Abstract
Alzheimer’s disease (AD) progression has been associated with the presence of brain-resident CD8+ T cells, and recent studies suggest a potential role of the CXCL16-CXCR6 axis in their recruitment to the brain. Here, we examined publicly available single-cell RNA sequencing datasets revealing that in the mouse brain, the receptor Cxcr6 is mainly expressed by CD8+ T cells, while the expression of its ligand Cxcl16 is predominantly observed in microglial cells. We found higher levels of Cxcl16 and Cxcr6 expression in APP/PS1 compared to wild-type mice. Furthermore, in vitro experiments using immortalized and primary murine cells suggested that Cxcl16 expression is driven by Aβ pathology. In contrast to our expectations, no changes in the number of Cxcr6+CD8+ cells was evident in the brains of microglia-depleted APP/PS1 mice, treated with the CSF1R antagonist PLX5622. This was related to an increased compensatory Cxcl16 expression by depletion-resistant microglia or by other brain-resident myeloid cells. Although we demonstrated a strong association between microglial Cxcl16 and AD pathology, PLX5622-sensitive microglia are dispensable in the recruitment of Cxcr6+CD8+ T cells to the brain of APP/PS1 mice. Future in vivo analysis will help to dissect the mechanism of CD8+ T cell recruitment to the brain.
Introduction
The classic neuropathological hallmarks of Alzheimer’s disease (AD) include the deposition of brain amyloid-β (Aβ) aggregates in the form of senile plaques and intracellular accumulation of hyperphosphorylated tau protein, which constitutes the characteristic neurofibrillary tangles1,2. However, neuroinflammation has been increasingly recognized to play crucial roles in AD progression3,4,5.
Seminal studies have demonstrated that, in addition to astrogliosis and microglial activation, other inflammatory processes provoke an adaptive immune response in AD. This includes the infiltration of CD8+ T cells into the disease-affected brain parenchyma6,7,8,9,10,11. Immunohistochemical (IHC) analysis of post-mortem brain specimens from AD-affected patients and AD-mouse models revealed a close contact of CD8+ T cells with microglia and neurons6,12. However, the molecular and cellular mechanism involved in the recruitment of CD8+ T cells to the brain parenchyma of AD individuals has not been clarified yet.
Recent data suggest the involvement of the CXCL16-CXCR6 axis in CD8+ T cell recruitment and accumulation in the diseased brain. The human transmembrane multi-domain chemokine (C-X-C motif) Ligand 16 (CXCL16), typically acts in its soluble form as a chemoattractant for Chemokine (C-X-C motif) Receptor 6 (CXCR6)-expressing lymphocytes13,14,15. In its membrane bound form, CXCL16 serves as a cell-cell-adhesion molecule binding to CXCR6 molecules expressed on the surface of attracted cells15,16. CXCR6, the only known receptor for CXCL16, is the classical homing receptor expressed on human tissue resident memory T (TRM) cells17.
Single-cell RNA sequencing (scRNA-seq) analysis of mice infected with central nervous system (CNS)-trophic virus, and cerebrospinal fluid (CSF) of cognitively impaired patients, indicated a potential contribution of the CXCL16-CXCR6 axis to the accumulation of CD8+ T cells in the brain18,19. Moreover, in a recent transcriptomic analysis on murine brain-isolated CD3+CD8+ T cells, we demonstrated higher Cxcr6 (the murine orthologue of CXCR6) expression levels compared to their blood counterparts. In particular, counting of hippocampal Cxcr6+CD8+ cells revealed higher numbers of this T cell subtype in APP/PS1 mice, a widely used animal model of AD20. Thus, CD8+ T cells expressing CXCR6 might infiltrate the brain via the CXCL16-CXCR6 axis as a potential molecular mechanism. Consequently, gaining insight into the cellular origin of CXCL16 within the brain would enhance our understanding of the mechanisms underlying the infiltration of parenchymal CD8+ T cells.
Here, we aim to assess the interplay between Cxcr6+CD8+ cells and Cxcl16 (the murine orthologue of CXCL16)-producing cells in the APP/PS1 brain. To this end, we analyzed previously published scRNA-seq murine datasets. We then performed quantitative PCR (qPCR), Western Blot (WB) and Enzyme-Linked Immunosorbent Assay (ELISA) on Aβ-treated immortalized and primary microglia and macrophage cell cultures, to assess Cxcl16 expression levels in in vitro AD-like conditions. Subsequently, we analyzed the brains of mice treated with PLX5622, a microglia-depleting agent inhibiting colony-stimulating factor 1 receptor (CSF1R/Csf1r), to investigate the role of microglia-produced Cxcl16 in the recruitment of Cxcr6+CD8+ cells in vivo. Finally, by taking advantage of other scRNA-seq datasets, we hypothesisze the existence of a compensatory mechanism that supports Cxcl16 expression following the pharmacological and genetic inhibition of Csf1r.
Results
The cellular sources of Cxcr6 and Cxcl16 in brain immune cells
We re-clustered scRNA-seq data of CD45+cells, sorted from brain tissue of APP/PS1 and WT animals21, using unsupervised graph-based clustering by Seurat (Fig. 1a). Cell type annotation was automatically performed using comprehensive immune system cell markers database22, (Fig. 1b). Cell clusters percentage revealed genotype-related differences, especially for the “Effector CD8+ T cells” cluster, which showed increased proportion in the APP/PS1 brain (Supplementary Figure S1a). Interestingly, re-clustering and automated cell-based annotation of CD45+ cells from 9- and 16-month-old APP/PS1 mice revealed a higher proportion of “Effector CD8+ T cells” in a later stage of the disease, confirming our recent data20 (Supplementary Figure S1b-d).
Cxcr6 and Cxcl16 expression in scRNA-seq from brain-derived CD45+ cells.
UMAP plots showed that in 16-month-old animals Cxcr6 transcript is mainly co-expressed together with the cytotoxic T cell marker Cd8a, while Cxcl16 expression is characteristic of brain macrophages identified by the expression of Aif1, which encodes for the microglial marker Iba1 (Fig. 1c). Violin plots further highlighted an increased presence of Cxcr6-expressing cells among the “Effector CD8+ T cells” (Wilcoxon test, p < 0.01) and the “Progenitor cells” (abbreviation for “Hematopoietic progenitor cells”, Wilcoxon test, p < 0.001) clusters in APP/PS1 mice compared to WT controls (Fig. 1d). Similarly, Cxcl16 expression is statistically significant expressed at higher levels in “Macrophages” cells from APP/PS1 compared to WT mice (Wilcoxon test, p < 0.05) (Fig. 1e).
Genotype specific differences are denoted also by the increased cell proportion and total amount of cells expressing Cxcr6 and Cxcl16 in APP/PS1 compared to WT brain (Supplementary Table S1−2).
Aβ pathology influences Cxcl16 expression in vitro and in vivo
To dissect the origin of Cxcl16 transcripts, we sub-clustered the “Macrophages” cell population, using previously established microglia/brain macrophage markers listed in the heat map (Supplementary Figure S2a-c). Violin plot revealed increased Cxcl16 expression levels in APP/PS1 “Disease-Associated Microglia” cell cluster (DAM, Wilcoxon test, p < 0.05) compared to WT (Fig. 2a), suggesting Cxcl16 as a signature marker of AD-related microglia states in APP/PS1 brain tissue.
Aβ pathology effect on Cxcl16 microglial expression in vitro and in vivo.
To explore the microglia-specific expression of Cxcl16 in the context of AD, we took advantage of a scRNA-seq dataset of CD11b+/CD45+ microglia isolated from the brain of the APP/PS1-L166P mouse model of AD, published by Sierksma et al., 202023. As expected, UMAP analysis revealed a marked genotype difference between APP/PS1-L166P and WT animals (Supplementary Figure S2d, e). Cxcl16 was expressed higher in all microglial subtypes of APP/PS1-L166P brain, with a statistically significant difference among “Cycling and proliferating microglia” (CPM, Wilcoxon test, p < 0.01), “Homeostatic microglia cluster 1” (HM.1, Wilcoxon test, p < 0.01), “Homeostatic microglia cluster 2” (HM.2, Wilcoxon test, p < 0.01), “Interferon-response microglia” (IRM, Wilcoxon test, p < 0.0001) and “Transitioning microglia” (TM, Wilcoxon test, p < 0.05) clusters (Fig. 2b), suggesting that Cxcl16 overexpression characterized both physiological and disease-associated microglial states in the APP/PS1-L166P model. The AD pathology-dependent microglial Cxcl16 expression is further supported by the IHC analysis of APP/PS1 dentate gyrus, which revealed the presence of Cxcl16 signal (white) in the proximity of Iba1+ microglia cells (red) around Aβ plaques (green) (Fig. 2c). To further understand whether Cxcl16 expression in microglial cells is affected by AD pathology, we exposed immortalized murine microglial (BV-2) cells to monomeric and aggregated Aβ42 peptides (Supplementary Figure S3a, the uncropped WB image is available in Supplementary Figure S4a). First, we confirmed a pro-inflammatory response to Aβ exposure of the BV-2 microglia, as demonstrated in the 2-fold increase of Tnf expression by qPCR (Friedman test, p = 0.0239, Dunn’s multiple comparisons test between vehicle- and aggregated Aβ42-treated cells, p = 0.0228, Supplementary Figure S3b). 24 h treatment with 1 µM of aggregated Aβ42 led to statistical significant increased expression of Cxcl16 transcript compared to vehicle-treated BV-2 cells (Friedman test, p = 0.0239, Dunn’s multiple comparisons test, p = 0.0342, Fig. 2d). To observe whether the change in mRNA expression is also reflected at the protein level, we performed WB analysis on monomeric and aggregated Aβ42-exposed BV-2 cells. Non-specific bands appeared in the WB analysis of the intracellular form of Cxcl16 (Supplementary Figure S4b). Subsequent experiments were therefore performed on the cell culture supernatant to analyse the secreted form of Cxcl16 (sCxcl16) (Fig. 2e, original WB images are available in Supplementary Figure S4b). Densitometric analysis revealed a statistically significant increase in sCxcl16 after the treatment with 1 µM aggregated Aβ42 (Friedman test, p = 0.0336, Dunn’s multiple comparisons test, p = 0.0429, Fig. 2f), suggesting an Aβ42 structure-dependent release of sCxcl16 from treated BV-2 cells.
