Aberrant adenosine A2A receptor signaling in the choroid plexus drives CSF hypersecretion and ventriculomegaly in hydrocephalus

Zheng, Wu; Hu, Lanxin; Yang, Yuwen; Li, Xuhang; Wu, Jia; Chen, Wangchao; Jiang, Yiwei; Chen, Yuhan; Wang, Hongcai; Li, Shiwei; Feng, Kuan; Zheng, Cheng; Wang, Yanjiang; Cunha, Rodrigo A.; Lv, Zhongyue; Chen, Jiang-Fan

doi:10.1038/s41467-025-66776-8

Download PDF

Article
Open access
Published: 07 December 2025

Aberrant adenosine A_2A receptor signaling in the choroid plexus drives CSF hypersecretion and ventriculomegaly in hydrocephalus

Wu Zheng ORCID: orcid.org/0000-0002-5091-9694^1,2^na1,
Lanxin Hu^1,3^na1,
Yuwen Yang^1,3^na1,
Xuhang Li¹,
Jia Wu⁴,
Wangchao Chen¹,
Yiwei Jiang¹,
Yuhan Chen¹,
Hongcai Wang⁵,
Shiwei Li⁵,
Kuan Feng⁵,
Cheng Zheng¹,
Yanjiang Wang⁶,
Rodrigo A. Cunha ORCID: orcid.org/0000-0003-2550-6422⁷,
Zhongyue Lv ORCID: orcid.org/0000-0002-8906-0629² &
…
Jiang-Fan Chen ORCID: orcid.org/0000-0002-0446-3956^1,3

Nature Communications volume 17, Article number: 83 (2026) Cite this article

5258 Accesses
3 Citations
Metrics details

Subjects

Hydrocephalus

Abstract

Hydrocephalus is a common and serious neuropathological condition characterized by disrupted cerebrospinal fluid (CSF) circulation, which lacks effective pharmacotherapy. Here, we demonstrate that adenosine A_2A receptor (A_2AR) signaling in the choroid plexus (ChP) is a trigger of hydrocephalus. Adenosine levels are increased in the CSF of hydrocephalus patients and mice, together with elevated ChP-A_2AR density. Accordingly, continued infusion of adenosine for 14 days or transgenic ChP-A_2AR overexpression is sufficient to drive CSF hypersecretion and ventriculomegaly. Conversely, selective knockdown of ChP-A_2AR reduces CSF production and ameliorates CSF hypersecretion and ventriculomegaly induced by autologous blood and kaolin. Furthermore, we unveil ChP-A_2AR signaling as a molecular mechanism linking brain insults with CSF hypersecretion through parallel PI3K/Akt-dependent activation of SPAK phosphorylation and NF-κB-dependent transcriptional regulation of ATP1A2. Lastly, the A_2AR antagonist KW6002 protects against hydrocephalus induced by autologous blood and kaolin, offering a novel treatment for hydrocephalus by repurposing the FDA-approved A_2AR antagonist istradefylline.

Targeting modulation of the choroid plexus blood-CSF barrier and CSF hypersecretion via lipid nanoparticle-mediated co-delivery of siRNA and resveratrol

Article Open access 10 July 2025

Paediatric hydrocephalus

Article 16 May 2024

Selective antagonism of adenosine A2A receptor reduces hypobaric hypoxia-induced neuroinflammation by inhibiting cGAS-STING pathway

Article Open access 05 January 2026

Introduction

Hydrocephalus is a common and serious neurological condition characterized by the enlargement of the cerebrospinal fluid (CSF)-filled ventricles in the brain, impairing brain development in children and causing neurodegeneration in adults¹. Although hydrocephalus is a potentially devastating condition, effective prevention strategies are lacking, and shunt revision surgery remains the mainstay of treatment (accounting for nearly one third of all neurosurgical procedures in the United States), despite frequent complications and high mortality¹. The efforts to develop alternative pharmacotherapies for hydrocephalus patients have largely failed due to an incomplete understanding of the pathogenesis of hydrocephalus.

The enlargement of CSF-filled ventricles is traditionally attributed to a decreased CSF reabsorption and an impediment in CSF outflow, although CSF hypersecretion by the choroid plexus (ChP) in response to infection, hemorrhage or traumatic insult, is increasingly recognized to play a role in the onset and progression of hydrocephalus^2,3. In particular, in the development of post-infectious (PIH) and post-hemorrhagic hydrocephalus (PHH), ChP triggers CSF hypersecretion in response to infection, hemorrhage, and traumatic insults. This secretory response may be initially adaptive to maintain homeostasis by clearing pathogenic organisms or debris, but becomes maladaptive eventually promoting ventriculomegaly and elevated intracranial pressure⁴. Accordingly, hypersecretion of CSF resulting from ChP hyperplasia or tumors is sufficient to cause non-obstructive hydrocephalus⁵. Several ChP-enriched epithelial transport proteins critically controlling CSF production were identified by genetic and pharmacological studies, including sodium-coupled bicarbonate transporters, AQP1 water-channels⁶ and particularly Na⁺/K⁺/2Cl⁻ cotransporters (NKCC1), which contribute to approximately half of the production of CSF⁷. And recent studies show that the targeted inhibition of SPAK (STE20/SPS1-related proline/alanine-rich kinase)-NKCC1 in the ChP alleviates kaolin-induced hydrocephalus, post-infectious and PHH^4,8,9.

While the ChP-enriched epithelial transport proteins controlling CSF production have been identified, the molecular link coupling brain infection and hemorrhage (trauma) with CSF hypersecretion and development of hydrocephalus is unknown. We now hypothesize that the ChP-adenosine-A_2AR signaling may serve as a molecular link between inflammation/infection/trauma and CSF hypersecretion. Adenosine is increasingly recognized as an important signal mediating adaptive stressful response, such as upon trauma, infection or inflammation¹⁰, which are the main triggers of ChP-mediated CSF hypersecretion in hydrocephalus. Accordingly, adenosine levels are markedly increased during transient forebrain ischemia and subsequent reperfusion^11,12, traumatic brain injury¹³ and brain infection¹⁴. Adenosine exerts its effects by activating G-protein coupled receptors, with a prominent role of A_2AR in the control of brain dysfunction¹⁵. Accordingly, we recently demonstrated that A_2AR expression in the ChP is induced in an experimental autoimmune encephalomyelitis model¹⁶ and the selective knockdown of A_2AR in the ChP prevents T-cell infiltration into the brain and mitigates immunologically-induced brain injury^16,17. Furthermore, we have also previously reported that astrocytic A_2AR are physically associated and control the activity of Na⁺-K⁺-ATPases¹⁸, which are a determinant of transmembrane fluxes in ChP epithelial cells¹⁹. This raises the exciting possibility that the aberrant adenosine-A_2AR signaling in ChP may serve as a molecular link between inflammation/infection/trauma with CSF hypersecretion via A_2AR-Na⁺-K⁺-ATPase interaction, which would prompt ChP-A_2AR as a potentially new target for the pharmacological treatment of hydrocephalus.

In this study, we demonstrated an aberrantly increased CSF-adenosine and ChP-A_2AR signaling in hydrocephalus mouse models and patients. The selective overexpression or knockdown of A_2AR and ATP1A2 in ChP, coupled with pharmacological treatment with A_2AR antagonists, uncovered ChP-A_2AR signaling as a molecular mechanism linking brain inflammation with CSF hypersecretion and ventriculomegaly via the SPAK/ATP1A2 cascade. This prompts a novel strategy for treating hydrocephalus with high translational potential by targeting the aberrant ChP-A_2AR signaling with the FDA-approved A_2AR antagonist KW6002.

Results

Adenosine signaling is aberrantly elevated in the CSF and ChP in hydrocephalus

To explore potential changes of adenosine levels within the CSF in the context of hydrocephalus pathology, we collected CSF samples from patients with different clinically defined conditions of hydrocephalus (n = 11) and age-matched controls with different neurological conditions without hydrocephalus (n = 8) (Table S1). HPLC analysis revealed that the median adenosine concentration in the CSF of controls was 0.242 μM (interquartile range: 0.029–0.337), and it was significantly elevated in the CSF of hydrocephalus patients, showing approximately a five-fold increase, with a median level of 1.038 μM (interquartile range: 0.673–3.204) (Fig. 1A–C).

