Intracranial inflammation and meningeal fibrosis are associated with perivascular changes, altered CSF tracer dynamics, and cognitive decline in a rat model of communicating hydrocephalus

Abstract

Background

Normal pressure hydrocephalus (NPH), a chronic communicating hydrocephalus characterized by ventriculomegaly and progressive cognitive decline, is increasingly prevalent in today’s rapidly expanding aging population. Although cerebrospinal fluid (CSF) diversion ameliorates symptoms, early diagnosis and effective intervention remain challenging due to incomplete understanding of the underlying pathophysiology. Disturbances in CSF circulation and perivascular transport have been implicated in NPH; however, their relationships to meningeal alterations and neurodegenerative changes remain to be clarified.

Methods

Chronic communicating hydrocephalus was induced in adult rats with parietal subarachnoid kaolin injection. Intracranial inflammation, meningeal alterations, perivascular changes, CSF tracer elimination dynamics, ventriculomegaly, and cognitive behaviors were examined two weeks, one month, and three months following kaolin treatment.

Results

Ventriculomegaly appeared late, three months following parietal subarachnoid kaolin injection, consistent with the characteristic of chronic communicating hydrocephalus. Interestingly, a series of intracranial pathological changes occurred as early as two weeks after kaolin induction, including meningeal thickening with increased collagen deposition, elevated interleukin-6 and CD163 expressions, astrocyte activation, expansion of perivascular spaces, downregulation of occludin, and increased and mislocalized aquaporin 4 expression. By one month, CSF tracer elimination was significantly reduced, accompanied by decreased expression of collagen IV, a key basal lamina component, and excitatory postsynaptic proteins postsynaptic density protein 95 and spinophilin. At three months, myelin basic protein expression was reduced, while α‑synuclein, an indicator of neurodegeneration, accumulated within cortical pyramidal neurons. Although recognition memory was retained up to three months following kaolin injection, spatial memory declined progressively from one to three months post-treatment.

Conclusions

This study demonstrates that in adult rats, parietal subarachnoid kaolin injection induces meningeal inflammation, cerebral pathological changes, altered CSF dynamics and ultimately communicating hydrocephalus with selective spatial memory impairment. This temporal sequence of events may provide insight into the pathological processes relevant to human NPH and inform future studies on CSF–perivascular transport alterations.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12987-026-00787-5.

Background

Normal pressure hydrocephalus (NPH) is a chronic neurological disorder characterized by ventriculomegaly and a gradual decline in cognitive and motor functions [1]. Although its clinical manifestations are well described, the underlying pathophysiology remains far from being understood. Both mechanical factors, such as ventricular enlargement-resulted brain mechanical distortion, and biochemical disturbances, such as impaired metabolic waste clearance, have been implicated in the functional deterioration associated with NPH. Unlike Alzheimer’s disease, NPH is distinctive in that the associated cognitive deficits can be partially reversed with CSF diversion surgery [2], underscoring the importance of early and accurate diagnosis. However, the lack of mechanistic insight continues to limit diagnostic precision and the advances of therapeutic strategies.

Increasing evidence suggests that impaired CSF absorption, rather than overproduction, plays a pivotal role in NPH development [3]. NPH is thus considered a form of adult chronic communicating hydrocephalus, in which CSF outflow is compromised after leaving the ventricles [4]. Current understanding of CSF dynamics has moved beyond the traditional concept that CSF is produced mainly by the choroid plexus, flows into the subarachnoid space, and discharges through arachnoid granulations. Nowadays, CSF and interstitial fluid are believed to leave the cranial compartment through multiple complementary pathways, including arachnoid villi, meningeal lymphatic vessels, perineural routes along cranial nerves and spinal nerve roots, and drainage across the cribriform plate into nasal lymphatics [5, 6]. Multiple CSF outflow routes are believed to contribute to human CSF circulation and drainage [5]. Perivascular pathways have been proposed as routes through which CSF may enter periarterial spaces, interact with interstitial compartments, and reach meningeal lymphatic vessels [7]. However, the proportion of brain parenchymal fluid that drains via such pathways directly into lymphatics without re-entering ventricular or subarachnoid CSF compartments remains uncertain. Aquaporin 4, a water channel localized to astrocytic endfeet at the brain parenchyma–vessel interface, is critical for CSF–interstitial fluid exchange, and its mislocalization disrupts brain waste clearance [8, 9]. Since many CSF and interstitial fluid outflow structures, such as arachnoid villi, meningeal lymphatics, and perivascular pathways, are embedded in or adjacent to the meninges, meningeal alterations such as fibrosis may influence CSF and perivascular fluid clearance. Consistent with this, imaging studies have reported perivascular structural alterations and altered CSF dynamics in patients with idiopathic NPH [10].

Meningeal alteration has been increasingly recognized as a contributor to chronic hydrocephalus. Fibrotic thickening of the dura and arachnoid mater have been documented in both experimental animals and human clinical studies following subarachnoid hemorrhage [11, 12]. Post-hemorrhagic meningeal fibrosis has been reported to be associated with impaired CSF absorption [12]. Aging, another major risk factor of idiopathic NPH, is accompanied by arachnoid hyperplasia, increased fibrosis, collagen disorganization, and up to 30% thickening of meninges [13, 14]. In a cynomolgus monkey model of experimentally induced subarachnoid hemorrhage, meningeal fibrosis, impaired CSF absorption, and development of communicating hydrocephalus were observed concurrently [15]. Taken together, these observations support the notion that meningeal alterations play a significant role in the development of communicating hydrocephalus. However, the relationships between meningeal pathology, altered CSF dynamics, and neurodegeneration remain to be resolved.

In this study, we used a rat model of kaolin-induced chronic communicating hydrocephalus that has been shown to recapitulate key pathological features of NPH [16, 17]. Although ventriculomegaly has been characterized in this model, the associated meningeal and perivascular alterations remain to be elucidated. Using this model, we explored the temporal relationships between meningeal and perivascular alterations, CSF flow dynamics changes, synaptic and neurodegenerative changes, and behavioral alterations during the development of ventriculomegaly. Intracranial inflammation, meningeal fibrosis, and associated glial activation were assessed by examining alterations of interleukin-6 (IL-6, a pro-inflammatory cytokine), glial fibrillary acidic protein (GFAP, an astrocytic marker), and ionized calcium-binding adapter molecule 1 (Iba1, a microglial marker). Perivascular and vascular alterations were evaluated by assessing changes in aquaporin 4 expression and localization, periarterial space dimension, Evans blue tracer elimination, cluster of differentiation 163 (CD163, a marker of perivascular macrophages) expression, and the expressions of blood–brain barrier proteins occludin (a tight-junction protein) and collagen IV (a basement membrane protein). Synaptic and neurodegenerative alterations were investigated by examining postsynaptic density protein 95 (PSD95) and spinophilin (both postsynaptic scaffolding proteins), myelin basic protein (MBP, an axonal myelination marker), and α-synuclein expressions. Finally, cognitive functions, novel object recognition and single-day Morris water maze performance, were also assessed. Paralleled examination of these parameters during the development of ventriculomegaly might shed light on the pathophysiological changes underlying the development of NPH.

Methods

Animal preparation

A total of 109 male Sprague-Dawley rats (8–12 weeks old; LASCO, Ilan, Taiwan) were used. Animals were housed and treated in accordance with the guidelines of the National Laboratory Animal Center. The experimental protocol was approved by the Animal Care and Use Committee of Tzu Chi University (approval numbers 111019 and 111085). Among these, 43 rats were used for histochemical, immunofluorescence, and immunohistochemical studies; 24 rats for intracisternal injection of Evans blue to assess CSF tracer elimination dynamics; and 42 rats for Western blot analyses.