To strengthen our hypothesis, we verified the expression level of Cxcl16 in a primary murine microglia cell culture system, since this more closely resembles the environment of the brain. Purity of mouse primary microglia cultures was assessed via Cd11b expression (Supplementary Figure S5a, b). In these experiments, we included a scrambled Aβ42 sequence to provide a more reliable control to evaluate the effect of Aβ42 peptide. Exposure to both monomeric and aggregated Aβ42, resulted in increased Tnf expression compared to scrambled Aβ42, with a statistical significant effect between scrambled and monomeric Aβ42 sequences (Friedman test, p < 0.0001, Dunn’s multiple comparisons test, p = 0.0212, Supplementary Figure S3c). Similarly, Cxcl16 transcript levels increased after treatment with monomeric and aggregated Aβ42. However, a statistically significant effect was only observed between scrambled and aggregated Aβ42 (Friedman test, p = 0.0002, Dunn’s multiple comparisons test, p = 0.0429, Fig. 2g), supporting the effect observed in BV-2 cells. We then quantified sCxcl16 in the supernatant of primary microglia treated with Aβ42 using ELISA, a more robust approach than WB for quantifying chemokines. The results indicated that both monomeric and aggregated forms induced a statistically significant increase in sCxcl16 when compared to scrambled Aβ42 sequence (Friedman test, p = 0.0018, Dunn’s multiple comparisons test, p = 0.0041 and p = 0.0081, respectively, Fig. 2h).
Further experiments were conducted using mouse bone marrow-derived macrophages (BMDMs), which express the typical macrophage markers CD11b and F4/80 (Supplementary Figure S5c, d), in order to establish whether the response to monomeric and aggregated Aβ42 is confined to CNS resident microglia cells. While these cells did not respond to Aβ42 in the same way as primary microglia in terms of Tnf expression (Friedman test, p = 0.2096, Supplementary Figure S3d), they exhibited significant Cxcl16 transcript upregulation upon exposure to monomeric and aggregated Aβ42 compared to the scrambled sequence (Friedman test, p = 0.0002, Dunn’s multiple comparisons test, p = 0.0044 and p = 0.0429, respectively, Fig. 2i). Furthermore, these gene expression changes were confirmed at the protein level. We observed a statistically significant increase in sCxcl16 in the supernatant of BMDMs exposed to both monomeric and aggregated Aβ42 when compared to scrambled control (Friedman test, p = 0.0024, Dunn’s multiple comparisons test, p = 0.0081 and p = 0.0041, respectively, Fig. 2j).
Overall, these in vitro experiments support the hypothesis that Cxcl16 expression in CNS-resident and peripheral macrophages depends on AD pathology.
Expression of Cxcl16 and Cxcr6 is elevated in APP/PS1 hippocampus and is not altered upon PLX5622 treatment
qPCR analysis on the hippocampus of 1-year-old mice revealed a significant upregulation of Cxcl16 in samples from the APP/PS1 mouse model compared to WT animals (Two-way ANOVA, p = 0.0029, Šídák’s multiple comparisons test between WT CTRL and APP/PS1 CTRL, p = 0.0171, Fig. 3a). Similar to Cxcl16, gene expression analysis indicated a significant overexpression of Cxcr6 in APP/PS1 hippocampus (Two-way ANOVA, p = 0.0327, Šídák’s multiple comparisons test between WT CTRL and APP/PS1 CTRL, p = 0.0239, Fig. 3b), supporting the role of the Cxcl16-Cxcr6 axis in APP/PS1 brain.
The role of the Cxcl16-Cxcr6 axis in the hippocampus of amyloidogenic mice after PLX5622-mediated microglia depletion.
To test the direct involvement of microglial cells in Cxcl16-mediated T cell recruitment, we performed qPCR analysis on the hippocampus of animals treated with PLX5622, a widely used microglia depleting agent24. Results indicated a slight reduction of Cxcl16 transcript (Two-way ANOVA, p = 0.0348, Fig. 3a), but no differences in Cxcr6 levels (Two-way ANOVA, p = 0.7042, Fig. 3b), upon PLX5622-treatment. Furthermore, post-hoc comparison analysis between control and PLX5622-treated APP/PS1 animals revealed a mild but not statistical significant reduction in Cxcl16 expression levels (Šídák’s multiple comparisons test, p = 0.095), implying a PLX5622-independent effect on its expression in the hippocampus of WT and APP/PS1 animals.
A weak correlation between Cxcl16 and Cxcr6 relative quantities was observed in control WT animals (Spearman R = 0.1429, p = 0.8028, Supplementary Figure S6a). Positive, although not significant, correlation was observed in PLX5622-treated WT and APP/PS1 mice (Spearman R = 0.8286, p = 0.0583, Supplementary Figure S6b, d). Nevertheless, the role of this chemokine-chemokine receptor axis in APP/PS1 hippocampus is supported by a strong positive correlation between Cxcl16 and Cxcr6 expression levels in APP/PS1 animals (Spearman R = 0.9429, p = 0.0167, Supplementary Figure S6c).
Cxcr6+CD8+ T cell number remain unchanged upon PLX5622 treatment in the hippocampus of APP/PS1 mice
To explore the potential involvement of microglia in regulating the number of brain-resident CD8+Cxcr6+ cells, we analyzed the hippocampal region of PLX5622-treated animals.
IHC analysis of CD8+Cxcr6+ cells (Fig. 3c) revealed genotype-dependent differences between APP/PS1 and WT animals (Two-way ANOVA, p = 0.0002, Šídák’s multiple comparisons test between WT CTRL and APP/PS1 CTRL, p = 0.0007, Supplementary Fig. 7a), as previously described20. However, we did not observe any differences in total CD8+Cxcr6+ cell numbers (Two-way ANOVA, p = 0.9125, Supplementary Figure S7a) and in the percentage of Cxcr6+ cells among total brain CD8+ cells upon PLX5622 treatment (Two-way ANOVA, p = 0.8619, Fig. 3d).
Pharmacological and genetic targeting of Csf1r expressing cells affects Cxcl16 expression in physiological conditions and in amyloidogenic brain tissue
To investigate the expression of Cxcl16 in mouse brain after microglia depletion, we analyzed scRNA-seq dataset of myeloid cells isolated from the brain of control and PLX5622-treated WT animals25. The analysis indicated a specific cell clustering organization characterizing PLX5622-treated mice, with a clear absence of the “Macrophages” cluster (Supplementary Figure 8a, b). Violin plot indicated a statistical significant increased expression level of Cxcl16 in “Interferon gamma signaling genes (ISG) expressing immune cells” (Wilcoxon test, p < 0.05) and “Neutrophils” (Wilcoxon test, p < 0.05) clusters of PLX5622-treated animals compared to untreated mice (Supplementary Figure 8c), pointing towards PLX5622-mediated overexpression of Cxcl16 also in other brain-resident, or brain-infiltrating, myeloid cells.
Furthermore, we analyzed a scRNA-seq dataset of brain cells from Murno et al., 202426. The mice included in the analysis have a deletion of the Fms intronic regulatory element (FIRE, Csf1rΔFIRE/ΔFIRE, herein abbreviated as KO), which is a highly conserved region that regulates CSF1R transcript elongation in macrophages. This mutation prevents differentiation of microglia and macrophages in the skin, kidney, heart, and peritoneum27,28. Although no genotype-dependent statistically significant differences in Cxcl16 expression were observed in 11–12-month-old animals (Supplementary Figure 9a-c), analysis of older mice (16–18-month-old) revealed elevated levels of Cxcl16 transcript expression in the myeloid cell compartment (including border-associated microglia (BAMs), dendritic cells (DCs) and monocytes (Mono), Wilcoxon test, p < 0.001), and in mural cells (Wilcoxon test, p < 0.0001), when compared with age-matched wild-type littermates (Csf1rWT/WT, referred to as WT, Supplementary Figure 9d-f).
To verify whether this compensatory expression of Cxcl16 also occurs in a brain affected by AD, we performed IHC analysis of Iba1+Cxcl16+ cells in the hippocampus of control and PLX5622-treated APP/PS1 mice (Fig. 3e). Since Cxcl16 signal was barely detectable in the brain tissue of either control and PLX5622-treated WT mice (Supplementary Figure 7b), Iba1+Cxcl16+ cells could not be quantified in WT mice. Quantification in brain tissue of APP/PS1 mice revealed an increased percentage of Cxcl16+ signals (white) among total brain Iba1+ cells (green) following PLX5622 treatment (Mann-Whitney test, p = 0.0087, Fig. 3f). Furthermore, we quantified the relative amount of Iba1+Cxcl16+ cells in close proximity to CD8+ cells (red, Fig. 3e). Therefore, we generated three-dimensional regions of interest (3D ROIs) for each channel (Supplementary Figure 7c) to measure the distance between CD8+ cells and Iba1+Cxcl16+ cells in x, y and z planes. By having defined a threshold of 100 μm distance, which is a measure previously reported in the literature to define T lymphocytes and Cxcl16 expressing cells proximity29, we found that the relative amount of Iba1+Cxcl16+ cells in close proximity with CD8+ cells did not change following PLX5622 treatment in APP/PS1 hippocampus (Mann-Whitney test, p = 0.8182, Supplementary Fig. 7d). In summary, while PLX5622 does not appear to influence the number of Iba1+Cxcl16+ cells in close proximity to CD8+ cells, it does lead to an increased percentage of Iba1+Cxcl16+ cells among the remaining Iba1+ cells.
