porat Z, Schwartz M (2015) Cerebral nitric oxide Nat Commun 10:5816. https:// doi. org/ 10. 1038/ s41467- 019- 13812-z represses choroid plexus NFkappaB-dependent gateway activity for 23. Heneka MT, Carson MJ, El Khoury J, Landreth GE, Brosseron F, Feinstein leukocyte trafficking. EMBO J 34:1816–1828. https:// doi. org/ 10. 15252/ DL, Jacobs AH, Wyss-Coray T, Vitorica J, Ransohoff RM et al (2015) Neuro - embj. 20159 1468 inflammation in Alzheimer’s disease. Lancet Neurol 14:388–405. https:// 6. Brait VH, Miro-Mur F, Perez-de-Puig I, Notario L, Hurtado B, Pedragosa J, doi. org/ 10. 1016/ S1474- 4422(15) 70016-5 Gallizioli M, Jimenez-Altayo F, Arbaizar-Rovirosa M, Otxoa-de-Amezaga A 24. Heneka MT, Golenbock DT, Latz E (2015) Innate immunity in Alzheimer’s et al (2019) CD69 plays a beneficial role in ischemic stroke by dampening disease. Nat Immunol 16:229–236. https:// doi. org/ 10. 1038/ ni. 3102 endothelial activation. Circ Res 124:279–291. https:// doi. org/ 10. 1161/ 25. Heneka MT, Kummer MP, Latz E (2014) Innate immune activation in neu- CIRCR ESAHA. 118. 313818 rodegenerative disease. Nat Rev Immunol 14:463–477. https:// doi. org/ 10. 7. Brkic M, Balusu S, Van Wonterghem E, Gorle N, Benilova I, Kremer A, 1038/ nri37 05 Van Hove I, Moons L, De Strooper B, Kanazir S et al (2015) Amyloid beta 26. Heppner FL, Ransohoff RM, Becher B (2015) Immune attack: the role of oligomers disrupt blood-CSF barrier integrity by activating matrix metal- inflammation in Alzheimer disease. Nat Rev Neurosci 16:358–372. https:// loproteinases. J Neurosci 35:12766–12778. https:// doi. org/ 10. 1523/ JNEUR doi. org/ 10. 1038/ nrn38 80 OSCI. 0006- 15. 2015 27. Hickman SE, Allison EK, El Khoury J (2008) Microglial dysfunction and 8. Carro E, Spuch C, Trejo JL, Antequera D, Torres-Aleman I (2005) Choroid defective beta-amyloid clearance pathways in aging Alzheimer’s disease plexus megalin is involved in neuroprotection by serum insulin-like mice. J Neurosci 28:8354–8360. https:// doi. org/ 10. 1523/ JNEUR OSCI. 0616- growth factor I. J Neurosci 25:10884–10893. https:// doi. org/ 10. 1523/ 08. 2008 JNEUR OSCI. 2909- 05. 2005 28. Holmes C (2013) Review: systemic inflammation and Alzheimer’s disease. 9. Chen YR, Glabe CG (2006) Distinct early folding and aggregation proper- Neuropathol Appl Neurobiol 39:51–68. https:// doi. org/ 10. 1111/j. 1365- ties of Alzheimer amyloid-beta peptides Abeta40 and Abeta42: stable 2990. 2012. 01307.x trimer or tetramer formation by Abeta42. J Biol Chem 281:24414–24422. 29. Holmes C, Cunningham C, Zotova E, Woolford J, Dean C, Kerr S, Culliford https:// doi. org/ 10. 1074/ jbc. M6023 63200 D, Perry VH (2009) Systemic inflammation and disease progression in 10. Cummings J, Lee G, Ritter A, Sabbagh M, Zhong K (2020) Alzheimer’s Alzheimer disease. Neurology 73:768–774. https:// doi. org/ 10. 1212/ WNL. disease drug development pipeline: 2020. Alzheimers Dement (N Y ) 0b013 e3181 b6bb95 6:e12050. https:// doi. org/ 10. 1002/ trc2. 12050 30. Hoogland IC, Houbolt C, van Westerloo DJ, van Gool WA, van de Beek 11. Dani M, Wood M, Mizoguchi R, Fan Z, Walker Z, Morgan R, Hinz R, Biju M, D (2015) Systemic inflammation and microglial activation: systematic Kuruvilla T, Brooks DJ et al (2018) Microglial activation correlates in vivo review of animal experiments. J Neuroinflammation 12:114. https:// doi. with both tau and amyloid in Alzheimer’s disease. Brain 141:2740–2754. org/ 10. 1186/ s12974- 015- 0332-6 https:// doi. org/ 10. 1093/ brain/ awy188 31. Jendresen C, Digre A, Cui H, Zhang X, Vlodavsky I, Li JP, Nilsson LNG (2019) 12. DeTure MA, Dickson DW (2019) The neuropathological diagnosis of Systemic LPS-induced Abeta-solubilization and clearance in AbetaPP- Alzheimer’s disease. Mol Neurodegener 14:32. https:// doi. org/ 10. 1186/ transgenic mice is diminished by heparanase overexpression. Sci Rep s13024- 019- 0333-5 9:4600. https:// doi. org/ 10. 1038/ s41598- 019- 40999-4 13. Esbjorner EK, Chan F, Rees E, Erdelyi M, Luheshi LM, Bertoncini CW, 32. Ji K, Akgul G, Wollmuth LP, Tsirka SE (2013) Microglia actively regulate the Kaminski CF, Dobson CM, Kaminski Schierle GS (2014) Direct observations number of functional synapses. PLoS ONE 8:e56293. https:// doi. org/ 10. of amyloid beta self-assembly in live cells provide insights into differ -1371/ journ al. pone. 00562 93 ences in the kinetics of Abeta(1–40) and Abeta(1–42) aggregation. Chem 33. Kadowaki H, Nishitoh H, Urano F, Sadamitsu C, Matsuzawa A, Takeda K, Biol 21:732–742. https:// doi. org/ 10. 1016/j. chemb iol. 2014. 03. 014 Masutani H, Yodoi J, Urano Y, Nagano T et al (2005) Amyloid beta induces 14. Frost GR, Jonas LA, Li YM (2019) Friend, foe or both? Immune activity neuronal cell death through ROS-mediated ASK1 activation. Cell Death in Alzheimer’s disease. Front Aging Neurosci 11:337. https:// doi. org/ 10. Differ 12:19–24. https:// doi. org/ 10. 1038/ sj. cdd. 44015 28 3389/ fnagi. 2019. 00337 34. Keaney J, Walsh DM, O’Malley T, Hudson N, Crosbie DE, Loftus T, Sheehan 15. Fujiyoshi M, Tachikawa M, Ohtsuki S, Ito S, Uchida Y, Akanuma S, Kamiie J, F, McDaid J, Humphries MM, Callanan JJ et al (2015) Autoregulated para- Hashimoto T, Hosoya K, Iwatsubo T et al (2011) Amyloid-beta pep- cellular clearance of amyloid-beta across the blood-brain barrier. Sci Adv tide(1–40) elimination from cerebrospinal fluid involves low-density 1:e1500472. https:// doi. org/ 10. 1126/ sciadv. 15004 72 lipoprotein receptor-related protein 1 at the blood-cerebrospinal fluid 35. Labzin LI, Heneka MT, Latz E (2018) Innate Immunity and Neurode- barrier. J Neurochem 118:407–415. https:// doi. org/ 10. 1111/j. 1471- 4159. generation. Annu Rev Med 69:437–449. https:// doi. org/ 10. 1146/ annur 2011. 07311.xev- med- 050715- 104343 16. Furube E, Kawai S, Inagaki H, Takagi S, Miyata S (2018) Brain Region- 36. Leng F, Edison P (2020) Neuroinflammation and microglial activation in dependent Heterogeneity and Dose-dependent Difference in Transient Alzheimer disease: where do we go from here? Nat Rev Neurol. https:// Microglia Population Increase during Lipopolysaccharide-induced Inflam-doi. org/ 10. 1038/ s41582- 020- 00435-y mation. Sci Rep 8:2203. https:// doi. org/ 10. 1038/ s41598- 018- 20643-3 37. Lopez-Rodriguez AB, Hennessy E, Murray CL, Nazmi A, Delaney HJ, Healy 17. Gandy S (2012) Lifelong management of amyloid-beta metabolism to D, Fagan SG, Rooney M, Stewart E, Lewis A et al (2021) Acute systemic prevent Alzheimer’s disease. N Engl J Med 367:864–866. https:// doi. org/ inflammation exacerbates neuroinflammation in Alzheimer’s disease: 10. 1056/ NEJMe 12079 95 IL-1beta drives amplified responses in primed astrocytes and neuronal 18. Glenner GG, Wong CW (1984) Alzheimer’s disease: initial report of the network dysfunction. Alzheimers Dement. https:// doi. org/ 10. 1002/ alz. purification and characterization of a novel cerebrovascular amyloid 12341 protein. Biochem Biophys Res Commun 120:885–890. https:// doi. org/ 10. 38. Marques F, Sousa JC (2015) The choroid plexus is modulated by various 1016/ S0006- 291X(84) 80190-4 peripheral stimuli: implications to diseases of the central nervous system. 19. Goate A, Chartier-Harlin MC, Mullan M, Brown J, Crawford F, Fidani L, Giuf- Front Cell Neurosci 9:136. https:// doi. org/ 10. 3389/ fncel. 2015. 00136 fra L, Haynes A, Irving N, James L et al (1991) Segregation of a missense 39. Marques F, Sousa JC, Coppola G, Geschwind DH, Sousa N, Palha JA, mutation in the amyloid precursor protein gene with familial Alzheimer’s Correia-Neves M (2009) The choroid plexus response to a repeated disease. Nature 349:704–706. https:// doi. org/ 10. 1038/ 34970 4a0 Xie  et al. acta neuropathol commun (2021) 9:163 Page 23 of 23 peripheral inflammatory stimulus. BMC Neurosci 10:135. https:// doi. org/ 58. Sheppard O, Coleman MP, Durrant CS (2019) Lipopolysaccharide-induced 10. 1186/ 1471- 2202- 10- 135 neuroinflammation induces presynaptic disruption through a direct 40. Michaud JP, Halle M, Lampron A, Theriault P, Prefontaine P, Filali M, action on brain tissue involving microglia-derived interleukin 1 beta. J Tribout-Jover P, Lanteigne AM, Jodoin R, Cluff C et al (2013) Toll-like recep - Neuroinflamm 16:106. https:// doi. org/ 10. 1186/ s12974- 019- 1490-8 tor 4 stimulation with the detoxified ligand monophosphoryl lipid A 59. Sierksma A, Lu A, Mancuso R, Fattorelli N, Thrupp N, Salta E, Zoco J, Blum improves Alzheimer’s disease-related pathology. Proc Natl Acad Sci USA D, Buee L, De Strooper B et al (2020) Novel Alzheimer risk genes deter- 110:1941–1946. https:// doi. org/ 10. 1073/ pnas. 12151 65110 mine the microglia response to amyloid-beta but not to TAU pathology. 41. Neher JJ, Neniskyte U, Hornik T, Brown GC (2014) Inhibition of UDP/P2Y6 EMBO Mol Med 12:e10606. https:// doi. org/ 10. 15252/ emmm. 20191 0606 purinergic signaling prevents phagocytosis of viable neurons by acti- 60. Sochocka M, Diniz BS, Leszek J (2017) Inflammatory response in the CNS: vated microglia in vitro and in vivo. Glia 62:1463–1475. https:// doi. org/ 10. friend or foe? Mol Neurobiol 54:8071–8089. https:// doi. org/ 10. 1007/ 1002/ glia. 22693s12035- 016- 0297-1 42. Nishihara H, Soldati S, Mossu A, Rosito M, Rudolph H, Muller WA, 61. Steeland S, Gorle N, Vandendriessche C, Balusu S, Brkic M, Van Cauwen- Latorre D, Sallusto F, Sospedra M, Martin R et al (2020) Human CD4(+) berghe C, Van Imschoot G, Van Wonterghem E, De Rycke R, Kremer A T cell subsets differ in their abilities to cross endothelial and epithelial et al (2018) Counteracting the effects of TNF receptor-1 has therapeutic brain barriers in vitro. Fluids Barriers CNS 17:3. https:// doi. org/ 10. 1186/ potential in Alzheimer’s disease. EMBO Mol Med 10:e8300. https:// doi. s12987- 019- 0165-2org/ 10. 15252/ emmm. 20170 8300 43. Orellana JA, Shoji KF, Abudara V, Ezan P, Amigou E, Saez PJ, Jiang JX, Naus 62. Storck SE, Meister S, Nahrath J, Meissner JN, Schubert N, Di Spiezio A, CC, Saez JC, Giaume C (2011) Amyloid beta-induced death in neurons Baches S, Vandenbroucke RE, Bouter Y, Prikulis I et al (2016) Endothelial involves glial and neuronal hemichannels. J Neurosci 31:4962–4977. LRP1 transports amyloid-beta(1–42) across the blood-brain barrier. J Clin https:// doi. org/ 10. 1523/ JNEUR OSCI. 6417- 10. 2011 Invest 126:123–136. https:// doi. org/ 10. 1172/ JCI81 108 44. Ott BR, Jones RN, Daiello LA, de la Monte SM, Stopa EG, Johanson CE, 63. Swartzlander DB, Propson NE, Roy ER, Saito T, Saido T, Wang B, Zheng Denby C, Grammas P (2018) Blood-cerebrospinal fluid barrier gradients H (2018) Concurrent cell type-specific isolation and profiling of mouse in mild cognitive impairment and Alzheimer’s Disease: relationship to brains in inflammation and Alzheimer’s disease. JCI Insight 3:e121109. inflammatory cytokines and chemokines. Front Aging Neurosci 10:245. https:// doi. org/ 10. 1172/ jci. insig ht. 121109 https:// doi. org/ 10. 3389/ fnagi. 2018. 00245 64. Tejera D, Mercan D, Sanchez-Caro JM, Hanan M, Greenberg D, Soreq H, 45. Palop JJ, Mucke L (2010) Amyloid-beta-induced neuronal dysfunction in Latz E, Golenbock D, Heneka MT (2019) Systemic inflammation impairs Alzheimer’s disease: from synapses toward neural networks. Nat Neurosci microglial Abeta clearance through NLRP3 inflammasome. EMBO J 13:812–818. https:// doi. org/ 10. 1038/ nn. 2583 38:e101064. https:// doi. org/ 10. 15252/ embj. 20181 01064 46. Park JC, Han SH, Mook-Jung I (2020) Peripheral inflammatory biomarkers 65. Thomson CA, McColl A, Graham GJ, Cavanagh J (2020) Sustained expo- in Alzheimer’s disease: a brief review. BMB Rep 53: 10–19. https:// doi. org/ sure to systemic endotoxin triggers chemokine induction in the brain 10. 5483/ BMBRep. 2020. 53.1. 309 followed by a rapid influx of leukocytes. J Neuroinflamm 17:94. https:// 47. Pascale CL, Miller MC, Chiu C, Boylan M, Caralopoulos IN, Gonzalez L, doi. org/ 10. 1186/ s12974- 020- 01759-8 Johanson CE, Silverberg GD (2011) Amyloid-beta transporter expres- 66. Vandenbroucke RE, Dejonckheere E, Van Lint P, Demeestere D, Van sion at the blood-CSF barrier is age-dependent. Fluids Barriers CNS 8:21. Wonterghem E, Vanlaere I, Puimege L, Van Hauwermeiren F, De Rycke R, https:// doi. org/ 10. 1186/ 2045- 8118-8- 21 Mc Guire C et al (2012) Matrix metalloprotease 8-dependent extracellular 48. Qin L, Wu X, Block ML, Liu Y, Breese GR, Hong JS, Knapp DJ, Crews FT matrix cleavage at the blood-CSF barrier contributes to lethality during (2007) Systemic LPS causes chronic neuroinflammation and progressive systemic inflammatory diseases. J Neurosci 32:9805–9816. https:// doi. neurodegeneration. Glia 55:453–462. https:// doi. org/ 10. 1002/ glia. 20467org/ 10. 1523/ JNEUR OSCI. 0967- 12. 2012 49. Reed-Geaghan EG, Croxford AL, Becher B, Landreth GE (2020) Plaque- 67. Walker KA, Ficek BN, Westbrook R (2019) Understanding the Role of associated myeloid cells derive from resident microglia in an Alzheimer’s Systemic Inflammation in Alzheimer’s Disease. ACS Chem Neurosci disease model. J Exp Med. https:// doi. org/ 10. 1084/ jem. 20191 374 10:3340–3342. https:// doi. org/ 10. 1021/ acsch emneu ro. 9b003 33 50. Ries M, Sastre M (2016) Mechanisms of Abeta clearance and degradation 68. Walker KA, Gottesman RF, Wu A, Knopman DS, Gross AL, Mosley TH Jr, by glial cells. Front Aging Neurosci 8:160. https:// doi. org/ 10. 3389/ fnagi. Selvin E, Windham BG (2019) Systemic inflammation during midlife and 2016. 