Induction of chronic communicating hydrocephalus

Chronic communicating hydrocephalus was induced by kaolin injection into the subarachnoid space overlying the parietal cortex with a protocol modified from those of Jusué-Torres et al. [16] and Li et al. [17]. Briefly, rats were anesthetized via intraperitoneal injection of tiletamine and zolazepam (Zoletil 50^®; Virbac, Carros, France; 30 mg/kg) and xylazine (Rompun^®; Bayer, Leverkusen, Germany; 10 mg/kg) before being mounted on a stereotaxic frame. A midline sagittal scalp incision was made, followed by a 2 mm diameter craniotomy over each cerebral hemisphere. The right craniotomy was 2.7 mm lateral to the sagittal suture and 4.55 mm posterior to bregma, while the left was 2.7 mm lateral to the sagittal suture and 2.35 mm posterior to bregma. To avoid brain parenchymal injury during kaolin injection, a shallow groove on the skull tangential to its surface was made posterior to the injection site to rest and guide the injection needle so that the needle remained steadily parallel to the cortical surface throughout the procedure. After exposing the dura, a small perforation was made on the dura to insert the tip of the injection needle. A total of 40 µl of sterilized kaolin suspension (250 mg/ml in 0.9% saline) was infused at 0.5 µl/sec using a 1-ml syringe with a 25-gauge needle mounted on an infusion pump (Legato 130, KD Scientific Inc., Holliston, MA, USA) into the subarachnoid space of the cerebral convexity in both hemispheres. A slow infusion rate was chosen to prevent reflux or leakage of the injected kaolin suspension. No visible cerebrospinal fluid leakage was observed during injection or wound closure afterward. Animals were then divided into three groups and allowed to survive for two weeks, one month, and three months, respectively. Sham-operated controls of each survival group received an equivalent amount of sterile saline injection.

Cognitive tests

Novel object recognition test

Novel object recognition test was conducted following the protocol of Reger et al. [18]. The testing chamber consisted of a black, opaque plastic arena (50 cm × 50 cm × 50 cm) placed in a dimly lit room. Behavioral activity and movement trajectories were recorded from above and analyzed using EthoVision 12 software (Noldus, Wageningen, Netherlands). The test consisted of two phases: habituation and testing. The habituation phase began six days before sacrifice and lasted four consecutive days, during which rats freely explored the empty arena for ten minutes, twice daily, with a 15-minute inter-trial rest in their home cages. On the fifth day, each rat was allowed 10-minute exploration of the empty arena, followed by a 15-minute rest in their home cages and then proceeded to the testing phase. The test included a familiarization trial followed by a test trial. During the familiarization trial, the rat was placed in the arena containing two identical objects and released with its back facing the objects. The rat was allowed to explore freely for five minutes before being returned to its home cage for a 15-minute rest. The two identical objects were then cleaned with 75% alcohol to remove any residual odors before conducting the test trial. The test trial followed the familiarization trial procedure, except that one of the familiar objects was replaced with a novel one. The location of the novel object was randomized. The recognition index (novel object exploration time / total exploration time of both objects) was calculated to assess object recognition.

Single-day Morris water maze test

Single-day Morris water maze test was conducted the day before sacrifice to assess spatial learning and memory. To confirm the rats’ ability to learn to swim toward a cued goal, a cued learning test was performed as a control one to seven days before kaolin injection.

The cued learning test was performed following that described by Vorhees and Williams [19]. A circular black plastic pool (180 cm in diameter) filled with water to a depth of 29 cm was placed in a room surrounded by curtains to minimize external distractions. A transparent acrylic platform (10 cm × 10 cm × 28 cm) was placed in the pool 1.0 cm below the water surface, with a flag extending 11 cm above the water. The test was conducted in two blocks in a single day, with a 30-minute rest interval between blocks. Each block consisted of four trials, during which the rat was placed at the edge of the pool, facing the center. In Block I, the starting positions and platform locations were paired as follows: north-southeast, east-northeast, south-southwest, and west-southeast. In Block II, the pairings were: south-northeast, north-northwest, west-northeast, and east-southeast. During each trial, the rat was allowed to swim freely to locate the submerged platform marked by the flag. The rat was gently guided to the platform if failed to locate it in two minutes. The rat was allowed to stay for 20 s on the platform after reaching it. Escape times from the four trials were averaged to determine the escape time for each block.

The single-day Morris water maze test was conducted as previously described [20]. The test was performed in the same circular pool as above but featured different visual cues positioned at the four cardinal directions around the pool. A hidden platform (10 cm × 10 cm × 28 cm) submerged 1.0 cm below the water surface without a flag was fixed at the center of southwest quadrant. The test consisted of three blocks conducted in a single day, with two-hour rest interval between blocks. Each block consisted of four trials, in which the rat was placed in the center of the pool, facing a different direction in each trial, and allowed to swim freely to locate the submerged platform. The rat was allowed to stay on the platform for 30 s after successful localization before the next trial. If it failed to find the platform in 60 s, the rats were gently guided to it and allowed to rest on the platform for 60 s before proceeding to the next trial. The escape times from the four trials were averaged to determine the escape time of each block of tests. The swimming trajectory was recorded and analyzed with EthoVision 12 software.

Histology

To assess the severity of ventriculomegaly, rats were anesthetized as described above and transcardially perfused with 4% paraformaldehyde in 0.1 M phosphate buffer (PB, pH 7.6) for 30 min. The brain and the overlying dura-arachnoid mater were carefully dissected, post-fixed in the same fixative for 4 h, and subsequently cryoprotected in 30% sucrose in 0.1 M PB. Coronal brain sections, 100 μm thickness, obtained at approximately − 0.26 mm relative to bregma were stained with Nissl’s stain and the Evans index was measured. Use of thick sections minimized disproportional tissue distortion in the process of mounting. The Evans index, a dimensionless measure, was defined as the ratio of the maximum ventricular diameter to the maximum brain diameter on the stained section.

To evaluate meningeal responses following kaolin injection, the dura-arachnoid mater was examined after Masson’s trichrome staining. After cryoprotection, the dura-arachnoid mater was sectioned into 16-µm-thick sagittal sections and processed with Masson’s trichrome stain kit (ab150686, Abcam, Cambridge, UK, USA) following manufacturer’s instructions. The maximum thickness of the dura-arachnoid mater of each rat was measured.

Western blot analysis

At the end of the survival, rats were deeply anesthetized as above and decapitated. The cerebral cortex, spanning approximately between bregma 1.0 mm to 5.0 mm, was dissected, flash frozen in liquid nitrogen, and stored at − 80 °C until analysis. Before analysis, tissue was homogenized at 4 °C in extraction buffer supplemented with protease inhibitors (#78510, Thermo Fisher Scientific, Rockford, IL, USA). The homogenates were centrifuged at 10,000×g for 10 min at 4 °C, and the supernatants were collected. Protein concentration was determined using BCA protein assay reagents (#23225, Thermo Fisher Scientific). Equal amounts of protein were separated by 8–15% SDS–polyacrylamide gel electrophoresis and transferred onto a polyvinylidene difluoride membrane (Bio-Rad, Hercules, CA, USA). The following primary antibodies were used to identify respective proteins: mouse anti-glial fibrillary acidic protein (GFAP) (MAB360, Merck Millipore, Billerica, MA, USA), goat anti-ionized calcium-binding adapter molecule 1 (Iba1) (ab5076, Abcam), rabbit anti-aquaporin 4 (ab259318, Abcam), mouse anti-interleukin-6 (IL-6) (ab9324, Abcam), rabbit anti-CD163 (ab182422, Abcam), rabbit anti-occludin (#71-1500, Thermo Fisher Scientific), rabbit anti-collagen IV (#PA5-104508, Thermo Fisher Scientific), mouse anti-postsynaptic density protein 95 (PSD95) (MAB1598, Merck Millipore), rabbit anti-spinophilin (14136 S, Cell Signaling Technology, Berkeley, CA, USA), mouse anti-myelin basic protein (MBP) (13344 S, Cell Signaling Technology), rabbit anti-α-synuclein (2642 S, Cell Signaling Technology), rabbit anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (2118 S, Cell Signaling Technology) and mouse anti-GAPDH (SC-166545, Santa Cruz Biotechnology, Santa Cruz, CA, USA). Corresponding secondary antibodies conjugated to horseradish peroxidase were used, and protein bands were visualized using an enhanced chemiluminescence system (RPN2106, Amersham Biosciences, Piscataway, NJ, USA). Band intensities were quantified using Gel-Pro Analyzer software (Media Cybernetics, Silver Spring, MD, USA). Protein levels were normalized to the internal gel loading control GAPDH. The results were expressed as fold of change relative to sham-operated controls.

