Adenosine A2A receptor-mediated interactions between Th1+ T cells and the choroid plexus epithelium via IFN-γ signalling control T-Cell infiltration in experimental autoimmune encephalomyelitis.

Introduction

Multiple sclerosis (MS) is an autoimmune disorder affecting the central nervous system (CNS) that is characterized by immune cell infiltration, inflammatory demyelination, and axonal injury [ 1 , 2 ]. To date, approximately ten drug disease-modifying therapies (DMTs) have received approval for MS treatment and have demonstrated varying degrees of therapeutic efficacy [ 3 ]. Despite their relatively good efficacy, the potential for serious complications (such as infections, bradycardia, heart blocks, macular oedema, and secondary autoimmune diseases) underscores the need for a better understanding of MS pathogenesis and the development of new effective therapies with fewer side effects.

Adenosine serves as a potent immune mediator, and its extracellular level increases rapidly under adverse conditions such as hypoxia, ischemia, or the inflammatory response [ 4 ]. When its level is increased, extracellular adenosine exerts broad immunomodulatory effects by activating four known G-protein-coupled receptors (namely, the A 1 , A 2A , A 2B , and A 3 receptor subtypes). Among these, the critical role of adenosine A 2A receptor (A 2A R) in the modulation of MS pathology has been extensively explored. Both A 2A R selective antagonists (KW6002, SCH58261 ) and an A 2A R nonselective antagonist (caffeine) have consistently demonstrated a protective effect against MS pathology in experimental autoimmune encephalomyelitis (EAE) model mice [ 5 – 8 ]. Thus, A 2A R antagonists may be novel therapies for MS because they indirectly target immune function while presumably being less toxic to cells than other MS treatments. In particular, the safety of KW6002, which was approved by the Food and Drug Administration (FDA) for the adjuvant treatment of Parkinson’s disease, increases our confidence in its use in the treatment of MS [ 9 ]. Unexpectedly, A 2A R knockout (A 2A R −/− ) mice exhibit severe EAE-related pathological changes, such as elevated immune infiltration, increased demyelination, and pronounced movement disorders [ 10 ]. The opposite effects of A 2A R antagonists and global A 2A R knockout on EAE pathology suggest distinct and potentially opposite effects of A 2A R on different cell types. Indeed, a study on bone marrow chimeric mice revealed that the deletion of A 2A R in nonhaematopoietic cells protected against the development of EAE, whereas the deletion of A 2A R in haematopoietic cells accelerated disease progression [ 11 ]. This highlights the contrasting effects of A 2A R signalling in immune and A 2A R signalling in nonimmune cells on EAE pathology. To explain the different effects of A 2A R knockout and A 2A R antagonists on EAE pathology, it is critical to uncover when immune and nonimmune cells interact and identify what molecular signalling pathway is involved in these interactions. Identifying the cells through which A 2A R exerts its various effects (i.e., protective and exacerbating effects) may provide a cellular basis for the protective effects of A 2A R antagonists against MS pathology.

The choroid plexus (ChP) serves as a crucial blood-cerebrospinal fluid barrier (BCSFB) that maintains a degree of immune privilege within the CNS by preventing the infiltration of peripheral immune cells into the cerebrospinal fluid (CSF) [ 12 ]. This barrier function of the ChP is critically important for CD4 + T-cell-mediated CNS neurodegeneration in MS. Accordingly, the initial infiltration of T cells across the ChP is considered the “first wave” of T-cell trafficking into the brain, and triggers blood-brain barrier (BBB) leakage and subsequent damage from further T-cell infiltration [ 13 ]. In a recent study, we demonstrated that the knockdown of A 2A R in the ChP epithelium delays EAE onset and mitigates EAE pathology by impeding T-cell infiltration across the ChP [ 5 ]. These findings suggest that blockade of A 2A R in the ChP epithelium protects against EAE pathology, suggesting that A 2A R antagonists may exert their therapeutic effect on EAE by blocking A 2A R in the ChP. The ChP epithelium has a well-acknowledged role as a gateway for immune cell invasion, and the resident or circulating immune cells in the ChP may interact with the ChP epithelium to modify ChP gateway activity. Thus, it is crucial to elucidate the A 2A R-mediated interactions between immune cells and other cells in the ChP to provide a cellular basis for the therapeutic effects of A 2A R antagonists on EAE pathology.

IFN-gamma (IFN-γ), the hallmark cytokine of Th1 + T cells, plays an important role in CD4 + T-cell-mediated EAE pathology [ 14 ]. Recently, IFN-γ secreted from immune cells was shown to act as a gatekeeper to control the infiltration of immune cells across the ChP by binding to IFN-γ receptor 1 (IFNGR1) and inducing the expression of chemokines and adhesion molecules (CAMs) located in the ChP epithelium [ 15 ]. Furthermore, A 2A R knockout in immune cells results in a dramatic increase in IFN-γ levels in EAE [ 11 ]. IFN-γ may be an ideal mediator to intervene A 2A R-mediated interactions between immune cells and the ChP epithelium to control T-cell infiltration in the ChP. Thus, we hypothesize that A 2A R-mediated interactions between CD4 + T cells and the ChP epithelium via IFN-γ signalling control gateway activity during the development of EAE pathology. A 2A R-mediated interactions between the ChP epithelium and residential/circulating Th1 + T cells may underlie the opposite effects of A 2A R antagonists and A 2A R knockout on EAE development.

This study sought to uncover the distinct and possibly opposing effects of A 2A R in the ChP epithelium and Th1 + T cells and the impact of their interactions on ChP gateway activity in EAE model mice. We demonstrated that A 2A R knockout in the ChP epithelium reduced ChP gateway activity, whereas A 2A R knockout in Th1 + T cells increased the proliferation of Th1 + T cells and increased IFN-γ secretion by Th1 + T cells. Furthermore, we showed that disrupting the interaction between Th1 + T cells and the ChP epithelium via inactivation of IFNGR1 in the ChP ameliorated EAE pathology and improved behaviour in A 2A R −/− mice. This study provides novel insights regarding longstanding questions about the interaction between A 2A R in immune (Th1 + T cells) and A 2A R in nonimmune cells (ChP epithelium) in EAE. These insights into the multifaceted roles of A 2A R signalling in the ChP in MS pathogenesis provide a cellular basis for the therapeutic efficacy of A 2A R antagonists in MS.

Materials and methods

Animals

We obtained adult male C57/B6J mice (22 ± 2 g) aged 8 weeks from SPF Biotechnology Co., Ltd. (Beijing, China). A 2A R −/− mice and A 2A R flox/flox mice were bred in-house as previously described [ 16 , 17 ]. The mice were housed in a controlled environment at a regulated temperature under specific pathogen-free conditions on a standard 12-h light/dark cycle. All experimental procedures involving animals were reviewed and approved by the Animal Use Ethics Committee of Wenzhou Medical University.