**Fig. 1: Aberrant adenosine signaling during hydrocephalus in the cerebrospinal fluid (CSF) and choroid plexus (ChP).**

Given that the ChP is surrounded by self-secreted CSF, we hypothesized that these increased CSF levels of adenosine might contribute to hydrocephalus pathology by activating adenosine receptors located in the ChP. Among adenosine receptors, A_2AR plays a crucial role in the control of inflammation²⁰ and brain trauma¹⁵, as well as in transmembrane transport²¹. Therefore, we resorted to a PHH model in mice through intracerebroventricular (ICV), based on injection of autologous blood, to investigate putative alterations of A_2AR. Five days after administration of autologous blood, MRI analysis revealed that PHH mice had larger ventricles compared to mice treated with PBS (Fig. 1D, E). In parallel, PHH mice also displayed increased CSF levels of adenosine (Fig. 1F) and a significant increase of A_2AR immunofluorescent staining in the ChP (Fig. 1G, H). To identify the cell types where A_2AR was upregulated in PHH mice, we observed that the increased A_2AR signals did not co-locate with IBA-1 (macrophages/microglia), CD31 (endothelial cells) or MHC-II (antigen-presenting cells, APCs) and were superimposed with intercellular cell adhesion molecule-1 (ICAM-1, ChP epithelium) staining (Fig. 1I, J). Similarly, we induced obstructive hydrocephalus in mice by ICV injection of kaolin (Fig. S1A, B) and found increased adenosine levels in CSF and elevated A_2AR signals in the ChP at 10 days after the injection (Fig. S1C–G). These findings collectively suggest that aberrant adenosine/A_2AR signaling in both the CSF and ChP is present and may play a role in the pathology of hydrocephalus.

ChP-specific overexpression of A_2ARs increases CSF secretion and triggers hydrocephalus

To ascertain the role of A_2AR signaling, we induced A_2AR overexpression in the ChP of wild-type (WT) mice by ICV injection of AAV2/5-CMV-A_2AR-HA (A_2AR-OE) or AAV2/5-CMV-EGFP viruses (Fig. 2A). After two weeks, through observation of EGFP and HA staining, we confirmed that the overexpression of A_2AR was restricted to the ChP tissue and co-localized perfectly with the HA-tag (Fig. 2B, C). Additionally, no HA signal, which reported the presence of the injected AAV2/5-CMV-A_2AR-HA, was observed in other brain regions (Figs. S2 and S3). Western blot analysis of ChP lysates further confirmed that ICV injection of AAV2/5-CMV-A_2AR-HA increased A_2AR protein levels (Fig. 2D, E).

Fig. 2: Overexpression of A2AR selectively in the choroid plexus (ChP) (A2AR-OE) leads to cerebrospinal fluid (CSF) hypersecretion and enlarged ventricles. — **Fig. 2: Overexpression of A_2AR selectively in the choroid plexus (ChP) (A_2AR-OE) leads to cerebrospinal fluid (CSF) hypersecretion and enlarged ventricles.**

Two weeks after ICV-injection of the viruses, MRI scans were conducted to quantify the size of the brain ventricles. Figure 2F, G shows that A_2AR overexpression selectively in the ChP resulted in enlarged brain ventricles without altering brain volume (Fig. 2H). Ex vivo DAPI staining of brain sections further confirmed the increased ventricular size induced by A_2AR overexpression (Fig. 2I, J). Accordingly, the measurement of CSF content demonstrated a significant increase in CSF secretion (Fig. 2K, L).

We next examined the status of CSF outflow in A_2AR-OE mice by administering the CSF tracer Evans Blue into the lateral ventricles of both A_2AR-OE and control mice. Our observations revealed unimpeded distribution of the dye through the cerebral aqueduct and all ventricles, indicating the presence of aqueduct patency and the maintenance of bulk CSF flow (Fig. 2M)³.

These findings provide robust support for the assertion that overexpression of A_2AR in the ChP enhances CSF secretion and leads to an increase in ventricular size.

ChP-specific knockdown of A_2ARs decreases CSF secretion and ventricular size

To further investigate the role of aberrant A_2ARs signaling in the ChP, we conducted a ChP-selective knockdown of A_2AR (A_2AR-Kd) through ICV-injection in WT mice of AAV2/5 carrying a shRNA targeting adora2a or a scramble control (AAV2/5-U6-shRNA(A_2AR)-Kd-CMV-EGFP or AAV2/5-U6-scramble-CMV-EGFP). Two weeks post-injection, we employed MRI to assess potential changes in ventricular size (Fig. 3A). To validate the ChP-specific A_2AR knockdown with our method, we tracked the fluorescence distribution of the injected AAV2/5 labeled with EGFP. EGFP signals were exclusively located in the ChP in both scramble and A_2AR-Kd groups (Fig. 3B). A qPCR analysis showed that the AAV2/5 carrying the shRNA targeting A_2AR, achieved a near 80% knockdown efficiency (Fig. 3C). To systematically assess the EGFP distribution, we also screened its potential expression in major organs and tissues (heart, liver, spleen, lung, kidney, dura mater and cervical lymph node). However, no EGFP signals were detected (Fig. S4). Strikingly, the specific knockdown of A_2AR in the ChP led to an almost complete loss of the ventricles (Fig. 3D, E), whereas total brain volume remained unchanged (Fig. 3F). This was confirmed with DAPI staining of brain sections, which also revealed a significant reduction in ventricular size upon A_2AR knockdown (Fig. 3G, H). When we measured the CSF content by collecting CSF from the cisterna magna during 15 min, we found that virtually no CSF could be extracted from the A_2AR-Kd mice (Fig. 3I, J).

Fig. 3: Choroid plexus (ChP)-selective knockdown of A2AR (A2AR-Kd) leads to reduced cerebrospinal fluid (CSF) secretion and ventricular size. — **Fig. 3: Choroid plexus (ChP)-selective knockdown of A_2AR (A_2AR-Kd) leads to reduced cerebrospinal fluid (CSF) secretion and ventricular size.**

Genetic deletion of A_2ARs in the ChP and pharmacological inhibition of the A_2ARs mitigate hydrocephalus induced by autologous blood and kaolin

Our findings strongly suggest that A_2AR signaling in the ChP plays a pivotal role in CSF secretion. Therefore, we hypothesized that deactivating A_2AR signaling could mitigate ventriculomegaly in hydrocephalus. Initially, we suppressed the expression of A_2AR in the ChP through ICV injection of AAV2/5-U6-shRNA (A_2AR)-Kd-CMV-EGFP virus in WT mice. After a two-week interval, we induced PHH model in the mice. After 5 days, we evaluated lateral ventricle volumes by MRI (Fig. 4A). A_2AR staining showed that the A_2AR-Kd virus effectively reduced the protein abundance of A_2AR in the ChP (Fig. 4B, C). The MRI data unequivocally showed that mice with A_2AR knockdown exhibited reduced ventriculomegaly, as compared to control mice (Fig. 4D). Specifically, the average volume of the ventricles in A_2AR-Kd mice was less than one-third of that observed in the hydrocephalus mice treated with mock virus (Fig. 4E). In addition, no discernable alterations of brain volume were noted (Fig. 4F).

Fig. 4: Target inactivation of A2AR in the choroid plexus (ChP) alleviates hydrocephalus pathology. — **Fig. 4: Target inactivation of A_2AR in the choroid plexus (ChP) alleviates hydrocephalus pathology.**

In our previous studies, we have established that the intraperitoneal injection (i.p.) of the A_2AR selective antagonist, KW6002, can exert a pharmacological effect on the ChP^16,22. Consequently, following the ICV injection of autologous blood into WT mice, the mice received a daily i.p. treatment with KW6002 (5 mg/kg) (Fig. 4G). After 5 days, MRI analysis revealed that KW6002 was effective in alleviating the autologous blood-induced ventricle enlargement (Fig. 4H, I) without altering brain volume (Fig. 4J).

Moreover, we also performed obstructive hydrocephalus induced with kaolin. The data showed that inactivation of A_2AR in ChP by genetic (Fig. S5A–F) or pharmacological (Fig. S5G–J) method also improved obstructive hydrocephalus. We detect the expression of other adenosine receptors (A₁, A_2B, A₃) in hydrocephalus (kaolin-induced and PHH). As the data showed, there were no obvious changes in those receptors (Fig. S6). These results strongly imply that the targeted modulation of the aberrant A_2AR signaling in the ChP ameliorates the morphological hallmark of hydrocephalus. This prompts the possibility that inactivating A_2AR specifically in the ChP could hold a therapeutic promise for hydrocephalus.

Aberrant A_2AR signaling in the ChP regulates CSF secretion by triggering a PI3K-dependent phosphorylation of SPAK

SPAK serves as a regulatory scaffold within a multi-ion transporter complex, located in the ChP epithelium, that critically regulates CSF production⁴. In the context of the post-infectious hydrocephalus, the active form of SPAK (p-SPAK) is upregulated within the ChP, and the knockout of SPAK attenuates inflammation (LPS)-induced elevations in both CSF secretion and ventriculomegaly⁴. Accordingly, we also observed an elevation of p-SPAK signals in a PHH model (Fig. 5A, B). Notably, we observed an enrichment of p-SPAK in the ChP of A_2AR-OE mice (Fig. 5C, D). Subsequent Western blot analysis confirmed that the overexpression of A_2AR increased the levels of p-SPAK (Fig. 5E, F) in the ChP.