Immunohistochemistry

Immunohistochemical analyses were performed on 25-µm-thick coronal brain sections to examine astrocytes, microglia, and α-synuclein. After cryoprotection, sections were treated with 1% H₂O₂ in 0.1 M PB for 1 h to quench endogenous peroxidase activity. To visualize astrocyte and α-synuclein protein, sections were blocked with 10% normal goat serum and 0.2% Triton X-100 in 0.1 M PBS for 1 h. For microglia, sections were blocked with 10% normal horse serum and 0.2% Triton X-100 in 0.1 M PBS for 1 h. After blocking, sections were incubated overnight at 4 °C with primary antibodies diluted in 0.1 M PBS. The primary antibodies used were mouse anti-GFAP antibody (MAB360; Millipore), mouse anti-α-synuclein (MA5-12272; Thermo Fisher Scientific), and goat anti-Iba1 (ab5076, Abcam), respectively. Following three rinses in PBS, the sections were incubated with secondary antibodies in 0.1 M PBS for 1 h followed by treatment in avidin–biotin-horseradish peroxidase reagents (PK-6100; Vector Laboratories). Biotinylated goat anti-mouse (AP132B; Millipore) and horse anti-goat antibodies (BA-9500; Vector Laboratories) were used accordingly. The sections were then reacted with 0.05% diaminobenzidine tetrahydrochloride and 0.01% H₂O₂ in 0.05 M Tris buffer (pH 7.4), mounted on slides, dehydrated, and cover-slipped.

Aquaporin 4 immunohistochemical staining and subsequent quantification

To identify cerebral aquaporin, the brain, spanning approximately 2.5 mm to 6.5 mm posterior to bregma and 0 to 4 mm lateral to bregma, was cryosectioned tangentially to the brain surface into 10-µm-thick horizontal sections. Four consecutive cerebral cortical sections, 100 μm apart, were blocked with 10% normal horse serum and 0.2% Triton X-100 in 0.1 M phosphate-buffered saline (PBS) for 1 h. After three PBS rinses, the sections were incubated overnight with rabbit anti-aquaporin 4 (ab259318; Abcam) in 0.1 M PBS at 4℃. After three PBS rinses, the sections were then incubated overnight with mouse anti-endothelial cell (ab9774; Abcam) at 4 ℃. Sections were reacted subsequently with VectaFluorTM Duet Immunofluorescence double labeling kit (DyLight^® 488 Anti-Rabbit IgG and DyLight^®594 Anti-Mouse IgG cocktail) (DK-8818; Vector Laboratories, Burlingame, CA, USA) for 1 h. Cell nuclei were visualized with 4’,6-diamidino-2-phenylindole (DAPI). After all processing, sections were mounted on glass slides with SlowFadeTM mountant (s36967; Invitrogen, Carlsbad, CA, USA). The sections were examined and recorded with NIS-Elements imaging software on a Nikon Eclipse Ni-U microscope with Nikon Digital Ds-Ri2 digital camera (NIS-Elements Advanced Research, Tokyo, Japan).

To evaluate changes in the perivascular space of the vessels penetrating into the cerebral cortex, fine arterioles with a near-circular cross-sectional profile in the tangentially prepared cerebral cortical sections were analyzed. They were divided arbitrarily into small (10–20 μm in diameter) and medium sizes (20–30 μm in diameter). The width of the perivascular space of each vessel was measured from two lines, horizontally and vertically crossing the center of the vessel, as follows. For either the vertical or horizontal measurement, the diameters of the outer boundary of endothelial staining and the vascular border of the aquaporin 4 staining were measured. The width of the perivascular space was taken as half the value of the difference between the two diameters. The width of the perivascular space of each vessel was then taken as the mean of the vertical and horizontal measurements of each vessel. Interestingly, the arterioles of the control rats had almost no measurable perivascular space, arguing that these vessels had no sizable tissue external to the endothelium. This also suggests that the space measured in the arterioles of the kaolin-treated rats likely reflected a close-to-real expansion of the true perivascular space.

The mean intensity of the aquaporin 4 immunofluorescence staining was semi-quantified using consistent intensity thresholds applied uniformly across all sections with Image Pro 10.0 (Media Cybernetics, MD, USA). For each rat, aquaporin 4 intensity at the perivascular site was measured in five randomly selected arterioles per tissue section. A total of 20 arterioles were analyzed per rat. The average intensity from these 20 arterioles represents the mean aquaporin 4 intensity at the perivascular site of each rat. Additionally, one to two regions of interest (ROIs, 20 μm × 20 μm) per tissue section, located 5 μm away from the perivascular site, were sampled to represent brain parenchymal aquaporin 4 expression intensity. A total of 4 to 8 ROIs were analyzed per rat. The average intensity of these ROIs was calculated to represent the mean brain parenchymal, away from the perivascular region, aquaporin 4 expression intensity of each rat.

Assessment of CSF tracer elimination dynamics

The fluid clearance function of rat brain was assessed by evaluating the clearance of the tracer Evans blue administered to the CSF following the methodology described by Maloveska et al. [21].

Surgical preparation

After anesthesia, the left femoral artery of the rat was exposed and a PE-50 polyethylene catheter (BD Intramedic™ Polyethene Tubing, Sparks, MD, USA) was inserted for blood sampling. The rat’s head was then secured in a stereotaxic frame, with the incisors positioned 20 mm below the horizontal interaural line. After a midline incision on the posterior scalp, the underlying muscles were bluntly dissected to expose the atlanto-occipital membrane. A 27-gauge needle connected to a 30-cm PE-20 polyethylene catheter was used to puncture the membrane. The other end of the catheter was connected to a 27-gauge open-ended flat needle and a 1-ml syringe attached to an infusion pump (Legato 130, KD Scientific Inc). With this setup, 60 µl of CSF was drained from each rat at a rate of 2 µl/min before Evans blue injection to minimize intracranial volume changes. Animals were carefully monitored and cared for during this surgical procedure with the depth of anesthesia continuously monitored and body temperature of the animal maintained at 37 °C with a heating pad.

Evans blue administration

A 30-cm PE-10 intrathecal tube, with one end cut at an angle and the other end connected to a 30-gauge open-ended flat needle and 1-ml syringe and linked to an infusion pump (Legato 130, KD Scientific Inc), was prepared. Following CSF drainage, the above-described PE-10 intrathecal tube was inserted into the cisterna magna and secured with super glue to prevent leakage. Evans blue (60 µl, 5% in saline; E2129, Sigma-Aldrich, St. Louis, MO, USA) was then injected into the cisterna magna at a rate of 2 µl/min. After injection, the PE-10 intrathecal tube was trimmed to 1 cm, and the free end was sealed with super glue. The overlying muscles and skin were then sutured sequentially.

Plasma collection and analysis

We followed the protocol described by Wang and Lai [22] to determine plasma Evans blue concentration. Blood samples (600 µl) were collected from the femoral artery 2 and 4 h after Evans blue injection and mixed with 10 µl of heparin (200 U/ml in Ringer-Locke solution, pH 7.4). The mixture was then centrifuged at 2000×g for 10 min at 4 °C, and the supernatant plasma was stored at − 80 °C until analysis. To analyze, plasma samples were mixed with 50% trichloroacetic acid (diluted in 0.9% saline; SI-T0699, Sigma-Aldrich) at 1:1 volume ratio. The mixture was then centrifuged at 10,000×g for 20 min at 4 °C to remove precipitated biomolecules. The absorbance of the supernatant at 620 nm was measured. Evans blue concentration was determined with a serial dilution standard curve and expressed as µg of dye per ml of the supernatant.

Tissue collection and analysis

The brain and deep cervical lymph nodes of the rats were harvested 4 h after Evans blue injection to determine the levels of Evans blue in the tissues. The left brain and cervical lymph nodes were immediately immersed in 4% paraformaldehyde in 0.1 M PB for three days for subsequent fluorescence analysis. The right deep cervical lymph nodes were weighed and immersed in formamide at 10:1 volume-to-weight ratio. The right hemisphere of the brain was homogenized in 600 µl artificial CSF, followed by the addition of 3 ml of formamide. The prepared tissue samples were incubated in a water bath at 55 °C for 48 h. Following incubation, the samples were centrifuged at 3200×g for 10 min. The supernatant was collected and stored at − 20 °C until subsequent analysis. Evans blue concentration was measured at 620 nm using a serial dilution standard curve and presented as µg of the dye per ml of the supernatant.