Discussion
The accumulation of CD8+ T cells in AD brain tissue is redefining our knowledge about the role of the adaptive immune system in the neurodegenerative CNS. Identifying the molecular mechanism leading to parenchymal infiltration of CD8+ T cells helps to explore their function in the pathological process of AD. In the present study, we assessed the role of microglial Cxcl16 in the putative recruitment of Cxcr6+CD8+ T cells to the brain of APP/PS1 animals.
A recent investigation has shown that CXCL16-CXCR6 signalling between a particular subset of monocytes (CD14−CD16+, also referred to as “nonclassical monocytes”30) and CD8+ T cells is specific of CSF from cognitively impaired individuals. The increased availability of CXCL16 in the CSF, as well as the upregulation of CXCR6 in CD8+ T cells, points to this axis as a key mechanism for T cell entry into the brain18. Moreover, scRNA-seq analysis on West-Nile Virus (WNV) CNS-infection further supports the role of the CXCL16-CXCR6 axis in CD8+ T cells brain infiltration. In particular, the authors observed that anti-Cxcl16 antibody treatment prevents CD8+CD103+ T cells accumulation in the forebrain of infected animals19. Besides the CNS, this chemokine-chemokine receptor axis plays a critical role in the recruitment of CD8+ TRM, as observed in the airways of influenza virus-infected animals31.
However, the role of CXCL16-CXCR6 axis is apparently not only restricted to CD8+ T lymphocytes recruitment to specific tissues, but it is also involved in the accumulation of neutrophils in the CSF during pneumococcal meningitis infection32, in the migration of brain tumor cells33, in the progression of glial tumors34 and in breast cancer-associated brain metastasis35.
Our study aimed to dissect the putative role of the CXCL16-CXCR6 axis in AD-specific recruitment of CD8+ T cells to the brain. Indeed, qPCR analysis revealed a strong positive correlation between Cxcr6 and Cxcl16 expression in the hippocampus of APP/PS1 mice, pointing towards an intimate relation between these molecules in AD pathology. Taking advantage of publicly available murine scRNA-seq datasets, we confirmed that Cxcr6 transcript is observed in lymphocytes expressing Cd8a and we discovered that Cxcl16 expression characterizes brain macrophages, in particular, Aif1 expressing cells. To prove the microglial origin of Cxcl16, we further sub-clustered “Macrophages” cells from Van Hove et al., 2019 and we observed that Cxcl16 expression levels are increased in APP/PS1-associated DAMs. The increased mRNA expression levels of Cxcl16 observed in the hippocampus of 12-month-old APP/PS1 animals are in line with the onset of CD8+ T cell accumulation in the brain parenchyma6. Cxcl16 overexpression in the APP/PS1 hippocampus is corroborated by the upregulation of its human orthologue in different brain areas of AD-affected patients and the positive correlation of Cxcl16 protein level with Aβ and tau pathology, in other AD animal models36.
Interestingly, the involvement of human CXCL16 in cognitive impairment and dementia is also supported by other studies on human subjects, showing an inverse correlation between CXCL16 plasma levels and total brain volumes37 and a direct correlation of its serum levels with the brain presence of large atherosclerotic plaque and micro-embolic signs38. Moreover, increased CXCL16/Cxcl16 expression has been found in brain tissue of multiple sclerosis (MS) patients and corresponding animal models39,40,41, and in brain areas affected by ischemic insults42.
The AD-specific upregulation of Cxcl16 in CPM, IRM and TM cell clusters in the dataset from Sierksma et al., 2020 suggests that the expression of mutant APP and PSEN1 transgenes is one of the main drivers of Cxcl16 production. This is corroborated by the elevated Cxcl16 signals observed around Aβ plaques in representative IHC images of the APP/PS1 hippocampal area in mice. The influence of Aβ pathology in microglial Cxcl16 expression is further supported by the statistically significant overexpression of Cxcl16 transcript in aggregated Aβ42-treated BV-2 cells and monomeric Aβ42-treated primary microglia. Interestingly, while monomeric Aβ42 had no effect on sCxcl16 in BV-2 cells, we observed a statistically significant increase in sCxcl16 in the supernatant of primary microglia treated with either monomeric or aggregated Aβ42. This finding further supports the specific expression of Cxcl16 in microglia associated with AD pathology. Although studies indicate that the release of other cytokines from stimulated primary microglia cells might depend on Aβ42 conformation43,44, further experiments are needed to establish whether Aβ conformational state influences the expression or membrane shedding of the CXCL16 protein. These experiments could involve using in vitro- or in vivo-derived Aβ conformers (e.g. oligomers, protofibrils and fibrils), as well as inhibiting ADAM10 and ADAM17, which are involved in the proteolytic cleavage of CXCL1645,46.
The major genotype-related differences in Cxcr6 expression characterized the “Effector CD8+ T cells” cluster from Van Hove et al., 2019. This observation is in line with the increased Cxcr6 transcript levels we observed in the hippocampus of APP/PS1 animals and its co-localization with CD8+ T cells populating APP/PS1 mouse brains20. Nevertheless, besides AD-related conditions, Cxcr6+CD8+ T lymphocytes were also found to be characteristic for mixed active/inactive lesions in MS patients47,48, glioblastoma-affected tissues49 and Herpes Simplex Virus-infected trigeminal ganglia50. However, Cxcr6 expression is not only restricted to CD8+ T cells. Indeed, our scRNA-seq analysis revealed its expression in other brain-derived CD45+ cells, such as “hematopoietic progenitor cells” (especially in APP/PS1 brain), “ISG expressing cells”, “memory CD4+ T cells”, “Natural killer cells” and “γδ-T cells”, suggesting it as a common marker for immune cells recruited to the brain.
To address the role of microglial Cxcl16 in the recruitment of Cxcr6+CD8+ T cells, we analyzed brain tissue from animals treated with PLX5622, a drug targeting and depleting Csf1r-expressing microglial cells24. In the hippocampus of PLX5622-treated APP/PS1 mice, we observed a mild but not statistically significant reduction of Cxcl16 levels compared to untreated animals. Moreover, we did not achieve statistically significant changes in Cxcr6 expression upon PLX5622 treatment. These gene expression data were validated via IHC analysis of Cxcr6+CD8+ cells in the hippocampal region from control and microglia-depleted mice. Results revealed that, while a consistent reduction of Iba1+ cells is evident, no changes in the amount of Cxcr6+CD8+ cells were observed in PLX5622-treated APP/PS1 animals. We would have expected microglia as the major attractant for CD8+ T cells, as recently demonstrated in a mouse model of tauopathy51, but PLX5622-mediated microglia depletion did not reduce the amount of hippocampal Cxcr6+CD8+ cells. Our results are partially supported by previous findings relative to PLX5622 treatment. Indeed, flow cytometry analysis of PLX5622-mediated microglia depletion in APP/PS1 mouse model revealed elevated total levels of CD3+CD8+ T cell in brain homogenates12. An increased number of CD8+ T cells was also observed within the brains of neurotropic hepatitis virus-infected mice, treated with PLX562252. Moreover, although a reduction in Cxcl16 transcript levels was present in the brain of PLX5622-treated WNV-infected animals, an unexpected increase in the CD8+ T cell count was observed53. The biological reason for the presence of CD8+ T cells in the brain of microglia-depleted animals is probably due to peripheral changes resulting from systemic PLX5622 treatment, potentially involving other chemokine-chemokine-receptor axes. Our data suggests that CD8+ TRM cells, characterized by the expression of Cxcr6, are not affected by microglia depletion via PLX5622.
With the aim of finding a molecular reason explaining the presence of Cxcr6+CD8+ cells in the parenchyma after PLX5622 treatment, we analyzed scRNA-seq data from brain-derived myeloid cells of wild type animals treated with PLX5622. Surprisingly, we observed a striking overexpression of Cxcl16 in various brain-infiltrating myeloid cell population following PLX5622 treatment (in particular, ISG expressing immune cells and neutrophils). These results, which partially counteract the mild reduction of Cxcl16 observed in qPCR analysis of the hippocampi from PLX5622-treated animals, demonstrate the higher accuracy of single-cell sequencing analysis compared to bulk gene expression studies.
However, since Zhan et al., 2020 dataset was restricted to brain myeloid cells (CD11b+), we cannot exclude that Cxcl16 can be expressed also by other cells of the CNS. Indeed, CXCL16 was observed also in human astrocytes, vascular cells, and endothelial cells46. Furthermore, in primary murine hippocampal culture, Cxcl16 is also expressed by neuronal cells42. We therefore analyzed a scRNA-seq dataset comprising brain cells from mice with a genetically ablated promoter involved in Csf1r transcription. Genotype-specific differences in Cxcl16 expression were only observed in myeloid and mural cells of older animals. This suggests that the effect of compensatory Cxcl16 expression following the genetic depletion of Csfr1+ cells depends on the age of the animals.
To strengthen the possibility of a compensatory mechanism involving the CXCL16-CXCR6 axis occurring upon PLX5622-mediated microglia depletion also in amyloidogenic brain, we conducted IHC analysis of Cxcl16 in the hippocampal area of control and PLX5622-treated APP/PS1 mice. Although no difference in the percentage of microglial (Iba1+) Cxcl16+ cells in the proximity of CD8+ cells between control and PLX5622 treated mice was observed, a significant increase in the relative amount of Iba1+Cxcl16+ cells was evident upon PLX5622 treatment. This indicates that, in addition to the putative activity of other myeloid cells, PLX5622 treatment increases Cxcl16 expression in the remaining microglia cells of APP/PS1 mouse brain.
Our current hypothesis is that within the context of AD, PLX5622 treatment not only eliminates Csf1r-expressing microglial cells responsible for most of Cxcl16 production, but at the same time induces expression of Cxcl16 in other brain-resident (e.g. PLX5622-resistant microglia) or newly recruited myeloid cells, or even non-myeloid cells of the CNS. This is supported by the statistically significant increase of Cxcl16 transcript and secreted protein levels observed in both monomeric and aggregated Aβ42-treated BMDMs. Therefore, even upon a PLX5622-mediated depletion of microglia, other cellular sources of Cxcl16 help to maintain the population of Cxcr6+CD8+ T cells within the brain of APP/PS1 mice (Fig. 4).