Fig. 5: A2AR signaling controls cerebrospinal fluid (CSF) secretion via SPAK phosphorylation in the choroid plexus (ChP). — **Fig. 5: A_2AR signaling controls cerebrospinal fluid (CSF) secretion via SPAK phosphorylation in the choroid plexus (ChP).**

It has been reported that the phosphorylation of SPAK is facilitated through the PI3K-Akt pathway²³ and adenosine can engage the PI3K-PKA-mTOR cascade via interaction with A_2AR^24,25. Thus, we postulated that A_2AR might regulate CSF secretion and ventricular size by augmenting the PI3K-PKA-dependent phosphorylation of SPAK. To test this hypothesis, we added either a selective PI3K inhibitor (GDC-0941) or a selective inhibitor of Akt (MK-2206) to mouse primary ChP epithelial cells that exhibited stable overexpression of A_2AR. After a 48-h interval, cells were collected for Western blot analysis, revealing a reduction in p-SPAK levels following either GDC-0941 or MK-2206 treatment (Fig. 5G–J).

Next, we designed an AAV2/5 vector carrying a shRNA targeting SPAK (AAV2/5-U6-shRNA (SPAK)-CMV-EGFP, SPAK-Kd) along with a control virus (AAV2/5-U6-scramble-CMV-EGFP, control). A virus combination (a mix of A_2AR-OE and SPAK-Kd virus) was subsequently injected into the lateral ventricle of WT mice (Fig. 5K). Mice injected with A_2AR-OE and AAV2/5-CMV-EGFP (control) virus were used as the control group. Two weeks later, we conducted p-SPAK staining, which confirmed that the shRNA targeting SPAK effectively reduced the protein level of p-SPAK by 60% (Fig. 5L, M), with immunostaining of both HA (A_2AR) and EGFP exclusively present in the ChP (Fig. 5N). The MRI analysis revealed that the suppression of p-SPAK in the ChP also mitigated the A_2AR-OE-induced ventriculomegaly (Fig. 5O–Q) and CSF hypersecretion (Fig. 5R, S). These findings further sustain that p-SPAK mediates the impact of A_2AR overexpression in excessive CSF secretion and enlarged brain ventricles.

A_2AR control CSF secretion via NF-κB-mediated transcriptional control of Na⁺-K⁺-ATPase (ATP1A2)

To further dissect the molecular mechanism underlying the A_2AR-mediated CSF secretion, we carried out a comparative RNA-seq analysis of ChP tissue harvested 14 days after ICV injections of either A_2AR-OE or control AAV into WT mice. We identified 67 differentially expressed genes (26 upregulated and 41 downregulated) in the A_2AR-OE group compared to the control group (Fig. 6A). The heatmap analysis of the top 50 genes visually illustrates significant alterations in transcript levels due to A_2AR overexpression (Fig. 6B). Gene ontology annotation revealed that the differentially expressed genes were predominantly enriched in ion transmembrane transporter activity, a process integral to CSF secretion (Fig. 6C). This substantiates the involvement of aberrant A_2AR signaling in controlling CSF production in the ChP.

Fig. 6: Overexpression of A2AR modulates the expression of ion channel proteins in the choroid plexus (ChP). — **Fig. 6: Overexpression of A_2AR modulates the expression of ion channel proteins in the choroid plexus (ChP).**

We further validated the RNA-seq data using qPCR. We selected certain candidate genes and found that the expression levels of atp1a2 and slc24a4 were reduced in the ChP of the A_2AR-OE group (Fig. 6D, E and S7). By contrast, overexpression of A_2AR did not alter the mRNA levels of atp1a1, atp1b1, and atp1b2 (Fig. 6F–H). Since ATP1A2 belongs to the superfamily of Na⁺-K⁺-ATPases and previous studies identified an ability of A_2AR to control Na⁺-K⁺-ATPase activity in astrocytes¹⁸, we hypothesized that A_2AR-ATP1A2 signaling could be involved in the CSF secretion in the ChP. Indeed, we observed that the overexpression of A_2AR led to a decrease in the protein levels of ATP1A2 (Fig. 6I, J). Building upon our previous study, where we found that A_2AR signaling in ChP epithelium regulates the expression of CCL20 by activating the NF-κB pathway¹⁶, we hypothesized that A_2AR might control the expression of ATP1A2 via the NF-κB pathway. To test this, we tested the effect of an inhibitor of NF-κB pathway transcriptional activity (JSH-23, 10 μM) in a stable mouse primary ChP epithelial cells, which overexpressed A_2AR. After a 48-h treatment, cells were harvested for Western blot analysis. Strikingly, JSH-23 treatment markedly increased ATP1A2 protein levels (Fig. 6K, L). This strongly implies that A_2AR regulates the expression of ATP1A2 in a manner dependent on the NF-κB pathway.

Although it has been reported that the protein levels of Na⁺-K⁺-ATPases are often decreased in models of hydrocephalus⁹, the precise role of Na⁺-K⁺-ATPase subunits, including ATP1A2, remains unknown in hydrocephalus. Thus, we sought to address this gap of knowledge by overexpressing ATP1A2 in the ChP. This was achieved through an ICV injection into WT mice of AAV2/5 carrying cDNA of atp1a2, which was labeled with a His-tag at the C-terminal (Fig. 7A). The successful overexpression of ATP1A2 in the ChP was confirmed by mCherry/His-tag staining and Western blot analysis (Fig. 7B–D). In addition, we checked that overexpression of ATP1A2 in ChP tissue did not alter the protein levels of A_2AR (Fig. 7E, F), which suggests that ATP1A2 may be a downstream gene of A_2AR.

Fig. 7: A2AR signaling in the choroid plexus (ChP) controls cerebrospinal fluid (CSF) secretion through transcriptional regulation of ATP1A2. — **Fig. 7: A_2AR signaling in the choroid plexus (ChP) controls cerebrospinal fluid (CSF) secretion through transcriptional regulation of ATP1A2.**

Two weeks after the ICV injection of AAV2/5-CMV-ATP1A2-His (ATP1A2-OE) or AAV2/5-CMV-mCherry (control) into WT mice, MRI was used to quantify lateral ventricle volume. Strikingly, the overexpression of ATP1A2 in the ChP resulted in smaller ventricles in comparison to control mice (Fig. 7G, H). There were no discernible changes in the overall brain volume (Fig. 7I). In line with the reduced ventricular sizes, ATP1A2 overexpression in the ChP resulted in decreased CSF content (Fig. 7J, K).

Since the data suggest that A_2AR overexpression (A_2AR-OE) regulates CSF secretion by suppressing ATP1A2 expression, we next attempted a co-overexpression of A_2AR and ATP1A2 in the ChP to validate this proposed mechanism. This was achieved through the simultaneous ICV injection of A_2AR-OE and ATP1A2-OE viruses into the lateral ventricle of WT mice (Fig. 7L). Mice receiving ICV injection of A_2AR-OE and AAV2/5-CMV-mCherry virus were considered as the control group. We observed that the HA signal (A_2AR) and His-tag signal (ATP1A2) were both selectively detected in the ChP (Fig. 7M). Two weeks after viral injection, MRI analysis revealed a reduction in ventricular size in mice with co-overexpression of A_2AR and ATP1A2, as compared to the control group (Fig. 7N–P). Similarly, the CSF volume in the co-overexpression group was lower than that in the controls (Fig. 7Q, R). These results demonstrate that overexpression of ATP1A2 can counteract the enlarged ventricle and excessive CSF secretion induced by overexpression of A_2AR. This might be relevant in hydrocephalus since we found that the protein level of ATP1A2 was partly restored after KW6002 treatment in the PHH model of hydrocephalus (Fig. S8).

Elevated adenosine in ventricles directly triggers a hydrocephalus phenotype by activating A_2AR signaling in the ChP

After unraveling the mechanism of A_2AR-mediated control of CSF secretion, we next attempted to experimentally link the abnormal adenosine levels in the CSF of hydrocephalus patients with the A_2AR-mediated CSF hypersecretion. To simulate the increased concentration of adenosine in hydrocephalus patients, we injected adenosine (or PBS as control) into the lateral ventricle of WT mice 4 times per day with an inter-injection interval of 6 h during 14 days, targeting a final concentration of 6 μM (comparing to the basal concentration of adenosine at 2 μM in normal mice, Fig. 1F) in a mouse CSF volume of 40 μL (Fig. 8A). Of note, the chronic administration of adenosine upregulated the protein levels of A_2AR in the ChP (Fig. 8B, C). The injected adenosine also promoted the phosphorylation of SPAK (Fig. 8D, E) and reduced the protein levels of ATP1A2 (Fig. 8F, G). Moreover, the chronic injection of adenosine resulted in enlarged ventricles (Fig. 8H, I) and CSF secretion (Fig. 8J, K). These data support the contention that elevated adenosine levels in the CSF may be a trigger of hydrocephalus. To grasp the source of adenosine, we tested the impact of impeding the extracellular formation of adenosine from adenine nucleotides, which is mediated by CD73. We found that the selective knockdown of CD73 in ChP did not affect ventricle size (Fig. S9), which indicates that the burst of extracellular adenosine in the ChP controlling CSF secretion is unlikely to be derived from ATP release and catabolism.