Fluorescence imaging of Evans blue

To identify the localization of Evans blue, the left brain and deep cervical lymph nodes were harvested and cryoprotected in 30% sucrose in 0.1 M PB for 3 days following fixation. The tissues were then cryosectioned, with the brain sectioned coronally, into 40-µm-thick sections. Sections were counterstained with DAPI for cell nuclei. Tissue sections were mounted on glass slides with VECTASHIELD^® antifade mounting medium (H-1000; Vector Laboratories) and examined with Nikon Eclipse Ni-U microscope and Nikon Digital Ds-Ri 2 digital camera (NIS-Elements Advanced Research) with NIS-Elements imaging software.

Statistical analysis

Data (mean ± SEM) were analyzed with Sigmaplot 14.0 (Grafiti LLC, Palo Alto, CA, USA). Normality was assessed with the Shapiro–Wilk test. For comparisons between hydrocephalic and sham-operated control rats under single experimental conditions, unpaired two-tailed Student’s t-tests were used when data showed normal distribution and with equal variances; Welch’s t-tests were applied when variances were unequal; Mann–Whitney rank-sum tests were used when normality was not met. Outcomes involving factorial structures were analyzed using two-way ANOVA with appropriate post-hoc tests. This included ventricular size (group × time), novel object recognition performance (group × time), and Morris water maze acquisition (group × time) data sets. For the Morris water maze, post-hoc comparisons across training blocks were performed using the Holm–Sidak all pairwise multiple comparison procedure. Effect sizes (Cohen’s d) were calculated for key comparisons, including ventricular size, Evans blue measurements, and behavioral outcomes. Statistical significance was taken when p < 0.05.

Results

Kaolin injection led to late-onset communicating hydrocephalus

To find out whether kaolin injection into the parietal subarachnoid space induced ventricular enlargement, we examined rats’ brains two weeks, one month, and three months after saline (Con) and kaolin (Hydro) injection. Representative coronal sections stained with Nissl are shown in Fig. 1A. Ventricular size was quantified by calculating the Evans index. A significant increase in the Evans index was detected only at three months following kaolin injection (Fig. 1B), indicating that parietal subarachnoid kaolin injection resulted in late-onset communicating hydrocephalus in rats.

Fig. 1.

Ventricular dilation following subarachnoid kaolin injection. (A) shows representative Nissl-stained, coronally sectioned whole brain micrographs of the kaolin-injected (Hydro) and saline-injected control (Con) 2 weeks (2 W) 1 month (1 M), and 3 months (3 M) post-injection. (B) shows the quantitative analyses of the Evans index of the brains of control and experimental rats. There was no significant difference in the Evans index of the saline-injected controls over time, therefore all data of the saline-injected, at different time points, were pooled and plotted as Con. n = 6–11 rats per group, *, P < 0.05 between the marked and the corresponding control. Scale bar = 5 mm in A. Cohen’s d value for the comparison between Con and Hydro at 3 months was 1.12

Kaolin injection led to delayed and sustained cognitive deficits

We then investigated whether cognitive impairment accompanied ventricular enlargement following kaolin injection. To assess short-term recognition memory, rats were subjected to novel object recognition test. Movement trajectories were comparable between kaolin and saline-treated rats (Fig. 2A; Hydro and Con, respectively). The recognition index, defined as the proportion of time spent exploring the novel object relative to the total exploration time, did not differ between groups (Fig. 2B), indicating that kaolin injection did not affect short-term recognition memory up to three months after treatment.

Fig. 2.

Changes in cognitive function in rats following subarachnoid kaolin injection. (A, B) show results of the novel object recognition test. (A) shows representative moving trajectories of the kaolin and saline-injected rats in the novel object and familiar object tests. (B) is the quantitative analyses of the recognition index of each group (n = 3–11 per group). (C–H) are single-day Morris water maze test results. (C) is the mean escape latencies in the cued-learning test in rats prior to either saline (Con) or kaolin (Hydro) injection (n = 28–34 per group). (D) is the swimming velocities of the rats during the test (n = 8–12 per group). (E) shows the representative cumulative trajectories of the swimming paths of the third trial block of each group. (F–H) show the mean escape latencies in the first, second, and third trial blocks in rats two weeks (F), one month (G), and three months (H) following injection. Con and Hydro represent saline and kaolin-injected rats, respectively. 2 W, 1 M, and 3 M indicate two weeks, one month, and three months following injection (n = 8–12 per group). *, P < 0.05 between the marked and the control in the same trial block. #, P < 0.05 and ##, P < 0.001 between the marked and the first trial block. $, P < 0.05 between the marked and the second trial block. Cohen’s d value for the comparison between C1M and H1M in block III was 1.09, and the effect sizes for the comparison between C3M and H3M in blocks II and III were 0.91 and 1.32, respectively

The animals’ spatial memory was then evaluated with the Morris water maze. Before injection, kaolin and saline-treated groups showed comparable performance in the cued learning task, indicating equivalent motivation and swimming ability (Fig. 2C). To minimize unwanted influences such as ventricular dilation or weight gain, a one-day water maze protocol consisted of three trial blocks with four trials per block was adopted. Swimming velocities were similar between groups (Fig. 2D). However, the cumulative swimming path of kaolin-treated rats appeared to be longer starting one month following injection (Fig. 2E, representative cumulative swimming trajectories from the third block of the test). The escape latency of the kaolin and saline groups was comparable two weeks after treatment (Fig. 2F). However, by one month after treatment, kaolin-treated rats exhibited a significantly longer mean escape latency in the third trial block than that of the saline controls (H1M vs. C1M, *P < 0.05; Fig. 2G). Spatial acquisition deficits became more pronounced three months post-treatment, when kaolin-treated rats showed significantly longer escape latencies in both the second and third trial blocks (H3M vs. C3M, *P < 0.05; Fig. 2H). In summary, parietal subarachnoid kaolin injection started to compromise rats’ spatial memory one month and worsened by three months post-treatment.

Kaolin injection quickly induced meningeal fibrosis and upregulated cerebral IL-6 expression

The delayed onset of communicating hydrocephalus at three months following kaolin injection provided an opportunity to examine molecular alterations preceding ventriculomegaly. Two weeks following kaolin injection, significant thickening of the meninges was detected (Fig. 3A). Masson’s trichrome staining revealed persistent meningeal thickening with increased collagen deposition, consistent with fibrosis, from two weeks to three months (Fig. 3A, B). Since fibrosis is often associated with inflammation, we analyzed cortical IL-6 expression. Western blot analysis showed significant upregulation of IL-6 beginning at two weeks and persisted up to three months following kaolin injection (Fig. 4A–C). Thus, kaolin injection quickly induced meningeal fibrosis and cerebral inflammation.

Fig. 3.

Masson’s trichrome staining of the meninges of rats following subarachnoid kaolin injection. (A) shows representative micrographs of the saline-injected control (Con) and kaolin-injected (Hydro) groups two weeks (2 W), one month (1 M), and three months (3 M) following injection. An apparent increase in collagen bundles (stained blue) was observed in the kaolin-injected groups. The brackets below the images of the 2 W group show examples of the meningeal thicknesses. (B) shows quantitative analyses of the meningeal thickness. H2W, H1M, and H3M stand for rats two weeks, one month, and three months following kaolin injection, whereas C2W, C1M, and C3M refer to saline-injected control rats at the corresponding time points. n = 4–8 rats per group, **, P < 0.001 between the marked and the corresponding control. Scale bar = 100 μm in A

Fig. 4.

Western blot analyses of IL-6 expressions in the rat cerebral cortex following subarachnoid kaolin injection. (A–C) Western blot analyses of IL-6 expressions at two weeks (A), one month (B), and three months (C) following kaolin injection. Representative blot images (top) and corresponding densitometric analyses (bottom) are shown for each panel. H2W, H1M, and H3M are rats two weeks, one month, and three months following kaolin injection, whereas C2W, C1M, and C3M are saline-treated control rats at the same time points. Arabic numerals denote individual animal identification numbers. GAPDH is the internal gel loading control. n = 4–8 rats per group. *, P < 0.05 between the marked and the corresponding control

Kaolin injection induced astrocytic activation and no apparent Iba-1 microglial response

Given the presence of meningeal fibrosis and brain inflammation, we next examined whether glial cells in the cerebral cortex responded to the insult. Western blot analysis revealed significant upregulation of GFAP from two weeks to three months following kaolin injection (Fig. 5A–C). In contrast, the expression of Iba1, a microglial marker, remained unchanged (Fig. 5D–F). Immunohistochemical staining further demonstrated markedly increased and thickened GFAP-positive astrocytic processes (Fig. 5G), whereas microglial density and morphology showed no significant changes at one month following kaolin injection (Fig. 5H). Thus, kaolin injection induced a pronounced astrocytic reaction with no overt Iba1-defined microglial changes in the cerebral cortex.