00160 cognitive change over 20 years: The ARIC Study. Neurology 92:e1256– 51. Riester K, Brawek B, Savitska D, Frohlich N, Zirdum E, Mojtahedi N, Heneka e1267. https:// doi. org/ 10. 1212/ WNL. 00000 00000 007094 MT, Garaschuk O (2020) In vivo characterization of functional states of 69. Wang Y, Jin S, Sonobe Y, Cheng Y, Horiuchi H, Parajuli B, Kawanokuchi cortical microglia during peripheral inflammation. Brain Behav Immun J, Mizuno T, Takeuchi H, Suzumura A (2014) Interleukin-1beta induces 87:243–255. https:// doi. org/ 10. 1016/j. bbi. 2019. 12. 007 blood-brain barrier disruption by downregulating Sonic hedgehog in 52. Saito T, Matsuba Y, Mihira N, Takano J, Nilsson P, Itohara S, Iwata N, Saido astrocytes. PLoS ONE 9:e110024. https:// doi. org/ 10. 1371/ journ al. pone. TC (2014) Single App knock-in mouse models of Alzheimer’s disease. Nat 01100 24 Neurosci 17:661–663. https:// doi. org/ 10. 1038/ nn. 3697 70. Whiten DR, Brownjohn PW, Moore S, De S, Strano A, Zuo Y, Haneklaus M, 53. Sasaguri H, Nilsson P, Hashimoto S, Nagata K, Saito T, De Strooper B, Hardy Klenerman D, Livesey FJ (2020) Tumour necrosis factor induces increased J, Vassar R, Winblad B, Saido TC (2017) APP mouse models for Alzheimer’s production of extracellular amyloid-β- and α-synuclein-containing aggre- disease preclinical studies. EMBO J 36:2473–2487. https:// doi. org/ 10. gates by human Alzheimer’s disease neurons. Brain Communications 2: 15252/ embj. 20179 7397 fcaa146. https:// doi. org/ 10. 1093/ brain comms/ fcaa1 46 54. Semmler A, Frisch C, Debeir T, Ramanathan M, Okulla T, Klockgether T, 71. You LH, Yan CZ, Zheng BJ, Ci YZ, Chang SY, Yu P, Gao GF, Li HY, Dong TY, Heneka MT (2007) Long-term cognitive impairment, neuronal loss and Chang YZ (2017) Astrocyte hepcidin is a key factor in LPS-induced neu- reduced cortical cholinergic innervation after recovery from sepsis in a ronal apoptosis. Cell Death Dis 8:e2676. https:// doi. org/ 10. 1038/ cddis. rodent model. Exp Neurol 204:733–740. https:// doi. org/ 10. 1016/j. expne 2017. 93 urol. 2007. 01. 003 72. Yuan P, Condello C, Keene CD, Wang Y, Bird TD, Paul SM, Luo W, Colonna 55. Semmler A, Okulla T, Sastre M, Dumitrescu-Ozimek L, Heneka MT (2005) M, Baddeley D, Grutzendler J (2016) TREM2 haplodeficiency in mice and Systemic inflammation induces apoptosis with variable vulnerability of humans impairs the microglia barrier function leading to decreased different brain regions. J Chem Neuroanat 30:144–157. https:// doi. org/ 10. amyloid compaction and severe axonal dystrophy. Neuron 92:252–264. 1016/j. jchem neu. 2005. 07. 003https:// doi. org/ 10. 1016/j. neuron. 2016. 09. 016 56. Sharma K, Wu LJ, Eyo UB (2020) Calming neurons with a microglial touch. Trends Neurosci 43:197–199. https:// doi. org/ 10. 1016/j. tins. 2020. 01. 008 Publisher’s Note 57. Sheng JG, Bora SH, Xu G, Borchelt DR, Price DL, Koliatsos VE (2003) Springer Nature remains neutral with regard to jurisdictional claims in pub- Lipopolysaccharide-induced-neuroinflammation increases intracellular lished maps and institutional affiliations. accumulation of amyloid precursor protein and amyloid beta peptide in APPswe transgenic mice. Neurobiol Dis 14:133–145. https:// doi. org/ 10. 1016/ s0969- 9961(03) 00069-x http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Acta Neuropathologica Communications Springer Journals

Loading next page...
 
/lp/springer-journals/low-grade-peripheral-inflammation-affects-brain-pathology-in-the-appnl-bzQLNRaAjQ
Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2021
eISSN
2051-5960
DOI
10.1186/s40478-021-01253-z
Publisher site
See Article on Publisher Site

Abstract

Alzheimer’s disease (AD) is a chronic neurodegenerative disease characterized by the accumulation of amyloid β (Aβ) and neurofibrillary tangles. The last decade, it became increasingly clear that neuroinflammation plays a key role in both the initiation and progression of AD. Moreover, also the presence of peripheral inflammation has been exten- sively documented. However, it is still ambiguous whether this observed inflammation is cause or consequence of AD pathogenesis. Recently, this has been studied using amyloid precursor protein (APP) overexpression mouse models of AD. However, the findings might be confounded by APP-overexpression artifacts. Here, we investigated the effect NL-G-F of low-grade peripheral inflammation in the APP knock-in (App ) mouse model. This revealed that low-grade peripheral inflammation affects (1) microglia characteristics, (2) blood-cerebrospinal fluid barrier integrity, (3) periph- eral immune cell infiltration and (4) Aβ deposition in the brain. Next, we identified mechanisms that might cause this effect on AD pathology, more precisely Aβ efflux, persistent microglial activation and insufficient Aβ clearance, neu- ronal dysfunction and promotion of Aβ aggregation. Our results further strengthen the believe that even low-grade peripheral inflammation has detrimental effects on AD progression and may further reinforce the idea to modulate peripheral inflammation as a therapeutic strategy for AD. Keywords: Low-grade peripheral inflammation, Brain barriers, Choroid plexus, Blood-CSF barrier, Alzheimer’s disease Introduction [1]. Worldwide, nearly 50 million people have AD or Alzheimer’s disease (AD) is a devastating age-related related dementia, and this number will multiply in the neurodegenerative disorder that is characterized by the next decades [1]. The speed of disease progression is sub - progressive and disabling deficits in cognitive functions jective to individual variability, but patients are estimated including reasoning, attention, judgment, comprehen- to live from a few up to 20 years after their diagnosis [1]. sion, memory and language. AD is the most common Next to dramatically affecting the life quality and expec - form of dementia and may contribute to 60–70% of cases tancy of patients, the disease also takes its toll on our healthcare system and is becoming one of the most eco- nomically taxing diseases in developed countries. Unfor- *Correspondence: Roosmarijn.Vandenbroucke@irc.VIB-UGent.be tunately, only symptomatic medication that is effective Lien Van Hoecke and Roosmarijn E. Vandenbroucke share senior authorship for some AD patients is available, but no cure nor treat- VIB-UGent Center for Inflammation Research, ment to reverse or even halt disease progression exists. Technologiepark-Zwijnaarde 71, 9052 Ghent, Belgium Full list of author information is available at the end of the article © The Author(s) 2021. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Xie et al. acta neuropathol commun (2021) 9:163 Page 2 of 23 It is already well-known for decades that the depo- the microglial cells. Microglia might also be primed in sition of amyloid-beta (Aβ) protein in senile plaques response to peripheral immune reaction [64]. outside neurons and the formation of neurofibrillary While numerous studies have looked into the pres- tangles (NFT) composed of hyperphosphorylated Tau ence of inflammation in AD using mouse models and (p-Tau) protein inside neurons result in the loss of syn- patient samples [46], only a limited amount of studies apses and neurodegeneration which ultimately leads to have looked into the direct impact of peripheral inflam - symptoms associated with AD [12]. The steady progress mation on AD pathology [29, 31, 68]. Among them is the in the understanding of the etiopathogenesis has led recent publication of Tejera et al. showing that peripheral to the evaluation of therapies aiming to reduce patho- inflammation alters Aβ pathology by negatively regulat - logical aggregates of either Aβ or p-Tau. Unfortunately, ing microglial clearance capacity [64]. Importantly, all none of these strategies has led to clinical success [10]. research performed so far on the effect of peripheral As a consequence, a more in-depth and more compre- inflammation on AD pathology is performed in mouse hensive understanding of the AD pathology is crucial models that overexpress the Aβ precursor protein (APP) for the development of novel effective therapies. in combinations with different familial AD (FAD) asso - Recently, emerging evidence suggests that innate ciated mutations in APP or presenilin 1 (PS1), such as immune activation plays a crucial role in the pathogen- Tg2576, APP/PS1, 5xFAD and 3xTg-AD [53]. These so esis and progression of AD [14, 24, 35]. For example, called first-generation transgenic mouse models exhibit genome-wide association studies (GWAS) have dem- AD pathology, but the overexpression may cause addi- onstrated that genes for immune receptors including tional phenotypes unrelated to AD. In addition, the CR1, CLU, CD33, and TREM2 are associated with AD mouse models use the neuron-specific Thy1-promotor to development[26]. More recently, Sierksma et  al. also overexpress the N-terminal truncated Aβ species, which identified SYK, GRN, SLC2A5, PYDC1, HEXB, and makes APP processing even less similar to the human BLNK as risk genes[59]. All these identified risk genes situation. Contrary, second-generation mouse models are involved in the regulation of the immune response utilize an APP knock-in strategy that closer represents within the central nervous system (CNS) but remark- the physiological accumulation of Aβ without pheno- ably also outside the CNS. Moreover, epidemiologi- types related to overexpression [52]. In this study, we are cal and translational research suggests that peripheral the first to report on the effects of low-grade peripheral inflammation may promote AD pathology [67]. All inflammation in the more representative second-genera - NL-G-F these findings support a substantial involvement of tion mouse models for AD, namely the App mouse both peripheral and central immune function in AD model. Our data reveal that, in agreement with the study pathogenesis. Consequently, understanding the con- of Tejera et  al. [64], microglial activity is also affected nections between the immune system and AD develop- using this AD mouse model. Besides, our results reveal ment might be key in our search for therapies against that upon low-grade peripheral inflammation there is an AD. influx of myeloid cells into the brain and a disruption of The principal resident immune cell of the CNS is the the blood-cerebrospinal fluid (CSF) barrier. Moreover, microglia. These phagocytotic cells are ubiquitously we demonstrate that not only microglial Aβ clearance distributed in the brain and patrol their assigned brain is affected by low-grade peripheral inflammation but regions for the presence of pathogens and cell debris also Aβ transport across the brain barriers and neuronal [23]. Moreover, microglial cells provide factors that functioning. support overall tissue maintenance and plasticity of neuronal circuits [32]. However, when homeostasis is Results disrupted, e.g. in response to inflammation, microglia Low‑grade peripheral inflammation induces adopt an activated state which is characterized by an neuroinflammation in second‑generation Alzheimer’s amoeba-like structure, an increase in proinflammatory disease mouse model. cytokine expression and a higher phagocytic activity During the last years it became increasingly clear that the [64]. If such an imbalance in homeostasis persists, the induction of peripheral inflammation causes an immune microglia cells trigger an exaggerated inflammatory response in the CNS [48, 54, 55], while its impact on e.g. response leading to a sustained exposure of neurons to Alzheimer-like pathology is less well studied. Here, we pro-inflammatory mediators, with neuronal dysfunc - used two i.p. injections (day 0 and day 7) of a low LPS tion and cell death as a consequence [25]. During aging dose (1.0  mg/kg body weight) to study how low-grade and neurodegeneration, microglia show enhanced sen- peripheral inflammation may affect AD pathology in sitivity to inflammatory stimuli, so called priming of NL-G-F 20–23 weeks old App mice 24 h (day 8) and 2 weeks (day 21) after the last LPS injection (Fig. 1a). Xie  et al. acta neuropathol commun (2021) 9:163 Page 3 of 23 Fig. 1 LPS induces transient peripheral inflammation and neuroinflammation. a Schematic representation of the experimental design. b Protein levels of IL-1β, TNF and IL-6 in plasma and hippocampus (n = 5–10). c Expression of the pro-inflammatory genes Il1β, Tnf and Il6 in hippocampus (n = 5–10). d Analysis of relative TLR4 activation by plasma and brain lysate (n = 9–21). Mean ± SEM, two-way ANOVA Bonferroni’s post hoc test for multiple comparisons. *p < 0.05, **p < 0.01, ***p < 0.001 NL-G-F Both WT and App mice show a drop in body Next, we looked into the effect of low-grade periph - weight and an increase in plasma IL-6 levels in eral inflammation on neuropathological changes in both NL-G-F response to both LPS injections (Additional file  4: WT and App mice. We investigated the effect on (1) Fig.  1). The IL-6 response is slightly more pronounced microglia characteristics, (2) influx in the CNS of periph - NL-G-F in the App compared to the wild type mice, and eral myeloid cells and (3) integrity of the blood-CSF in all case we observed a less strong inflammatory barrier. response to the second LPS injection. Figure  1b shows elevated levels of the pro-inflammatory cytokines Low‑grade peripheral inflammation