Fig. 8: ICV-injection of adenosine triggered a hydrocephalus phenotype via activation of A2AR signaling in the choroid plexus (ChP). — **Fig. 8: ICV-injection of adenosine triggered a hydrocephalus phenotype via activation of A_2AR signaling in the choroid plexus (ChP).**

Finally, we tested the impact of adenosine treatment during 14 days in mice that were first subject to a knockdown of A_2AR selectively in the ChP (A_2AR-Kd), as confirmed by the presence of EGFP signals restricted to the ChP (Fig. 8L). DAPI staining showed that the knockdown of A_2AR in ChP prevented the phenotype of enlarged ventricles caused by adenosine (Fig. 8M, N). Conversely, we found that the injection of the selective A_2AR agonist (CGS21680) induced CSF hypersecretion and enlarged ventricles (Fig. S10). This confirms our hypothesis that the elevated adenosine in the CSF triggered a hydrocephalus pathology by aberrant A_2AR signaling in the ChP.

Discussion

The present study establishes that elevated adenosine in the CSF and aberrant A_2AR signaling in the ChP are key triggers of CSF hypersecretion and ventriculomegaly in hydrocephalus. This new adenosine hypothesis of the pathogenesis of hydrocephalus is based on the present observations that: (1) adenosine levels were increased in the CSF of patients with hydrocephalus and in mouse hydrocephalus model, which also displayed an increased A_2AR density in the ChP; (2) the overexpression of A_2AR selectively in the ChP was sufficient to induce a hydrocephalus-like phenotype, characterized by CSF hypersecretion and ventriculomegaly; (3) the targeted knockdown of A_2AR in the ChP mitigated hydrocephalus; (4) likewise, the CSF hypersecretion and ventriculomegaly was ameliorated by systemic administration of the A_2AR antagonist KW6002; (5) A_2AR controlled the expression and density of active forms of proteins critical for CSF production in the ChP, namely phosphorylated SPAK and ATP1A2; (6) ventricular infusion of adenosine during 14 days triggered ChP-A_2AR upregulation, SPAK phosphorylation and ATP1A2 downregulation, leading to CSF hypersecretion and ventricle enlargement, an effect abolished by A_2AR knockdown selectively in the ChP. Overall, this demonstrates that the aberrantly increased adenosine-A_2AR signaling within the CSF-ChP axis is a critical event in hydrocephalus pathology. This key role of aberrant CSF-adenosine and ChP-A_2AR in the control of CSF hypersecretion, allows understanding the previously reported ability of caffeine (non-selective adenosine receptor antagonist) to improve developmental motor activities and reflexes in a mouse model of infant hydrocephalus²⁶ and the efficacy of diuretin, a sodium salicylate form of theobromine, which is equipotent to caffeine at adenosine receptors²⁷, in 4 cases of pediatric hydrocephalus²⁸. However, we are not aware of epidemiological studies linking caffeine consumption and hydrocephalus, which ought to be considered in the future.

The present identification that elevated CSF levels of adenosine and aberrant A_2AR signaling in the ChP are key mediators of hydrocephalus also provides a straightforward rationale to understand how inflammation, hemorrhage, and stroke are major risk factors for the development of hydrocephalus¹. RNA analysis of the CSF of patients with subarachnoid hemorrhage and post-hemorrhagic and autologous blood-induced post-hemorrhage hydrocephalus identified immune cell infiltration, chemokine receptor signaling and secretory activity, and ligand-gated ion channel activity as the main molecular pathways^29,30. Previous studies suggested that activation of Toll-like receptor 4 (TLR4) by proinflammatory cytokines and of the transient receptor potential vanilloid 4 (TRPV4) channel by the serum lipid lysophosphatidic acid could mediate CSF hypersecretion via the NKCC1^4,9,30. In accordance with the involvement of inflammation as a trigger of hydrocephalus, early clinical trials showed that dexamethasone improved survival in young and adults suffering from tuberculous meningitis, presumably by reducing hydrocephalus and preventing infarction, based on the reduction of CSF hypersecretion and intra-cranial pressure³¹. Yet, blood or its component, hemin, does not activate the NKCC1 and NKA activity³⁰, indicating that molecular signals, other than TLR4, may be responsible for sensing inflammatory/hemorrhage insults and activating NKCC1 and CSF hypersecretion. Adenosine emerges as a likely candidate to link not only inflammation but also other conditions, such as stroke or trauma, to the hypersecretion of CSF and hydrocephalus. In fact, adenosine levels are significantly augmented upon brain trauma or inflammation^11,12,13,14 since extracellular adenosine is an important intercellular signal, mediating adaptive responses to stressful conditions, such as upon trauma, infection or inflammation¹⁰. The CSF-hypersecretion and enlarged ventricle induced by autologous blood, overexpression of A_2ARs in ChP and kaolin all led to uniformly elevated adenosine levels in the ChP and CSF. This positions the increased adenosine levels as a key signal to link noxious brain conditions with the flow and composition of the extracellular medium in the brain parenchyma by altering the production of CSF upon activation of A_2ARs in the ChP. This distinctive role of ChP-A_2AR suggests that the A_2AR is not a direct secretary protein like NKCC1, ATP1A2, and SPAK, which demonstrates a polarized localization in the ChP to control CSF production. Instead, we specifically demonstrate that ChP-A_2AR is an indirect controller of CSF secretion through newly unveiled intracellular signaling mechanisms (PI3K/Akt and NF-κB) acting on targets with a polarized localization, namely SPAK and ATP1A2. Thus, A_2ARs are not required to display a polarized localization at the apical side of ChP to control CSF production. It should be noted that A_2AR overactivation leading to hydrocephalus may be a maladaptive attempt to extreme a physiological process of adaptation of CSF flow to metabolic fluctuations of brain function, whereby adenosine generation is intrinsically associated with increased cellular metabolism³² and the physiological activation of A_2AR in the ChP provides a homeostatic increase of CSF production to reset the extracellular medium of the brain parenchyma. This physiological and homeostatic role of adenosine production and A_2AR activation is in agreement with the presently observed decreased CSF production and reduced ventricular volumes upon ChP-specific knockdown of A_2AR and the previously reported ability of acutely administered caffeine (an adenosine receptor antagonist) to decrease CSF production³³ and intracranial pressure in control animals³⁴. This homeostatic role is possibly driven by relatively small fluctuations of the extracellular levels of adenosine and by the low density of A_2AR in the ChP in physiological conditions; in contrast, the large production of adenosine in conditions of inflammation, trauma or ischemic stroke^11,12,13,14 leads to an up-regulation of A_2AR and the over-activation of this adenosine-A_2AR signaling system becomes paramount in generating CSF hyperproduction leading to hydrocephalus. Interestingly, we now observed that the increase of extracellular adenosine triggers A_2AR up-regulation as previously observed in models of Parkinson’s disease³⁵, which might be a general process involved in neurodegeneration whose mechanism remains unclear. Lastly, since adenosine generation is intrinsically associated with cellular metabolism and homeostasis³², the adenosine-mediated control of CSF secretion is expected to be ontogenically controlled in a homeostatic manner, providing new insights into the control of the CSF composition by development, metabolism, circadian rhythms, and hormones. An aspect that remains to be clarified is the source of the adenosine entraining the increased CSF accumulation and ventricular enlargement. Since inflammation, a known trigger of hydrocephalus, involves a robust release of ATP³⁶, one hypothesis would be that ATP-derived adenosine might be the source of the observed A_2AR overactivation leading to hydrocephalus. The lack of the effect of ChP knockdown of CD73 on ventricle size indicates that the surge of extracellular adenosine in the ChP in PHH is unlikely derived from extracellular conversion of ATP into adenosine. An alternative is that the burst of extracellular adenosine controlling CSF secretion might be derived from a direct outflow of adenosine through equilibrated nucleoside transporters from the cells in the ChP. However, both the cellular sources and the mechanism of increased adenosine levels in the ChP leading to hydrocephalus remains to be investigated.