Fig. 5.

Western blot analyses and immunohistochemical staining of GFAP and Iba1 in the rat cerebral cortex following subarachnoid kaolin injection. (A–C) show the Western blots and corresponding analyses of the GFAP expressions two weeks, one month, and three months following kaolin injection (n = 3–6 per group). (D–F) show the Western blots and corresponding analyses of Iba1 at similar time points (n = 4–6 per group). Representative blot images are shown at the top, with the corresponding densitometric analysis plotted in the lower part of each panel. H2W, H1M, and H3M represent rats at two weeks, one month, and three months following kaolin injection, whereas C2W, C1M, and C3M indicate saline-injected control rats at the corresponding time points. Arabic numerals denote individual animal identification numbers. GAPDH is the internal gel loading control. *, P < 0.05 between the marked and the corresponding control. (G, H) show the representative images of GFAP (G) and Iba1 (H) immunohistochemical staining in the superficial layers (upper panel) and deep layers (lower panel) of the cerebral cortex in rats one month following kaolin injection. Scale bar = 50 μm

Kaolin injection altered cerebral cortical aquaporin 4 expression

Activation of astrocytes in the cerebral cortex prompted us to investigate whether aquaporin 4 expression, predominantly in astrocytes, was altered. Western blot analysis revealed a sustained and significant increase in cortical aquaporin 4 expression from two weeks to three months following kaolin injection (Fig. 6A–C). Immunohistochemical staining showed that in both kaolin and saline-treated rats, aquaporin 4 was highly expressed in structures immediately surrounding blood vessels (Fig. 7A), consistent with its preferential localization to astrocytic endfeet. Quantitative analysis revealed no significant difference in the mean perivascular aquaporin 4 intensities between kaolin and control groups (Fig. 7B). However, kaolin-treated rats showed increased aquaporin 4 immunoreactivity in the cortical parenchyma beyond the perivascular location (Fig. 7C), likely the labeling of astrocytic processes away from blood vessels. Quantitative analysis confirmed that in kaolin-treated rats, parenchymal aquaporin 4 expression increased at two weeks and remained elevated up to three months following treatment (Fig. 7D). Thus, kaolin injection led to increased aquaporin 4 expression in astrocytes at sites away from their perivascular endfeet. This resulted in a loss of the polarized distribution of aquaporin 4 in astrocytes, which is higher at the perivascular endfeet.

Fig. 6.

Western blot analyses of aquaporin 4 expression in the cerebral cortex of rats following subarachnoid kaolin injection. (A–C) show Western blot analyses of aquaporin 4 two weeks, one month, and three months following kaolin injection, respectively (n = 5–6 per group). Representative blot images are shown at the top, while the corresponding densitometric analysis plotted in the lower part of each panel. H2W, H1M, and H3M are rats at two weeks, one month, and three months following kaolin injection, while C2W, C1M, and C3M represent control rats at the corresponding time points following saline injection. Arabic numerals denote the identification numbers of individual animals. GAPDH is the internal gel loading control. *, P < 0.05 between the marked and the corresponding control

Fig. 7.

Aquaporin 4 expressions in the rat cerebral cortex following subarachnoid kaolin injection. (A) shows the representative double-labeling immunohistochemical fluorescence labeling around cerebral blood vessels with aquaporin 4 expression (green) and the vascular endothelial cells (red). (B) Histogram plot of the mean intensity of the perivascular aquaporin 4 expression in each group (n = 4–6 per group). (C) demonstrates the representative staining of the aquaporin 4 in the brain parenchyma of the cerebral cortex. Insets are enlarged views of the dotted enclosed square areas. (D) plots the mean aquaporin 4 expression intensity in randomly selected 20 μm × 20 μm regions of the brain parenchyma, located 5 μm away from the vessel (n = 4–6 per group). Scale bar = 10 μm in (A) and (C). H2W, H1M, and H3M represent rats two weeks, one month, and three months following kaolin injection, whereas C2W, C1M, and C3M stand for saline-treated control rats at the corresponding time points. *, P < 0.05 between the marked and the corresponding control

Kaolin injection upregulated CD163 and reduced blood–brain barrier protein expression

To learn more about the inflammatory consequences following kaolin injection, we investigated the expressions of the perivascular macrophage marker CD163, tight junction protein occludin, and basement membrane component collagen IV to evaluate cerebral cortical perivascular macrophage status and blood–brain barrier integrity following treatment. Western blot analysis revealed a significant upregulation of CD163 at two weeks and persisting up to three months following kaolin injection (Fig. 8A–C). In contrast, occludin expression declined at two weeks and remained reduced to three months following kaolin injection (Fig. 8D–F). On the other hand, collagen IV showed a delayed reduction in expression, first detected at one month and persisted to three months following treatment (Fig. 8G, I). Together, these findings indicate that kaolin injection activated perivascular macrophages and slowly impaired blood–brain barrier integrity in the cerebral cortex.

Fig. 8.

Western blot analyses of the CD163, tight junction protein occludin, and the key basement membrane component collagen IV expressions in the rat cerebral cortex following subarachnoid kaolin injection. (A–C) are Western blot analyses of cerebral cortical CD163 expressions two weeks (A), one month (B), and three months (C) following kaolin injection. (D–F) and (G, I) are Western blot analyses of occludin and collagen IV expressions and at the corresponding time points as indicated. In each panel, representative blot images are shown on the top, while the corresponding densitometric analysis at the bottom. H2W, H1M, and H3M are rats two weeks, one month, and three months following kaolin injection, while C2W, C1M, and C3M are from saline-treated control rats at each corresponding time point. Arabic numerals denote individual animal identification numbers. GAPDH is the internal gel loading control. n = 4–8 rats per group. *, P < 0.05 between the marked and the corresponding control

Kaolin injection resulted in perivascular alterations and altered CSF tracer elimination

Meningeal inflammation may alter meningeal lymphatics to potentially influence CSF outflow dynamics. To investigate this, we examined whether kaolin injection affected the perivascular space surrounding cortical penetrating vessels. Figure 9A shows representative double-immunofluorescence images of the cross-sections of fine intracerebral arterioles stained for endothelium (red) and aquaporin 4 (green). Measurement of the distance between the endothelial lining of the vessel and the astrocytic endfeet was taken as an estimate of the size of the perivascular space. Changes in the width of this space between kaolin-treated and control rats are likely to reflect real perivascular space widening since in control rats, these fine arterioles had almost negligible perivascular space. Quantitative analysis showed a significant widening of the perivascular space around small (10–20 μm) and medium-sized (20–30 μm) arterioles as early as two weeks following kaolin injection (Fig. 9B and C, respectively). This expansion persisted up to three months following kaolin injection, the latest time point examined, consisted with the sustained dilation of the cerebral cortical periarterial space following kaolin injection.

Fig. 9.

Widening of the perivascular space of cerebral cortical vessels in rats following subarachnoid kaolin injection. (A) shows representative merged micrographs of the cross-section profiles of small-sized (top row) and medium-sized arterioles (bottom row) from tangential cortical sections double stained with endothelium marker (red) and aquaporin 4 (astrocytic endfeet marker; green). Perivascular space is the unstained (dark) space between the endothelium and astrocytic perivascular endfeet. Scale bar = 10 μm. (B, C) are quantitative analyses of the perivascular space width of small and medium-sized arterioles, respectively (n = 4–7 per group), as described in the Materials and Methods. *, P < 0.05 and **, P < 0.001 between the marked and the corresponding control. H2W, H1M, and H3M represent rats two weeks, one month, and three months following kaolin injection, respectively, while C2W, C1M, and C3M are saline-treated control rats at the corresponding time points

To further characterize the impact of kaolin-induced meningeal and perivascular changes, we quantified the elimination of exogenously administered Evans blue from the CSF to the plasma and cervical lymph nodes, as well as its residual accumulation in the brain (Fig. 10A–F). In control rats, Evans blue introduced into the cisterna magna was cleared into cervical lymph nodes in 10 min (data not shown), representing intact brain lymphatic clearance. In contrast, kaolin-injected rats exhibited progressively graver impairment of this clearance capacity over time. Two weeks following kaolin injection, plasma Evans blue levels measured two and four hours after cisterna magna application were slightly lower than those of control rats but did not reach statistical significance (Fig. 10A). In one-month post-kaolin treatment rats, however, plasma Evans blue levels at both time points, i.e., two and four hours after been introduced, were significantly reduced compared to the control groups (Fig. 10B). These findings indicate altered CSF tracer transport to the systemic circulation. In consistence with these findings, Evans blue trapped in the cervical lymph nodes of one-month saline-treated rats (C1M), in dark blue color, was clearly visible to the naked eye, whereas those from the kaolin-treated rats (H1M) showed faint coloration (Fig. 10C, first column). This gross observation was supported by reduced red fluorescence intensity of Evans blue in lymph node cross-sections from kaolin-treated rats (Fig. 10C, second to fifth columns). Direct biochemical quantification further confirmed a significant reduction of Evans blue in the cervical lymph nodes of the one-month kaolin-treated rats compared to the control groups (Fig. 10D). Conversely, the retention of Evans blue in the brain was significantly greater in one-month kaolin-treated rats than saline controls, as indicated by the stronger red fluorescence in the cortical tissue underneath the subarachnoid space (Fig. 10E). Figure 10E also shows that the dye had permeated into the brain tissue underneath the surrounding subarachnoid space more extensively in the one-month kaolin-treated than the control rats. Biochemical quantification corroborated this finding, with increased Evans blue content in the brains of the kaolin-treated rats (Fig. 10F). Together, these findings show subarachnoid kaolin injection caused cerebral inflammation and meningeal fibrosis within two weeks of treatment. By one month, altered CSF tracer elimination were observed.

Fig. 10.

Effects of subarachnoid kaolin injection on CSF tracer elimination. (A, B) show Evans blue levels in the plasma of rats two weeks and one month following kaolin injection (n = 5–7 per group). Measurement was performed at two (2 h) and four (4 h) hours after introducing Evans blue into the cisterna magna. (C) shows representative images of the deep cervical lymph nodes, including macroscopic views of the whole intact nodes in situ (first column, outlined with dotted lines), and bright-field (second column) and fluorescence (Evans blue (red), DAPI (blue), and merged of Evans blue and DAPI) images of whole lymph node section. (D) shows quantitative analyses of Evans blue levels in the deep cervical lymph nodes of rats two weeks and one month following kaolin injection (n = 4–7 per group). (E) shows paired Evans blue, DAPI, and merged fluorescence images of a brain hemisphere from an one-month saline- and kaolin-treated animal, respectively. (F) shows quantitative analyses of Evans blue levels in the brain tissues at two weeks and one month following kaolin injection (n = 4–7 per group). Scale bars = 1 mm for the gross view of lymph nodes, 500 μm for lymph node cross sections, and 1000 μm for the brain sections. *, P < 0.05 between the marked and the corresponding control. H2W and H1M stand for rats two weeks and one month following kaolin injection. C2W and C1M denote saline-treated control rats at the same time points. Cohen’s d value for the comparison of plasma Evans blue levels between C1M and H1M at 2 and 4 h after Evans blue injection was − 1.82 and − 1.51, respectively. The effect sizes for the comparison of Evans blue levels in the lymph node and brain tissue between C1M and H1M were − 1.11 and 1.43, respectively

Kaolin injection downregulated excitatory synaptic and myelination protein expressions

Given the observed changes in CSF tracer elimination, we next investigated whether kaolin injection affected cortical neuronal connectivity. The expressions of three key neuronal structural proteins representing excitatory synaptic connectivity and axonal output were investigated. Western blot analysis showed significant reductions in PSD95, a glutamatergic excitatory postsynaptic marker, and spinophilin, a dendritic spine marker, beginning at one month and persisted to three months following kaolin injection (Fig. 11A–F), whereas MBP, an axonal myelination marker, was significantly reduced only at three months following kaolin injection (Fig. 12A–C). These findings suggest a late-onset temporal cascade of brain connectivity changes following kaolin injection, started with a reduction of dendritic excitatory inputs followed later by impaired axonal output of cerebral cortical neurons.

Fig. 11.

Western blot analyses of PSD95 and spinophilin in the rat cerebral cortex following subarachnoid kaolin injection. (A–C) are Western blot and analyses of PSD95 expressions two weeks (A), one month (B), and three months (C) following kaolin injection. (D–F) are Western blot analyses of spinophilin expressions at the same time points. Representative blot images (top) and corresponding densitometric analysis (bottom) are shown for each panel. H2W, H1M, and H3M represent rats two weeks, one month, and three months following kaolin injection, while C2W, C1M, and C3M are saline-treated control rats at the same time points. Arabic numerals denote individual animal identification numbers. GAPDH is the internal gel loading control. n = 3–7 rats per group. *, P < 0.05 between the marked and the corresponding control

Fig. 12.

Western blot and analyses of the MBP expression in the rat cerebral cortex following subarachnoid kaolin injection. (A–C) are Western blot and analyses of MBP two weeks (A), one month (B), and three months (C) following kaolin injection. In each panel, representative blot images are presented at the top, while the corresponding densitometric analysis at the bottom. H2W, H1M, and H3M represent rats two weeks, one month, and three months following kaolin injection, whereas C2W, C1M, and C3M are saline-treated control rats at the same time points. Arabic numerals denote individual animal identification numbers. GAPDH is the internal gel loading control. n = 5–7 rats per group. *, P < 0.05 between the marked and the corresponding control

Kaolin injection upregulated α-synuclein expression in cortical pyramidal neurons

To find out whether kaolin injection caused neuronal degenerative changes, the expression of α-synuclein in the cerebral cortical neurons was investigated. Western blotting showed a significantly increase of cortical α-synuclein expression late, three months, following kaolin injection (Fig. 13A–C). Immunohistochemical examination revealed that the increase was localized to the cell bodies (arrowheads) and thick dendritic processes (arrows) of cortical pyramidal neurons (Fig. 13D, H3M versus C3M). Together these findings suggest intracellular accumulation of α-synuclein in cerebral cortical output neurons, particularly in the cell bodies and proximal dendrites. Thus, kaolin injection ultimately, three months after treatment, caused the degeneration of cortical pyramidal neurons.

Fig. 13.

Changes in α-synuclein expression in the rat cerebral cortex following subarachnoid kaolin injection. (A–C) are Western blot and analyses of α-synuclein expressions two weeks, one month, and three months following kaolin injection. In each panel, representative blot images are on the top, while corresponding densitometric analysis at the bottom. H2W, H1M, and H3M represent rats two weeks, one month, and three months following kaolin injection, whereas C2W, C1M, and C3M are saline-treated control rats at the corresponding time points. Arabic numerals denote individual animal identification numbers. GAPDH is the internal gel loading control. n = 4–7 rats per group. *, P < 0.05 between the marked and the corresponding control. (D) shows representative immunohistochemical staining of α-synuclein in the rat cerebral cortex three months following kaolin (H3M) and saline injection (C3M). Arrows indicate the apical dendrites of cortical pyramidal neurons, while arrowheads denote pyramidal neuronal cell bodies. Scale bar = 50 μm

Discussion

Figure 14 summarizes the temporal cascade of molecular and functional changes that we revealed in the present rat model of parietal subarachnoid kaolin injection that induced a late-onset communicating hydrocephalus. Meningeal fibrosis accompanied by inflammatory responses, including increased CD163 and IL-6 expressions, reactive astrogliosis, and elevated aquaporin 4 expression with redistribution were apparent two weeks following kaolin injection. At the same time, signs of loss of blood–brain barrier integrity and marked expansion of the cortical parenchymal periarterial space became evident. By one month following kaolin injection, CSF tracer elimination was significantly reduced. This was accompanied by decreased expression of collagen IV, a key basement membrane component, to further undermine the integrity of blood–brain barrier. At this time, cognitive deficits, in parallel with reductions in excitatory synaptic markers PSD95 and spinophilin, emerged. By three months following kaolin injection, ventriculomegaly became evident in the presence of reduced MBP expression, suggestive of compromised cortical axonal output, and α-synuclein accumulation in cortical pyramidal neurons, indicative of cerebral neurodegeneration to further compromise cortical function.