A working hypothesis for the effect of PLX5622 in the recruitment of Cxcr6+CD8+ T cells to the brain of APP/PS1 mice.
Limitations of the study
Our findings suggest a correlation between the CXCL16-CXCR6 axis and AD pathology, but it remains unknown whether modulating this axis would affect the microglia-mediated CD8+ T cell trafficking in the brain or influence AD progression, as recently questioned54. Cxcl16 has been already proposed to exert neuroprotective and anti-inflammatory activities in the brain55,56, pointing towards a Cxcl16-stimulation strategy for AD. Moreover, although CD8+ T cell ablation led to increased expression of genes related to neuronal and synapse formation6, recent findings suggested a protective role for brain CD8+ T cells and Cxcr6 in mouse AD pathogenesis57.
Beside our current hypothesis, the absence of any changes in Cxcr6+CD8+ cells in the brain of microglia-depleted animals could be explained by our in vivo experimental settings. Four weeks of PLX5622 treatment might not be sufficient to trigger a reduction of microglial-derived Cxcl16 levels and, therefore, to decrease the presence of Cxcr6+CD8+ cells in the hippocampus. Alternatively, the Cxcr6+CD8+ cells observed in the microglia-depleted brains were residing in the brain already before microglia depletion. Furthermore, sex-dependent differences in the efficiency of PLX5622 have been reported in literature58,59. Consequently, the potential impact of gender bias in the microglia Cxcl16-mediated recruitment of Cxcr6+CD8+ cells in the brain of treated animals cannot be excluded. The limited number of animals in each experimental group from our PLX5622 in vivo treatment does not allow us to consider this factor in the present study.
While the potential for phenotypic changes in brain-resident T cells following PLX5622 treatment cannot be discounted, the present study was not designed to investigate the immunological signature of these lymphocytes following microglia depletion in AD. A substantial body of research suggests that PLX5622-dependent depletion of microglia exerts an influence in the quantity and phenotype of T cells in various disease contexts, including viral infections, MS and the aging process52,53,60,61,62,63,64. Consequently, further single-cell experiments on an AD transgenic mouse model are required to ascertain whether PLX5622-mediated depletion of microglia affects T cell activation status in AD.
Furthermore, we do not dismiss the potential role of other chemokine-chemokine receptor axis and different cell types, beyond microglia, in the recruitment of CD8+ T lymphocytes to the brain. Recently, the Cxcl13-Cxcr5 axis has been demonstrated to be essential in the recruitment of CD8+ T cells from injured neurons65. However, this axis is also involved in recruiting other lymphocytic cells to the CSF66,67, suggesting a pleiotropic function of Cxcl13-Cxcr5 communication in disease and pathology. Interestingly, a recent neuronal-microglia-CD8+ T cells co-culture system revealed key roles for the C-X-C motif chemokine ligand 10 (CXCL10) and its receptor, CXCR3, in regulating T cell infiltration in cell cultures derived from AD patients68. On the other hand, the CXCL16-CXCR6 axis could play multiple roles in the calls-to-arms of different peripheral cells, other than CD8+ T lymphocytes, towards AD-affected brain tissue.
Conclusions
Understanding the molecular mechanism involved in the accumulation of adaptive immune cells in neurodegenerative-affected brains is essential to further explore their functional role in disease progression. Here, we hypothesize a putative involvement of the CXCL16-CXCR6 axis in AD pathology. Although our data suggest an active contribution of microglial cells in Cxcl16 production in AD, IHC analysis of the hippocampi from PLX5622-treated APP/PS1 animals revealed no changes in the recruitment of CD8+Cxcr6+ cells to the brain parenchyma.
Alternative models of microglia depletion or modulation of the CXCL16-CXCR6 axis in AD pathology could help to understand the contribution of CXCL16-CXCR6 immune axis in the recruitment of CD8+ T cells to the brain.
Methods
ScRNA-seq datasets analysis
ScRNA-seq datasets were analyzed with Seurat package (version 5.0.1), following the guided clustering tutorial . Data from Van Hove et al., 201921 and Zhan et al., 202025 were retrieved using the Read10X function, while the read_tsv function was used to access count data and metadata from Sierksma et al., 202023. With CreateSeuratObject, we generated the related Seurat objects. In Van Hove et al., 2019, cells from 16-month-old wild type (WT) and APP/PS1 mice (6945 total cells) were split, according to genotype, and then cells from 16-month-old APP/PS1 animals were merged together with 9-month-old animals (7831 total cells). Sierksma et al., 2020 dataset was subset only for 11-month-old mice expressing APP Swedish and PSEN1-L166P transgenes (herein abbreviated as APP/PS1-L166P) and respective WT animals (3821 total cells). Zhan et al., 2020 dataset included three biological replicates which were merged together in two groups (namely PLX5622 and CTRL, 26054 total cells). The quality control metrics of each dataset were visualized with violin plot. To exclude empty droplets with few detected genes and multiplets with aberrantly high gene count, datasets were subdivided to contain more than 200 and less than 5,000 unique genes (nFeature_RNA) and less than 30,000 UMI counts (nCount_RNA). A threshold of 5% was set for the amount of mitochondrial genes. We normalized the feature expression levels by the total expression for each cell with the LogNormalize function. To calculate a subset of features exhibiting high cell-to-cell variation in the datasets, the FindVariableFeatures function was used. As a standard pre-processing step prior to dimensional reduction, the ScaleData function linearly transformed the expression of each gene. Principal Component Analysis (PCA) was run on the scaled highly variable features. After ranking principal components using the percentage of variance revealed by the ElbowPlot function, cell clustering was performed with the FindNeighbors function, based on the first 30 principal components (PCs), and the FindClusters function with resolution set as 1. Dimensional reduction was performed using the RunUMAP function accounting 30 PCs. Uniform Manifold Approximation and Projection (UMAP) visualization was plotted using DimPlot, with “umap” as reduction option. FeaturePlot was used to present features expression level at single cell level on UMAP, displaying cluster-specific expression genes. Automated cell type annotation22 was applied to Van Hove et al., 2019 and Zhan et al., 2020 datasets. With the auto_detect_tissue_type function, “immune system” has been selected as the tissue representing the datasets with the highest probability. The assigned clusters were visualized by grouping them according to the newly generated “clustomclassif” metadata identity. After automated cell type annotation, the “Unknown” cluster of myeloid cells from the Zhan et al., 2020 dataset was removed for an ease of representation. Total cell clusters proportion plots were generated with the RColorBrewer package. The table representing the percentage of cells expressing determined transcripts from the Van Hove et al., 2019 dataset, was generated using the output from Percent_Expressing and PrctCellExpringGene functions. To dissect the “Macrophages” cell cluster from the Van Hove et al., 2019 dataset, we subset the original clusters and we used the FindAllMarkers function to rename them according to established scRNA-seq microglia/macrophages markers69,70,71. We created the heat map of the clusters representing “Macrophages” cells, showing the main four differentially expressed genes for each cluster using the DoHeatmap function. Microglia cells from the Sierksma et al., 2020 dataset were visualized using features from the already established “cell.states” metadata identity. VlnPlot function was used to represent the expression levels of selected transcript grouping data by the cell types and splitting for either genotype (Van Hove et al., 2019 and Sierksma et al., 2020 datasets) or treatment (Zhan et al., 2020 dataset).
The scRNA-seq dataset from Munro et al., 202426 was analyzed using the SingleCellExperiment package (version 1.31.1), in accordance with the user manual. The dataset included three biological replicates of 11–12- and 16–18-month-old Csf1rWT/WT (referred to as WT) and Csf1rΔFIRE/ΔFIRE (herein abbreviated as KO) mice. Data were retrieved using readRDS function. The SingleCellExperiment object was visualized using genotype and features from the established cell types metadata (i.e. “clusters_named”) as variables via plotReducedDim function from scater package. Finally, ggcells function was used to plot the expression levels of the Cxcl16, grouping the data by the cell types and splitting by genotype.
Animals
APP Swedish and PSEN1-dE9 mice, expressing a chimeric mouse/human mutant amyloid precursor protein (Mo/HuAPP695swe) and a mutant human presenilin 1 (and PSEN1-dE9), both directed to CNS neurons under the prion protein promoter (available by Jackson Laboratory, http://www.jax.org/strain/005864, RRID: MMRRC_034832-JAX), were used. Mice were housed at the Paracelsus Medical University Salzburg under standard conditions at a temperature of 22 °C and a 12-h light/dark cycle with ad libitum access to standard food and water. Local ethical committees (BMWFW-66.019/0032-WF/V/3b/2016) approved animal care, handling, genotyping, and experiments.
For this study, 12-month-old female and male animals were used and treated for 28 days with PLX5622 chow12. Age-matched non-transgenic mice, derived from the breeding of APP Swedish PS1 dE9 (herein abbreviated as APP/PS1) were used as control animals (herein referred as WT). All animals were adapted to control chow two weeks before introducing the PLX5622 chow. Thus, after computer based randomisation, four experimental groups were generated: WT and APP/PS1 mice, which received control chow (WT CTRL and APP/PS1 CTRL, respectively), together with WT and APP/PS1 mice, which received the PLX5622 chow for a total of 28 days (WT PLX5622 and APP/PS1 PLX5622, respectively). The researchers experimenters could not be blinded because the chow were added to the cages to which the pre-defined different animal groups belonged to. No inclusion and exclusion criteria were defined a priori.