This study also advances prospects of hydrocephalus therapy by identifying aberrant A_2AR signaling as the key trigger of two parallel molecular and signaling mechanisms: (1) a control of the phosphorylation of the SPAK through the recruitment of PI3K-Akt, as previously demonstrated in other cell types, namely in endothelial cells³⁷; (2) a modification of gene expression, namely of ATP1A2, through the control of NF-κB, as previously shown in other cell types³⁸. We now show that there is a simultaneous need of altering the efficiency of ionic transmembrane flow thought the SPAK as well as a decreased energy-driven exchange of cations through ATP1A2 in the ChP to sustain an increased CSF production: in fact, the pharmacological or genetic correction of each of the two pathways operated by aberrant A_2AR signaling is sufficient to prevent the condition of hydrocephalus triggered by A_2AR overexpression. Our current data converge with previous RNA-seq analyses pinpointing ligand-gated ion channel activity as the final common pathway for CSF hypersecretion in hydrocephalus; our findings fully align with previous reports of an increased SPAK phosphorylation and decreased protein levels of Na⁺-K⁺-ATPase or decreased phosphorylated Na⁺-K⁺-ATPase in rats with post-hemorrhagic or post-ischemic hydrocephalus⁹ and also with the postulated role of TLR4-NF-κB signaling and of SPAK kinase in coupling the immune response in the ChP through its interactions with cytokine receptors and ion transport proteins^4,39. A major advance provided by the present study was the identification of the causal role selectively of ATP1A2 in the development of hydrocephalus, rather than other Na⁺-K⁺-ATPase isozymes⁴⁰, all identified in the brush border membrane of the ChP⁴¹ and with different sensitivities to ouabain. This selective A_2AR-ATP1A2 interaction is reminiscent of our previous demonstration that A_2AR signaling in astrocytes regulates glutamate uptake via the regulation of ATP1A2 activity¹⁸. Thus, our finding that ChP-A_2AR controls ATP1A2 prompts the new and exciting possibility of managing hydrocephalus by indirectly and selectively controlling of ATP1A2 amongst the different Na⁺-K⁺-ATPase isozymes with different pharmacological properties and functional efficiencies⁴². Although we demonstrated that A_2AR-mediated changes in molecular signals (p-SPAK and ATP1A2) underlying hydrocephalus, our limited current knowledge of the mechanisms underlying hydrocephalus makes it possible that maybe there might exist are other pathways involved in hydrocephalus in parallel with A_2AR signals. This makes the robust size effect of A_2AR surprising in view of the numerous pathways converging on the control of CSF dynamics and ventricle size. Although we have now provided detailed support for the involvement of an A_2AR-SPAK/ATP1A2 pathway, it is possible that other pathways may be simultaneously affected by A_2AR, such as slc24a4 or slc17a8 and eventually aquaporins. Importantly, some of these pathways may be controlled at the protein rather than at the transcriptomic level as identified for aquaporin-4⁴³, and future work should be aimed at detailing the likely multiple pathways affected by A_2AR to understand its striking and robust of ventricle volume. This effect of A_2AR on ventricle size was not previously noted, probably because the impact of A_2AR blockade in the ChP to decrease ventricle size was larger in physiological conditions than in the models of hydrocephalus, the same occurring in TBI³ and EAE models¹⁶.

Our present proposal of the adenosine hypothesis of hydrocephalus prompts the tempting contention that pharmacologically targeting aberrant A_2AR signaling in ChP may be a novel and superior therapeutic strategy with high translational potential for controlling hydrocephalus. Thus, the systemic administration of NKCC1 inhibitors may preferentially interfere with NKCC1 either peripherally or within the brain vasculature, likely plaguing the efficacy and patient tolerability of NKCC1 inhibitors (such as acetazolamide and furosemide) for hydrocephalus treatment. Indeed, hemorrhage-induced CSF hypersecretion and hydrocephalus are sensitive to intrathecal delivery, but not to systemic exposure, of the SPAK-NKCC1 inhibitor, bumetanide⁹. In this regard, ChP-A_2AR signaling may represent the most promising target for the treatment of hydrocephalus, in view of the localized aberrant increase of adenosine in the CSF of patients with hydrocephalus, in line with the elevated levels of inosine, an adenosine metabolite, in the CSF of infants with hydrocephalus⁴⁴. Notably, we demonstrated that the systemic administration of KW6002 ameliorated the hydrocephalus phenotypes. This is of particular relevance since there is a dire need for effective pharmacotherapy for hydrocephalus, and repurposing the FDA-approved A_2AR antagonist, KW6002⁴⁵, emerges as a new therapeutic option to manage hydrocephalus. Conversely, the benefit of KW6002 in Parkinson’s disease could be dampened by its reduction of CSF dynamics, reducing glymphatic activity and aggregate clearance to bolster PD pathology. This might be counteracted by the parallel effect of KW6002 in the ChP to increase the tightness of the ChP barrier, as shown in a EAE model^16,46, reducing immune cell infiltration and dampening PD pathology. Careful monitoring of the impact of KW6002 on PD disease progression, considering its multiple actions on the ChP, including CSF production and increased integrality of the ChP tight junctions, are warranted to clarify the exact overall effect.

Methods

Study approval

All experiments adhered to the guidelines outlined in the US National Institutes of Health Guide for the Care and Use of Laboratory Animals. The animal experiments were approved by the Animal Experimentation Regulation of Wenzhou Medical University (xmsq2021-0012). Additionally, the collection of CSF samples from patients was approved by the Human Ethics Committee of Ningbo University (KY2023SL185-01). Written informed consent was obtained from all human participants for the use of their samples and data in research. However, specific consent for the publication of potentially identifiable information was not obtained. Therefore, to protect participant privacy, all individual data (such as exact age and the hospital name) have been presented in an aggregated form (age data are presented as mean ± SD) to ensure that no individual participant can be identified.

Human subjects

Human CSF samples were collected with written informed consent from the department of Neurology and Neurosurgery at a single tertiary medical center in Ningbo, China, following ethical approval (KY2023SL185-01). To prevent potential adenosine degradation, the collected human CSF samples were rapidly frozen in liquid nitrogen and subsequently stored at −80 °C. The remaining samples were stored at −80 °C. Due to the small volume of the CSF samples, they are generally not shared.

Quantification of adenosine levels in human CSF using ultra-high performance liquid chromatography (UPLC)

To construct the adenosine standard curve, we initially dissolved adenosine powder (Sigma, A9251-25G) into the mobile phase (50 mmol/L NaH₂PO₄ and methanol in a volume ratio of 7:3) to obtain a series of adenosine solutions with varying concentrations between 0.0625 and 5 μmol/L. We measured the absorbance of these standard samples at 254 nm using a UPLC instrument (Thermo, Accela 1250). The mobile phase flowed at a rate of 400 μL/min, while the column temperature was maintained at 30 °C. Ultimately, we generated a linear standard curve (R² > 0.999).

A double-blind method was used to assess adenosine concentrations in the collected human CSF samples. Following the collation of case data and the assignment of sample numbers, the CSF samples were transferred to the tester. After a preliminary evaluation of all patient CSF samples, the tester divided them into high and low-concentration groups. For the high-concentration group, 20 μL of 0.025% dipyridamole (Sigma, D9766-1G) and 800 μL of iced methanol were added to 200 μL of CSF samples, which were then placed on ice for 15 min. The samples then underwent centrifugation at 15,294 x g for 15 min at 4 °C to remove any debris and were filtered through 0.22 μm sterile polyether sulfone membranes. The samples were dried using nitrogen gas, and the resulting powder was reconstituted in 400 μL of mobile phase (diluted 2×). For the low-concentration group, 60 μL of 0.025% dipyridamole and 2400 μL of iced methanol were added to 600 μL of CSF samples, which were also kept on ice for 15 min. These samples were subjected to the same centrifugation and filtration processes as above. After drying, the powder was redissolved in 300 μL of mobile phase (concentrated 2×). Then, the absorbance of these processed samples was measured after UPLC separation and the data were analyzed using MATLAB software.

Animals

Wild-type (WT) C57BL/6 mice (male and female), aged 6–8 weeks and weighing 22–25 g, were purchased from SPF Biotechnology Co., Ltd. (Beijing, China). The mice were housed in an environment maintained at 25 °C with a standard 12-h light and 12-h dark cycle and were provided with ad libitum access to both food and water.

Isolation of the ChP

The mice were subjected to deep anesthesia using 1% sodium pentobarbital and subsequently perfused with pre-cooled PBS. The entire brain was then carefully extracted and quickly placed on an ice slab. Under a microscope, the ChP was quickly dissected from the lateral, third, and fourth ventricles and then transferred into 1.5 mL tubes. These samples were rapidly frozen using liquid nitrogen and stored in a −80 °C for subsequent experimental procedures.

Immunofluorescence

The coronal brain sections (20 μm thick) were initially rinsed three times with PBS, followed by permeabilization in a solution containing PBS with 0.3% Triton X-100 (Roche Diagnostics GmbH, 11332481001), 1% bovine serum albumin (BSA, from Amresco, 0332-100G) and 5% normal donkey serum (Jackson, 017-000-121) at room temperature for 2 h. Next, the sections were incubated overnight at 4 °C with the following primary antibodies against: A_2AR (Frontier Institute, A2A-Go-Af700, 1:200), p-SPAK (Millipore, 07-2273, 1:300), IBA-1 (Wako, 019-19741, 1:400), CD31 (Bio-techne, AF3628, 1:200), ICAM-1 (Abcam, ab179707, 1:400), MHC-Ⅱ (Invitrogen, 2028-06-19, 1:200), HA-tag (Cell Signaling Technology, #3724, 1:200), ATP1A2 (Proteintech, 16836-1-AP, 1:200), and 488-conjugated His-tag (Proteintech, CL488-66005, 1:100). The sections then underwent three PBS wash cycles before incubation with secondary antibodies coupled to either Alexa Fluor 594 or 555, for 2 h at room temperature (except for the His-tag mAb). Fluorescence images were acquired using a confocal microscope (LSM900, Zeiss), and subsequent analysis of fluorescence intensity was performed with the ImageJ software.

Quantitative real-time PCR (qPCR)

RNA was extracted using the Trizol reagent, and cDNA was synthesized with the cDNA synthesis kit (Vazyme, R323-01). PCR amplification was carried out employing the SYBR Green kit (Takara, RR82LR). The expression levels were normalized to the reference gene (ppia) using the 2^−ΔΔCt method. The primer sequences were displayed in Table S2.