affects microglia IL-1β, TNF and IL-6 in the plasma 24 h after LPS injec- characteristics. tion (Fig.  1b). In contrast, all cytokines were back at Microglia, the brain-resident immune cells, are the key baseline levels 2  weeks later. Also, neuroinflammation players in regulating central inflammation. Tejera et  al. was observed by an increase in protein and mRNA lev- recently described that microglial cells of WT mice show els of IL-1β and TNF, but not IL-6, in the hippocampus morphological signs of activation 24  h after LPS injec- (Fig.  1b, c). Similar to the peripheral cytokine levels, tion [64]. Here, we elaborated further on the effect of also the increase in brain cytokine levels was transient low-grade peripheral inflammation on brain inflamma - NL-G-F as this increase was detectable  24  h after LPS injec- tion in WT mice and App mice by studying micro- tion while the levels were normalized both in WT mice glia proliferation and activation. Moreover, we compared NL-G-F and App mice 2 weeks later. The increase in brain microglia characteristics when peripheral inflammation inflammation at 24 h is not due to the presence of LPS is still ongoing (24 h after the last low dose LPS injection) in the brain as we could not detect TLR4 activation in to when peripheral inflammation is resolved (2  weeks brain lysates, despite the fact that LPS is still present in after the last LPS injection). Our analysis revealed more NL-G-F the blood at that timepoint (Fig. 1d). microglia and increased proliferation in App mice Xie et al. acta neuropathol commun (2021) 9:163 Page 4 of 23 Low‑grade peripheral inflammation induces leukocyte compared to WT mice at baseline. Moreover, LPS chal- trafficking to the brain lenge further increases both of these parameters. In addi- NL-G-F Next to brain resident microglial cells, also infiltrating tion, the relative effect of LPS on WT and App mice NL-G-F leukocytes can contribute to neuroinflammation. To is similar, but LPS stimulated App mice display with investigate the infiltration of peripheral macrophages the highest number of IBA1 + microglia cells, the high- in the brain after LPS injection, brain sections were est rate of proliferation, the shortest dendrite length, the immunostained for IBA1 and TMEM119. TMEM119 smallest number of branch points and the smallest vol- is specifically expressed on microglial cells, but not on ume. All of this is consistent with the increasing number IBA1 infiltrating macrophages [16, 22]. As shown in of microglia and their amoeboid stage (Fig.  2a–d). The Fig.  3a, b, no macrophage infiltration was observed in significant increase of Ki67 microglial cells was only NL-G-F the cortex or hippocampus 24  h after LPS stimulation, observed in the cortex of App mice 2  weeks after NL-G-F neither in WT nor App mice. In contrast, 2 weeks peripheral inflammation and in the hippocampus of WT later, a clear increase in infiltrating macrophages mice 24  h after the last LPS injection (Fig.  2a–d). Espe- was visible in LPS stimulated mice compared to con- cially microglial cells that are not located in the area of trol mice. This was again observed in both WT and Aβ plaques showed a high proliferation upon peripheral NL-G-F App mice. Although we observed an increased inflammation (Fig. 2a–d). infiltration of peripheral immune cells into the brain Also, the microgliosis marker Aif1 and the microglial upon LPS challenges, the infiltration is still very mod - activation marker CD69 were significantly upregulated in est. In addition, perivascular macrophages (PVMs) are the hippocampus 24  h after LPS treatment both in WT + − NL-G-F also IBA1 and TMEM119 and these cells have been and in App mice (Fig. 2e). This increase is transient shown to increase in neurodegenerative disease models as no differences were observed 2  weeks after the last [20]. Therefore, other techniques which can accurately LPS treatment. distinguish invading peripheral monocytes from brain Despite the fact that peripheral inflammation and neu - endogenous PVMs and microglia, for example using roinflammation are not detectable  2  weeks after LPS genetic labeling of different myeloid populations [49], injection based on cytokine levels (Fig.  1b), the quanti- should be used to further validate this result. tative morphometric three-dimensional (3D) measure- Furthermore, we checked the expression of leuko- ments of IBA1 microglial cells revealed a significant cyte trafficking molecules in the brain during periph - decrease in the number of segments, branch and termi- NL-G-F eral inflammation and we show that exposure to LPS nal points in hippocampus of WT and App mice significantly increases the expression levels of integrin 2 weeks after peripheral inflammation compared to their ligand (Icam1) and chemokines (Ccl2 and Cxcl10) in respective controls (Fig. 2f, g). However, the shorter pro- the hippocampus and/or choroid plexus (CP) of WT cesses and smaller volumes were only observed in WT NL-G-F and App mice 24  h after the second LPS injec- mice injected with LPS compared to WT mice injected NL-G-F tion. However, only Ccl2 expression in the CP remained with PBS. Interestingly, the App mice also showed increased 2  weeks after LPS stimulation. In addition, more activated microglia compared with WT mice in the gene expression of Icam1, Ccl2 and Cxcl10 was basal conditions (Fig.  2g). Additionally, we also inves- more significantly upregulated in response to low- tigated the morphological changes of microglia in the grade peripheral inflammation in the CP compared cortex (Additional file  4: Fig. 2). WT mice showed more to the hippocampus (Fig.  3c, d). Taken together, these pronounced changes in all the examined parameters, findings indicate that low-grade peripheral inflamma - namely dendrite length, number of segments, branch tion induces immune cell infiltration into the brain. and terminal points and cell volume 2 weeks after the last LPS injections compared to the control condition, while NL-G-F this was not observed in App mice. (See figure on next page.) Fig. 2 Low-grade peripheral inflammation affects microglia proliferation and activation. a Representative images of IBA1, 6E10 and Ki67 staining in hippocampus. Scale bar: 100 μm and 20 μm (insert). b Quantification of microglial proliferation in hippocampus (n = 4–6). c Representative images of IBA1, 6E10 and Ki67 staining in cortex. Scale bar: 20 μm. d Quantification of microglial proliferation in cortex (n = 4–6). d Gene expression of microgliosis marker Aif1 and Cd69 in hippocampus (n = 5–10). f Representative 3D reconstruction images of IBA1 microglia from hippocampus 2 weeks after LPS stimulation. Scale bars: 20 µm. g Imaris-based quantification of cell morphology of IBA1 microglia in hippocampus. Each symbol represents one mouse, each mouse was randomly selected with 3–5 cells outside the Aβ for analysis (n = 4–5). Mean ± SEM, two-way ANOVA Bonferroni’s post hoc test for multiple comparisons. *p < 0.05, **p < 0.01 Xie  et al. acta neuropathol commun (2021) 9:163 Page 5 of 23 Fig. 2 (See legend on previous page.) Xie et al. acta neuropathol commun (2021) 9:163 Page 6 of 23 Loss of blood‑CSF barrier integrity in response and OCLN (Additional file  4: Fig. 4a-c). Also, a fragmented to low‑grade peripheral inflammation border staining of ZO-1 and OCLN was observed upon We have previously shown that high dose LPS has detri- IL-1β treatment (Additional file  4: Fig.  4b). These results mental effects on blood-CSF barrier integrity, while the suggest that the proinflammatory cytokine IL-1β is suffi - effect on blood–brain barrier (BBB) was less pronounced cient to induce loss of blood-CSF barrier integrity, but only [2, 66]. To investigate whether low-grade peripheral induces limited differences in the expression of TJ proteins inflammation alters the tight junction (TJ) complex at the and genes in primary CP epithelial cells. CP epithelial cells, the cellular localization of the TJ pro- teins was evaluated by immunocytochemistry and confo- Low‑grade peripheral inflammation results in a higher NL‑G‑F cal microscopy. In control conditions, E-cadherin (CDH1), amyloid deposition over time in App mice Occludin (OCLN), Claudin-1 (CLDN1) and Claudin-5 It is already well-known for decades that the deposition (CLDN5) immunoreactivity appeared as a near continu- of Aβ protein in senile plaques outside neurons results in ous staining at the apical cell border (Fig.  4a). Upon low- the loss of synapses and neurodegeneration which ulti- grade peripheral inflammation, a loss of these TJ proteins mately leads to symptoms associated with AD [12]. Next immunostaining was observed, leading to a fragmented to this, peripheral inflammation has been associated with border and diffuse distribution staining, although the AD [28]. Here, we analyzed whether low-grade inflam - NL-G-F NL-G-F effect on App mice was more pronounced (Fig.  4a). mation has an impact on Aβ pathology in App mice. NL-G-F In addition, immunofluorescence quantitative analysis In unchallenged App mice, we observed an confirmed lower expression of these TJ proteins in the CP increased Aβ deposition at both examined timepoints. NL-G-F of LPS injected WT and App mice compared with Interestingly, 6E10 staining of Aβ revealed that the the corresponding PBS injected mice. However, the lower amount of Aβ plaques is increased in both hippocam- expression of TJ proteins only showed a significant down - pus and cortex upon low-grade peripheral inflammation NL- regulation of CLDN5 2 weeks after LPS injection in App (Fig. 5a, b) and this increase is significant 2 weeks after the G-F NL- mice (Fig.  4b). Consistently, the gene expression levels second LPS injection compared to PBS injected App G-F showed downregulation trends after LPS challenge and mice. In addition to the amount of Aβ aggregation, NL-G-F the App mice seemed more susceptible to disruption also the degree of plaque compactness and the surface by systemic LPS than the WT mice (Fig. 4c). These results area reflect AD pathology [72]. Morphometric analysis show that peripheral inflammation has more pronounced of Aβ stained brain sections revealed an increase in small effects on redistribution of TJ proteins and less on their plaques (< 10 µm ) in the cortex, while in the hippocam- gene expression. Additionally, we also investigated the TJ pus all different sizes of plaques were increased (< 10 µm , 2 2 protein ZO-1 but no differences were visible either on pro - 10–20 µm and > 20 µm ) (Fig.  5c). In agreement with tein level nor on mRNA level, at least not at the examined the Aβ plaque analysis, we also observed a significant time points (Additional file  4: Fig. 3a-c). Collectively, these increase in both soluble and insoluble Aβ in the cortex 1-40 NL-G-F NL-G-F results suggest that the blood-CSF barrier in App of App mice 2 weeks after LPS stimulation (Fig. 5d). mice is more vulnerable to low-grade peripheral inflam - For Aβ only the soluble fraction was increased 2 weeks 1-42 mation compared to their WT counterparts. after LPS stimulation (Fig.  5d). However, the levels of Aβ IL-1β is a major pro-inflammatory cytokine released peptides are not significantly changed 24 h after LPS stim - from activated microglia and have been demonstrated that ulation (Fig. 5d). Taken together, these results suggest that IL-1β treatment increases BBB permeability in  vitro [69]. low-grade peripheral inflammation affects Aβ deposition NL-G-F Here, we hypothesize that IL-1β may also directly affect in App mice, which may be the result of aberrant Aβ blood-CSF barrier permeability. To this end, we studied clearance from the brain. the effect of IL-1β on blood-CSF barrier integrity using primary CP epithelial cells followed by transepithelial elec- Low‑grade peripheral inflammation disturbs Aβ transport trical resistance (TEER) measurements and TJ proteins across the choroid plexus (CP) epithelial cells staining. This revealed that IL-1β treatment significantly Increased levels of Aβ deposition in the brain may be the reduced TEER and TJ proteins expression including ZO-1 result of impaired Aβ clearance that on its term can be the (See figure on next page.) Fig. 3 Low-grade peripheral inflammation induces leukocytes trafficking to the brain. a Representative images of IBA1 and TMEM119 staining in + - hippocampus and cortex. Scale bar: 20 μm. b Quantification of IBA1 TMEM119 macrophages in hippocampus (left graph) and cortex (right graph). Each symbol represents one mouse. Expression of the gene Icam1, Ccl2 and Cxcl10 in hippocampus c and in CP d. Mean of 5 ± SEM, two-way ANOVA Bonferroni’s post hoc test for multiple comparisons. *p < 0.05, **p < 0.01, ****p < 0.0001 Xie  et al. acta neuropathol commun (2021) 9:163 Page 7 of 23 Fig. 3 (See legend on previous page.) Xie et al. acta neuropathol commun (2021) 9:163 Page 8 of 23 result of a disturbed balance of transport across the brain To address whether the reduced Aβ transport can be barriers and/or defective degradation of Aβ aggregates. linked to a reduction in LRP2 transport of the Aβ, we first First, we investigated the effect of low-grade peripheral investigated the expression of LRP2 upon IL-1β stimula- inflammation on Aβ transcytosis across the blood-CSF bar - tion. We showed that there is a reduction in LRP2 expres- rier. As shown in Fig. 6a, Aβ CSF levels were increased sion by CP epithelial cells upon stimulation with IL-1β 1-42 NL-G-F in App mice 24  h after LPS stimulation. Addition- (Fig.  6g, h). The LRP2-dependency of this reduced trans - ally, the level of Aβ in plasma was significantly lower in port was further studied using a blocking anti-LRP2 anti- 1-42 NL-G-F App mice at 24 h, but not at 2 weeks post LPS chal- body. Interestingly, sequential treatment of CP epithelial lenge compared to PBS injected mice (Fig.  6a). To ascer- cells with IL-1β and anti-LRP2 showed no additive effect tain whether this effect correlates with Aβ efflux from the on Aβ transport (Fig.  0.6f ). Importantly, neither the anti- CSF via the blood-CSF barrier, we analyzed the expression LRP2 antibody nor the IgG control had detrimental effects of genes responsible for Aβ influx and efflux in the CP. In on blood-CSF barrier integrity (Additional file  4: Fig. 4a, b). line with the changes of Aβ in CSF and plasma, the expres- Altogether, these data indicate that the pro-inflammatory sion of the Aβ efflux transporters P-gp and Lrp2 showed cytokine IL-1β blocks Aβ transport at least partially by increased trends 2  weeks post LPS injection compared inhibiting LRP2. NL-G-F to the 24  h timepoint in App mice, and only Lrp2 showed significant change (Fig.  6b). However, no signifi - Low‑grade peripheral inflammation affects Aβ cant differences were observed compared to PBS injected phagocytosis by microglial cells NL-G-F App mice. Moreover, protein levels of LRP2 shows Next to Aβ efflux, also defective phagocytotic clearance by a similar trend as the mRNA expression profile (Fig.  6c). microglial cells can explain the increased Aβ deposition Additionally, it has been reported that also LRP1 plays an in the brain [50]. As shown in Fig.  2, low-grade periph- important role in Aβ transport across CP epithelial cells eral inflammation significantly increases microglia prolif - [15, 47]. Therefore, we looked into the expression of LRP1 eration, mainly of the non-plaque-associated microglia. in the CP but no obvious changes were seen 24 h after LPS Moreover, the number of microglia around Aβ plaques NL-G-F stimulation either at protein nor at mRNA-level in the is reduced 24  h after LPS stimulation in App mice NL-G-F App mice (Additional file 4 : Fig. 5a-c). Taken together, (Fig. 7a, b). We also found more microglia migration to the these data suggest that low-grade peripheral inflammation vessel at that timepoint (Additional file  4: Fig. 7a, b). Based disturbs Aβ efflux from the CSF into the blood mainly. on these findings, the question arises if the reduced num - Next, we examined whether the observed IL-1β increase ber of microglial cells around Aβ plaques has an effect on in hippocampus during low-grade peripheral inflammation Aβ engulfment and clearance. To study this, we quantified (Fig.  1b) can explain the effect seen on Aβ transport. To the amount of internalized Aβ and observed a decrease in address this, we established an in vitro primary CP epithe- Aβ engulfment by microglial cells 24  h after LPS stimula- NL-G-F lial transport system using fluorescently labelled Aβ [34, tion in App mice (Fig. 7a, c). Furthermore, we exam- 1-42 62]. The addition of fluorescently labelled Aβ monomers ined CD68 phagocytic microglial cells by performing a 1-42 to the apical side of the epithelial cells allows the cells to co-staining of CD68, Aβ and IBA1 on hippocampus and transport the Aβ to the basolateral side. The rate of Aβ cortex (Fig. 7d). A transient increase in CD68 expression is 1-42 transport can be calculated by measuring the transported observed 24 h after LPS challenge in the hippocamps and NL-G-F Aβ in the basolateral compartment. To eliminate the the cortex of App mice. Interestingly, we observed 1-42 effects of paracellular diffusion of Aβ , scrambled Aβ that less CD68 microglial cells are recruited to the Aβ 1-42 1-42 NL-G-F was added to the cell cultures together with Aβ and the plaques in App mice after LPS challenge, especially at 1-42 transcytosis quotient (TQ) of Aβ was normalized to the the 24-h timepoint (Fig. 7d, e). 1-42 diffusion rate of scrambled Aβ . Confirming our previ - Next, we investigated whether the changes in Aβ engulf- 1-42 ous findings, TQ of Aβ was reduced upon IL-1β stimu- ment can be explained by the activation status of the 1-42 lation of CP epithelial cells (Fig. 6f ). IL-1β did not alter the microglial cells in response to low-grade peripheral inflam - viability of primary CP epithelial cells, even at concentra- mation. To study this, we examined the correlation of the tions ten times greater than that which caused changes in activation state of the microglia and the distance to Aβ the levels of Aβ transport (Additional file 4 : Fig. 6). plaques during low-grade peripheral inflammation. In the (See figure on next page.) Fig. 4 Characterization blood-CSF barrier integrity during peripheral immune challenge. a, b Representative images of CP and quantification of the percentage red staining of stained for CDH1, OCLN, CLDN1 and CLDN5. The dotted line indicates the ependymal cells that line the ventricle (n = 4–6). Scale bar: 20 μm. c Expression of the genes Cdh1, Ocln, Cldn1 and Cldn5 in hippocampus (n = 5). Mean ± SEM, two-way ANOVA Bonferroni’s post hoc test for multiple comparisons. *p < 0.05 Xie  et al. acta neuropathol commun (2021) 9:163 Page 9 of 23 Fig. 4 (See legend on previous page.) Xie et al. acta neuropathol commun (2021) 9:163 Page 10 of 23 NL-G-F Fig. 5 Low-grade peripheral inflammation affects Aβ deposition in App mice. a Representative images of 6E10 staining in hippocampus and NL-G-F cortex of App mice. Scale bar: 100 μm. b Quantification of Aβ plaque area and number (n = 5–6). Hippocampus (left two graphs); cortex (right two graphs). c Quantification of Aβ plaque size distribution. Hippocampus (left); cortex (right); (n = 5–6). d Soluble and insoluble Aβ and Aβ 1-40 1-42 levels in prefrontal cortex tissues (n = 5). Mean ± SEM, two-way ANOVA Bonferroni’s post hoc test for multiple comparisons. *p < 0.05, **p < 0.01 hippocampus, our analysis shows that if the microglial of segments, branch points and terminal points (Fig.  7f, cell is in close proximity to Aβ deposits, the LPS stimula- g). In contrast, peripheral LPS challenge induces a more tion does not lead to further morphological changes of the pronounced microglial activation with increasing distance microglia in terms of dendrite length, cell volume, numbers of the microglia to Aβ plaques compared to PBS injected Xie  et al. acta neuropathol commun (2021) 9:163 Page 11 of 23 NL-G-F App mice (Fig.  7f, g). These morphological changes Correspondingly, coimmunostaining of IBA1 and NeuN are reflected by a reduction in dendrite length, cell volume, showed a higher percentage of microglia and neurons number of segments, branch points and terminal points. that are in close contact with each other in response to The same morphological changes of microglia in the cortex low-grade peripheral inflammation (Additional file  4: were observed upon low-grade peripheral inflammation Fig. 10a, b). (Additional file 4 : Fig. 8). Next, we studied whether peripheral inflammation Altogether, these results show that peripheral inflam - can affect Aβ aggregation by affecting neuron func - mation affects microglial activation, migration and motil - tion. Firstly, we performed live cell imaging to moni- ity and reduces microglial Aβ phagocytosis. tor the morphologic changes and Aβ biodistribution in primary neuron cultures after IL-1β treatment. Com- Low‑grade peripheral inflammation induces neuronal pared to neuronal cell cultures without IL-1β treatment dysfunction and Aβ (Additional file  1: movie 1), the addition of Aβ The phagocytotic ability of a microglial cell is not only affects neuronal function that promotes Aβ aggregation important to clear Aβ aggregates, but is also beneficial close to the soma (Additional file  2: movie 2). Pretreat- for the clearance of apoptotic neurons [23, 36]. However, ment with IL-1β leads to neuronal dysfunction with a activation of microglia upon peripheral inflammation has reduction in dendrites and a similar Aβ as in the condi- shown to cause microglial phagocytosis of healthy neu- tion where only Aβ was added (without the pretreatment rons and synapses that lead to neuronal loss and dysfunc- with IL-1β) (Additional file  3: movie 3). However, it can’t tion [41, 58]. With this in mind, we tested whether an be excluded that Aβ aggregation induces fluorescence enhanced microglial activation and impaired microglial quenching of HF488-Labeled Aβ , reflected by a reduc - 1-42 phagocytosis induced by low-grade peripheral inflamma - tion in fluorescence intensity [13]. To circumvent this NL-G-F tion causes neuronal and synaptic loss in App mice. phenomenon and study the impact of IL-1β on Aβ aggre- In line with a previous study [71], we observed increased gation in neurons, we used fluorescence lifetime imaging trends in cell and neuronal death in the hippocampus (FLIM) of HF488-Labeled Aβ to detect Aβ aggrega- 1-42 and cortex after LPS challenge. Yet, neuronal death only tion as described previously [13]. The fluorescence life - reaches statistical significance in WT mice 24  h after time depends on conformational changes associated LPS challenge (Additional file  4: Fig.  9a-c). Interestingly, with Aβ aggregation and decreases with the increase of NL-G-F the App mice showed a significant higher percent - Aβ aggregation. As shown in Fig. 9a–c, we observed less age of neuronal death in the hippocampus compared to and slower Aβ aggregation in the condition without cells WT mice 2  weeks after PBS injection (Additional file  4: compared to Aβ in the presence of neuronal cells. Inter- Fig. 9a-c). estingly, IL-1β treatment further significantly enhances In the cortex, immunostainings of the presynap- the rate of aggregation in the latter condition. Taken tic marker synaptophysin (SYP) revealed a decrease in together, these data indicate that neuronal dysfunction expression 2 weeks after low-grade peripheral inflamma - affected by peripheral inflammation may play a role in NL-G-F tion in App mice, but no significant differences at accelerating or exacerbating AD pathology. NL-G-F the 24  h timepoint compared to PBS injected App mice. In WT mice, no LPS-induced obvious differences Discussion in SYP expression were observed (Fig.  8a, b). Next, we Already more than 30 years ago, Aβ was for the first time used western blotting to detect the differences of SYP isolated and proposed to play an important role in AD expression in the hippocampus 2  weeks after LPS stim- pathogenesis [18, 19]. Initial efforts in the development ulation. As displayed in Fig.  8c, d, this confirmed the of a curative AD treatment mainly focused on strate- decrease in SYP expression in the hippocampus upon gies to lower Aβ levels and decrease toxic Aβ aggregates. NL-G-F low-grade peripheral inflammation in App mice. Unfortunately, these attempts are without any clinical (See figure on next page.) Fig. 6 Low-grade peripheral inflammation affects Aβ transport across the blood-CSF barrier in vivo and in vitro. a Quantification of Aβ and 1-40 Aβ in CSF and plasma (n = 5–8). b Expression of the genes P-gp and Lrp2 in CP (n = 5). c Representative images of LRP2 immunostaining of CP. 1-42 Scale bar: 20 μm. d Quantification of the relative red staining of LRP2 in CP (n = 5 per group). e Schematic diagram of Aβ transcytosis analysis in an 5-FAM TAMRA CP epithelial transport system using fluorescently labeled Aβ peptides. f movement of scrambled Aβ (left) and Aβ (middle) from 1-42 1-42 the basolateral to the apical chamber, and normalized Aβ transcytosis quotient of Aβ from the basolateral to the apical chamber (right) (n = 3). 