Mouse model of EAE

The EAE model was induced following a previously established procedure [ 5 , 18 ]. The mice received subcutaneous injections of an emulsion containing MOG 35−55 peptide (1 mg/mL in PBS, Anaspec) and complete Freund’s adjuvant (CFA, Sigma). In addition, pertussis toxin (PTX, 200 ng) was administered intravenously (200 µL in PBS) on days 0 and 2 after treatment.

EAE scores, which reflected the condition of the tail and limbs, were assessed daily on the basis of clinical signs using a scale ranging from 0 to 15. The onset of EAE was defined as the first day when the EAE score was ≥ 1, and the progression stage encompassed the duration between EAE onset and the day when the EAE score reached its peak. For the tail, 0 represents no signs, 1 represents half paralysis, and 2 reflects complete paralysis. For each limb, 0 represents no signs, 1 represents weak or altered gait, 2 represents paresis, while a score of 3 reflects a fully paralyzed limb. The EAE score for each mouse is the sum of the scores for the tail and all four limbs. Thus, a mouse with complete paralysis of all four limbs can achieve a score of 14, and mortality is scored as 15.

A 2A R antagonist treatment

For in vivo treatment with the A 2A R antagonist KW6002, the mice received daily intraperitoneal injections of either vehicle or KW6002 at a dose of 5 mg/kg. KW6002 was prepared in a solution comprising 15% DMSO, 15% castor oil, and 70% sterile PBS. The vehicle control was 15% DMSO diluted in sterile PBS and castor oil.

Isolation of the ChP

The mice were deeply anaesthetized with 1% pentobarbital sodium, followed by perfusion with prechilled sterile PBS. The entire brain was then carefully dissected. With the aid of a microscope, the ChP was rapidly extracted from the lateral ventricle, third ventricle, and fourth ventricle. These isolated ChPs were then carefully placed in 1.5-mL EP tubes and stored at -80 °C for subsequent experiments.

ChP-specific knockdown of IFNGR1 with AAV2/5-shRNA

For ICV injection of AAV2/5-shRNA, anaesthetized mice were carefully secured onto a stereotaxic frame. After the cranial surface was exposed, 2 µL of AAV2/5-U6-EGFP (2.7 × 10 12 ) or AAV2/5-U6-shRNA-IFNGR1 (3.1 × 10 12 ) was precisely injected into the unilateral lateral ventricle (AP: -0.58; ML: +1.10; DL: -2.20) of male A 2A R −/− mice at a rate of 0.4 µL/min. The needle was left in place for an additional 5 min before being carefully withdrawn. The mice were then gently placed on a warming blanket for recovery.

Collection of CSF

On day 14 after EAE induction, anaesthetized mice were carefully secured to a stereotaxic apparatus. After the cerebellomedullary cistern was exposed, CSF (5 µL per mouse) was collected by puncturing the meninges with a capillary glass pipette and placed on ice for subsequent flow cytometry (FCM) analysis.

ICV injection of Th1 + T cells in vivo

The experiment was conducted following a previously reported procedure [ 19 ]. First, TAT-Cre recombinant proteins (Millipore) were injected into the lateral ventricles (AP: -0.58; ML: +1.10; DL: -2.20) of A 2A R flox/flox mice. Subsequently, Th1 + T cells were isolated from the spleens of A 2A R +/+ or A 2A R −/− mice using FCM and activated with PMA/lonomycin for 24 h. The activated Th1 + T cells (2 × 10 5 cells) were then injected into the lateral ventricles (AP: -0.58; ML: -1.10; DL: -2.20) of A 2A R flox/flox mice pretreated with TAT-Cre. To inhibit IFN-γ signalling, A 2A R flox/flox mice were pretreated (i.p.) with 500 µg (per mouse) of IFN-γ neutralizing antibody (clone XMG1.2, BIOXCELL) one day prior to ICV injection of Th1 + T cells. After 24 h, the mice were perfused with PBS, and then immunofluorescent staining with anti-intercellular cell adhesion molecule-1 (ICAM-1) and anti-CD4 antibodies was performed.

Quantitative realtime PCR (qPCR)

Following perfusion of the mice with sterile PBS, the ChP was promptly extracted from the ventricles. RNA was isolated using TRIzol (Invitrogen), and cDNA synthesis was carried out using a PrimeScript™ 1st Strand cDNA Synthesis Kit (Takara, 6110 A). Quantitative PCR was conducted using SYBR-Green Premix Ex Taq (Takara) on a real-time PCR system (CFX96; Bio-Rad). The following primers were used: ppia (forward 5′-AGCATACAGGTCCTGGCATCTTGT-3′; reverse 5′-CAAAGACCACATGCTTGCCATCCA-3′), icam-1 (forward 5′-AGATCACATTCACGGTGCTGGCTA-3′; reverse 5′-AGCTTTGGGATGGTAGCTGGAAGA-3′), ifngr1 (forward 5′-ATGGAACTGATTCATGAATGT-3′; reverse 5′-TTAGGCTTTGTTGTTGTTGTC-3′), ccl5 (forward 5′-GCTGCTTTGCCTACCTCTCC-3′; reverse 5′-TCGAGTGACAAACACGACTGC-3′), and ccl20 (forward 5′-GCCTCTCGTACATACAGACGC-3′; reverse 5′-CCAGTTCTGCTTTGGATCAGC-3′). The 2 −ΔΔCt method was employed to calculate relative gene expression.

Immunofluorescence

Brain slices (30 μm thick, including the ChP) and spinal cord Sections (30 μm thick) were thoroughly washed with PBS three times, followed by a 2-h incubation in PBS containing 0.3% Triton X-100 and 0.5% BSA. The slices were subsequently incubated with a rabbit anti-ICAM1 antibody (1:100, Abcam), rabbit anti-IBA1 antibody (1:200, Wako), rat anti-CD68 antibody (1:100, Bio-Rad), rabbit anti-MBP antibody (1:50, Cell Signalling Technology), mouse anti-NF-H antibody (1:400, Cell Signalling Technology), rabbit anti-Claudin1 antibody (1:200, Invitrogen), or rabbit anti-IFN-γR1 antibody (1:200, Abcam) for 14 h. After three rounds of rinsing for 10 min, the slices were incubated with the corresponding secondary antibodies at room temperature for 2 h. Finally, DAPI staining was performed for 10 min.

The ChP, which was isolated from the lateral ventricle, was fixed in 4% PFA for 1 h. Subsequently, the ChP tissues were stained with a rat anti-CD4 antibody (1:200, Invitrogen). After washing, the ChP tissues were incubated with the appropriate secondary antibodies at room temperature for 2 h, followed by DAPI staining. Finally, the ChP tissues were flat-mounted onto glass slides. The stained sections were thoroughly examined and scanned using a confocal inverted fluorescence microscope (LSM900, Zeiss). IMARIS 9.0 software was used for three-dimensional reconstruction of microglial morphology.