Western blotting

ChP tissues or cells were lysed using RIPA buffer supplemented with protease inhibitors (Bimake, B14001) and a phosphatase inhibitor cocktail (Bimake, B15001). After homogenization, the samples were centrifuged at 10,621 x g for 10 min to remove cellular debris. Protein concentration was measured using the BCA assay kit (Beyotime, P0010). The samples were denatured in 5× sodium dodecyl-sulfate (Beyotime, P0015) by heating at 90 °C for 8 min. Equal amounts of proteins were loaded and separated by electrophoresis on either a 10% gradient polyacrylamide gel (Vazyme, E303-01) or an 8% gradient polyacrylamide gel (Vazyme, E302-01). The gel was then transferred onto a 0.45 µm PVDF membrane (GE Amersham, 10600023) at 300 mA for 1–2 h in ice water. The membrane was then blocked with Tris buffered saline with 0.1% Tween 20 (TBST) and 1% BSA (BBI, A600332-0100) and incubated overnight at 4 °C with primary antibodies against: β-actin (Proteintech, 66009-1-Ig, 1:2000), A_2AR (A2A-Go-Af700, Frontier Institute, 1:200), p-SPAK (Millipore, 07-2273, 1:1000) or ATP1A2 (Invitrogen, PA5-101902, 1:600). Finally, the membrane was incubated with horseradish peroxidase-conjugated secondary antibodies for 90 min at room temperature. The membrane was visualized using a molecular imager (Bio-Rad, ChemiDoc XRS+).

Targeted knockdown or overexpression of genes in the ChP

Mice were anesthetized using 1% pentobarbital sodium, and their top fur was trimmed. Erythromycin eye ointment was applied to prevent light-induced damage. Subsequently, the mice were immobilized in a stereotaxic apparatus, and 2 μL of various AAV2/5 constructs were separately injected into the lateral ventricles at the following coordinates: AP: −0.58; ML: −1.10; DL: −2.20. The injection rate was set at 400 nL/min, and the needle remained in place for an additional 20 min before withdrawal. The AAV2/5 constructs used were as follows: AAV2/5-U6-shRNA (A_2AR)-CMV-EGFP (6.27 × 10¹² vg/mL), AAV2/5-U6-scramble-CMV-EGFP (6.21 × 10¹² vg/mL), AAV2/5-CMV-A_2AR-HA (5.19 × 10¹² vg/mL), AAV2/5-CMV-EGFP (5.54 × 10¹² vg/mL), AAV2/5-U6-shRNA (SPAK)-CMV-EGFP (5.22 × 10¹² vg/mL), AAV2/5-CMV-ATP1A2-His (5.77 × 10¹² vg/mL) and AAV2/5-CMV-mCherry (5.80 × 10¹² vg/mL). For the co-expression of two viruses in the ChP, a mixture of 4 μL of various viruses was injected into the lateral ventricles at the coordinates mentioned above. The combinations of different viruses administered simultaneously were as follows: AAV2/5-CMV-A_2AR-HA and AAV2/5-U6-shRNA (SPAK)-CMV-EGFP, AAV2/5-CMV-A_2AR-HA and AAV2/5-U6-scramble-CMV-EGFP, AAV2/5-CMV-A_2AR-HA and AAV2/5-CMV-ATP1A2-His, AAV2/5-CMV-A_2AR-HA and AAV2/5-CMV-mCherry. The volume ratio of the two viruses in each mixture was 1:1. All mice were allowed a two-week recovery period after viral injection before subsequent experiments.

Model of post-hemorrhagic hydrocephalus (PHH)

The PHH model was performed as previously described^47,48. In brief, 20 μL of autologous blood or sterile PBS were administered into the right lateral ventricle (AP: −0.58; ML: +1.10; DL: −2.20) at a rate of 5 μL/min. Then, the needle was left in place for 10 min and carefully removed at a rate of 1 mm/min. The incision in the skin was then sealed using medical glue (Vetbond, 3M 1469sb), and the mice were placed on a warm heating pad for recovery.

Model of obstructive hydrocephalus

The mice were anesthetized with 1% pentobarbital sodium. Subsequently, a 3 μL injection of either 20% kaolin (Sigma-Aldrich, 1332-58-7) or sterile PBS was administered into the right lateral ventricle (AP: −0.58; ML: +1.10; DL: −2.20) at a rate of 500 nL/min, and the needle was left in place for 7 min before being carefully withdrawn. The incision in the skin was then sealed using medical glue, and the mice were placed on a warm heating pad for the recovery period.

ICV injection of Evans blue dye

Briefly, 3 μL of Evans blue dye (0.5%, Sigma-Aldrich, E2125) were injected into the right lateral ventricle (AP: −0.58; ML: +1.10; DL: −2.20) using a steady rate of 400 nL/min, keeping the Hamilton syringe in place for 30 min before withdrawal. The mice were then placed on a warm heating pad for an additional hour to ensure adequate circulation of the dye within the CSF spaces. Mice were then transcardially perfused, and their brains were harvested and sagitally cut. Images were then captured to assess the patency of CSF circulation.

CSF collection and measurement

CSF was collected from the anesthetized mice by inserting a glass capillary into the cisterna magna and allowing it to sit for 15 min. The length of the glass capillary filled with CSF was promptly measured using a ruler.

Drug administration

KW6002 (Sundia MediTech Company Ltd), a selective A_2AR antagonist, was dissolved in a 1:1 premixture of dimethyl sulfoxide (DMSO) (Sigma-Aldrich, 276855) and corn oil (Thermo Scientific, 8001-30-7). This mixture was further diluted in saline (in a 3:7 volume ratio) to achieve complete dissolution of the KW6002 suspension. Three days after the ICV injection of kaolin, the mice received daily intraperitoneal injections of either the vehicle or KW6002 (5 mg/kg).

Primary culture of ChP epithelial cells

The ChP cell cultures were produced as described in our previous study¹⁶. Briefly, following a perfusion with PBS, the ChPs were dissected, digested with trypsin (0.25%) for 20 min and dissociated by pipetting. The cells were cultured in L-polylysine-coated 24-well plates for 5 days in Dulbecco’s modified Eagle’s medium (DMEM)/Ham’s F12 medium supplemented with 10% fetal bovine serum, 1 mM L-glutamine, 1 mM sodium pyruvate, penicillin/streptomycin (100 U/mL), insulin (5 ng/mL), 20 μM arabinosylcytosine, sodium selenite (5 ng/mL) and epidermal growth factor (10 ng/mL).

Over-expression of A_2AR in the primary ChP epithelial cells

In brief, the primary ChP epithelial cells were transfected with the constructed lentivirus (CMV-A_2AR-3Flag-Ubi-ZsGreen-IRES-Puromycin). After 48 h, the cells were screened with puromycin (1 μg/mL, 2.5 μg/mL). The ZsGreen-positive cells were then used for follow-up assays.

Effects of PI3K/Akt or NF-kB inhibitors

In brief, the primary ChP epithelial cells, stably transfected to overexpress A_2AR, were treated with either JSH-23 (10 μM, a NF-κB transcriptional activity inhibitor, Sigma-Aldrich, 749886-87-1), MK-2206 (1 μM, a selective Akt inhibitor, MedChemExpress, HY-108232), GDC-0941 (0.5 μM, a selective PI3K inhibitor, MedChemExpress, HY-50094) or vehicle. After 48 h, the cells were collected for Western blot analysis.

Whole-genome RNA sequencing (RNA-seq)

Illumina library construction adhered to the manufacturer’s instructions (Illumina, USA) at ApexBio Technology LLC (Shanghai, China). Subsequently, reads were aligned to the mouse genome (mm10) using STAR version 2.5. The aligned reads were quantified using the Homer software (analyze repeats). Differentially expressed genes were identified by DEseq2⁴⁹ with five replications for each sample. A cut-off criterion of log₂|fold-change|>1.2 with a P < 0.05 was applied. The false discovery rate was obtained by correcting the P value of the hypothesis testing results of differentially expressed genes using the Benjamin Hochberg method. To analyze gene ontology and KEGG pathway categories, ClusterProfiler⁵⁰ was employed.

ICV-infusion of adenosine

A dosing administration device (Kedou Brain-Computer Technology) was implanted into the right lateral ventricle and immobilized on the surface of mouse sculp by the instant adhesive gel. One week later, adenosine solution (80 μM) was prepared with 0.5% methyl cellulose (Sigma-Aldrich, M0512-100G) and filtered through the 0.22 μm sterile PES membrane (JETBIOFIL, 230701-052-A). CGS21680 (80 μM, ab12053) was prepared with dimethyl sulfoxide (Beyotime, AT038-500ml) and filtered through the 0.22 μm sterile PES membrane. Adenosine or CGS21680 was immediately infused into the mouse lateral ventricle via the implanted cannula (Kedou Brain-Computer Technology) connected to the immobilized sculp device by the syringe pump (R462, RWD) at the rate of 0.4 μL/min for 5 min. A similar infusion was performed 4 times a day for 2 weeks.