Tissue collection
After four weeks of treatment, mice were anesthetized by intraperitoneal injection of a xylazine (5.36 mg/mL final, Xylapan 20 mg/mL, Vetoquino Österreich GmbH, Z.-Nr. 8–00359), ketamine (20.5 mg/mL final, Ketamidor 100 mg/mL, VetViva Richter GmbH, Z.-Nr. 8–01141),
and acepromazine maleate (0.27 mg/mL final, Vanastress 10 mg/mL, VANA GmbH, E1123-S.1216, Z.-Nr. 8–00442) mixture. Afterwards, their thoracic cavity was opened with an incision caudal to the sternum. Animals were manually perfused through the left ventricle with 1× ice cold Hank`s Balanced Salt Solution (Thermo Fisher Scientific, #14175-053) containing 15 mM HEPES (Thermo Fisher Scientific, #15630-106) and 0.5% glucose (Sigma-Aldrich, #G7021) to rinse out the blood. Mice were decapitated and brains were extracted from the skull. Hippocampus from one brain hemisphere was immediately transferred to RNAlater® (Sigma-Aldrich, #R0901) and stored at −80 °C. The other brain hemisphere was immersed in 4% paraformaldehyde (in 0.1 M sodium phosphate solution, pH = 7.4, Sigma-Aldrich, #S9390) for fixation over night before being washed in 1× PBS and transferred into 30% sucrose (AppliChem, #A2211) for cryoprotection. When fully soaked with sucrose, brain hemispheres were cut in 40 μm slices on dry ice using a sliding microtome (Leica) dividing them in representative tenths of the brain. Sections were stored at −20 °C in cryoprotectant solution (Ethylene Glycol, Sigma-Aldrich, #100949, Glycerol, AppliChem, #131339.1211, 0.1 M phosphate buffer, Sigma-Aldrich, #D8537, in a 1:1:2 volume ratio).
BV-2 cells culture
Murine microglial BV-2 cells (originally obtained from Banca Biologica and Cell Factory, IRCCS Azienda Ospedaliera Universitaria San Martino, Genova, Italy, RRID: CVCL_0182) were maintained in Dulbecco’s Modified Eagle’s Medium (DMEM) with high glucose (Thermo Fisher Scientific, #11995065) and DMEM low glucose (Sigma-Aldrich, #FG0415) in a 1:1 ratio, supplemented with 10% fetal bovine serum (FBS, Thermo Fisher Scientific, #A5256701) and Penicillin-Streptomycin (Pan Biotech, # P06-07100), under standard culture conditions (95% relative humidity with 5% CO2, at 37 °C).
Primary microglia isolation and maintenance
The isolation of primary microglial cells was achieved using P0-P2 pups of the C57BL/6J strain, following established protocols72,73,74,75. The animal breeding, handling, genotyping, and procedures were approved by local authorities (BMBFW-66.019/0011-WF/V/3b/2016 and BMBFW: 2020 − 0.827.682). Following decapitation, the brains were extracted from the skull and placed in DMEM with the addition of 10% FBS and Penicillin-Streptomycin (henceforth referred to as ‘glia medium’). After the removal of cerebellum and olfactory bulbs, the forebrains were divided in the right and left hemispheres, and the meninges were carefully removed to prevent fibroblasts growth in the culture. The tissues were placed in 50 mL Falcon tubes containing glia medium. The cell suspension was passed through a 100 μm cell strainer (Corning, #734–2762) by means of gentle trituration with a 5 mL pipette and a serological glass pipette, in order to remove any larger debris. The cell suspension from three forebrains was seeded in T75 cell culture flasks (TPP, #90076) that had been coated with 5 µg/mL poly-D-lysine (PDL) (EMD Millipore, #A-003-E in H2O). The cells were cultivated in glia medium in accordance with standard culture conditions. The following day, and subsequently every fourth day, the medium was replaced while the remaining media was collected, and filtered sterile with 22 μm syringe filter (TPP, #99722) and used as “conditioned” media. In the context of each media change, the new media was supplemented with 50% of the fresh glia media and 50% of conditioned media. Following a 14-day culture period, a confluent layer of mixed glia cells was observed, comprising astrocytes, oligodendrocytes, ependymal cells, and microglia. In order to dislodge the loosely adherent microglia from the confluent cell layer, the culture flasks were subjected to shaking process (“shake-off”) at a frequency of 110 rpm for a duration up to 90 min at 37 °C. Afterwards, the supernatant containing microglia cells was used for further experiments, with the cells being maintained under standard culture conditions. The culture of mixed glia could be sustained for a period of up to five “shake-offs” by the incorporation of a mixture of fresh and conditioned media.
Establishment of mouse bone marrow-derived macrophages
BMDMs were prepared and cultured from C57BL/6J mice , using established procedures adapted from published protocols76,77. In brief, following cervical dislocation, any remaining muscle tissue and tendons on the tibia and femur were cleared with gauze. The bone marrow progenitor cells were then flushed with ice cold 1× PBS and passed through the 100 μm and the 70 μm cell strainers (Corning, #734–2762 and #734–2761). The cells were then centrifuged at 250×g, 4 °C for 5 min. Red blood cells were removed with Red cell lysis buffer (from Adult Brain Dissociation Kit, Miltenyi Biotec, #130-107-677). The cell suspension was then centrifuged again (250×g, 4 °C for 5 min). The cells were seeded in glia medium containing 10 ng/mL macrophage colony-stimulating factor (M-CSF, Peprotech, #315-02). The cultures were maintained for 7 days with a media change on day 4 under standard culture conditions.
Flow cytometry
To monitor the purity of the primary murine culture, microglia cells after “shake-off” and BMDMs were analyzed by flow cytometry. 1*105 cells were washed with 1 mL 1× PBS, resuspended with 1 mL of flow buffer (2% (w/v) bovine serum albumin (BSA), Sigma-Aldrich, #A9647, and 200 mM ethylenediaminetetraacetic acid, Sigma-Aldrich, #E1644, in PBS) and stained with a 1:10,000 dilution of Live/Dead 488 viability dye (Thermo Fisher Scientific, # L34969). The cells were incubated at room temperature (RT) in the dark for 30 min. After washing with flow buffer, Mouse BD Fc Block™ (1:100, BD Pharmingen™ Purified Rat Anti-Mouse CD16/CD32, BD Biosciences, #553142) was added and the cells were incubated again at RT in the dark for 5 minutes. This was followed by immunostaining with anti-rat CD11b APC (1:50, clone M1/70; BioLegend, #101211) and anti-rat F4/80 PerCP-Cyanine5.5 (1:75, clone BM8, Invitrogen™, #45-4801-80) for 30 min at 4 °C in the dark. After washing with 2 mL flow buffer, the cells were resuspended in 300 µL flow buffer and analyzed using a BD Accuri C6 flow cytometer (BD Biosciences). The data were analyzed using BD Accuri C6 software (BD Biosciences).
Cell culture experiments
BV-2 cells were seeded at a density of 2.5*104 cells per well in a 6-well tissue culture plate (TPP, #92406). Two days later, the cells were treated with either 1 µM of monomeric or aggregated (after an overnight incubation at 37 °C) human Aβ42 (rPeptide, #A-1002-1), for 24 h in FBS-depleted medium, in duplicate. The controls received either medium only (herein referred as “Medium”) or the solution used to dissolve Aβ42 peptide (1% NH4OH, herein referred as “Vehicle”, corresponding to 0.0045% NH4OH, Sigma-Aldrich, #221228). For the primary microglia experiments, 6.6 *104 cells per well were seeded in a 96-well tissue culture plate (TPP, #92696). BMDMs were seeded at a density of 1.5*104 cells per well in a 96-well tissue culture plate (Nunc UpCell Surface, ThermoFisher, #174897). One day (for primary microglia) or 7 days (for BMDMs) after seeding, cells were treated with Medium, Vehicle, 1 µM of monomeric or aggregated human Aβ42, or 1 µM of scrambled human Aβ42 sequence (rPeptide, #A-1004-1).
RNA isolation and reverse transcription
BV-2 cells were lysed in 500 µL TRI® Reagent (Sigma-Aldrich, #T9424) and subsequently pooled with the respective technical duplicate well for a total volume of 1 mL. Brain tissue was homogenized in 200 µL TRI® Reagent with pellet pestles (Bartelt, #7.620842) and 800 µL were added after the homogenization step. For phase separation, 150 µL of 1-bromo-3-chloropropane (Sigma-Aldrich, #B9673) was added, vortexed and centrifuged (12,000×g for 15 min, at 4 °C). After transferring the aqueous phase into a new tube, 1 µL GlycoBlue™ (Thermo Fisher Scientific, #AM9516) and 500 µL 2-Propanol (Sigma-Aldrich, #1096342511) were added and, after vortexing, the samples were centrifuged (12,000×g for 10 min, at 4 °C). The pellet was washed with 1 mL 75% ethanol, dried and re-suspended in 30 µL RNase-free water (pre-warmed to 55 °C, from Promega, #A6102). Total RNA concentrations were determined with a NanoVue plus (GE Healthcare). cDNA was synthesized using GoScript Reverse Transcriptase Mix including a 1:1 ratio of Oligo(dT) (Promega, #A2790) and random primers (Promega, #A2800).