H&E staining

H&E staining was performed in accordance with previously established methods [ 20 ]. To evaluate morphological alterations and assess immune cell infiltration in the spinal cord tissues of EAE model mice in each group, paraffin-embedded sections of the spinal cord were subjected to H&E staining. The average number of infiltrating immune cells was determined by subtracting the average number of cell nuclei observed under high magnification in each experimental group from that in the control group. The tissue sections were analysed semi-quantitatively according to the following criteria: 0 = absence of inflammation; 1 = presence of cells infiltrating solely around blood vessels and the meninges; 2 = mild cell infiltration within the parenchyma (1–10 infiltrating cells/section); 3 = moderate cell infiltration within the parenchyma (11–100 infiltrating cells/section); and 4 = extensive cell infiltration within the brain parenchyma (more than 100 infiltrating cells/section) [ 6 ].

FCM analysis of immune cells

On days 14 after EAE induced, the mice were perfused with sterile PBS, after which their spleens and spinal cords were carefully isolated. Cell suspensions were prepared from spleen and spinal cord tissues, with red blood cells being lysed using ACK buffer (BL503b, Biosharp). After blocking with 10 µL of Clear Back FCR blocking reagent (MTG-001, MBL) for 5 min at room temperature, the cells were subsequently incubated at 4 °C for 30 min with the following fluorescently labelled monoclonal antibodies: FVS780 (565388, BD Horizon), CD3-FITC (11-0032-82, Invitrogen), CD4-APC (14-0042-82, Invitrogen), CD183-BV421 (126529, Biolegend), and CD196-PE (129804, Biolegend). The isotype control antibodies used in this study are IgG Isotype Control-BV421 (400935, Biolegend) and IgG Isotype Control-PE (400907, Biolegend). After centrifugation, 200 µL of PBS containing 1% BSA was added to facilitate cell resuspension. Finally, the stained cells were individually assessed using a flow cytometer (Attune NXT, Invitrogen), and the data were processed with FlowJo version 10 software.

Analysis of IFN-γ levels by enzyme-linked immunosorbent assay (ELISA)

On the 14th day following EAE induction, the mice were anaesthetized with 1% pentobarbital sodium for the purpose of isolating peripheral blood. Blood serum was then extracted to assess the levels of IFN-γ via ELISA (abs520007, Absin).

Measurement of IFN-γ levels by the ELISPOT assay

Mouse spleens were isolated to prepare splenocyte suspensions. After the samples were passed through a 40-µm cell strainer, the cells were incubated with fluorescent antibodies (CD3-FITC, CD4-APC, and CD183-BV421) at 4 °C for 30 min. Th1 + T cells (identified as CD3 + CD4 + CD183 + cells) were sorted using a high-speed flow cytometer (Moflo Astrios EQ). Following the sorting process, the suspended cells were subjected to on-machine detection to verify the Th1 + T-cell purity (> 97%).

To assess IFN-γ expression in Th1 + T cells isolated from the different groups, a mouse IFN-γ precoated ELISPOT assay kit (2210005, Dakewe) was used. Initially, Th1 + T-cell samples from each group were resuspended in serum-free cell culture medium. Subsequently, 10 µL of the positive stimulant PMA/lonomycin and 100 µL of the resuspended cell suspension (concentration of 1 × 10 5 cells/mL) were added to each well of the assay plate. Once all the cell samples and stimulants were added, the plate was incubated at 37 °C in a 5% CO 2 cell culture incubator for 24 h. Following incubation, AEC colour developing solution was added to each experimental well, and the parameters of interest were measured using a plate reader (Mabtech IRIS).

Trans-epithelial electrical resistance (TEER)

The TEER was measured as described in the previous study [ 15 ]. In brief, ChP epithelial cells obtained from A 2A R flox/flox mice (postnatal day 7) were seeded in the upper chamber of a Transwell plate (pore size of 5 μm and diameter of 6.5 mm) to establish an epithelial cell barrier. After being cultured for a period of 5 days, 2 µL of TAT-Cre (Millipore) or sterile PBS was added to the upper chamber. Following a 24-h incubation period, activated Th1 + T cells (concentration of 1 × 10 5 cells), with or without an IFN-γ neutralizing antibody (20 µg/mL, clone XMG1.2, BIOXCELL), were added to the upper chamber. After a cocultivation period of 48 h, the TEER was measured using the Millipore Millicell ERS-2 system (MERS00002).

T-cell migration assay

As previously described, ChP epithelial cells obtained from A 2A R flox/flox mice at postnatal day 7 were seeded in the upper chamber of a Transwell plate (pore size of 5 μm and diameter of 6.5 mm). After a 5-day incubation, 2 µL of TAT-Cre or sterile PBS was added to the upper chamber. Following a 24-h incubation period, activated Th1 + T cells (1 × 10 5 cells), with or without an IFN-γ neutralizing antibody (20 µg/mL), were added to the upper chamber. After a coculture period of 48 h, the Th1 + T cells and medium in the upper chamber were aspirated. Then, CD4 + T cells (1 × 10 5 cells) from the spleens of A 2A R +/+ mice were isolated using Anti-Mouse CD4 Magnetic Particles (551539, BD) and added to the upper chamber. Concurrently, CXCL12 (1 µg/mL) was added to the lower chamber. After an additional 24-h incubation period, the CD4 + T cells in the lower chamber were stained with crystal violet. Finally, the cells were visualized and counted using an optical microscope.

Statistical analysis

Repeated measures two-way analysis of variance (ANOVA) was used to analyse the differences in neurological deficit scores across the different groups throughout the course of EAE. Statistical analysis of molecular expression data from tissues and cells from multiple groups was conducted using one-way ANOVA and Tukey’s multiple comparisons test. Comparisons between two groups, including the A 2A R −/− and IFNGR1-Kd groups, during the course of EAE were performed using Student’s t tests.

Statistical analyses were performed using GraphPad Prism 9.0. Statistical significance was established at P  < 0.05. The data are expressed as the mean ± SEM or mean ± SD.

Results

Global A 2A R knockout increased CD4 + T-cell infiltration across the ChP in EAE model mice

To explore the potential interactions between CD4 + T cells and the ChP epithelium in the context of EAE, we first examined the infiltration of CD4 + T cells across the ChP of A 2A R −/− mice. Fourteen days after EAE induction, the mice were perfused and ChP tissues were isolated for CD4 immunostaining. Our analyses revealed an increased number of CD4-positive cells in the ChP of A 2A R −/− mice compared with the ChP of their wild-type (A 2A R +/+ ) littermates (Fig.  1 A-B). However, a decrease in the number of CD4-positive cells in the ChP was detected in EAE model mice treated with the A 2A R antagonist (KW6002) compared with those in the vehicle group (Fig.  1 C-D). These findings reveal the different effects of global A 2A R knockout and A 2A R antagonists during EAE progression.