Magnetic resonance imaging (MRI)

Mice were subjected to imaging using a small animal MRI system (Bruker BioSpec 94/20 USR, 9.4 T) after being anesthetized with isoflurane. To maintain body temperature, a warm pad was utilized. All axial images were obtained under the following conditions: TE/TR = 33/2500; Ave = 4; RARE = 8; slice thickness = 0.3 mm. Brain sizes and ventricle volumes were calculated through manual segmentation with the assistance of ImageJ or 3Dslicer software. It is worth noting that ventricle areas were not excluded during the measurement of brain sizes.

Statistical analysis

All statistical analyses were performed with GraphPad Prism 9.0. The data are screened for normality using the Kolmogorov-Smirnov test and are presented as mean ± s.e.m. Two-tailed Mann–Whitney U test was used to compare the data of adenosine levels in hydrocephalus patients and controls, as the data sets did not follow normal distribution (the data are presented as median and interquartile range). Two-tailed Student’s t test was used in other data. No adjustments for multiple comparisons were applied in the initial analysis.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

RNA-seq datasets were deposited into BioProject database under number PRJNA1353615. All sequencing data is publicly available. Source data are provided with this paper.

References

Bramall, A. N., Anton, E. S., Kahle, K. T. & Fecci, P. E. Navigating the ventricles: novel insights into the pathogenesis of hydrocephalus. EBioMedicine 78, 103931 (2022).
PubMed PubMed Central Google Scholar
Saunders, N. R., Dziegielewska, K. M., Fame, R. M., Lehtinen, M. K. & Liddelow, S. A. The choroid plexus: a missing link in our understanding of brain development and function. Physiol. Rev. 103, 919–956 (2023).
PubMed Google Scholar
Zhang, Z. et al. NLRP3 inflammasome-mediated choroid plexus hypersecretion contributes to hydrocephalus after intraventricular hemorrhage via phosphorylated NKCC1 channels. J. Neuroinflammation 19, 163 (2022).
PubMed PubMed Central Google Scholar
Robert, S. M. et al. The choroid plexus links innate immunity to CSF dysregulation in hydrocephalus. Cell 186, 764–785.e721 (2023).
PubMed PubMed Central Google Scholar
Karimy, J. K. et al. Inflammation in acquired hydrocephalus: pathogenic mechanisms and therapeutic targets. Nat. Rev. Neurol. 16, 285–296 (2020).
PubMed PubMed Central Google Scholar
Bothwell, S. W., Janigro, D. & Patabendige, A. Cerebrospinal fluid dynamics and intracranial pressure elevation in neurological diseases. Fluids Barriers CNS 16, 9 (2019).
PubMed PubMed Central Google Scholar
Steffensen, A. B. et al. Cotransporter-mediated water transport underlying cerebrospinal fluid formation. Nat. Commun. 9, 2167 (2018).
ADS PubMed PubMed Central Google Scholar
Xu, H. et al. Choroid plexus NKCC1 mediates cerebrospinal fluid clearance during mouse early postnatal development. Nat. Commun. 12, 447 (2021).
ADS PubMed PubMed Central Google Scholar
Karimy, J. K. et al. Inflammation-dependent cerebrospinal fluid hypersecretion by the choroid plexus epithelium in posthemorrhagic hydrocephalus. Nat. Med. 23, 997–1003 (2017).
PubMed Google Scholar
Chen, J. F., Eltzschig, H. K. & Fredholm, B. B. Adenosine receptors as drug targets-what are the challenges? Nat. Rev. Drug Discov. 12, 265–286 (2013).
PubMed PubMed Central Google Scholar
Park, T. S., Van Wylen, D. G., Rubio, R. & Berne, R. M. Increased brain interstitial fluid adenosine concentration during hypoxia in newborn piglet. J. Cereb. Blood Flow. Metab. 7, 178–183 (1987).
PubMed Google Scholar
Phillis, J. W., Walter, G. A. & Simpson, R. E. Brain adenosine and transmitter amino acid release from the ischemic rat cerebral cortex: effects of the adenosine deaminase inhibitor deoxycoformycin. J. Neurochem. 56, 644–650 (1991).
PubMed Google Scholar
Headrick, J. P., Bendall, M. R., Faden, A. I. & Vink, R. Dissociation of adenosine levels from bioenergetic state in experimental brain trauma: potential role in secondary injury. J. Cereb. Blood Flow. Metab. 14, 853–861 (1994).
PubMed Google Scholar
Yoshida, Y. et al. Adenosine and neopterin levels in cerebrospinal fluid of patients with neurological disorders. Intern. Med. 38, 133–139 (1999).
PubMed Google Scholar
Cunha, R. A. How does adenosine control neuronal dysfunction and neurodegeneration? J. Neurochem. 139, 1019–1055 (2016).
PubMed Google Scholar
Zheng, W. et al. Choroid plexus-selective inactivation of adenosine A(2A) receptors protects against T cell infiltration and experimental autoimmune encephalomyelitis. J. Neuroinflammation 19, 52 (2022).
PubMed PubMed Central Google Scholar
Mills, J. H., Alabanza, L. M., Mahamed, D. A. & Bynoe, M. S. Extracellular adenosine signaling induces CX3CL1 expression in the brain to promote experimental autoimmune encephalomyelitis. J. Neuroinflammation 9, 193 (2012).
PubMed PubMed Central Google Scholar
Matos, M., Augusto, E., Agostinho, P., Cunha, R. A. & Chen, J. F. Antagonistic interaction between adenosine A2A receptors and Na+/K+-ATPase-α2 controlling glutamate uptake in astrocytes. J. Neurosci. 33, 18492–18502 (2013).
PubMed PubMed Central Google Scholar
Brown, P. D., Davies, S. L., Speake, T. & Millar, I. D. Molecular mechanisms of cerebrospinal fluid production. Neuroscience 129, 957–970 (2004).
PubMed Google Scholar
Sitkovsky, M. V. et al. Physiological control of immune response and inflammatory tissue damage by hypoxia-inducible factors and adenosine A2A receptors. Annu. Rev. Immunol. 22, 657–682 (2004).
PubMed Google Scholar
Bucheimer, R. E. & Linden, J. Purinergic regulation of epithelial transport. J. Physiol. 555, 311–321 (2004).
PubMed Google Scholar
Ye, M. et al. Adenosine A(2A) receptor controls the gateway of the choroid plexus. Purinergic Signal. 19, 135–144 (2023).
PubMed Google Scholar
Nishida, H. et al. Phosphatidylinositol 3-kinase/Akt signaling pathway activates the WNK-OSR1/SPAK-NCC phosphorylation cascade in hyperinsulinemic db/db mice. Hypertension 60, 981–990 (2012).
PubMed Google Scholar
Shi, L. et al. Adenosine interaction with adenosine receptor A2a promotes gastric cancer metastasis by enhancing PI3K-AKT-mTOR signaling. Mol. Biol. Cell 30, 2527–2534 (2019).
PubMed PubMed Central Google Scholar
Mori, Y., Higuchi, M., Masuyama, N. & Gotoh, Y. Adenosine A2A receptor facilitates calcium-dependent protein secretion through the activation of protein kinase A and phosphatidylinositol-3 kinase in PC12 cells. Cell Struct. Funct. 29, 101–110 (2004).
PubMed Google Scholar
Olopade, F., Femi-Akinlosotu, O., Ibitoye, C. & Shokunbi, T. Probing caffeine administration as a medical management for hydrocephalus: an experimental study. Pediatr. Neurol. 135, 12–21 (2022).
PubMed Google Scholar
Valada, P., Alçada-Morais, S., Cunha, R. A. & Lopes, J. P. Thebromine targets adenosine receptors to control hippocampal neuronal function and damage. Int. J. Mol. Sci. 23, 10510 (2022).
PETERMAN, M. G. THE KETOGENIC DIET IN THE TREATMENT OF EPILEPSY: A PRELIMINARY REPORT. Am. J. Dis. Child. 28, 28–33 (1924).
Google Scholar
Lolansen, S. D. et al. Inflammatory markers in cerebrospinal fluid from patients with hydrocephalus: a systematic literature review. Dis. Markers 2021, 8834822 (2021).
PubMed PubMed Central Google Scholar
Lolansen, S. D. et al. Posthemorrhagic hydrocephalus associates with elevated inflammation and CSF hypersecretion via activation of choroidal transporters. Fluids Barriers CNS 19, 62 (2022).
PubMed PubMed Central Google Scholar
Thwaites, G. E. et al. Dexamethasone for the treatment of tuberculous meningitis in adolescents and adults. N. Engl. J. Med. 351, 1741–1751 (2004).
PubMed Google Scholar
Cunha, R. A. Adenosine as a neuromodulator and as a homeostatic regulator in the nervous system: different roles, different sources and different receptors. Neurochem. Int. 38, 107–125 (2001).
PubMed Google Scholar
Han, M. E. et al. Regulation of cerebrospinal fluid production by caffeine consumption. BMC Neurosci. 10, 110 (2009).
PubMed PubMed Central Google Scholar
Israelsen, I. M. E., Westgate, C. S. J., Kamp-Jensen, C., Jensen, R. H. & Eftekhari, S. Effects of caffeine on intracranial pressure and pain perception in freely moving rats. Headache 63, 1220–1231 (2023).
PubMed Google Scholar
Meng, F. et al. CD73-derived adenosine controls inflammation and neurodegeneration by modulating dopamine signalling. Brain 142, 700–718 (2019).
PubMed Google Scholar
Di Virgilio, F., Vultaggio-Poma, V., Falzoni, S. & Giuliani, A. L. Extracellular ATP: a powerful inflammatory mediator in the central nervous system. Neuropharmacology 224, 109333 (2023).
PubMed Google Scholar
Ahmad, A., Schaack, J. B., White, C. W. & Ahmad, S. Adenosine A2A receptor-dependent proliferation of pulmonary endothelial cells is mediated through calcium mobilization, PI3-kinase and ERK1/2 pathways. Biochem. Biophys. Res. Commun. 434, 566–571 (2013).
ADS PubMed PubMed Central Google Scholar
Kermanian, F., Soleimani, M., Ebrahimzadeh, A., Haghir, H. & Mehdizadeh, M. Effects of adenosine A2a receptor agonist and antagonist on hippocampal nuclear factor-kB expression preceded by MDMA toxicity. Metab. Brain Dis. 28, 45–52 (2013).
PubMed Google Scholar
Yan, Y. & Merlin, D. Ste20-related proline/alanine-rich kinase: a novel regulator of intestinal inflammation. World J. Gastroenterol. 14, 6115–6121 (2008).
PubMed PubMed Central Google Scholar
Blanco, G. & Mercer, R. W. Isozymes of the Na-K-ATPase: heterogeneity in structure, diversity in function. Am. J. Physiol. 275, F633–F650 (1998).
PubMed Google Scholar
Baasch Christensen, I. et al. Multiple Na, K-ATPase subunits colocalize in the brush border of mouse choroid plexus epithelial cells. Int. J. Mol. Sci. 22, 1569 (2021).
Krivoĭ, I. I. [Regulatory function of the Na,K-ATPase alpha2 isoform]. Biofizika 57, 771–788 (2012).
PubMed Google Scholar
Dias, L. et al. Impact of glucocorticoid-associated stress-like conditions on aquaporin-4 in cultured astrocytes and its modulation by adenosine A(2A) receptors. J. Neurochem. 169, e16299 (2025).
PubMed Google Scholar
Schmidt, H., Siems, W. G., Grune, T. & Grauel, E. L. Concentration of purine compounds in the cerebrospinal fluid of infants suffering from sepsis, convulsions and hydrocephalus. J. Perinat. Med. 23, 167–174 (1995).
PubMed Google Scholar
Chen, J. F. & Cunha, R. A. The belated US FDA approval of the adenosine A(2A) receptor antagonist istradefylline for treatment of Parkinson’s disease. Purinergic Signal. 16, 167–174 (2020).
PubMed PubMed Central Google Scholar
Qi, C. et al. Adenosine A(2A) receptor-mediated interactions between Th1(+) T cells and the choroid plexus epithelium via IFN-gamma signalling control T-Cell infiltration in experimental autoimmune encephalomyelitis. Cell Commun. Signal 23, 94 (2025).
PubMed PubMed Central Google Scholar
Zhu, W. et al. Changes in motor function, cognition, and emotion-related behavior after right hemispheric intracerebral hemorrhage in various brain regions of mouse. Brain Behav. Immun. 69, 568–581 (2018).
PubMed PubMed Central Google Scholar
Shen, D. et al. Ferroptosis in oligodendrocyte progenitor cells mediates white matter injury after hemorrhagic stroke. Cell Death Dis. 13, 259 (2022).
PubMed PubMed Central Google Scholar
Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550 (2014).
PubMed PubMed Central Google Scholar
Wu, T. et al. clusterProfiler 4.0: a universal enrichment tool for interpreting omics data. Innovation 2, 100141 (2021).
PubMed PubMed Central Google Scholar