Fig. 14.

Schematic summary of all the changes identified in the present study following parietal convexity subarachnoid kaolin injection. The timeline illustrated is arbitrary and not to scale

Early responses: meningeal and brain inflammatory reactions

Two weeks after kaolin injection, an apparent fibrosis of the overlying dura-arachnoid mater and widening of brain parenchymal periarterial space were noted. These changes could have affected CSF drainage and brain interstitial fluid clearance.

The pathological consequences of an expansion of the brain parenchymal periarterial space might be explained through several interrelated mechanisms. First, periarterial space expansion may alter vascular geometry and the structural organization of perivascular compartments. Such structural changes could be associated with inflammatory responses and might influence local CSF distribution. Consistent with this, we found increased expression of CD163, GFAP, and IL-6 in the brain, suggestive of activation of perivascular macrophages and astrocytes which might lead to the release of proinflammatory mediators involving in phagocytic clearance [23]. Second, expansion of the periarterial space may modify local perivascular microenvironment. This could influence endothelial homeostasis and vascular barrier efficacy. In this regard, endothelial barrier integrity is tightly regulated by local microenvironmental factors [24], and endothelial dysfunction has been associated with changes in tight-junction proteins, including occludin [25]. In the present study, decreased occludin expression was observed to associate with periarterial space expansion, suggesting alterations of the perivascular microenvironment. Third, astrocytes at the vascular interface are known to be highly mechanosensitive, mechanical deformation could activate Piezo1- and TRPV4-mediated calcium influx in astrocytes to drive their reactive transformation characterized by GFAP upregulation and proinflammatory cytokine release [26]. Our findings of astrocytic hypertrophy, upregulated GFAP expression, and elevated brain IL-6 levels are in accord with this regulatory process. Taken together, our results support the notion that kaolin injection-induced expansion of brain parenchymal periarterial space could contribute to brain inflammation and the compromise of blood–brain barrier integrity.

Interestingly, classical hydrocephalus models induced by intraventricular or cisterna magna kaolin injection have consistently been reported to trigger early microglial proliferation [27, 28]. In the present model, however, we observed no upregulation of Iba1, a well-established microglial marker, from two weeks to three months after treatment, during which GFAP, IL-6, and CD163 expressions were upregulated. Besides perivascular macrophages, CD163 is known to be expressed in M2-like microglia, which could participate in scavenging, extracellular matrix remodeling, and anti-inflammatory responses [29]. Thus, CD163 upregulation could more likely reflect an enhanced activation of perivascular macrophages or alternatively, a phenotypic shift toward M2-like polarization, presumably in response to the astrocytic IL-6 signaling [30]. In this study, although the absence of overt changes in Iba1 expression argues against a robust microglial activation following kaolin treatment, subtle or noncanonical microglial responses cannot be excluded. Together, our findings argue that parietal subarachnoid kaolin injection elicits an immune response primarily characterized by astrocytic activation, perivascular macrophage involvement, and no overt Iba1-defined microglial upregulation. Interestingly, increased CD163 expression has been reported to associate with perivascular changes, β-amyloid deposition, and blood–brain barrier disruption [31, 32]. In this regard, enhanced CD163 expression following kaolin injection might reflect a compensatory response mediated by perivascular macrophages and M2-like microglia to counteract impaired perivascular clearance and vascular stress. In the light of the above, Iba1 alone may be insufficient to delineate the relative contributions of astrocytes, microglia, and perivascular macrophages in hydrocephalus pathology; incorporation of more specific microglial markers such as CD68 might be an alternative to resolve this issue in future studies.

Aquaporin 4, a water channel protein highly expressed in astrocytes, had been shown to mediate glutamate-induced astrocyte swelling [33]. In the present study, parietal subarachnoid kaolin injection induced the hypertrophy of astrocytes, thickening of astrocytic processes, and upregulation of aquaporin 4 expression, consistent with features of glutamate-induced astrocytic cytotoxic edema [33]. Interestingly, the increase of aquaporin 4 in astrocytes extended to compartments beyond the perivascular endfeet, indicating a loss of its typical polarized localization in control astrocytes. Similar aquaporin 4 mislocalization had been reported in both experimental models and human cases of hydrocephalus, where it was associated with brain interstitial fluid accumulation [34]. In the light of this, activation of astrocytes with increased and redistributed aquaporin 4 expression could have contributed to the fluid imbalance in the hydrocephalic brain.

Together, the group of early reactions that we disclosed two weeks following parietal subarachnoid kaolin injection, including meningeal fibrosis, periarterial space expansion, reduced occludin expression, astrocytic hypertrophy, and aquaporin 4 mislocalization, might compound with each other to compromise brain interstitial fluid–CSF exchange, metabolic waste clearance, and CSF drainage.

Mid-term responses: perivascular alterations, reduced CSF tracer elimination, and cognitive deficits

One month following kaolin injection, perivascular alterations were observed in association with compromised CSF tracer elimination. Clearance of the exogenously introduced Evans blue from the CSF to the plasma and cervical lymph nodes was compromised and its retention in the brain increased (Fig. 10A–F). This was accompanied by a reduction of collagen IV expression, which had been shown to cause basement membrane remodeling to increase blood–brain barrier permeability [35]. Defective basement membrane could allow the extravasation of plasma proteins and fluids to favor fluid retention in the brain [36, 37]. Interestingly, collagen IV had been shown to exhibit species and stage-dependent alterations in Alzheimer’s disease, decreased in transgenic mice but increased in human postmortem tissue [38, 39]. With the short lifespan of rodents, reduction of collagen IV might represent an early pathological event contributing to β-amyloid retention. In this study, decrease in collagen IV concurred with the emergence of brain clearance impairment, suggesting that loss of collagen IV might serve as an early indicator of blood–brain barrier disruption and CSF drainage dysfunction in association with meningeal fibrosis. Notably, in the present study, the reduction of collagen IV was preceded by an increase in brain CD163 expression, raising the possibility that perivascular macrophages might play a role in collagen IV degradation [40].

Behaviorally, the single-day Morris water maze we adopted revealed spatial memory impairment in our kaolin-treated rats one month following kaolin treatment, while novel object recognition memory remained intact. This dissociation could have reflected brain region-specific vulnerability. Spatial memory depends heavily on motor and sensory cortices located underneath the parietal subarachnoid kaolin injection site, whereas novel object recognition memory relies more on the perirhinal cortex, located farther away from the present model’s kaolin application site and thus less affected [41]. In support of this, we identified concurrent reduction in the expressions of spinophilin and PSD95 in both the motor and sensory cortices (Fig. 14). Spinophilin and PSD95 are excitatory postsynaptic scaffolding proteins essential for synaptic stability [42, 43], and their reduction had been shown to associate with synaptic degeneration and cognitive deficits in Alzheimer’s models [44, 45]. Our findings thus suggest that parietal subarachnoid kaolin injection could impair spatial but not recognition memory through region-specific synaptic vulnerability mediated by the loss of excitatory synaptic proteins. In this regard, alteration of CSF dynamics at this stage may change local molecular environment, such as the generation of β-amyloid and inflammatory cytokines, both linked to dendritic spine alterations and plastic synaptic changes [7], as we found in the present study.

Late consequences: neurodegeneration and ventriculomegaly

By three months following kaolin injection, MBP expressed in the affected brain became reduced, potentially reflecting axonal demyelination (Fig. 14). This delayed response involved oligodendrocytes and might reflect the cumulative effect of intracranial inflammation, altered CSF dynamics, and brain extracellular environment changes. A disturbed neural microenvironment at this stage was further supported by the accumulation of α-synuclein in cortical pyramidal neurons (Fig. 13A–D). Although neurons possess intrinsic α-synuclein degradation pathways, such as the ubiquitin–proteasome system and the chaperone-mediated autophagy [46, 47], persistent disturbances in CSF dynamics may compromise these cellular protective mechanisms and potentially contribute to α-synuclein accumulation and aggregation. In this regard, dendritic spines appeared to be particularly vulnerable to such disturbances as we found sustained reductions in spinophilin and PSD95 preceded α-synuclein accumulation in cortical pyramidal neurons. In the present study, reductions in spinophilin and PSD95 expressions were observed alongside altered CSF dynamics starting one month after kaolin injection. These may be associated with the progressive decline of spatial memory performance from one to three months after kaolin injection (Fig. 11A–F).