In EAE model mice, global inactivation of the A 2A R gene increased ChP gateway activity and T-cell infiltration across the ChP. A Immunofluorescence staining showing CD4 expression at 14 days after MOG 35−55 treatment in both A 2A R +/+ and A 2A R −/− mice. Scale bar, 100 μm ( n  = 6/group). B Quantitative analysis of the CD4 + T-cell number on the basis of the staining images in Fig. 1A ( n  = 6/group). Statistical analysis was conducted using an unpaired t test. C Immunofluorescence staining of CD4 at 14 days following MOG 35−55 treatment in mice treated with KW6002 or vehicle. Scale bar, 100 μm ( n  = 6/group). D Quantitative analysis of the CD4 + T-cell number on the basis of the staining images in Fig. 1C ( n  = 6/group). Unpaired t test. E Staining for ICAM-1 at 14 days following MOG 35−55 treatment in both A 2A R +/+ and A 2A R −/− mice. Scale bar, 100 μm ( n  = 6/group). F Analysis of the fluorescence intensity of ICAM-1 based on the staining images in Fig. 1E. Unpaired t test. G-J Elevated mRNA levels of icam-1 , ccl5 , ccl20 , and ifngr1 in the ChP of A 2A R −/− mice compared with their A 2A R +/+ littermates 14 days after EAE induction. Unpaired t test ( n  = 5/group). K FCM analysis of T-cell infiltration in the CSF of A 2A R −/− or A 2A R +/+ mice at 14 days after EAE induction. L Fourteen days after EAE induction, FCM analysis revealed that the infiltration of CD4 + T cells into the CSF was greater in A 2A R −/− mice than in A 2A R +/+ mice. Unpaired t test ( n  = 6/group). M-N Fourteen days after EAE induction, FCM analysis revealed that the infiltration of Th1 + or Th17 + T cells into the CSF was greater in A 2A R −/− mice than in A 2A R +/+ mice. Unpaired t test ( n  = 6/group). O Measurement of IFN-γ levels in the blood serum of A 2A R +/+ and A 2A R −/− mice 14 days after EAE induction. Unpaired t test ( n  = 5/group). P Measurement of IFN-γ levels in the blood serum of mice treated with KW6002 or vehicle 14 days after EAE induction. Unpaired t test ( n  = 5/group). Significance levels are indicated as follows: * p  < 0.05, and ** p  < 0.01. Note: The error bars represent the SEM.

CAMs play crucial roles in the migration of lymphocytes. In this study, immunofluorescence staining revealed an increase in the protein expression of ICAM-1 in the ChP of A 2A R −/− mice compared with the ChP of their A 2A R +/+ littermates on day 14 after EAE induction (Fig.  1 E-F). Concurrently, qPCR analysis demonstrated significant upregulation of icam-1 , ccl5 , ccl20 , and ifngr1 mRNA in the ChP of A 2A R −/− mice (Fig.  1 G-J). These results suggest impaired gateway function of the ChP in A 2A R −/− mice, facilitating T-cell infiltration into the brain via the ChP.

Next, we examined T-cell infiltration in the CSF of A 2A R −/− mice with EAE. Fourteen days after MOG 35−55 treatment, CSF was collected from A 2A R −/− mice and A 2A R +/+ mice. FCM analysis revealed that there were more CD4 + T cells, including Th1 + and Th17 + T cells, in the CSF of A 2A R −/− mice than in the CSF of A 2A R +/+ mice (Fig.  1 K-N, Figure S1 ). This increased infiltration across the ChP of A 2A R −/− mice implies a greater likelihood of interactions between CD4 + T cells and the ChP epithelium.

Given that Th1 + T cells are the primary producers of IFN-γ, the increase in the number of Th1 + T cells likely contributes to elevated IFN-γ levels in the circulatory system during EAE progression. On the 14th day following EAE induction, IFN-γ levels in the peripheral blood serum were found to be significantly greater in mice with global A 2A R knockout than in control mice by ELISA, whereas IFN-γ levels in the blood serum of antagonist-treated mice were significantly lower than those in the vehicle group (Fig.  1 O-P). These findings suggest that IFN-γ secreted by Th1 + cells may be a key regulator of the opposite regulatory effects of A 2A R antagonists and global A 2A R knockout on EAE development.

A 2A R-deficient Th1 + T cells from the spleen enhanced cell proliferation and IFN-γ secretion

A previous study indicated that IFN-γ can diminish the gateway activity of the ChP [ 15 ]. Thus, we propose that CD4 + T cells may regulate ChP gateway activity via IFN-γ secretion. To assess the impact of A 2A R deficiency on CD4 + T cells in the context of EAE, we conducted FCM analyses of spleens isolated fourteen days after EAE induction. Compared with T cells from A 2A R +/+ mice, A 2A R-deficient CD4 + T cells exhibited enhanced proliferation in response to MOG 35 − 55 treatment (Fig.  2 A-B, Figure S2 ). Additionally, A 2A R deficiency led to an increase in the numbers of Th1 + and Th17 + T cells (Fig.  2 C-D).

A 2A R-deficient T cells in the spleen increased cell proliferation and IFN-γ secretion. A Fourteen days after EAE induction, spleens from both A 2A R +/+ and A 2A R −/− mice were subjected to FCM to assess variations in the populations of CD4 + , Th1 + , and Th17 + T cells ( n  = 6/group). B-D FCM analysis was used to quantify the proportions of CD4 + , Th1 + , and Th17 + T cells, as presented in Fig. 2A. Statistical evaluation was conducted using an unpaired t test ( n  = 6/group). E Assessment of the purity of Th1 + T cells isolated from the spleen tissues of A 2A R +/+ and A 2A R −/− mice using FCM ( n  = 5/group). F An ELISPOT was conducted to evaluate IFN-γ secretion by activated Th1 + T cells from the spleens of A 2A R +/+ and A 2A R -/- mice ( n  = 5/group). G-I Statistical analysis of the ELISPOT assay results shown in Fig. 2F. The volume of spots indicates the total IFN-γ content ( G ), the number of IFN-γ-positive spots represents the number of IFN-γ-secreting cells ( H ), and the average volume of the spots reflects the mean amount of IFN-γ secreted per Th1 + T-cell ( I ). Unpaired t test ( n  = 5/group). Significance levels are indicated as follows: * p  < 0.05, ** p  < 0.01, and *** p  < 0.001. Note: The error bars represent the SEM.