Gene expression analysis of primary murine microglia and BMDMs was performed using the TaqMan™ Fast Advanced Cells-to-CT™ Kit (Thermo Fisher Scientific, #A35377), following the manufacturer’s instructions up to the stage of synthesising cDNA.
qPCR
Quantitative gene expression analysis (qPCR) was performed using TaqMan real time-PCR technology. Technical duplicates containing 10 ng (for hippocampal tissue) or 7.5 ng (for BV-2 cells) of cDNA were amplified with the GoTaq Probe qPCR Master Mix (Promega, #A6102) using a two-step cycling protocol (95 °C for 15 s, 60 °C for 60 s; 40 cycles using Bio-Rad CFX 96 Cycler) in LabQ 96-well PCR plate (LabShop Online, #PS3441-00 C), sealed with Real-Time PCR plate Sealing Film (LabShop Online, #PSPETST100). 2 µL of the Cells-to-CT reverse transcription solution from primary murine microglia and BMDMs were used. The following validated exon-spanning gene expression assays were employed: G6pdx (Integrated DNA Technologies, Mm.PT.58.13826440) and Ywhaz (Integrated DNA Technologies, Mm.PT.39a.22214831), as validated candidate housekeepers, Cxcl16 (Integrated DNA Technologies, Mm.PT.56a.42520449), Cxcr6 (Integrated DNA Technologies, Mm.PT.58.8341778), and Tnf (Thermofisher, Mm00443258_m1) for target gene analysis.
qPCR analysis was performed for cDNA from hippocampus of twenty-four animals (n = 6 WT CTRL; n = 6 WT PLX5622; n = 6 APP/PS1 CTRL; n = 6 APP/PS1 PLX5622) and from five independent biological experiments on BV-2 cells (n = 5 Medium-; n = 5 Vehicle-; n = 5 1 µM monomeric Aβ42-; n = 5 1 µM aggregated Aβ42-treated cells), on primary murine microglia and BMDMs (n = 5 Medium-; n = 5 Vehicle-; n = 5 1 µM scramble Aβ42-; n = 5 1 µM monomeric Aβ42-; n = 5 1 µM aggregated Aβ42-treated cells). The mRNA relative quantities was calculated with a modified version of 2−ΔΔCT method78 taking into account the efficiency (E) of each exon-spanning gene expression assay (target genes: Cxcr6, Cxcl16, Tnf; reference genes: G6pdx and Ywhaz), using the following Eq. (1):
Firstly, we calculated the efficiency of each gene expression assays using titration of pooled sample from cDNA of WT and APP/PS1 brains and of BV-2, murine primary microglia and BMDM cells. Then, we calculated the mean E reference gene CT (reference gene) of G6pdx and Ywhaz for each sample. The relative quantities of Cxcr6 and Cxcl16 expression levels in the brain tissue were obtained dividing the average E target geneCT (target gene, calibrator)/E reference gene CT (reference gene, calibrator) of the “calibrator” WT control samples (n = 6) by the E target geneCT (target gene, test)/E reference gene CT (reference gene, test) of each sample. For the relative quantities of Cxcl16 and Tnf in the cell cultures (i.e. BV-2 cells, murine primary microglia and BMDM), the average E target geneCT (target gene, calibrator)/E reference gene CT (reference gene, calibrator) of the “calibrator” medium-treated cells (n = 5) by the E target geneCT (target gene, test)/E reference gene CT (reference gene, test) of vehicle- and Aβ42-treated cells. To make the plots easier to consult, avoiding negative data points, values of relative quantities smaller than 1 were not converted using the equation − 1/relative quantities.
Western blot
2 mL of FBS-depleted conditioned medium from technical duplicates of BV-2 cells was centrifuged at 300×g for 5 min. The supernatants were then filtered (0.22 μm, TPP, #99722) to remove cellular debris. Proteins were precipitated with ice-cold acetone after 1 h at −20 °C and followed by centrifugation (15,000×g for 10 min, at 4 °C). The residual protein pellets were resuspended in 30 µL ddH2O, and their concentration was determined using the Bicinchoninic acid method by measuring absorbance at 562 nm on a microplate reader (Mithras LB 940, Berthold Technologies GmbH & Co.KG). 4× loading buffer (125 mM Tris pH 6.8, Fisher Scientific, #BP1521, 6% SDS, AppliChem, #A2572, 20% Glycerol, 10% β-Mercaptoethanol, Sigma-Aldrich, #444203 and 0.002% Bromophenol blue, Supelco, #1.08122) was added to 60 µg of total proteins from cells supernatants in 1:4 ratio and subsequently boiled 7 min at 95 °C. 2 µg of monomeric or aggregated Aβ42 were used to check their aggregation status in WB. 20 µg of total protein from BV-2 cell lysate were prepared after lysis with ice-cold RIPA buffer (20 mM Tris, pH 7.5, Fisher Scientific, #BP1521, 100 mM NaCl, VWR, #27810.364, 1% Triton ™ X-100, Sigma-Aldrich, #T9284 0.5% sodium deoxycholate, Sigma-Aldrich, #D6750, 0.1% SDS).
Samples were loaded onto 4–20% Mini-PROTEAN TGX Stain-free Gel (Bio-Rad, #4568096) and separated by electrophoresis with PowerPC® Basic (Bio-Rad). The gels were transferred to PVDF membrane using the Trans-Blot® Turbo Transfer pack (Bio-Rad). After stain-free signals acquisition, membranes were incubated for 1 h in 5% milk in 1× TSB-T (0.05% Tween® 20, Sigma-Aldrich, #P1379, in 1× TBS) to block unspecific binding sites, and subsequently incubated overnight at 4 °C with anti-Cxcl16 antibody (1:350, Antibodies–online.com, #AA 85–200) or anti-Aβ (1:1,000, clone 6E10, Covance, #SIG-39300). After three washings in 1× TBS-T, membranes were then incubated for 1 h with goat anti-rabbit HRP secondary antibody (1:10,000, Cell Signalling Technology, #7074) or goat anti-mouse HRP secondary antibody (1:10,000, Life Tech, #A10677). Reactions were visualized by chemiluminescence on ChemiDoc™ MP Imaging System (Bio-Rad) using Clarity™ Western ECL substrate (Bio-Rad, #1705061). Densitometric analysis of sCxcl16 bands was carried out using ImageLab software (Bio-Rad) and normalized with the signals from the corresponding bands of the stain-free blot.
ELISA
Analysis of sCxcl16 in the supernatant from primary microglia and BMDMs was performed using a commercial mouse Cxcl16 ELISA Kit (Thermo Fisher Scientific, #EMCXCL16), according to the manufacturer’s instructions. The cell supernatant samples (stored at −80 °C until following collection) were diluted 1:10 for the analysis. Absorbance at 450 nm was measured using a microplate reader (Mithras LB 940, Berthold Technologies GmbH & Co.KG). The resulting values were plotted on a standard curve, prepared with serial dilution of recombinant mouse Cxcl16, to calculate the sCxcl16 concentration (pg/mL) in the cell supernatants.
Fluorescence IHC
IHC of mouse tissue was performed on free-floating sections as previously described6,20. Briefly, antigen retrieval was performed depending on the used primary antibody by steaming the sections for 15 min in 1× citrate buffer (pH 6.0, Sigma-Aldrich., #C9999). To block unspecific binding sites, slices were incubated for 1 h in blocking solution (1% BSA, Sigma-Aldrich, # A9647, 0.2% fish skin gelatine, Sigma-Aldrich, #G7765, 0.1% Tween® 20 in 1× PBS). Brain slices were incubated on a shaker overnight at RT with the following primary antibodies: rat anti-CD8a (1:100, Invitrogen, #14–0195-82), rabbit anti-Cxcr6 (1:500, Thermofisher, #PA5-79117), goat anti-Iba1 (1:500, Abcam, #ab5076), rabbit anti-Cxcl16 (1:100, antibodies.online, #ABIN686572). Sections were extensively washed in 1× PBS and incubated for 4 h at RT in appropriate fluorescent-labelled secondary antibodies: donkey anti-goat Alexa 488 (Invitrogen, #A11055), donkey anti-rat Rhodamine Red (Szabo-Scandic, #JAC712295150), donkey anti-goat Rhodamine Red (Jackson, 108758), donkey anti-rabbit Alexa 647 (Invitrogen, #A31573) (all diluted 1:500), DAPI (1 mg/ml, 1:2000, Sigma-Aldrich, #D9542), for nucleus counterstaining and Amytracker 520 (1:1,200, Biozol, #EBB-A520) for Aβ plaques staining. The sections were extensively washed in 1× PBS and mounted onto microscope glass slides (Superfrost™ Plus, Thermo Fischer Scientific, #22037246). Brain sections were cover slipped semidry in ProLong Gold antifade mountant (Thermo Fischer Scientific, # P36930).
Confocal microscopy and image processing
For fluorescence imaging, the confocal laser scanning microscope LSM 710 from Zeiss was used. Images were taken with the ZEN 2011 SP3 software (black edition, Zeiss). For representative pictures of CD8+Cxcr6+, Iba1+ cells, or Cxcl16+ signals, were obtained as confocal z stacks in ×40 or ×63 oil magnification of granule cell and polymorph layers of the dentate gyrus in hippocampal region. Images were combined to merged maximum intensity projections and edited as well as processed with the ImageJ/Fiji software (version 1.53q) and Microsoft PowerPoint.
For quantification of Cxcr6+CD8+ and Cxcl16+Iba1+ cells were counted in at least nine images (320 × 320 μm) from three to four different hippocampal sections (including the dentate gyrus and Cornu Ammonis 1 (CA1) to CA3 regions), with three to five images per section, using ImageJ’s cell counter plugin. This resulted in an average of ∼50 analyzed cells per animal.
An ImageJ macro was used to quantify the percentage of Cxcl16+Iba1+ cells and analyse their distance (in 3D) to CD8+ cells. Firstly, scale bar was converted to pixels per µm (using a conversion factor of 2.8927 px/µm). Brightness and contrast adjustments were performed differently for each channel. After filtering the images with the Gaussian Blur 3D plugin, the 3D object counter plugin provided a list of voxels for each channel. The 3D Manager plugin then automatically selected segments for each channel (a minimum of 20 for Iba1 channel, a minimum of 10 for CD8 channel, a minimum of 10 for Cxcl16 channel and a maximum of 255 for every channel). 3D ROIs corresponding to CD8+ cells were selected manually. The 3D ROIs for Iba1 and Cxcl16 were automatically identified by setting a voxel volume greater than 500 for the former and a voxel volume ranging from 4 to 300 for the latter. Finally, the “Measure” extension of the 3D Manager plugin was used to automatically compute the distances between all objects.
Subsequent analysis was conducted in R (version 4.4.2) using the dplyr package. ImageJ output files were retrieved using the read.csv function. Firstly, Iba1 cells located within a 15 μm border-to-border distance of Cxcl16 ROIs (considered to be double-positive Iba1+Cxcl16+ cells) was quantified. Then the percentages over the number of total Iba1+ ROIs was calculated. Their distance from CD8+ ROIs was then measured. The analysis included the cells with a border-to-border distance lower than 100 μm, as recently defined for measuring the proximity of lymphocytes to myeloid cells expressing Cxcl16 in the brain tissue29.