1-42 g Representative images of LRP2 staining in primary CP epithelial cells. h Quantification of the relative red staining of LRP2 in primary CP epithelial cells (n = 3). Mean ± SEM, two-way ANOVA Bonferroni’s post hoc test for multiple comparisons (a, b, d), one-way ANOVA Bonferroni’s post hoc test for multiple comparisons f Nonparametric Mann–Whitney U test h. *p < 0.05, **p < 0.01, ***p < 0.001 Xie et al. acta neuropathol commun (2021) 9:163 Page 12 of 23 Fig. 6 (See legend on previous page.) Xie  et al. acta neuropathol commun (2021) 9:163 Page 13 of 23 success so far, indicating that there is an urgent need for peripheral immune stimulation in contrast to microglia a more in-depth understanding of the onset and pro- located at a greater distance from the plaques; both in the NL-G-F gression of AD pathology. Over the last years, it became hippocampus and cortex of App mice. Indeed, dis- increasingly clear that the innate immune system plays tant microglia showed sustained changes in morphology, a crucial role in the pathogenesis of AD, even in the reflected by a reduction in dendrite length, cell volume, early clinically silent period [17, 24, 64]. Considerable number of segments, branch points and terminal points. efforts have been devoted to understand the interactions These changes were rather moderate 24  h after the last between the peripheral immune system and the CNS [37, LPS challenge but further increased after the second 51, 63] in APP overexpression mouse models. However, LPS challenge. The fact that this is not the case for the the underlying mechanisms are not yet fully understood. microglia in direct contact with the Aβ plaques might With this manuscript, we are the first to report the effects be explained by the fact that these microglia are already of LPS-induced low-grade peripheral inflammation on activated and may not further react to an additional trig- AD pathology in a second-generation mouse model for ger. Next, we showed that the microgliosis marker Aif1 NL-G-F AD, namely the App mouse model. The use of this and microglial activation marker Cd69 are transiently more representative mouse model may help to give a bet- increased in the hippocampus after LPS challenge. This ter understanding of the deleterious impact of peripheral is however in contrast to a previous study that found that immune activation on AD development. Aif1 is significantly downregulated in microglial cells Tejera et al. recently reported that peripheral and CNS upon LPS treatment but is unchanged in the whole brain inflammation related symptoms return to basal condi - [63]. Another study reported an increase in Cd69 expres- tions 10  days after a low dose LPS challenge [64]. Based sion after exposure of glial cells to LPS or IL-4, but the -/- on this study, we used two low-dose LPS injections microglia isolated from CD69 mice still exhibited an separated by 7  days to mimic two discrete infectious enhanced production of proinflammatory mediators [6]. events associated with low-grade peripheral inflamma - Al these results together suggest that an increase in cell tion. Importantly, IL-6 plasma level analysis revealed a number may be responsible for the maintenance or slight NL-G-F slightly stronger inflammatory response in the App increase of Aif1 mRNA levels and CD69 reported in our mice compared to wild type controls and showed in both study. Moreover, CD69 exerts some negative regulation genotypes a reduced impact of the second compared to of inflammation upon exposure to stimulus [6]. the first LPS injection, suggesting partial desensitiza - The expression levels of the pro-inflammatory cytokines tion. Next, we investigated both the short and long-term IL-1β and TNF in the brain increase transiently after LPS effects by analyzing the response 24 h and 2 weeks after challenge. These results are consistent with the results the last LPS injection, respectively. of previous studies in first-generation AD mouse mod - Microglial cells are the most prominent immune cells els. In those studies, LPS treatment of APP overexpress- of the CNS and they continually survey and maintain ing mice induces an increased expression of cytokines homeostasis in the brain [23]. Any changes in the CNS [37, 64]. However, the extent of the LPS effects are dif - lead to activation, proliferation, and morphological ferent from our results. For example, a recent study in 5 changes of these cells [60]. On the one hand, activated and 15 months old APP/PS1 mice combined with a single microglia can clear neuronal damage by phagocytosis injection of LPS shows high pro-inflammatory cytokine while on the other hand, activated microglia will also levels 2 days after LPS injection which returns to baseline release molecules that can initiate neuroinflammation levels 8 days later [64]. The differences between our results [23]. In concordance with previous findings [64], we and this study can be explained by a different experimen - observed that microglia located at the site of Aβ depo- tal setup, including LPS type (ultrapure LPS from S. typh- sition don’t show striking morphological changes upon imurium vs. LPS from S. abortus equi, number of LPS (See figure on next page.) NL-G-F Fig. 7 Low-grade peripheral inflammation impairs microglial phagocytosis of Aβ in App mice. a Representative images of IBA1 and 6E10 staining in hippocampus and cortex. The dotted circle shows the border of Aβ plaques. Scale bar: 20 μm. b Quantification of the IBA1 microglia within the Aβ plaque surface. Hippocampus (left); cortex (right) (n = 5–6; each symbol represents the average from 4–6 plaques in one mouse). c Quantification of the percentage of internalized Aβ in the hippocampus (left graph) and cortex (right) (n = 5–6). d Representative images of IBA1, 6E10 and CD68 staining in hippocampus and cortex. Arrow points to the CD68 microglia. Scale bar in hippocampus: 100 μm and 20 μm (insert); Scale bar in cortex: 20 μm. e Quantification of CD68 microglia (n = 5) in hippocampus (left two graphs) and cortex (right two graphs). f Representative images of relationship between microglial morphology and its distance to Aβ deposition. Scale bars: 20 µm. g Imaris-based quantification of cell morphology of IBA1 microglia in hippocampus. Co-staining of IBA1, CD68 and 6E10 was performed on 5 µm paraffin sections. Each symbol represents the average in one cell and 10–15 cells are analyzed per mouse (n = 4–5). Mean ± SEM, two-way ANOVA Bonferroni’s post hoc test for multiple comparisons (b, c, e), nonparametric Mann–Whitney U test g. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 Xie et al. acta neuropathol commun (2021) 9:163 Page 14 of 23 Fig. 7 (See legend on previous page.) Xie  et al. acta neuropathol commun (2021) 9:163 Page 15 of 23 Fig. 8 Low-grade peripheral inflammation and Aβ induces neuronal dysfunction. a Representative images of SYP staining in the cortex. Scale bar: 10 μm. b Quantification of relative SYP area in cortex (n = 5). c Western blot of SYP protein levels in hippocampus 2 weeks after LPS challenge. d Densitometric analysis of relative protein levels of SYP normalized to β-actin (n = 5). Scale bar: 50 μm. Mean ± SEM, two-way ANOVA Bonferroni’s post hoc test for multiple comparisons. *p < 0.05, **p < 0.01 injections (single vs. twice, once a week). The different the brain. So far, different mechanisms are proposed types of LPS may induce varying degrees of inflammatory among which: transmission by the vagal afferents, periph - response and repeated LPS challenge causes less inflam - eral mediators crossing/transporting via circumventricu- matory response than the first LPS injection. In addition, lar organs (CVOs) or the BBB, or by signaling through IL-1β is produced predominantly by microglia and plays the endothelial cells of the BBB [30]. More recently, the an important role in microglia activation and proliferation CP epithelial cells that form the blood-CSF barrier are [36]. Under chronic peripheral inflammation, microglia identified as a novel player in this regard [2, 38]. Here, may reach an effector phenotype that express low levels we show that in WT mice, low-grade peripheral inflam - of cytokines [51]. In our study, we performed two LPS mation has only limited effects on the blood-CSF barrier injections to mimic long-term chronic peripheral inflam - integrity, both at short and long-term. This is in contrast mation and the second stimulus may lead to an exagger- to our previous study in which we showed that a single ated microglial response. This may explain the results of high dose LPS challenge causes severe blood-CSF bar- sustained microglial activation but transient changes in rier disturbance [66]. However, long-term peripheral inflammation level in our study. inflammation did induce significant changes in TJ pro - A pressing question is how pathological insults that teins expression and therefore in blood-CSF barrier NL-G-F reside in circulation and peripheral organs can get into integrity in App mice, especially 2  weeks after LPS (See figure on next page.) Fig. 9 IL-1β stimulation increases Aβ aggregation in primary neuron culture. a Schematic diagram of kinetics of Aβ aggregation in primary neuron HF488 after stimulation with IL-1β. b Representative images of different conditions at 0, 12 and 24 h after adding PBS/Aβ oligomers. The images 1-42 show an overlay of the color-coded FLIM image and the transmitted image. Scale bar: 20 μm. c Quantification of HF488 fluorescence lifetime of HF488 Aβ oligomers (n = 4). Mean ± SEM, nonparametric Mann–Whitney U test. *p < 0.05 1-42 Xie et al. acta neuropathol commun (2021) 9:163 Page 16 of 23 Fig. 9 (See legend on previous page.) Xie  et al. acta neuropathol commun (2021) 9:163 Page 17 of 23 stimulation. While we previously showed that intracer- peripheral immune stimulation affects Aβ pathology. ebroventricular injection of Aβ oligomers contributes This revealed that low-grade peripheral inflammation 1-42 NL-G-F to loss of blood-CSF barrier integrity [7, 61], App significantly increases Aβ aggregation and in particu - mice showed only a trend in increased blood-CSF barrier lar the formation of small size plaques. Previous stud- leakiness compared to WT mice. This might be explained ies demonstrated that low dose LPS injection(s) lead to by the low Aβ levels (around 0.15  ng/ml) in the CSF increased Aβ deposition in APP/PS1 mice [40, 64] while 1-42 NL-G-F of App mice which is not sufficient to significantly other studies failed to show this LPS effect in first-gen - disrupt blood-CSF barrier integrity compared to the dose eration AD mouse models [31]. The varied outcomes of Aβ oligomers (125 ng/ml) injected in our previous could be due to the experimental setup, including mouse 1-42 studies. In addition, in  vitro experiments revealed that genetic background, age-at-onset of plaque formation, IL-1β exposure of primary CP epithelial cells resulted in age at LPS stimulation, number of injections, LPS dose, a significant barrier disruption as previously observed route of administration, and the time between injec- upon TNF stimulation [5]. These results suggest that tion and sacrifice. In our study, we found that sustained there are mechanisms in the brain to protect the blood- inflammation stimulates microglia to further transform CSF barrier from peripheral inflammation. into a phagocytic phenotype associated with proliferation Loss of brain barrier integrity may lead to peripheral and morphological changes. This is in line with a previ - immune cell infiltration in the brain via paracellular path - ous study which indicated that peripheral inflammation ways. In addition, these cells can also cross the brain bar- reduces the microglial Aβ clearance [64]. riers by the expression of integrin ligands and chemokines In addition, analysis of microglia biodistribution in the by epithelial cells [5]. We identified an increased expres - brain suggests that, shortly after induction of periph- sion of Icam1, Ccl2 and Cxcl10, which are crucial for leu- eral inflammation, microglia migrate to blood vessels, a kocyte trafficking [5 ], in the CP and hippocampus of WT phenomenon that has been observed before [22]. Con- NL-G-F and App mice 24 h after LPS stimulation. These find - sequently, this migration leads to fewer microglia avail- ing are consistent with previous reports [39, 65]. It is there- able for Aβ phagocytosis. Two weeks after induction of fore important to further understand whether the repeated peripheral inflammation, however, more microglia were exposure to a peripheral stimulus ultimately results in the again observed close to the Aβ plaques. Nonetheless, Aβ entry of immune cells into the brain. Based on immuno- aggregation further continued in response to the chronic histochemical stainings, we observed an increase in leu- low-grade peripheral inflammation, which might be the NL-G-F kocyte infiltration in the brain of both WT and App result of the chronically activated microglial cells which mice 2 weeks after the last LPS stimulation. Although we are no longer able to process Aβ [3, 27]. Altogether, our observed an increased infiltration of peripheral immune data indicate that peripheral inflammation induces dys - cells into the brain upon LPS challenges, the infiltration is functional microglia that may account for the increased still very modest. In addition, perivascular macrophages Aβ deposition which we observed in the brains of LPS + − NL-G-F (PVMs) are also IBA1 and TMEM119 and these cells challenged App mice. have been shown to increase in neurodegenerative disease Aβ clearance from the brain is also mediated by a com- models [20]. Therefore, other techniques which can accu - bination of transcellular transport mechanisms across rately distinguish invading peripheral monocytes from the blood–brain and blood-CSF barriers [8, 62]. LRP2 is brain endogenous PVMs and microglia, for example using an efflux transporter expressed at the blood-CSF barrier genetic labeling of different myeloid populations [49], and is involved in the elimination of Aβ across CP epithe- should be used to further validate this result. Our results lial cells [8]. According to our results low-grade chronic are in agreement with a previous study that showed that peripheral inflammation induces a short-term effect on a single low dose LPS challenge is sufficient to induce leu - Aβ transporter expression levels associated with changes kocyte infiltration in the brain of aged APP/PS1 mice [64]. in levels of peripheral and central inflammation. In Moreover, a sustained exposure triggers a constant influx agreement with this, Ott and colleagues previously sug- of leukocytes [65] and this can be caused by brain barrier gested that changes in blood-CSF barrier transport are disruption. On the other hand, one recent study indicates related to the expression of inflammatory cytokines and that