While IFN-γ levels are elevated in A 2A R −/− mice during EAE, the underlying cause, whether it is an increased number of Th1 + T cells or increased IFN-γ secretion from each cell, remains unclear. To further explore the potential effects of A 2A R on IFN-γ secretion, Th1 + T cells were isolated from the spleens of both A 2A R −/− and A 2A R +/+ mice 14 days after EAE induction using FCM (Fig.  2 E). The purity of the sorted Th1 + T cells was approximately 97.56% (Table S1). These sorted cells (1 × 10 4 cells per well) were subsequently activated in vitro, and IFN-γ secretion from these cells was subsequently quantified using an ELISPOT assay (Fig.  2 F). Notably, the A 2A R −/− group presented larger spots than the A 2A R +/+ group did, indicating that they produced a larger amount of IFN-γ in response to the same stimulus (Fig.  2 G). Furthermore, the increase in the number of IFN-γ + spots (representing the number of Th1 + T cells that secrete interferon) in the A 2A R −/− group suggested that the absence of A 2A R led to increase the number of Th1 + T cells under equivalent stimulation conditions (Fig.  2 H). Notably, the average spot area (representing IFN-γ secretion from an individual cell) was also larger in the A 2A R −/− group than in the A 2A R +/+ group (Fig.  2 I). This finding implies that individual A 2A R-deficient Th1 + T cells possess an increased capacity to secrete IFN-γ when stimulated.

A 2A R-deficient ChP epithelial cells reduced ChP gateway activity whereas A 2A R-deficient Th1 + T cells increased ChP gateway activity via IFN-γ signalling

To elucidate the cellular basis for the modulation of ChP gateway activity by A 2A R, we conducted an experiment in which Th1 + T cells were cocultured with ChP epithelial cells (Fig.  3 A-B). We first isolated ChP epithelial cells from A 2A R flox/flox mice and cultured them in the upper chamber of a Transwell plate. After 5 days of culture, we added TAT-Cre recombinase to the culture medium to achieve specific knockdown of A 2A Rs in the ChP epithelial cells. As expected, an increase in TEER was observed in the TAT-Cre-treated ChP epithelial cells (Fig.  3 C). However, when TAT-Cre-treated ChP epithelial cells from A 2A R flox/flox mice were cocultured with splenic Th1 + T cells from A 2A R +/+ mice for 24 h, we observed a significant decrease in TEER, indicating disrupted ChP barrier integrity (Fig.  3 C). This reduction in TEER was more pronounced when ChP epithelial cells were cocultured with activated Th1 + T cells from A 2A R −/− mice than when they were cocultured with Th1 + T cells from A 2A R +/+ mice (Fig.  3 C). Furthermore, the application of an IFN-γ neutralizing antibody abolished the reduction in TEER induced by Th1 + T cells from A 2A R −/− mice (Fig.  3 C).

A 2A R-deficient ChP epithelial cells reduced ChP gateway activity whereas A 2A R-deficient splenic Th1 + T cells increased ChP gateway activity via IFN-γ signalling. A The sorting strategy used to isolate Th1 + T cells from the spleens of A 2A R +/+ and A 2A R -/- mice via FCM. B Schematic diagram of the coculture system of Th1 + T cells and ChP epithelial cells in vitro . ChP epithelial cells were isolated from A 2A R flox/flox mice (P7) and seeded into the upper chambers. Twenty-four hours later, TAT-Cre or PBS was added. After culture for another 5 days, preactivated Th1 + T cells from A 2A R +/+ or A 2A R −/ − mice were added to the upper chambers. Forty-eight hours later, the TEER was measured. C Statistical analysis of the TEER values in Fig. 3B. TEER analysis revealed that activated Th1 + T cells from A 2A R −/− mice significantly reduced the TEER of the ChP epithelium to a greater extent than Th1 + T cells from A 2A R +/+ mice did, predominantly through IFN-γ secretion. One-way ANOVA and Tukey’s multiple comparisons test were used ( n  = 5/group). D Transwell experiments revealed that A 2A R knockdown in the ChP epithelium, achieved through TAT-Cre treatment, reduced CD4 + T-cell migration across the ChP. However, this decreased ChP gateway activity was negated by activated Th1 + T cells from both A 2A R +/+ and A 2A R −/− mice, which increased CD4 + T-cell migration ( n  = 5/group). E The number of migrated CD4 + T cells was analysed from the data in Fig. 3D. One-way ANOVA and Tukey’s multiple comparisons test were used ( n  = 5/group). F-I The expression patterns of ccl5 , ccl20 , icam-1 and ifngr1 in the ChP epithelial cells following coculture were analysed from the data in Fig. 3B ( n  = 5/group). Significance levels are marked as follows: * p  < 0.05, ** p  < 0.01, and *** p  < 0.001. Note: The errors bars represent the SEM.

We further conducted a Transwell assay in which activated Th1 + T cells were cocultured with ChP epithelial cells from A 2A R flox/flox mice (pretreated with PBS or Cre recombinase) for 24 h; then, the Th1 + T cells were removed, and an equivalent number of isolated CD4 + T cells were added to the upper chambers. After 24 h of incubation, we observed reduced migration of CD4 + T cells to the lower chamber in the TAT-Cre-treated ChP epithelial cell coculture group than in the PBS-treated ChP epithelial cell coculture group (Fig.  3 D-E). Additionally, qPCR analysis revealed a significant decrease in mRNA levels of icam-1 , ccl5 , ccl20 , and ifngr1 in the TAT-Cre-treated ChP epithelial cells compared to PBS-treated ChP epithelial cells (Fig.  3 F-I). These findings suggest that targeted knockdown of A 2A R in the ChP epithelium attenuates ChP gateway activity, thereby impeding the transmigration of immune cells. Importantly, when ChP epithelial cells were cocultured with Th1 + T cells from A 2A R +/+ mice, there was an increase in cell migration to the lower chamber and significant upregulation of icam-1 , ccl5 , ccl20 , and ifngr1 mRNA expression in the ChP epithelial cells, as shown in Fig.  3 D-I. These results suggest that activated Th1 + T cells from A 2A R +/+ mice disrupted ChP gateway function and increased the migration of CD4 + T cells across the ChP epithelium. In contrast, compared with those from A 2A R +/+ mice, activated Th1 + T cells from A 2A R −/− mice impaired ChP gateway function and increased immune cell migration capability to a greater extent (Fig.  3 D-I). However, the impairment of ChP gateway function and increase in migration induced by Th1 + T cells from A 2A R −/− mice were effectively mitigated upon the addition of an IFN-γ neutralizing antibody (Fig.  3 D-I).