Download references

Acknowledgements

This research received financial support from the following sources: the Research Fund for International Senior Scientists (Grant No. 82150710558 to J.F.C.), the National Natural Science Foundation of China (Grant No. 31800903 to W.Z.), the Zhejiang Provincial Natural Science Foundation of China (Grant No. LMS25H090008 to W.Z.), the Science & Technology Initiative STI2030-Major Projects (Grant No. 2021ZD0203400 to J.F.C.), the Department of Science and Technology of Zhejiang Province (Grant No. 2023ZY1011 to W.Z.), the Ningbo Major Research and Development Plan Program (Grant No. 2023Z196 to W.Z.), the Ningbo Medical Science and Technology Project (Grant No.2023Y03 to Z.Y.L.) and the Ningbo Medical and Health Leading Academic Discipline Project (Grant No. 2022‑F04 to H.C.W.).

Author information

These authors contributed equally: Wu Zheng, Lanxin Hu, Yuwen Yang.

Authors and Affiliations

The Eye-Brain Research Center and Molecular Neuropharmacological Lab, State Key Laboratory of Ophthalmology, Optometry and Visual Science, Eye Hospital, Wenzhou Medical University, Wenzhou, Zhejiang, China
Wu Zheng, Lanxin Hu, Yuwen Yang, Xuhang Li, Wangchao Chen, Yiwei Jiang, Yuhan Chen, Cheng Zheng & Jiang-Fan Chen
Department of Neurology, the Affiliated Lihuili Hospital, Ningbo University, Ningbo, Zhejiang, China
Wu Zheng & Zhongyue Lv
Oujiang Laboratory (Zhejiang Laboratory for Regenerative Medicine, Vision and Brain Health), School of Ophthalmology & Optometry and Eye Hospital, Wenzhou Medical University, Wenzhou, Zhejiang, China
Lanxin Hu, Yuwen Yang & Jiang-Fan Chen
Key Laboratory for Laboratory Medicine, Ministry of Education, Zhejiang Provincial Key Laboratory of Medical Genetics, School of Laboratory Medicine and Life Science, Wenzhou Medical University, Wenzhou, Zhejiang, China
Jia Wu
Department of Neurosurgery, Ningbo Medical Center Lihuili Hospital, Ningbo University, Ningbo, Zhejiang, China
Hongcai Wang, Shiwei Li & Kuan Feng
The Department of Neurology, Daping Hospital, The Arm Medical University, Chongqing, China
Yanjiang Wang
CNC-Center for Neuroscience and Cell Biology, Faculty of Medicine, University of Coimbra, Coimbra, Portugal
Rodrigo A. Cunha

Authors

Wu Zheng
View author publications
Search author on:PubMed Google Scholar
Lanxin Hu
View author publications
Search author on:PubMed Google Scholar
Yuwen Yang
View author publications
Search author on:PubMed Google Scholar
Xuhang Li
View author publications
Search author on:PubMed Google Scholar
Jia Wu
View author publications
Search author on:PubMed Google Scholar
Wangchao Chen
View author publications
Search author on:PubMed Google Scholar
Yiwei Jiang
View author publications
Search author on:PubMed Google Scholar
Yuhan Chen
View author publications
Search author on:PubMed Google Scholar
Hongcai Wang
View author publications
Search author on:PubMed Google Scholar
Shiwei Li
View author publications
Search author on:PubMed Google Scholar
Kuan Feng
View author publications
Search author on:PubMed Google Scholar
Cheng Zheng
View author publications
Search author on:PubMed Google Scholar
Yanjiang Wang
View author publications
Search author on:PubMed Google Scholar
Rodrigo A. Cunha
View author publications
Search author on:PubMed Google Scholar
Zhongyue Lv
View author publications
Search author on:PubMed Google Scholar
Jiang-Fan Chen
View author publications
Search author on:PubMed Google Scholar

Contributions

W.Z. and J.F.C. designed and conceived the research; W.Z., L.X.H., Y.W.Y., X.H.L., W.C.C. and Y.W.J. prepared the tables and figures; W.Z., L.X.H., Y.W.Y., X.H.L., J.W., W.C.C., Y.W.J., Y.H.C. and C.Z. performed the experiments and analysis; Z.Y.L., H.C.W., Y.J.W., S.W.L. and K.F. collected the CSF from hydrocephalus patients and controls. W.Z., R.A.C. and J.F.C. wrote and revised the manuscript.

Corresponding authors

Correspondence to Wu Zheng, Zhongyue Lv or Jiang-Fan Chen.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Miriam Echevarria and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

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

Supplementary information

Supplementary Information (download PDF )

Reporting Summary (download PDF )

Transparent Peer Review file (download PDF )

Source data

Source Data (download XLSX )

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, 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 you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. 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-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Zheng, W., Hu, L., Yang, Y. et al. Aberrant adenosine A_2A receptor signaling in the choroid plexus drives CSF hypersecretion and ventriculomegaly in hydrocephalus. Nat Commun 17, 83 (2026). https://doi.org/10.1038/s41467-025-66776-8

Download citation

Received: 24 February 2025
Accepted: 14 November 2025
Published: 07 December 2025
Version of record: 05 January 2026
DOI: https://doi.org/10.1038/s41467-025-66776-8