leukocytes can also pass directly through epithelial chemokines, such as higher IL-1β and TNF in serum and cells [42]. Our findings suggest that leukocytes can adapt CSF of mild cognitive impairment patients are associated their behavior to different circumstances as there is no with low efficiency of transport small and much larger obvious barrier damage in WT mice but still immune cell molecules in the blood-CSF barrier [44]. IL-1β is pro- infiltration. duced predominantly by microglia and plays an impor- As microglial activation is associated with Aβ deposi- tant role in microglia activation and proliferation [36]. To tion [11], we tested whether in our set-up the low-grade specifically examine the role of IL-1β in the regulation of Xie et al. acta neuropathol commun (2021) 9:163 Page 18 of 23 Aβ efflux, we tested Aβ transcytosis in  vitro after IL-1β [53], which overcome intrinsic drawbacks of the APP stimulation. We show that IL-1β not only disrupts the overexpression mouse models by utilizing an App knock- barrier integrity but also affects Aβ transcytosis by inhib - in strategy were generated to overproduce pathogenic iting, in part, the LRP2 transporter. Indeed, we observed Aβ such as Aβ without overexpressing APP. In agree- 1-42 that inhibition of LPR2 in CP epithelial cells reduces Aβ ment with previous studies in APP overexpression mouse NL-G-F efflux from CSF to blood and sequential treatment of models [31, 40, 57, 64], App mice showed higher anti-LRP2 and IL-1β did not have additive effect on Aβ Aβ deposition during low-grade peripheral inflam - transport. Altogether, our findings indicate that low- mation. Importantly, we identified several pathways grade peripheral inflammation affects Aβ pathology by which are activated upon peripheral inflammation and multiple pathways including microglial phagocytosis and subsequently contribute to aggravation of AD pathol- Aβ transcytosis. ogy, including sustained microglial activation, neuronal Previous evidence suggests that Aβ plays a vital role dysfunction and Aβ efflux from the brain. Our results in the induction of neuronal dysfunction from synapses strengthen previous findings in APP overexpressing mice toward neural networks [45] and Aβ is considered to and emphasize the potential vital role of preventing and/ 1-42 be the most neurotoxic form of Aβ [9]. In concordance or intervening with peripheral inflammation to impact with this, we found that the soluble Aβ levels were AD development and progression. 1-42 elevated 2  weeks after low-grade peripheral inflamma - tion. Interestingly, we identified synapse loss at the same Methods timepoint and showed that in vitro treatment of primary Mice neurons with Aβ oligomers also induces synapse loss. NL-G-F 1-42 The generation of App mice carrying Arctic, Swed- These results suggest that Aβ may mediate synaptic 1-42 ish, and Beyreuther/Iberian mutations was described activity and lead to synapse loss. previously [52]. The colony of these mice was main - Here, we also show that low-grade peripheral inflam - tained in our animal house at VIB Center for Inflamma - mation transient induces total cell and neuron death in tion Research. C57BL/6  J mice were used as a control. NL-G-F NL-G-F WT and App mice. In addition, the App mice Mice were kept in individually ventilated cages under a show a higher percentage of neuronal cell death com- 12-h dark/12-h light cycle in specific pathogen-free ani - pared to WT mice, which might be due to Aβ pathology mal facility and received food and water ad libitum. Both induced cell death as demonstrated in previous stud- male and female mice were used and all mice were aged ies [33, 43]. Microglial cells are important players in the between 20–23  weeks at the start of the experiment. maintenance and plasticity of neuronal circuits, con- All animal studies were conducted in compliance with tributing to the protection and remodeling of synapses. governmental and EU guidelines for the care and use The ability of microglial phagocytosis is essential for the of laboratory animals and were approved by the ethical clearance of apoptotic neurons and regulation of neu- committee of the Faculty of Sciences, Ghent University, ronal activity [23, 36]. In our study, we observed a higher Belgium. percentage of microglia that are in contact with neurons in short term and long term low-grade peripheral inflam - Reagents mation, which might lead to the removal of pre-synaptic LPS from S. abortus equi (L-5886) was obtained from input from neurons as described by Sharma et  al. [56]. Sigma-Aldrich. Recombinant mouse IL-1β was pro- The accumulation of Aβ around synapses has also been duced in E. coli and purified by VIB protein core. IL-1β shown to induce microglial complement activation in had specific activity of 6.12 × 10   IU/mg and no detect- neuronal synapses and subsequent microglial engulfment able endotoxin contamination. Synthetic Aβ , HiLyte [4]. Recent study report that TNF induces extracellular 1-42 Fluor 488-labeled (AS-60479), TAMRA-labeled (AS- Aβ production and aggregation in neuronal cell cultures 60476) and scrambled Aβ , 5-FAM labeled (AS-60892) [70]. Therefore, we hypothesize that pro-inflammatory 1-42 at the N-terminus were purchased from Anaspec. Aβ cytokine IL-1β may be also involved in Aβ aggregation 1-42 monomers and oligomers were prepared according to the by affecting neuronal functions. In line with this assump - manufacturer’s instructions. tion, we found that IL-1β induces accelerated Aβ aggre- gation in neurons. These findings indicate that peripheral inflammation influences neuronal functioning which Induction of low‑grade inflammation subsequently plays an important role in AD pathogenesis. LPS injections (1.0  mg/kg body weight, i.p.) were per- In conclusion, our study sheds light on how low-grade formed on day 0 and 7 as displayed in Fig.  1a. Control peripheral inflammation affects AD pathogenesis in a mice received i.p. PBS injections. Depending on the NL-G-F new mouse model of AD, namely the App model experiment, the mice were sacrificed 24  h or 2  weeks Xie  et al. acta neuropathol commun (2021) 9:163 Page 19 of 23 after the second LPS/PBS injection. Body weight, tem- are given as relative expression values normalized to the perature loss, and sickness behavior were checked the geometric mean of the housekeeping genes. Primers used first 3 days after LPS injection. for qPCR are depicted in Additional file 4: Table 1. TLR4 activation analysis Measurement of Aβ by ELISA The level of TLR4 activation in plasma and brain was The Aβ was extracted and measured using a standard measured using the HEK-Blue mTLR4 assay (InvivoGen) protocol as described previously [61]. Briefly, cortex according to the manufacturer’s instructions. HEK-Blue samples were homogenized in Tissue Protein Extraction mTLR4 cells were seeded at 25,000 cells per well in a Buffer containing complete protease inhibitor (Thermo 96-well plate in detection medium. The brain sample was Scientific) and phosphatase inhibitor tablets (Sigma- homogenized in PBS supplemented with 1% penicillin/ Aldrich) using a Precellys (Bertin Technologies) and streptomycin (Gibco), and brain supernatant was col- subsequently centrifuged at 5000  g for 5  min at 4  °C. lected by centrifugation at 20,000g in a microcentrifuge Supernatant was collected and centrifuged at 4 °C for 1 h at 4 °C for 15 min. The next day, cells were incubated for at 100,000 g (TLA-100 Rotor; Beckman Coulter). Super- 12  h with the different samples (20  μl plasma with 1/5 natant containing soluble Aβ was removed and stored at dilutions and 20  μl brain supernatant corresponding to − 80 °C. The pellet was further processed in GuHCl solu - 50  μg total protein), followed by absorption measure- tion containing complete protease inhibitor, sonicated, ment of the culture medium at 655  nm (iMark Micro- vortexed, incubated for 60  min at 25  °C and centrifuged plate Absorbance Reader, Bio-Rad) and relative TLR4 at 70,000  g for 20  min at 4  °C. Supernatant containing activation was calculated. insoluble Aβ was 12 times diluted with GuHCl diluent and immediately frozen at -80  °C. The levels of Aβ 1-42 Cytokine measurements and Aβ in plasma, CSF, and cortex lysates were deter- 1-40 The hippocampus was dissected and immediately snap- mined using a sandwich ELISA assay. Briefly, the sample- frozen in liquid nitrogen and stored at −  80  °C until detection mixtures were added to the ELISA plate coated use. After homogenizing with lysis buffer (0.5% CHAPS with anti-Aβ (1.5 μg/ml; JRF/cAb42/26) or anti-Aβ 1-42 1-40 in PBS supplemented with complete protease inhibitor antibody (1.5 μg/ml; JRF/cAb40/28) and incubated ON at (Thermo Scientific)), supernatant was collected by cen - 4 °C with slow shaking. Absorption at 450 nm was meas- trifugation at 20,000g in a microcentrifuge at 4  °C for ured after adding substrate solution (BD Biosciences 15 min. Protein concentration of the samples was meas- OptEIA ) followed by stopping buffer (1 M H SO ). The 2 4 ured using the Pierce BCA Protein Assay Kit (Thermo amount of Aβ was determined with GraphPad Prism 8.0 Scientific). Cytokines in plasma and hippocampus lysates using a nonlinear regression model. were measured using V-PLEX Custom Mouse Cytokine Kit for IL-1β, TNF and IL-6 (Meso Scale Discovery) Primary mouse CP epithelial cell isolation according to the manufacturer’s instructions. Signals Primary mouse CP epithelial cells were isolated from were measured on a MESO QuickPlex SQ 120 reader P2-P4 WT pups as previously described [2]. CP tissue (Meso Scale Discovery). was isolated from the lateral and fourth ventricles, pooled and digested with pronase (53,702, Sigma-Aldrich) RNA extraction and RT‑qPCR analysis for 7  min. For monolayer cultures, cells were plated in Samples were snap-frozen in liquid nitrogen and stored 24-well Transwell polyester inserts (pore size, 0.4  μm; at -80 °C until further use. After homogenizing the tissue surface area, 33.6 mm ; Corning) coated with laminin with TRIzol (Invitrogen) in a tube containing zirconium (L2020, Sigma-Aldrich). Cells were grown in DMEM/ oxide beads on a TissueLyser (QIAGEN), chloroform F12 supplemented with 10% FBS, 2  mM L-Glutamine was added and the homogenate was separated into 3 (Gibco), 1% penicillin/streptomycin at 37  °C and 5% phases by centrifugation at 20,000g in a microcentri- CO for 1 to 2  weeks until their TEER values reached a fuge at 4  °C for 15  min. Total RNA was extracted from plateau. the upper aqueous phase using Aurum total RNA kit (Bio-Rad) according to the manufacturer’s instructions. In vitro transcytosis of Aβ 1‑42 The concentration of total RNA was determined by the The Aβ transcytosis across the epithelial monolayers was Nanodrop-1000 (Thermo Scientific) and total RNA was measured in triplicate based on the protocol previously reverse-transcribed into cDNA with SensiFAST cDNA described in [21, 34, 62]. In brief, mouse CP epithelial Synthesis Kit (Bioline). qPCR was performed on the cells (5 × 10 cells per well) were grown to confluence Roche LightCycler 480 System (Applied Biosystems) and treated with PBS or IL-1β (2.5  ng/ml) in the basal ™ ® using SensiFAST SYBR No-ROX Kit (Bioline). Results side of Transwell, followed by 1  h blockage with 15  µg/ Xie et al. acta neuropathol commun (2021) 9:163 Page 20 of 23 ml anti-LRP1 or 15  µg/ml control IgG (10500C, Invitro- the FIFO image mode using 256 time channels (Becker gen). After washing the wells with fresh medium, 1  μM & Hickl GmbH, Berlin). All acquisitions were done TAMRA 5−FAM Aβ monomers and 1 μM scrambled Aβ with the 820  nm laser line of the MaiTai DeepSee mul- 1-42 1-42 monomers in assay medium (DMEM, 5% FBS, 25  mM tiphoton laser with a repetition rate of 80 MHz. A Plan- HEPES (Gibco), 2  mM L-Glutamine) were added to the Apochromat 40x/1.4 Oil DIC objective lens was used and apical chamber of each well. Subsequently, the cells were the scanning resolution was set at 512 × 512 pixels. The incubated at 37  °C. Samples were taken from the baso- selected positions were followed for a total of 24  h and lateral chamber and replaced with fresh medium every every 2 h a FLIM image was captured. All TCSPC images 20  min for 2  h and then transferred to 96-well plates were processed initially using SPCImage version 5.2 (Nunc). The fluorescence was determined using FLU - (Becker & Hickl) and fitted as mono-exponential decays. Ostar Omega reader (BMG LABTECH) at an excitation Further image analysis was carried out using NIH Image J wavelength of 485  nm and an emission wavelength of software. Statistical analysis was performed in GraphPad 520  nm. Relative fluorescence units were converted to Software. The average lifetimes in 4 images with healthy values of ng/ml, using corresponding standard curves, cells were computed and these were then averaged to and were corrected for background fluorescence and obtain the mean lifetime (± SEM) for each particular serial dilutions over the course of the experiment. Trans- time point and treatment. Significant differences between port of Aβ across the monolayer was calculated asTQ treatments were obtained by means of comparison using 1-42 using the following formula: nonparametric Mann–Whitney U test. TQ = Aβ (dQ/dT )/Scrambled Aβ (dQ/dT ). 