ICV injection of activated Th1 + T cells reversed the attenuation of ChP gateway activity caused by A 2A R knockdown in the ChP via IFN-γ signalling

To further elucidate the effects of A 2A R-mediated interactions between ChP epithelial cells and Th1 + T cells in controlling ChP gateway activity in intact animals, we first injected TAT-Cre recombinant proteins into the lateral ventricles of A 2A R flox/flox mice. Two weeks after the injection, we observed a marked reduction in ICAM-1 protein expression in the ChP of the mice treated with TAT-Cre compared with that of the PBS-treated control mice (Fig.  4 A-B). Two weeks following ICV injection of TAT-Cre, we isolated Th1 + T cells from the spleen tissues of A 2A R +/+ and A 2A R −/− mice using FCM. These Th1 + T cells were subsequently reactivated in vitro for 24 h. The activated Th1 + T cells were then injected into the lateral ventricles of A 2A R flox/flox mice that had been pretreated with TAT-Cre as described. Twenty-four hours later, ICV injection of activated Th1 + T cells (A 2A R +/+ ) resulted in increased ICAM-1 protein levels and infiltration of CD4 + T cells into the ChP. Furthermore, activated Th1 + T cells (A 2A R −/− ) triggered increased ICAM-1 protein levels and increased CD4 + T-cell infiltration into the ChP (Fig.  4 A-D). Similarly, qPCR analysis revealed that compared with the control group, which received ICV injection of sterile PBS, the TAT-Cre-treated group presented decreased mRNA levels of icam-1 , ccl5 , ccl20 , and ifngr1 in the ChP (Fig.  4 E-H). Injection of activated Th1 + T cells (A 2A R +/+ ) into the lateral ventricle led to increased mRNA levels of icam-1 , ccl5 , ccl20 , and ifngr1 , whereas activated Th1 + T cells from A 2A R −/− mice induced a more pronounced increase in the levels of these mRNAs (Fig.  4 E-H).

A 2A R-deficient Th1 + T cells impaired ChP barrier function in vivo. A-B Immunostaining for ICAM-1 in brain slices from A 2A R flox/flox mice after ICV injection of PBS, activated Th1 + T cells from A 2A R +/+ or activated Th1 + T cells from A 2A R −/− mice and 24 h after IP injection of an IFN-γ neutralizing antibody ( n  = 6/group). C-D Immunostaining for CD4 in brain slices from A 2A R flox/flox mice after ICV injection of PBS, activated Th1 + T cells from A 2A R +/+ or activated Th1 + T cells from A 2A R −/− mice and 24 h after IP injection of an IFN-γ neutralizing antibody ( n  = 6/group). E-H The mRNA levels of ccl5 , ccl20 , icam-1 and ifngr1 in the ChP epithelium following ICV injection of PBS, activated Th1 + T cells from A 2A R +/+ or activated Th1 + T cells from A 2A R −/− mice and 24 h after IP injection of an IFN-γ neutralizing antibody ( n  = 5/group). Significance levels are marked as follows: * p  < 0.05, ** p  < 0.01, and *** p  < 0.001. Note: The error bars represent the SEM.

Research has indicated that IFN-γ modulates ChP gateway activity by interacting with IFN-γ receptor in the ChP epithelium [ 15 ]. Importantly, pretreatment with an anti-IFN-γ neutralizing antibody (i.p.) abolished T-cell infiltration induced by activated Th1 + T cells (A 2A R −/− ) (Fig.  4 A-D). Similarly, injection of an anti-IFN-γ antibody inhibited the inflammatory response induced by activated Th1 + T cells (A 2A R −/− ) in the ChP (Fig.  4 E-H).

Together, these in vitro and in vivo findings suggest that knockdown of A 2A R in the ChP epithelium reduced ChP gateway activity. In contrast, A 2A R-deficient CD4 + T cells, mainly Th1 + T cells, reversed the effect of ChP epithelium-specific A 2A R knockdown on ChP gateway activity, primarily through increased secretion of IFN-γ.

ChP-specific knockdown of IFN-γ signalling (IFNGR1) in A 2A R-deficient mice attenuated EAE symptoms

To further investigate the role of A 2A R-mediated interactions between ChP epithelial cells and immune cells in the ChP in the development of EAE pathology, we blocked IFN-γ signalling specifically in the ChP by injecting AAV2/5-U6-shRNA (IFNGR1)-EGFP or a control virus into the lateral ventricles of A 2A R −/− mice. As demonstrated in Fig.  5 A, EGFP expression was exclusively observed in the ChP. Two weeks after injection, the mice were treated with the MOG   35−55 peptide. We observed a significant reduction in the average EAE score in the IFNGR1-Kd group compared with the control group (Fig.  5 B). Additionally, the onset of EAE was delayed in the IFNGR1-Kd group (Fig.  5 B). Immunostaining for IFNGR1 showed approximately 60% efficiency in shRNA-mediated knockdown (Fig.  5 C-D). H&E staining of the spinal cord revealed reduced immune infiltration in the IFNGR1-Kd group (Fig.  5 E-F). MBP staining also revealed that the knockdown of IFNGR1 significantly mitigated EAE-induced demyelination in A 2A R −/− mice (Fig.  6 A-B), confirming the protective effect of IFNGR1-Kd in A 2A R-deficient mice.

Alleviation of EAE pathology in A 2A R −/− mice via ChP-specific knockdown of IFNGR1. A Representative fluorescence image showing A 2A R −/− mice two weeks following ICV injection of AAV2/5-shRNA-EGFP or a control (scramble) virus. EGFP fluorescence was exclusively observed in the ChP tissue. B The EAE score was reduced in A 2A R −/− mice due to specific IFNGR1 knockdown ( n  = 6/group). C Immunostaining for IFNGR1 revealed a significant reduction in IFNGR1 protein levels in A 2A R −/− mice 14 days after EAE induction following AAV2/5-shRNA-EGFP injection ( n  = 6/group). D Analysis of the IFNGR1 fluorescence intensity from the images in Fig. 5C. Unpaired t test ( n  = 6/group). E H&E staining revealed decreased immune cell infiltration into the spinal cords in A 2A R −/− mice compared with control mice (A 2A R −/− mice treated with ICV injection of the control virus) on day 14 following EAE induction ( n  = 5/group). F Semiquantitative analysis of the degree of immune cell infiltration from the images in Fig. 5E. Unpaired t test ( n  = 5/group). Significance levels are indicated as follows: ### p  < 0.001, * p  < 0.05, ** p  < 0.01, and *** p  < 0.001. Note: the error bars represent the SEM or SD (Fig. 5B).

Reduction in neuroinflammation in A 2A R-deficient EAE model mice via ChP-specific knockdown of IFNGR1. A Representative images of MBP and NF-H staining showing severe demyelination in A 2A R-deficient mice 14 days after EAE induction and a notable reduction in demyelination following ChP-specific knockdown of IFNGR1. MBP was used as a myelin marker, and NF-H was used as a neuronal marker. B Statistical analysis of MBP coverage from the data shown in Fig. 6A. Unpaired t test ( n  = 6/group). C Claudin-1 staining indicating that ChP-specific knockdown of IFNGR1 in A 2A R −/− mice prevented the loss of Claudin-1 in the ChP at 14 days after EAE induction ( n  = 6/group). D Quantitative assessment of the intensity of Claudin-1 staining in the images in Fig. 6C. Analysis was conducted using an unpaired t test ( n  = 6/group). E IBA-1 and CD68 staining demonstrating a significant reduction in CD68 expression in IBA-1-positive cells in A 2A R −/− mice following ChP-specific IFNGR1 knockdown ( n  = 6/group). F Statistical analysis of IBA-1 and CD68 double-positive cell counts from the data in Fig. 6E. Unpaired t test ( n  = 6/group). G Three-dimensional reconstruction of microglial morphology showing a reduced cell diameter in microglia in the ChP in A 2A R-deficient mice with IFNGR1 knockdown compared with control mice ( n  = 40 cells/group). H-J Quantitative analysis of microglial diameter, branch number, and dendritic length on the basis of the 3D reconstructions in Fig. 6G. Unpaired t test ( n  = 40 cells/group). Significance levels are marked as follows: ** p  < 0.01 and *** p  < 0.001. Note: the error bars represent the SEM.