Statistics
All graphing and statistical analysis were performed using GraphPad Prism 9 or R. scRNA-seq data represented in violin plots were analyzed using stat_compare_means function with Wilcoxon test. Data from qPCR and IHC analysis were tested for normal distribution with D’Agostino & Pearson test. To identify outliers, the ROUT method was applied. For hippocampal gene expression of Cxcr6 and Cxcl16 transcripts and for IHC Cxcr6+CD8+ cells quantification, data were analyzed with Two-way ANOVA and Šídák’s multiple comparisons test. Linear regression was performed for Cxcr6 and Cxcl16 transcript levels and tested with Spearman R correlation analysis. For Cxcl16 gene expression, WB and ELISA analysis in in vitro cell culture experiments, Friedman test with Dunn`s multiple comparison (between vehicle and monomeric or aggregated Aβ42-treated cells in BV-2, and between scrambled Aβ42- and monomeric or aggregated Aβ42-treated cells in primary microglia and BMDMs) was used. The IHC quantification of Cxcl16+Iba1+ cells was assessed with Mann-Whitney test. The level of statistical significance was set at 0.05 for all analyses. For scRNA-seq data analysis, p values were reported as < 0.05, < 0.01, < 0.001 or < 0.0001. p values were not indicated for non-statistically significant differences. For all the other analyses, p values representing statistical significant differences in Two-way ANOVA, in Friedman, in Mann-Whitney and in multiple comparison tests are shown in the plot. The data were represented as individual values alone or together with the mean and standard error of the mean.
Data availability
The mouse scRNA-seq datasets used in this study derived from Brain Immune Atlas website (https://www.brainimmuneatlas.org/index.php) and the Gene Expression Omnibus repository, under the accession numbers GSE128855, GSE142267, GSE150169, and GSE267545. Any additional data supporting the findings of this study are included in the article and related supplementary information files.
References
Jucker, M. & Walker, L. C. Self-propagation of pathogenic protein aggregates in neurodegenerative diseases. Nature 501 (7465), 45–51 (2013).
DeTure, M. A. & Dickson, D. W. The neuropathological diagnosis of alzheimer’s disease. Mol. Neurodegener. 14 (1), 32 (2019).
Heneka, M. T. et al. Neuroinflammation in alzheimer’s disease. Lancet Neurol. 14 (4), 388–405 (2015).
McManus, R. M. & Heneka, M. T. Role of neuroinflammation in neurodegeneration: new insights. Alzheimers Res. Ther. 9 (1), 14 (2017).
Leng, F. & Edison, P. Neuroinflammation and microglial activation in alzheimer disease: where do we go from here? Nat. Rev. Neurol. 17 (3), 157–172 (2021).
Unger, M. S. et al. CD8 + T-cells infiltrate alzheimer’s disease brains and regulate neuronal- and synapse-related gene expression in APP-PS1 Transgenic CD8 + T-cells infiltrate alzheimer’s disease brains and regulate neuronal- and synapse-related gene expression in APP-PS1 Transgenic mice. Brain Behav. Immun. 89, 67–86 (2020).
Gate, D. et al. Clonally expanded CD8 T cells patrol the cerebrospinal fluid in alzheimer’s disease. Nature 577 (7790), 399–404 (2020).
Togo, T. et al. Occurrence of T cells in the brain of alzheimer’s disease and other neurological diseases. J. Neuroimmunol. 124 (1–2), 83–92 (2002).
Rogers, J. et al. Expression of immune system-associated antigens by cells of the human central nervous system: relationship to the pathology of alzheimer’s disease. Neurobiol. Aging. 9 (4), 339–349 (1988).
Merlini, M. et al. Extravascular CD3 + T cells in brains of alzheimer disease patients correlate with Tau but not with amyloid pathology: an immunohistochemical study. Neurodegener Dis. 18 (1), 49–56 (2018).
Itagaki, S., McGeer, P. L. & Akiyama, H. Presence of T-cytotoxic suppressor and leucocyte common antigen positive cells in alzheimer’s disease brain tissue. Neurosci. Lett. 91 (3), 259–264 (1988).
Unger, M. S. et al. Microglia prevent peripheral immune cell invasion and promote an anti-inflammatory environment in the brain of APP-PS1 Transgenic mice. J. Neuroinflammation. 15 (1), 274 (2018).
Matloubian, M. et al. A transmembrane CXC chemokine is a ligand for HIV-coreceptor Bonzo. Nat. Immunol. 1 (4), 298–304 (2000).
Wilbanks, A. et al. Expression cloning of the STRL33/BONZO/TYMSTRligand reveals elements of CC, CXC, and CX3C chemokines. J. Immunol. 166 (8), 5145–5154 (2001).
Nakayama, T. et al. Cutting edge: profile of chemokine receptor expression on human plasma cells accounts for their efficient recruitment to target tissues. J. Immunol. 170 (3), 1136–1140 (2003).
Shimaoka, T. et al. Cell surface-anchored SR-PSOX/CXC chemokine ligand 16 mediates firm adhesion of CXC chemokine receptor 6-expressing cells. J. Leukoc. Biol. 75 (2), 267–274 (2004).
Mabrouk, N. et al. CXCR6 expressing T cells: functions and role in the control of tumors. Front. Immunol. 13, 1022136 (2022).
Piehl, N. et al. Cerebrospinal fluid immune dysregulation during healthy brain aging and cognitive impairment. Cell 185 (26), 5028–5039e13 (2022).
Rosen, S. F. et al. Single-cell RNA transcriptome analysis of CNS immune cells reveals CXCL16/CXCR6 as maintenance factors for tissue-resident T cells that drive synapse elimination. Genome Med. 14 (1), 108 (2022).
Altendorfer, B. et al. Transcriptomic profiling identifies CD8 + T cells in the brain of aged and alzheimer’s disease Transgenic mice as Tissue-Resident memory T cells. J. Immunol. 209 (7), 1272–1285 (2022).
Van Hove, H. et al. A single-cell atlas of mouse brain macrophages reveals unique transcriptional identities shaped by ontogeny and tissue environment. Nat. Neurosci. 22 (6), 1021–1035 (2019).
Ianevski, A., Giri, A. K. & Aittokallio, T. Fully-automated and ultra-fast cell-type identification using specific marker combinations from single-cell transcriptomic data. Nat. Commun. 13 (1), 1246 (2022).
Sierksma, A. et al. Novel alzheimer risk genes determine the microglia response to amyloid-β but not to TAU pathology. EMBO Mol. Med. 12 (3), e10606 (2020).
Spangenberg, E. et al. Sustained microglial depletion with CSF1R inhibitor impairs parenchymal plaque development in an alzheimer’s disease model. Nat. Commun. 10 (1), 3758 (2019).
Zhan, L. et al. A MAC2-positive progenitor-like microglial population is resistant to CSF1R inhibition in adult mouse brain. Elife 9, e51796 (2020).
Munro, D. A. D. et al. Microglia protect against age-associated brain pathologies. Neuron 112 (16), 2732–2748e8 (2024).
Rojo, R. et al. Deletion of a Csf1r enhancer selectively impacts CSF1R expression and development of tissue macrophage populations. Nat. Commun. 10 (1), 3215 (2019).
Munro, D. A. D. et al. CNS macrophages differentially rely on an intronic. Development 147(23), dev194449 (2020).
Lu, I. N. et al. The CXCL16/CXCR6 axis is linked to immune effector cell-associated neurotoxicity in chimeric antigen receptor (CAR) T cell therapy. Genome Med. 17 (1), 71 (2025).
Narasimhan, P. B. et al. Nonclassical monocytes in health and disease. Annu. Rev. Immunol. 37, 439–456 (2019).
Wein, A. N. et al. CXCR6 regulates localization of tissue-resident memory CD8 T cells to the airways. J. Exp. Med. 216 (12), 2748–2762 (2019).
Woehrl, B. et al. CXCL16 contributes to neutrophil recruitment to cerebrospinal fluid in Pneumococcal meningitis. J. Infect. Dis. 202 (9), 1389–1396 (2010).
Adamski, V. et al. The chemokine receptor CXCR6 evokes reverse signaling via the transmembrane chemokine CXCL16. Int. J. Mol. Sci. 18(7), 1468 (2017).
Hattermann, K. et al. The CXCL16-CXCR6 chemokine axis in glial tumors. J. Neuroimmunol. 260 (1–2), 47–54 (2013).
Chung, B. et al. Human brain metastatic stroma attracts breast cancer cells via chemokines CXCL16 and CXCL12. NPJ Breast Cancer. 3, 6 (2017).
Li, X. et al. Convergent transcriptomic and genomic evidence supporting a dysregulation of CXCL16 and CCL5 in alzheimer’s disease. Alzheimers Res. Ther. 15 (1), 17 (2023).
Fang, Y. et al. Association between inflammatory biomarkers and cognitive aging. PLoS One. 17 (9), e0274350 (2022).
Jin, G. The relationship between serum CXCL16 level and carotid vulnerable plaque in patients with ischemic stroke. Eur. Rev. Med. Pharmacol. Sci. 21 (17), 3911–3915 (2017).
Romero-Suárez, S. et al. The central nervous system contains ILC1s that differ from NK cells in the response to inflammation. Front. Immunol. 10, 2337 (2019).
Wojkowska, D. W. et al. Interactions between neutrophils, Th17 cells, and chemokines during the initiation of experimental model of multiple sclerosis. Mediators Inflamm. 2014, p590409 (2014).
le Blanc, L. M. et al. CXCL16 is elevated in the cerebrospinal fluid versus serum and in inflammatory conditions with suspected and proved central nervous system involvement. Neurosci. Lett. 397 (1–2), 145–148 (2006).