Although we observed a progressive sequence of inflammatory, perivascular, and CSF-related alterations starting from two weeks onward to the development of ventriculomegaly (Fig. 15), a direct causal relationship between the early perivascular pathology and the delayed ventricular enlargement remains to be demonstrated. In view of recent analyses of CSF physiology, ventricular enlargement is more likely the outcome of multifactorial processes rather than solely from impaired CSF outflow [5].

Fig. 15.

A pair of cartoon diagrams compare putative CSF–perivascular transport pathways and meningeal structures between kaolin-injected (hydrocephalic; left) and normal brain (right). In the hydrocephalic brain, kaolin injection was associated with: ❶ Meningeal thickening with increased collagen deposition: this potentially impaired CSF outflow through arachnoid granulations and meningeal lymphatics. ❷ CSF stasis: resulted in periarterial space expansion (or vice versa) and perivascular macrophage activation. ❸ Downregulation of blood–brain barrier protein expressions: this impaired blood–brain barrier and might further exacerbate neuroinflammation. ❹ Disruption of the polarized distribution of aquaporin 4 in astrocytes: in normal brain, aquaporin 4 concentrated on astrocytic endfeet; reactive astrogliosis and increased expression of aquaporin 4 in location away from the perivascular endfeet disrupted this normally polarized distribution of aquaporin 4. This could have disrupted the polarized flow of brain interstitial fluid along the perivascular space for the clearance of metabolic waste. ❺ Neuronal retention of α-synuclein: altered aquaporin 4 polarization and changes in CSF elimination dynamics may change brain extracellular environment to result in the accumulation of neurodegeneration-related proteins such as α-synuclein. In contrast, in the normal brain (right half): brain waste clearance function was ensured by intact meningeal lymphatics and arachnoid granulations. Proper perivascular space and polarized localization of aquaporin 4 at the astrocytic perivascular endfeet work with arterial pulsation to facilitate the directional clearance dynamics of interstitial fluid/CSF flow (long arrows) in normal brain

Clinical implications: relevance to human normal pressure hydrocephalus

Although overly simplified, the rough heuristic of one rat day to 30 human days [48] suggests that the timing of the observed pathological changes—including perivascular alterations, changes in CSF dynamics, and cognitive impairment—may precede ventriculomegaly by several years in humans. This delayed presentation of ventricular enlargement is broadly consistent with the clinical course of NPH, where ventriculomegaly develops gradually over several years. In addition, like what we have observed in the present rat model, late-stage NPH in humans has been reported to exhibit perivascular alterations, loss of polarized perivascular aquaporin 4 distribution in astrocytes, and increased α-synuclein levels in the CSF and brain [49–51]. Besides these, the region-specific reduction of synaptic proteins and cognitive deficits demonstrated in the present model mirror the selective vulnerability of frontal and motor cortices in NPH patients [52]. The late-stage intracellular accumulation of α-synuclein in cortical pyramidal neurons observed in the present model raises the possibility that persistent alterations in CSF dynamics may be associated with neurodegenerative protein accumulation. This interpretation is consistent with reports describing Parkinsonian features in some patients with NPH [53]. Collectively, our findings support the present kaolin-induced meningeal fibrosis model as a relevant preclinical approach to recapitulate, both temporally and regionally, the pathological changes in human NPH.

Limitations

Several limitations in the present study should be acknowledged. First, we did not perform longitudinal intracranial pressure measurements in this model. Although NPH is typically characterized by normal range cerebrospinal fluid pressure with lumbar puncture, continuous monitoring show intermittent fluctuations in intracranial pressure [4]. In experimental kaolin-induced communicating hydrocephalus models, ventricular enlargement and meningeal fibrosis are the commonly reported structural features, while systematic longitudinal intracranial pressure monitoring has been less frequently integrated into these experimental studies [16, 17, 54]. Future studies incorporating repeated intracranial pressure assessments over time may provide further insight into intracranial pressure dynamics during the development of adult hydrocephalus. Second, administration of Evans blue into the cisterna magna to evaluate CSF tracer elimination does not allow specific assessment of parenchymal solute movement within perivascular pathways [55]. In this study, Evans blue introduced into the cisterna magna appeared to have diffused in the subarachnoid space and into distant cerebral tissue underneath the subarachnoid space in a matter of hours, especially in the kaolin-treated rat brain (Fig. 10E). Its clearance into the plasma and cervical lymph nodes appeared to argue that it was discharged through the CSF and lymphatic pathways (Fig. 10A–D). The increase in its retention in the brains (Fig. 10E, F) of the kaolin-treated rats could reflect either enhanced entry, reduced efflux, and/or altered CSF–interstitial exchange. In this regard, our Evans blue findings demonstrate reduced CSF tracer elimination; however, these data are insufficient to support a specific impairment of parenchymal clearance pathways in kaolin-treated rats. Third, unlike the rapid and dramatic expansion of ventricular size in acute obstructive hydrocephalus [16, 17, 54], the mild ventriculomegaly (Evans index ≈ 0.35) that appeared late in our model is consistent with the progressive and late dilatation of ventricles in other subarachnoid space kaolin injection studies [16, 17, 54]. Additional morphometric indices, such as ventricular volume or corpus callosum angle in cases where brain MRI images were available could help to further delineate ventriculomegaly severity assessment. Fourth, we chose to study male, but not female rats, for estrogen shows neuroprotective effects [56, 57] and dendritic spines of female rat cortical pyramidal neurons undergo cyclic modulation during the estrous cycle [58, 59]. Finally, the absence of a sham-operated control group undergoing catheterization without CSF drainage or Evans blue injection limits us from excluding potential procedural effects associated with tracer administration into the cisterna magna for the assessment of CSF tracer elimination. However, since all animals in the present study were subjected to the same surgical protocol, a procedure effect, if present, would be expected to affect all groups equally and couldn’t be solely responsible for the phenomenon observed.

Conclusion

This study identified a temporal sequence of pathophysiological changes beginning with intracranial inflammation and meningeal fibrosis, followed by communicating hydrocephalus and the emergence of cognitive deficits after parietal subarachnoid kaolin injection (Fig. 15). In this process, astrocytic and perivascular alterations preceded changes in CSF tracer elimination. Subsequent reductions in excitatory synaptic proteins, cognitive deficits, accumulation of neurodegeneration-associated proteins, axonal demyelination, and ventricular enlargement were observed over time. These findings raise the possibility that alterations in CSF dynamics may be associated with the temporal progression of intracranial inflammation, meningeal fibrosis, and ventricular enlargement. The cascade of events revealed provide insight into how inflammation could have driven the impairment of CSF and brain interstitial fluid clearance to culminate in chronic communicating hydrocephalus. Our findings provide a framework for future studies aiming at exploring therapeutic strategies targeting meningeal and perivascular alterations in NPH.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

CD163

Cluster of differentiation 163

CSF

Cerebrospinal fluid

DAPI

4’,6-diamidino-2-phenylindole

GAPDH

Glyceraldehyde-3-phosphate dehydrogenase

GFAP

Glial fibrillary acidic protein

Iba1

Ionized calcium-binding adapter molecule 1

IL-6

Interleukin-6

MBP

Myelin basic protein

NPH

Normal pressure hydrocephalus

PB

Phosphate buffer

PBS

Phosphate-buffered saline

PSD95

Postsynaptic density protein 95

ROI

Regions of interest

Author contributions

C.L.W. and L.J.C. performed the experiments and collected the data. G.F.T., S.T.T., and L.J.C. supervised the study. L.J.C. and G.FT. wrote the original draft. All authors contributed to the conception and design of the study and reviewed and approved the final manuscript.

Funding

The work was supported by grants from Buddhist Tzu Chi Medical Foundation and Tzu Chi Medical Mission Project (TCMMP; grant numbers TCMMP 112-01-02 and TCMF-MP 113–01–02) and by the National Science and Technology Council (NSTC), Taiwan (grant number NSTC 113-2320-B-320-007).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Animals were housed and treated in accordance with the guidelines of the National Laboratory Animal Center. The experimental protocol followed the “3Rs” principle in animal research and was approved by the Animal Care and Use Committee of Tzu Chi University (approval numbers 111019 and 111085).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.