Moreover, we noted the near disappearance of the tight junction protein Claudin-1 in the ChP epithelium of A 2A R-deficient mice on day 14 following EAE induction. This loss was notably reversed following the concurrent knockdown of IFNGR1 (Fig.  6 C-D). Inflammation in the spinal cord was assessed through immunostaining for IBA-1 and CD68. The results indicated pronounced activation of IBA-1-positive microglia/macrophages (IBA-1 and CD68 double-positive cells) in A 2A R −/− mice on day 14 following EAE induction, whereas only minimal cells positive for both IBA-1 and CD68 were detected in A 2A R-deficient mice with ChP-specific IFNGR1 knockdown (Fig.  6 E-F). The 3D reconstruction of microglial morphology further confirmed that IFNGR1 knockdown reduced the soma size, whereas increased the dendritic length and branch numbers, suggesting a shift towards a resting state (Fig.  6 G-J). These findings collectively suggest that inhibiting the interaction between CD4 + T cells and the ChP epithelium via the knockdown of IFNGR1 in the ChP protected against EAE pathology in A 2A R-deficient mice.

IFNGR1 knockdown in the ChP of A 2A R-deficient mice reduced T-cell infiltration during EAE pathology

The infiltration of pathogenic CD4 + T cells into the CNS is a primary contributor to neuronal damage in EAE. Fourteen days after EAE induction, we observed substantial infiltration of CD4 + T cells in the ChP of A 2A R-deficient mice. Notably, this infiltration was markedly diminished following the targeted knockdown of IFNGR1 in these mice (Fig.  7 A-B). IFNGR1 knockdown also reduced the protein expression of ICAM-1 within the ChP of A 2A R-deficient mice (Fig.  7 C-D).

IFNGR1 knockdown in the ChP curtails CD4 + T-cell infiltration in A 2A R-deficient EAE mice. A Immunostaining evidence of reduced infiltration of CD4 + T cells across the ChP in A 2A R −/− mice following IFNGR1 knockdown 14 days after EAE induction ( n  = 6/group). B Statistical analysis of the CD4 + T-cell number from the data in Fig. 7A using an unpaired t test ( n  = 6/group). C Demonstration of decreased ICAM-1 protein expression in the ChP in A 2A R −/− mice following IFNGR1 knockdown at 14 days following EAE induction ( n  = 6/group). D Quantitative assessment of the ICAM-1 staining intensity from the data in Fig. 7C; statistical analysis was conducted using an unpaired t test ( n  = 6/group). E Isolation and analysis of immune cells from the spinal cord at 14 days after EAE induction using FCM. F FCM analysis indicating that IFNGR1 knockdown in the ChP led to reduced CD4 + T-cell infiltration into the spinal cords of A 2A R −/− mice. Unpaired t test ( n  = 6/group). G-H FCM data showing that IFNGR1 knockdown in the ChP reduced the infiltration of Th1 + or Th17 + T cells into the spinal cords of A 2A R −/− mice. Unpaired t test ( n  = 6/group). Significance levels are indicated as follows: ** p  < 0.01 and *** p  < 0.001. Note: the error bars represent the SEM.

We also examined the infiltration dynamics of CD4 + T cells in the spinal cord at 14 days after EAE induction. FCM analysis revealed a significant number of CD4 + T cells in the spinal cord in A 2A R −/− mice compared with A 2A R +/+ mice. Importantly, IFNGR1 knockdown resulted in an approximately 50% reduction in the infiltration of CD4 + T cells in the spinal cord (Fig.  7 E-F, Figure S3 ). Moreover, we assessed the impact of ChP-specific IFNGR1 knockdown on the infiltration of pathogenic CD4 + T-cell subpopulations, specifically Th1 + and Th17 + T cells. Consistent with our earlier findings [ 21 ], IFNGR1 knockdown in the ChP decreased the numbers of both Th1 + and Th17 + T cells in the spinal cords of A 2A R-deficient mice (Fig.  7 G-H).

We next further explored the effect of IFNGR1 knockdown in the ChP on immune cell infiltration in the CSF of A 2A R −/− mice. FCM analysis of the CSF revealed that IFNGR1 knockdown in the ChP reduced the numbers of CD4 + T, Th1 + T and Th17 + T cells in the CSF of A 2A R −/− mice at 14 days after EAE induction (Fig.  8 A-D).

IFNGR1 knockdown in the ChP protected against the infiltration of CD4 + T cells into the CSF of A 2A R-deficient EAE model mice. A Isolation and analysis of immune cells in the CSF at 14 days after EAE induction using FCM. B IFNGR1 knockdown in the ChP reduced the infiltration of CD4 + T cells into the CSF of A 2A R −/− mice on day 14 following EAE induction. Unpaired t test ( n  = 6/group). C-D IFNGR1 knockdown in the ChP reduced the infiltration of Th1 + or Th17 + T cells into the CSF of A 2A R −/− mice on day 14 following EAE induction. Unpaired t test ( n  = 6/group). Significance levels are marked as follows: * p  < 0.05 and *** p  < 0.001. Note: the error bars represent the SEM.

Discussion

The most important finding of this study is that A 2A R-mediated interactions between Th1 + T cells and the ChP epithelium play a critical role in controlling T-cell infiltration across the ChP and EAE pathology. The distinct roles of A 2A R in immune cells (Th1 + cells) and nonimmune cells (ChP epithelial cells) and their interaction are highlighted by two separate findings. First, coculture of splenic Th1 + cells and ChP epithelial cells from A 2A R flox/flox mice treated with TAT-Cre to achieve ChP-specific A 2A R knockdown revealed that A 2A R knockout in ChP epithelial cells strengthened the ChP barrier and attenuated T-cell migration. Moreover, in intact animals, A 2A R knockdown in the ChP epithelium alone reduced ChP gateway activity, attenuated T-cell infiltration and attenuated the expression of leukocyte trafficking-related genes. The finding that A 2A R knockdown in the ChP reduced ChP gateway activity aligns with our previous results regarding EAE pathology [ 5 , 22 ] and is also consistent with the ability of A 2A R antagonists to increase BBB integrity [ 23 ] and with reducing inflammatory responses and pathological lesions by A 2A R knockdown in endothelial cells in stroke model mice [ 24 ]. On the other hand, coculture of ChP epithelial cells with activated splenic Th1 + T cells from A 2A R +/+ mice increased ChP permeability and increased T-cell infiltration and the expression of leukocyte trafficking-related molecules. This observation aligns with studies showing that activated T cells in the CSF can regulate T-cell migration across the ChP, which is dependent on chemokines and ICAM-1 [ 19 ]. Importantly, the co-culture ChP epithelial cells with activated splenic Th1 + T cells from A 2A R −/− mice countered A 2A R-deficient ChP epithelium effect on ChP permeability with increased T cell infiltration and expression of leukocyte trafficking molecules. Moreover, increased IFN-γ secretion and increased ChP gateway activity were observed after ICV injection of Th1 + T cells from A 2A R −/− mice compared with after ICV injection of Th1 + T cells from A 2A R +/+ mice. A 2A R-deficient splenic Th1 + T cells reversed the beneficial effects of A 2A R knockout in the ChP epithelium (i.e., reduced ChP gateway activity to resist infiltration by immune cells), thereby exacerbating EAE pathology. Thus, maintaining A 2A R function in the immune system is required to suppress the proinflammatory effects of immune cells and preserve ChP gateway activity. This may further explain the protective effects of A 2A R agonists in the early phase of EAE [ 25 ], as they decrease the proliferation of T cells, subsequently reducing the opportunity for interactions between T cells and the ChP.