1−42 1−42 Immunohistochemistry where dQ/dT (μg/s) is the rate of appearance of Aβ on For immunostainings on brain sections, mice were tran- the receiver side after application and is calculated by scardially perfused with ice-cold 4% PFA in PBS. Subse- plotting the cumulative amount (Q) versus time(s). quently, brains were carefully extracted from the skull and split into two halves (mid-sagittal). The right hemi - Primary neuron culture spheres were embedded in Frozen Section Medium Cortical neuronal cells were prepared from embryonic (Thermo Scientific) immediately in cryomolds (Sakura) (E16-E18) mouse pups. To dissociate the cortex tissue, that were frozen on dry ice, and stored at − 80  °C until the tissue was finely chopped and digested in papain further use. The left hemispheres were post-fixed over - (Sigma-Aldrich) for 20  min. Neurons were triturated night 4% PFA in PBS at 4 °C. After dehydration, samples gently with a pipette and passed through the 70  µm cell were embedded in paraffin in cryomolds and stored at strainer (BD Falcon) before plating them onto Poly-d-Ly - RT until further use. The brains were cut into 5 μm slices sine (Thermo Scientific) coated 8-well chamber (iBidi) at paraffin sections (HM 340 E, Thermo Scientific) or 20 μm a density of 1 × 10 cells per well. Cultures were main- cryosections (CryoStar NX70, Thermo Scientific). For tained in Neurobasal medium (Thermo Scientific) con - immunofluorescence staining, sections were permeabi - taining 1% penicillin/streptomycin, 0.5 mM GlutaMAX lized in PBS containing 0.3–0.5% Triton X-100. Following I Supplement (Thermo Scientific) and 2% B27 supple - blocking with goat immunomix (GIM) (5% goat serum, ment (Thermo Scientific) at 37  °C with 5% CO . The 0.1% BSA, 0.3–0.5% Triton X-100 in PBS) at RT for 1  h, absence of astrocytes (< 2%) was confirmed by the virtual sections were incubated with primary Abs in GIM at 4 °C absence of glial fibrillary acidic (GFAP) protein immu - overnight. After washing with PBS, sections were stained nostaining (data not shown). with fluorophore-conjugated secondary Abs in PBS or PBS containing 0.1% Triton X-100 at RT for 1–1.5  h. FLIM Counterstaining was done with Hoechst reagent (Sigma- FLIM experiments were carried out using HF488-labeled Aldrich, 1:1000 in PBS). For immunostainings on pri- Aβ oligomers according to a standard protocol as mary cells, cells were fixed with 2% PFA for 20  min on 1-42 described previously [13]. The images were recorded by ice. Next, cells were washed three times with PBS and time-correlated single photon counting (TCSPC) using permeabilized with 0.1% Triton X-100 for 10 min on ice. Zeiss LSM780 confocal microscope equipped with a Samples were washed with blocking buffer (1% BSA in MaiTai DeepSee multiphoton laser (SpectraPhysics) and PBS) and incubated with primary Abs (diluted in block- a modular FLIM system (Becker & Hickl). The emitted ing buffer) for 2 h at RT/overnight at 4 °C. After washing photons were detected with a hybrid detection module with PBS, cells were stained with fluorophore-conjugated (HPM-100–40). Data was recorded by a Simple-Tau sys- secondary Abs in PBS for 1 h at RT. Next, samples were tem (M1-SPC-150 FLIM module, Becker & Hickl) with counterstained with Hoechst and the sections were the instrument recording software SPCM version 9.52 in Xie  et al. acta neuropathol commun (2021) 9:163 Page 21 of 23 mounted. Confocal laser scanning microscopy was per- proximity of Aβ plaques between 2 groups were ana- formed using Zeiss LSM780 or Zeiss Axioscan Z.1. lyzed using nonparametric Mann–Whitney U test. Com- parisons of percentages of body weight loss between 2 Image analysis groups were analyzed using 2-way repeated measures Quantification was performed using ImageJ soft - ANOVA. Comparison of 2 factors was analyzed using ware (version 1.53c, National Institutes of Health) and 2-way ANOVA with Bonferroni’s post hoc test for mul- reported as area fraction of region of interest. The tiple comparisons for parametric data. GraphPad Prism expression levels of TJ proteins and transporter proteins 8.0 was used for statistical analysis. Differences were con - were calculated as relative protein positive signal inten- sidered significant at P < 0.05. sity divided by protein positive area. Infiltrating immune + − cells were calculated as the number of IBA1 TMEM119 Supplementary Information cells. For inside or outside Aβ plaque-associated micro- The online version contains supplementary material available at https:// doi. org/ 10. 1186/ s40478- 021- 01253-z. glia quantification, IBA1 cells differentiated by whether they had contact with Aβ plaques and were manually Additional file 1. Representative movie of primary neuron culture after counted. For quantification of the Aβ internalization treatment with PBS for 5 h and traced for 24.h + + ratio, the area of Aβ-internalized microglia (Aβ IBA1 ) Additional file 2. Representative movie of primary neuron culture after was normalized to the Aβ-positive area. The expression treatment with PBS for 5 h, followed by theaddition of Aβ HF488 1-42 of synaptic molecule expression was calculated as per- oligomers and traced for 24 h. centage of SYP within the Hoechst area. Additional file 3. Representative movie of primary neuron culture after treatment with IL-1β for 5 h, followed by theaddition of Aβ HF488 1-42 oligomers and traced for 24 h. Microglia 3D reconstruction Additional file 4. Supplementary Figures S1-S12 and Appendix Tables For 3D reconstruction of microglia, 50  μm vibratome S1-S2. brain sections were stained overnight with anti-IBA1 Ab at 4 °C, followed by secondary Abs for 2 h at RT. Z-stack Acknowledgements images were taken with a Zeiss LSM 880 with Fast Airys- We thank M. Mercken (Johnson & Johnson Pharmaceuticals Research and can (Zeiss, Germany), using a Plan-Apochromat 40 × 1.3 Development, Beerse, Belgium) and Bart De Strooper ( VIB-KU Leuven) for providing the antibodies against Aβ (JRF/ABN/24, JRF/cAB40/28, and JRF/ oil DIC UV-IR M27 objective. The 3D reconstructions cAB42/26) for the Aβ ELISA. We thank the VIB BioImaging Core for training, and measurements were done by filament tracing algo - support and access to the instrument park. This work was supported by the rithm from Imaris software (Bitplane). Research Foundation-Flanders (FWO), Chinese Scholarship Council (CSC), The Foundation for Alzheimer’s Research Belgium (SAO-FRA) and the Baillet Latour Fund. The graphical abstract was created with BioRender.com. Western blot Authors’ contributions Hippocampus extracts were prepared in 0.5% CHAPS J.X., N.G., L.V.H., and R.E.V. designed experiments. J.X., N.G., C.V., G.V.I., E.V.W, buffer containing complete protease inhibitor (Thermo C.V.C. and L.V.H performed experiments. J.X., L.V.H., and R.E.V. wrote the manu- Scientific) and centrifuged for 15 min at 20,000g at 4 °C, script. All authors read and approved the final manuscript. and protein concentration was determined using a Pierce BCA Protein Assay Kit (Thermo Scientific). Supernatants Declarations were denatured in 5xLaemmli buffer, separated by SDS- Competing interests PAGE gel electrophoresis, and transferred to nitrocel- The authors have declared that no conflict of interest exists. lulose. Following blocking with Odyssey Blocking buffer Author details (LI-COR Biosciences), the membrane was first incubated VIB-UGent Center for Inflammation Research, Technologiepark -Zwijnaarde 71, with primary Abs, and then with fluorophore-conjugated 9052 Ghent, Belgium. Department of Biomedical Molecular Biology, Ghent secondary Abs. Protein bands were visualized Odyssey Fc University, 9000 Ghent, Belgium. VIB BioImaging Core, VIB, Ghent, Belgium. Imaging System (LI-COR Biosciences) and quantification Received: 26 May 2021 Accepted: 1 September 2021 was done in Image Studio (LI-COR Biosciences). Statistics Data are shown as mean ± SEM. 2-tailed Student’s t test References for parametric data was used to compare 2 groups. To 1. Alzheimer’s disease facts and figures. Alzheimers Dement (2020). https:// doi. org/ 10. 1002/ alz. 12068 compare 3 or more groups, 1-way ANOVA with Bon- 2. Balusu S, Van Wonterghem E, De Rycke R, Raemdonck K, Stremersch S, ferroni’s post hoc test for multiple comparisons for Gevaert K, Brkic M, Demeestere D, Vanhooren V, Hendrix A et al (2016) parametric data was used. Comparisons of the relation- Identification of a novel mechanism of blood-brain communication during peripheral inflammation via choroid plexus-derived extracellular ship between the activation state of microglia and the Xie et al. acta neuropathol commun (2021) 9:163 Page 22 of 23 vesicles. EMBO Mol Med 8:1162–1183. https:// doi. org/ 10. 15252/ emmm. 20. Gomez-Nicola D, Schetters ST, Perry VH (2014) Differential role of CCR2 in 20160 6271 the dynamics of microglia and perivascular macrophages during prion 3. Bard F, Cannon C, Barbour R, Burke RL, Games D, Grajeda H, Guido T, Hu disease. Glia 62:1041–1052. https:// doi. org/ 10. 1002/ glia. 22660 K, Huang J, Johnson-Wood K et al (2000) Peripherally administered anti- 21. Gu H, Zhong Z, Jiang W, Du E, Dodel R, Liu J, Farlow MR, Zheng W, Du Y bodies against amyloid beta-peptide enter the central nervous system (2014) The role of choroid plexus in IVIG-induced beta-amyloid clearance. and reduce pathology in a mouse model of Alzheimer disease. Nat Med Neuroscience 270:168–176. https:// doi. org/ 10. 1016/j. neuro scien ce. 2014. 6:916–919. https:// doi. org/ 10. 1038/ 7868204. 011 4. Bartels T, De Schepper S, Hong S (2020) Microglia modulate neurode- 22. Haruwaka K, Ikegami A, Tachibana Y, Ohno N, Konishi H, Hashimoto A, generation in Alzheimer’s and Parkinson’s diseases. Science 370:66–69. Matsumoto M, Kato D, Ono R, Kiyama H et al (2019) Dual microglia effects https:// doi. org/ 10. 1126/ scien ce. abb85 87 on blood brain barrier permeability induced by systemic inflammation. 5. Baruch K, Kertser A, Porat Z, Schwartz M (2015) Cerebral nitric oxide Nat Commun 10:5816. https:// doi. org/ 10. 1038/ s41467- 019- 13812-z represses choroid plexus NFkappaB-dependent gateway activity for 23. Heneka MT, Carson MJ, El Khoury J, Landreth GE, Brosseron F, Feinstein leukocyte trafficking. EMBO J 34:1816–1828. https:// doi. org/ 10. 15252/ DL, Jacobs AH, Wyss-Coray T, Vitorica J, Ransohoff RM et al (2015) Neuro - embj. 20159 1468 inflammation in Alzheimer’s disease. Lancet Neurol 14:388–405. https:// 6. Brait VH, Miro-Mur F, Perez-de-Puig I, Notario L, Hurtado B, Pedragosa J, doi. org/ 10. 1016/ S1474- 4422(15) 70016-5 Gallizioli M, Jimenez-Altayo F, Arbaizar-Rovirosa M, Otxoa-de-Amezaga A 24. Heneka MT, Golenbock DT, Latz E (2015) Innate immunity in Alzheimer’s et al (2019) CD69 plays a beneficial role in ischemic stroke by dampening disease. Nat Immunol 16:229–236. https:// doi. org/ 10. 1038/ ni. 3102 endothelial activation. Circ Res 124:279–291. https:// doi. org/ 10. 1161/ 25. Heneka MT, Kummer MP, Latz E (2014) Innate immune activation in neu- CIRCR ESAHA. 118. 313818 rodegenerative disease. Nat Rev Immunol 14:463–477. https:// doi. org/ 10. 7. Brkic M, Balusu S, Van Wonterghem E, Gorle N, Benilova I, Kremer A, 1038/ nri37 05 Van Hove I, Moons L, De Strooper B, Kanazir S et al (2015) Amyloid beta 26. Heppner FL, Ransohoff RM, Becher B (2015) Immune attack: the role of oligomers disrupt blood-CSF barrier integrity by activating matrix metal- inflammation in Alzheimer disease. Nat Rev Neurosci 16:358–372. https:// loproteinases. J Neurosci 35:12766–12778. https:// doi. org/ 10. 1523/ JNEUR doi. org/ 10. 1038/ nrn38 80 OSCI. 0006- 15. 2015 27. Hickman SE, Allison EK, El Khoury J (2008) Microglial dysfunction and 8. Carro E, Spuch C, Trejo JL, Antequera D, Torres-Aleman I (2005) Choroid defective beta-amyloid clearance pathways in aging Alzheimer’s disease plexus megalin is involved in neuroprotection by serum insulin-like mice. J Neurosci 28:8354–8360. https:// doi. org/ 10. 1523/ JNEUR OSCI. 0616- growth factor I. J Neurosci 25:10884–10893. https:// doi. org/ 10. 1523/ 08. 2008 JNEUR OSCI. 2909- 05. 2005 28. Holmes C (2013) Review: systemic inflammation and Alzheimer’s disease. 9. Chen YR, Glabe CG (2006) Distinct early folding and aggregation proper- Neuropathol Appl Neurobiol 39:51–68. https:// doi. org/ 10. 1111/j. 1365- ties of Alzheimer amyloid-beta peptides Abeta40 and Abeta42: stable 2990. 2012. 01307.x trimer or tetramer formation by Abeta42. J Biol Chem 281:24414–24422. 29. Holmes C, Cunningham C, Zotova E, Woolford J, Dean C, Kerr S, Culliford https:// doi. org/ 10. 1074/ jbc. M6023 63200 D, Perry VH (2009) Systemic inflammation and disease progression in 10. Cummings J, Lee G, Ritter A, Sabbagh M, Zhong K (2020) Alzheimer’s Alzheimer disease. Neurology 73:768–774. https:// doi. org/ 10. 1212/ WNL. disease drug development pipeline: 2020. Alzheimers Dement (N Y ) 0b013 e3181 b6bb95 6:e12050. https:// doi. org/ 10. 1002/ trc2. 12050 30. Hoogland IC, Houbolt C, van Westerloo DJ, van Gool WA, van de Beek 11. Dani M, Wood M, Mizoguchi R, Fan Z, Walker Z, Morgan R, Hinz R, Biju M, D (2015) Systemic inflammation and microglial activation: systematic Kuruvilla T, Brooks DJ et al (2018) Microglial activation correlates in vivo review of animal experiments. J Neuroinflammation 12:114. https:// doi. with both tau and amyloid in Alzheimer’s disease. Brain 141:2740–2754. org/ 10. 1186/ s12974- 015- 0332-6 https:// doi. org/ 10. 