Rosito, M. et al. Trasmembrane chemokines CX3CL1 and CXCL16 drive interplay between neurons, microglia and astrocytes to counteract pMCAO and excitotoxic neuronal death. Front. Cell. Neurosci. 8, 193 (2014).
Sondag, C. M., Dhawan, G. & Combs, C. K. Beta amyloid oligomers and fibrils stimulate differential activation of primary microglia. J. Neuroinflammation. 6, 1 (2009).
Gouwens, L. K. et al. Amyloid-β42 protofibrils are internalized by microglia more extensively than monomers. Brain Res. 1648 (Pt A), 485–495 (2016).
Gough, P. J. et al. A disintegrin and metalloproteinase 10-mediated cleavage and shedding regulates the cell surface expression of CXC chemokine ligand 16. J. Immunol. 172 (6), 3678–3685 (2004).
Ludwig, A. et al. Enhanced expression and shedding of the transmembrane chemokine CXCL16 by reactive astrocytes and glioma cells. J. Neurochem. 93 (5), 1293–1303 (2005).
Fransen, N. L. et al. Tissue-resident memory T cells invade the brain parenchyma in multiple sclerosis white matter lesions. Brain 143 (6), 1714–1730 (2020).
Hsiao, C. C. et al. White matter lesions in multiple sclerosis are enriched for CD20. Eur. J. Immunol. 51 (2), 483–486 (2021).
Kollis, P. M. et al. Characterising distinct migratory profiles of infiltrating T-Cell subsets in human glioblastoma. Front. Immunol. 13, 850226 (2022).
Unger, P. A. et al. T-cells in human trigeminal ganglia express canonical tissue-resident memory T-cell markers. J. Neuroinflammation. 19 (1), 249 (2022).
Chen, X. et al. Microglia-mediated T cell infiltration drives neurodegeneration in tauopathy. Nature 615 (7953), 668–677 (2023).
Mangale, V. et al. Microglia influence host defense, disease, and repair following murine coronavirus infection of the central nervous system. Glia 68 (11), 2345–2360 (2020).
Spiteri, A. G. et al. PLX5622 reduces disease severity in lethal CNS infection by Off-Target Inhibition of peripheral inflammatory monocyte production. Front. Immunol. 13, 851556 (2022).
Costa, M. R. Switch of innate to adaptative immune responses in the brain of patients with alzheimer’s disease correlates with tauopathy progression. NPJ Aging. 10 (1), 19 (2024).
Di Castro, M. A. et al. The chemokine CXCL16 modulates neurotransmitter release in hippocampal CA1 area. Sci. Rep. 6, 34633 (2016).
Lepore, F. et al. CXCL16/CXCR6 axis drives Microglia/Macrophages phenotype in physiological conditions and plays a crucial role in glioma. Front. Immunol. 9, 2750 (2018).
Su, W. et al. CXCR6 orchestrates brain CD8+ T cell residency and limits mouse alzheimer’s disease pathology. Nat. Immunol. 24 (10), 1735–1747 (2023).
Sharon, A., Erez, H. & Spira, M. E. Significant Sex Differences in the Efficacy of the CSF1R Inhibitor-PLX5622 on Rat Brain Microglia Elimination 15, Pharmaceuticals, Basel, (2022).
Johnson, N. R. et al. CSF1R inhibitors induce a sex-specific resilient microglial phenotype and functional rescue in a tauopathy mouse model. Nat. Commun. 14 (1), 118 (2023).
Lei, F. et al. CSF1R Inhibition by a small-molecule inhibitor is not microglia specific; affecting hematopoiesis and the function of macrophages. Proc. Natl. Acad. Sci. U S A. 117 (38), 23336–23338 (2020).
Sanchez, J. M. S. et al. The CSF1R-Microglia axis has protective Host-Specific roles during neurotropic picornavirus infection. Front. Immunol. 12, 621090 (2021).
Wheeler, D. L. et al. Microglia are required for protection against lethal coronavirus encephalitis in mice. J. Clin. Invest. 128 (3), 931–943 (2018).
Montilla, A. et al. Microglia and meningeal macrophages depletion delays the onset of experimental autoimmune encephalomyelitis. Cell. Death Dis. 14 (1), 16 (2023).
Groh, J. et al. Microglia activation orchestrates CXCL10-mediated CD8. Nat. Neurosci. 28 (6), 1160–1173 (2025).
Zhou, L. et al. Reversible CD8 T cell-neuron cross-talk causes aging-dependent neuronal regenerative decline. Science 376 (6594), eabd5926 (2022).
Harrer, C. et al. The CXCL13/CXCR5 immune axis in health and disease-implications for intrathecal B cell activities in neuroinflammation. Cells 11(17), 2649 (2022).
Harrer, C. et al. The CXCL13/CXCR5-chemokine axis in neuroinflammation: evidence of CXCR5 + CD4 T cell recruitment to CSF. Fluids Barriers CNS. 18 (1), 40 (2021).
Jorfi, M. et al. Infiltrating CD8+ T cells exacerbate alzheimer’s disease pathology in a 3D human neuroimmune axis model. Nat. Neurosci. 26 (9), 1489–1504 (2023).
Mancuso, R. et al. Xenografted human microglia display diverse transcriptomic states in response to Alzheimer’s disease-related amyloid-β pathology. Nat. Neurosci. 20, e090394 (2024).
Sala Frigerio, C. et al. The major risk factors for alzheimer’s disease: Age, Sex, and genes modulate the microglia response to Aβ plaques. Cell. Rep. 27 (4), 1293–1306e6 (2019).
Chen, Y. & Colonna, M. Microglia in Alzheimer’s disease at single-cell level. are there common patterns in humans and mice?. J. Exp. Med. 218(9), e20202717 (2021).
Färber, K. & Kettenmann, H. Physiology of microglial cells. Brain Res. Rev. 48, 133–143 (2005).
Frei, K. et al. Astrocyte-derived Interleukin 3 as a growth factor for microglia cells and peritoneal macrophages. J. Immunol. 137 (11), 3521–3527 (1986).
Giulian, D. & Baker, T. J. Characterization of ameboid microglia isolated from developing mammalian brain. J. Neurosci. 6 (8), 2163–2178 (1986).
Prinz, M. et al. Microglial activation by components of gram-positive and -negative bacteria: distinct and common routes to the induction of ion channels and cytokines. J. Neuropathol. Exp. Neurol. 58 (10), 1078–1089 (1999).
Lacey, D. C. et al. Defining GM-CSF- and macrophage-CSF-dependent macrophage responses by in vitro models. J. Immunol. 188 (11), 5752–5765 (2012).
Toda, G. et al. Preparation and culture of bone marrow-derived macrophages from mice for functional analysis. STAR. Protoc. 2 (1), 100246 (2021).
Livak, K. J. & Schmittgen, T. D. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta delta C(T)) method. Methods 25 (4), 402–408 (2001).
Acknowledgements
The authors want to thank the Microscope Core facility at PMU for their support of this work. MZ sincerely thanks Dr. Massimiliano Cocca for his technical suggestions on the analysis of scRNA-seq datasets, Giulia Pruonto, MSc, for her suggestions in the preparation of murine BMDMs, Dr. Andrea Zurl to provide critical comments to the experimental plan, Dr. Andrea Harrer and Dr. Bruno Benedetti for the fruitful discussions. This work is supported by the Austrian Science Fund (FWF) grant (P 35417-B) “Defining the role of CD8+ T-cells in Alzheimer’s disease”, and through funding of the PMU Research support Fund (PMU-FFF) Stand-Alone-Project “Molecular profiling of CD8+ T-cells in Alzheimer’s disease: new immune mediators in the brain?” (E-20/32/169-UNG).
Author information
Authors and Affiliations
Contributions
Conceptualization: Marco Zattoni, Sabine Bernegger, Barbara Altendorfer and Ludwig Aigner (MZ, SB, BA and LA conceived the study); Methodology: Marco Zattoni, Sabine Bernegger, Sofia Weinbender, Heike Mrowetz, Ariane Benedetti and Michael Stefan Unger (MZ and AB established murine primary cultures. MZ and SW performed qPCR, WB and ELISA analysis. SB performed the histological examination of the brain. MSU performed animal experiments and, together with HM, sacrificed and performed tissue collection); Formal analysis and investigation: Marco Zattoni, Sabine Bernegger and Rodolphe Poupardin (MZ and RP analyzed and interpreted the scRNA-seq datasets. SB and MZ performed the histological analysis. MZ analyzed qPCR, WB and ELISA data); Writing - original draft preparation: Marco Zattoni (MZ wrote the manuscript); Writing - review and editing: Marco Zattoni, Sabine Bernegger, Sofia Weinbender, Barbara Altendorfer, Heike Mrowetz, Ariane Benedetti, Rodolphe Poupardin, Michael Stefan Unger and Ludwig Aigner (all authors read and approved the final manuscript); Funding acquisition: Ludwig Aigner; Resources: Barbara Altendorfer (BA provided Aβ42 peptides) and Ludwig Aigner; Supervision: Michael Stefan Unger and Ludwig Aigner.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Ethics
All animal experiments were conducted in compliance with national ethical guidelines.
Approval for animal experiments
All animal experiments were approved by the Austrian Ministry of Science and Research (BMBFW-66.019/0011-WF/V/3b/2016 and BMBFW: 2020 − 0.827.682) and were performed in accordance with relevant named guidelines and regulations. Furthermore, the authors complied with the ARRIVE guidelines, as the “ARRIVE 10 essentials” are described in details in the “Methods” section.
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Below is the link to the electronic supplementary material.
Rights and permissions
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://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Zattoni, M., Bernegger, S., Weinbender, S. et al. The involvement of microglia and the CXCL16-CXCR6 axis in the recruitment of CD8+ T cells to an amyloidogenic mouse brain. Sci Rep 15, 38221 (2025). https://doi.org/10.1038/s41598-025-22137-5
Received:
Accepted:
Published:
Version of record:
DOI: https://doi.org/10.1038/s41598-025-22137-5