Collectively, these findings provide evidence that A 2A R-deficient Th1 + T cells exacerbate EAE pathology and increase immune cell infiltration in the ChP in A 2A R-deficient mice, whereas A 2A R-deficient ChP epithelial cells exhibit neuroprotective effects and attenuate immune cell infiltration in the ChP. The identification of the cell types in the ChP involved in the opposing effects of A 2A R antagonists and A 2A R −/− on EAE development provides a cellular basis for the protective effects of A 2A R antagonists against MS pathology. This finding also suggests that targeted delivery to the ChP may represent an important therapeutic avenue to enhance the protective effect of A 2A R antagonists against EAE/MS pathology.

Another important finding is that IFN-γ signalling is the key mediator of the A 2A R-mediated interaction between Th1 + T cells and the ChP epithelium in the control of EAE pathology. IFN-γ, the hallmark cytokine of Th1 + T cells, plays an important role in CD4 + T cell-mediated EAE pathology [ 14 ]. Furthermore, IFN-γ binds to IFNGR1 and induces the expression of CAMs in the ChP epithelium [ 15 ]. This allows IFN-γ to act as an ideal mediator of the interactions between immune cells and the ChP epithelium. Furthermore, A 2A R knockout in immune cells results in a dramatic increase in IFN-γ levels in EAE [ 11 ]. This finding was supported by our demonstration that A 2A R knockout elevated the IFN-γ level in the blood serum on day 14 after EAE induction, whereas the A 2A R antagonist KW6002 decreased the IFN-γ level in the blood serum on day 14 following EAE induction. These findings indicate that KW6002 may attenuate the infiltration of T cells and protect against MS/EAE pathology by restraining IFN-γ secretion and reducing the binding of IFN-γ to IFNGR1 in the ChP, resulting in a strengthened ChP barrier. Finally, by knocking down IFNGR1 in the ChP, we observed a significant decrease in EAE onset and significant alleviation of EAE pathology in A 2A R-deficient mice. Taken together, these findings establish that IFN-γ signalling is the key mediator of A 2A R-mediated interactions between Th1 + T cells and the ChP epithelium.

IFN-γ is increased in MS patients, and IFN-γ is believed to be harmful in the context of the disease [ 26 ]. Like A 2A R, IFN-γ also has extremely complex and contradictory effects on EAE pathology according to previous studies. For example, silencing IFN-γ in astrocytes alleviates EAE pathology, whereas silencing IFN-γ in microglia leads to more severe symptoms [ 14 ]. Accordingly, knockout of IFN-γ or IFNGR1 also exacerbates the pathological progression of EAE [ 27 , 28 ]. Our data indicated that inactivation of IFN-γ signalling in the ChP was protective against EAE. Thus, selectively inhibiting IFN-γ signalling in the ChP may be a promising strategy for MS treatment. Rather than directly targeting IFN-γ with an anti-IFN-γ antibody, indirect modulation of IFN-γ with A 2A R antagonists (which apparently preferentially target IFN-γ signalling in the ChP) may be beneficial for MS patients. Although the effects of A 2A R-mediated interactions between Th1 + T cells and the ChP via IFN-γ signalling on EAE pathology were assessed in this study, the precise mechanism by which A 2A R deficiency promotes the production of IFN-γ whereas A 2A R antagonist inhibits it during EAE has yet to be elucidated and needs future investigation.

Conclusion

We elucidated the critical role of A 2A R-mediated interactions between Th1 + T cells and the ChP via IFN-γ in EAE. A 2A R knockout in the ChP epithelium reinforced the ChP barrier, whereas A 2A R-deficient Th1 + T cells destroyed the ChP barrier via increased secretion of IFN-γ. This study further described the dual role of A 2A R signalling in immune modulation and emphasized that targeted inactivation of A 2A R in the ChP was a promising strategy for MS treatment.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1

Supplementary Material 2

Supplementary Material 3

Acknowledgements

Not applicable.

Abbreviations

Multiple sclerosis

Choroid plexus

Experimental autoimmune encephalomyelitis

Central nervous system

Adenosine A 2A receptor

Intracerebroventricular

Myelin oligodendrocyte glycoprotein

Blood-brain barrier

Cerebrospinal fluid

Complete Freund’s adjuvant

Pertussis toxin

Repeated-measurements

Intercellular cell adhesion molecule-1

Blood-CSF barrier

Food and Drug Administration

Trans-epithelial electrical resistance

Interferon gamma

IFN-γ receptor 1

Disease-modifying therapy

Enzyme-linked immunosorbent assay

Analysis of variance

Chemokines and adhesion molecules

Flow cytometry

Author contributions

W.Z., J.F.C., and J.Q. conceptualized and designed the experiments and contributed to writing the manuscript; C.X.Q., Y.W.Y., P.T., C.Z., X.H.L., and N.J. conducted the experiments; W.Z., and J.F. C. further refined the manuscript.

Funding

This research was funded by the National Natural Science Foundation of China (Grant No. 31800903 and No. 82150710558) and the Natural Science Foundation of Zhejiang province, China (Grant No. LMS25H090008 and No. LQ18H090007).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

The animal experiments conducted in this study received approval from the Institutional Ethics Committee for Animal Use in Research and Education at Wenzhou Medical University, China.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Chenxing Qi, Yuwen Yang and Ping Tang have contributed equally to this work.

Contributor Information

Jia Qu, Email: jiaqu@163.com.

Jiang-Fan Chen, Email: chenjf@wmu.edu.cn.

Wu Zheng, Email: zheng123wu@163.com.

References

Associated Data

Supplementary Materials

Supplementary Material 1

Supplementary Material 2

Supplementary Material 3

Data Availability Statement

No datasets were generated or analysed during the current study.