1093/ brain/ awy188 31. Jendresen C, Digre A, Cui H, Zhang X, Vlodavsky I, Li JP, Nilsson LNG (2019) 12. DeTure MA, Dickson DW (2019) The neuropathological diagnosis of Systemic LPS-induced Abeta-solubilization and clearance in AbetaPP- Alzheimer’s disease. Mol Neurodegener 14:32. https:// doi. org/ 10. 1186/ transgenic mice is diminished by heparanase overexpression. Sci Rep s13024- 019- 0333-5 9:4600. https:// doi. org/ 10. 1038/ s41598- 019- 40999-4 13. Esbjorner EK, Chan F, Rees E, Erdelyi M, Luheshi LM, Bertoncini CW, 32. Ji K, Akgul G, Wollmuth LP, Tsirka SE (2013) Microglia actively regulate the Kaminski CF, Dobson CM, Kaminski Schierle GS (2014) Direct observations number of functional synapses. PLoS ONE 8:e56293. https:// doi. org/ 10. of amyloid beta self-assembly in live cells provide insights into differ -1371/ journ al. pone. 00562 93 ences in the kinetics of Abeta(1–40) and Abeta(1–42) aggregation. Chem 33. Kadowaki H, Nishitoh H, Urano F, Sadamitsu C, Matsuzawa A, Takeda K, Biol 21:732–742. https:// doi. org/ 10. 1016/j. chemb iol. 2014. 03. 014 Masutani H, Yodoi J, Urano Y, Nagano T et al (2005) Amyloid beta induces 14. Frost GR, Jonas LA, Li YM (2019) Friend, foe or both? Immune activity neuronal cell death through ROS-mediated ASK1 activation. Cell Death in Alzheimer’s disease. Front Aging Neurosci 11:337. https:// doi. org/ 10. Differ 12:19–24. https:// doi. org/ 10. 1038/ sj. cdd. 44015 28 3389/ fnagi. 2019. 00337 34. Keaney J, Walsh DM, O’Malley T, Hudson N, Crosbie DE, Loftus T, Sheehan 15. Fujiyoshi M, Tachikawa M, Ohtsuki S, Ito S, Uchida Y, Akanuma S, Kamiie J, F, McDaid J, Humphries MM, Callanan JJ et al (2015) Autoregulated para- Hashimoto T, Hosoya K, Iwatsubo T et al (2011) Amyloid-beta pep- cellular clearance of amyloid-beta across the blood-brain barrier. Sci Adv tide(1–40) elimination from cerebrospinal fluid involves low-density 1:e1500472. https:// doi. org/ 10. 1126/ sciadv. 15004 72 lipoprotein receptor-related protein 1 at the blood-cerebrospinal fluid 35. Labzin LI, Heneka MT, Latz E (2018) Innate Immunity and Neurode- barrier. J Neurochem 118:407–415. https:// doi. org/ 10. 1111/j. 1471- 4159. generation. Annu Rev Med 69:437–449. https:// doi. org/ 10. 1146/ annur 2011. 07311.xev- med- 050715- 104343 16. Furube E, Kawai S, Inagaki H, Takagi S, Miyata S (2018) Brain Region- 36. Leng F, Edison P (2020) Neuroinflammation and microglial activation in dependent Heterogeneity and Dose-dependent Difference in Transient Alzheimer disease: where do we go from here? Nat Rev Neurol. https:// Microglia Population Increase during Lipopolysaccharide-induced Inflam-doi. org/ 10. 1038/ s41582- 020- 00435-y mation. Sci Rep 8:2203. https:// doi. org/ 10. 1038/ s41598- 018- 20643-3 37. Lopez-Rodriguez AB, Hennessy E, Murray CL, Nazmi A, Delaney HJ, Healy 17. Gandy S (2012) Lifelong management of amyloid-beta metabolism to D, Fagan SG, Rooney M, Stewart E, Lewis A et al (2021) Acute systemic prevent Alzheimer’s disease. N Engl J Med 367:864–866. https:// doi. org/ inflammation exacerbates neuroinflammation in Alzheimer’s disease: 10. 1056/ NEJMe 12079 95 IL-1beta drives amplified responses in primed astrocytes and neuronal 18. Glenner GG, Wong CW (1984) Alzheimer’s disease: initial report of the network dysfunction. Alzheimers Dement. https:// doi. org/ 10. 1002/ alz. purification and characterization of a novel cerebrovascular amyloid 12341 protein. Biochem Biophys Res Commun 120:885–890. https:// doi. org/ 10. 38. Marques F, Sousa JC (2015) The choroid plexus is modulated by various 1016/ S0006- 291X(84) 80190-4 peripheral stimuli: implications to diseases of the central nervous system. 19. Goate A, Chartier-Harlin MC, Mullan M, Brown J, Crawford F, Fidani L, Giuf- Front Cell Neurosci 9:136. https:// doi. org/ 10. 3389/ fncel. 2015. 00136 fra L, Haynes A, Irving N, James L et al (1991) Segregation of a missense 39. Marques F, Sousa JC, Coppola G, Geschwind DH, Sousa N, Palha JA, mutation in the amyloid precursor protein gene with familial Alzheimer’s Correia-Neves M (2009) The choroid plexus response to a repeated disease. Nature 349:704–706. https:// doi. org/ 10. 1038/ 34970 4a0 Xie  et al. acta neuropathol commun (2021) 9:163 Page 23 of 23 peripheral inflammatory stimulus. BMC Neurosci 10:135. https:// doi. org/ 58. Sheppard O, Coleman MP, Durrant CS (2019) Lipopolysaccharide-induced 10. 1186/ 1471- 2202- 10- 135 neuroinflammation induces presynaptic disruption through a direct 40. Michaud JP, Halle M, Lampron A, Theriault P, Prefontaine P, Filali M, action on brain tissue involving microglia-derived interleukin 1 beta. J Tribout-Jover P, Lanteigne AM, Jodoin R, Cluff C et al (2013) Toll-like recep - Neuroinflamm 16:106. https:// doi. org/ 10. 1186/ s12974- 019- 1490-8 tor 4 stimulation with the detoxified ligand monophosphoryl lipid A 59. Sierksma A, Lu A, Mancuso R, Fattorelli N, Thrupp N, Salta E, Zoco J, Blum improves Alzheimer’s disease-related pathology. Proc Natl Acad Sci USA D, Buee L, De Strooper B et al (2020) Novel Alzheimer risk genes deter- 110:1941–1946. https:// doi. org/ 10. 1073/ pnas. 12151 65110 mine the microglia response to amyloid-beta but not to TAU pathology. 41. Neher JJ, Neniskyte U, Hornik T, Brown GC (2014) Inhibition of UDP/P2Y6 EMBO Mol Med 12:e10606. https:// doi. org/ 10. 15252/ emmm. 20191 0606 purinergic signaling prevents phagocytosis of viable neurons by acti- 60. Sochocka M, Diniz BS, Leszek J (2017) Inflammatory response in the CNS: vated microglia in vitro and in vivo. Glia 62:1463–1475. https:// doi. org/ 10. friend or foe? Mol Neurobiol 54:8071–8089. https:// doi. org/ 10. 1007/ 1002/ glia. 22693s12035- 016- 0297-1 42. Nishihara H, Soldati S, Mossu A, Rosito M, Rudolph H, Muller WA, 61. Steeland S, Gorle N, Vandendriessche C, Balusu S, Brkic M, Van Cauwen- Latorre D, Sallusto F, Sospedra M, Martin R et al (2020) Human CD4(+) berghe C, Van Imschoot G, Van Wonterghem E, De Rycke R, Kremer A T cell subsets differ in their abilities to cross endothelial and epithelial et al (2018) Counteracting the effects of TNF receptor-1 has therapeutic brain barriers in vitro. Fluids Barriers CNS 17:3. https:// doi. org/ 10. 1186/ potential in Alzheimer’s disease. EMBO Mol Med 10:e8300. https:// doi. s12987- 019- 0165-2org/ 10. 15252/ emmm. 20170 8300 43. Orellana JA, Shoji KF, Abudara V, Ezan P, Amigou E, Saez PJ, Jiang JX, Naus 62. Storck SE, Meister S, Nahrath J, Meissner JN, Schubert N, Di Spiezio A, CC, Saez JC, Giaume C (2011) Amyloid beta-induced death in neurons Baches S, Vandenbroucke RE, Bouter Y, Prikulis I et al (2016) Endothelial involves glial and neuronal hemichannels. J Neurosci 31:4962–4977. LRP1 transports amyloid-beta(1–42) across the blood-brain barrier. J Clin https:// doi. org/ 10. 1523/ JNEUR OSCI. 6417- 10. 2011 Invest 126:123–136. https:// doi. org/ 10. 1172/ JCI81 108 44. Ott BR, Jones RN, Daiello LA, de la Monte SM, Stopa EG, Johanson CE, 63. Swartzlander DB, Propson NE, Roy ER, Saito T, Saido T, Wang B, Zheng Denby C, Grammas P (2018) Blood-cerebrospinal fluid barrier gradients H (2018) Concurrent cell type-specific isolation and profiling of mouse in mild cognitive impairment and Alzheimer’s Disease: relationship to brains in inflammation and Alzheimer’s disease. JCI Insight 3:e121109. inflammatory cytokines and chemokines. Front Aging Neurosci 10:245. https:// doi. org/ 10. 1172/ jci. insig ht. 121109 https:// doi. org/ 10. 3389/ fnagi. 2018. 00245 64. Tejera D, Mercan D, Sanchez-Caro JM, Hanan M, Greenberg D, Soreq H, 45. Palop JJ, Mucke L (2010) Amyloid-beta-induced neuronal dysfunction in Latz E, Golenbock D, Heneka MT (2019) Systemic inflammation impairs Alzheimer’s disease: from synapses toward neural networks. Nat Neurosci microglial Abeta clearance through NLRP3 inflammasome. EMBO J 13:812–818. https:// doi. org/ 10. 1038/ nn. 2583 38:e101064. https:// doi. org/ 10. 15252/ embj. 20181 01064 46. Park JC, Han SH, Mook-Jung I (2020) Peripheral inflammatory biomarkers 65. Thomson CA, McColl A, Graham GJ, Cavanagh J (2020) Sustained expo- in Alzheimer’s disease: a brief review. BMB Rep 53: 10–19. https:// doi. org/ sure to systemic endotoxin triggers chemokine induction in the brain 10. 5483/ BMBRep. 2020. 53.1. 309 followed by a rapid influx of leukocytes. J Neuroinflamm 17:94. https:// 47. Pascale CL, Miller MC, Chiu C, Boylan M, Caralopoulos IN, Gonzalez L, doi. org/ 10. 1186/ s12974- 020- 01759-8 Johanson CE, Silverberg GD (2011) Amyloid-beta transporter expres- 66. Vandenbroucke RE, Dejonckheere E, Van Lint P, Demeestere D, Van sion at the blood-CSF barrier is age-dependent. Fluids Barriers CNS 8:21. Wonterghem E, Vanlaere I, Puimege L, Van Hauwermeiren F, De Rycke R, https:// doi. org/ 10. 1186/ 2045- 8118-8- 21 Mc Guire C et al (2012) Matrix metalloprotease 8-dependent extracellular 48. Qin L, Wu X, Block ML, Liu Y, Breese GR, Hong JS, Knapp DJ, Crews FT matrix cleavage at the blood-CSF barrier contributes to lethality during (2007) Systemic LPS causes chronic neuroinflammation and progressive systemic inflammatory diseases. J Neurosci 32:9805–9816. https:// doi. neurodegeneration. Glia 55:453–462. https:// doi. org/ 10. 1002/ glia. 20467org/ 10. 1523/ JNEUR OSCI. 0967- 12. 2012 49. Reed-Geaghan EG, Croxford AL, Becher B, Landreth GE (2020) Plaque- 67. Walker KA, Ficek BN, Westbrook R (2019) Understanding the Role of associated myeloid cells derive from resident microglia in an Alzheimer’s Systemic Inflammation in Alzheimer’s Disease. ACS Chem Neurosci disease model. J Exp Med. https:// doi. org/ 10. 1084/ jem. 20191 374 10:3340–3342. https:// doi. org/ 10. 1021/ acsch emneu ro. 9b003 33 50. Ries M, Sastre M (2016) Mechanisms of Abeta clearance and degradation 68. Walker KA, Gottesman RF, Wu A, Knopman DS, Gross AL, Mosley TH Jr, by glial cells. Front Aging Neurosci 8:160. https:// doi. org/ 10. 3389/ fnagi. Selvin E, Windham BG (2019) Systemic inflammation during midlife and 2016. 00160 cognitive change over 20 years: The ARIC Study. Neurology 92:e1256– 51. Riester K, Brawek B, Savitska D, Frohlich N, Zirdum E, Mojtahedi N, Heneka e1267. https:// doi. org/ 10. 1212/ WNL. 00000 00000 007094 MT, Garaschuk O (2020) In vivo characterization of functional states of 69. Wang Y, Jin S, Sonobe Y, Cheng Y, Horiuchi H, Parajuli B, Kawanokuchi cortical microglia during peripheral inflammation. Brain Behav Immun J, Mizuno T, Takeuchi H, Suzumura A (2014) Interleukin-1beta induces 87:243–255. https:// doi. org/ 10. 1016/j. bbi. 2019. 12. 007 blood-brain barrier disruption by downregulating Sonic hedgehog in 52. Saito T, Matsuba Y, Mihira N, Takano J, Nilsson P, Itohara S, Iwata N, Saido astrocytes. PLoS ONE 9:e110024. https:// doi. org/ 10. 1371/ journ al. pone. TC (2014) Single App knock-in mouse models of Alzheimer’s disease. Nat 01100 24 Neurosci 17:661–663. https:// doi. org/ 10. 1038/ nn. 3697 70. Whiten DR, Brownjohn PW, Moore S, De S, Strano A, Zuo Y, Haneklaus M, 53. Sasaguri H, Nilsson P, Hashimoto S, Nagata K, Saito T, De Strooper B, Hardy Klenerman D, Livesey FJ (2020) Tumour necrosis factor induces increased J, Vassar R, Winblad B, Saido TC (2017) APP mouse models for Alzheimer’s production of extracellular amyloid-β- and α-synuclein-containing aggre- disease preclinical studies. EMBO J 36:2473–2487. https:// doi. org/ 10. gates by human Alzheimer’s disease neurons. Brain Communications 2: 15252/ embj. 20179 7397 fcaa146. https:// doi. org/ 10. 1093/ brain comms/ fcaa1 46 54. Semmler A, Frisch C, Debeir T, Ramanathan M, Okulla T, Klockgether T, 71. You LH, Yan CZ, Zheng BJ, Ci YZ, Chang SY, Yu P, Gao GF, Li HY, Dong TY, Heneka MT (2007) Long-term cognitive impairment, neuronal loss and Chang YZ (2017) Astrocyte hepcidin is a key factor in LPS-induced neu- reduced cortical cholinergic innervation after recovery from sepsis in a ronal apoptosis. Cell Death Dis 8:e2676. https:// doi. org/ 10. 1038/ cddis. rodent model. Exp Neurol 204:733–740. https:// doi. org/ 10. 1016/j. expne 2017. 93 urol. 2007. 01. 003 72. Yuan P, Condello C, Keene CD, Wang Y, Bird TD, Paul SM, Luo W, Colonna 55. Semmler A, Okulla T, Sastre M, Dumitrescu-Ozimek L, Heneka MT (2005) M, Baddeley D, Grutzendler J (2016) TREM2 haplodeficiency in mice and Systemic inflammation induces apoptosis with variable vulnerability of humans impairs the microglia barrier function leading to decreased different brain regions. J Chem Neuroanat 30:144–157. https:// doi. org/ 10. amyloid compaction and severe axonal dystrophy. Neuron 92:252–264. 1016/j. jchem neu. 2005. 07. 003https:// doi. org/ 10. 1016/j. neuron. 2016. 09. 016 56. Sharma K, Wu LJ, Eyo UB (2020) Calming neurons with a microglial touch. Trends Neurosci 43:197–199. https:// doi. org/ 10. 1016/j. tins. 2020. 01. 008 Publisher’s Note 57. Sheng JG, Bora SH, Xu G, Borchelt DR, Price DL, Koliatsos VE (2003) Springer Nature remains neutral with regard to jurisdictional claims in pub- Lipopolysaccharide-induced-neuroinflammation increases intracellular lished maps and institutional affiliations. accumulation of amyloid precursor protein and amyloid beta peptide in APPswe transgenic mice. Neurobiol Dis 14:133–145. https:// doi. org/ 10. 1016/ s0969- 9961(03) 00069-x

Journal

Acta Neuropathologica CommunicationsSpringer Journals

Published: Oct 7, 2021

Keywords: Low-grade peripheral inflammation;  Brain barriers; Choroid plexus;  Blood-CSF barrier; Alzheimer's disease

References