Introduction
Ischemic leukoencephalopathy refers to a wide range of lesions within the deep perforating arteriole in the hemispheres. The cognitive dysfunction caused by ischemic leukoencephalopathy is the main cause of vascular dementia [
1]. Amelioration of cognitive impairment caused by persistent hypoperfusion can significantly improve the survival quality of chronic leukodystrophy patients [
2]. However, at present, there is insufficient understanding of the pathogenesis of cognitive dysfunction caused by chronic hypoperfusion white matter lesions (CHWMLs), which greatly hinders the development of therapeutic drugs.
In recent years, the adenosine A
2A receptor (A
2AR) has been extensively studied in various disease models due to its multiple biological effects, such as pro-inflammation, anti-inflammation, neuroprotection [
3‐
5]. In central nervous system diseases, adenosine A
2AR has a dual role, with a protective role in some diseases [
6‐
8] and aggravation of damage in other diseases [
9,
10]. The effect of adenosine A
2AR on cognitive function is also contradictory. It has been reported that inhibition of neuron adenosine A
2AR can improve cognitive dysfunction caused by Parkinson’s disease [
11]. Conditional knockout of astrocyte A
2AR can promote long-term memory retention in Alzheimer’s disease patients [
7]. Adenosine A
2AR aggravates intracranial inflammation through activating microglia, leading to cognitive dysfunction in mice with acute ischemia and hypoxia [
12]. However, it has also been reported that promoting adenosine A
2AR accumulation in the hippocampus can promote cognitive improvement in schizophrenia [
13], that adenosine A
2AR knockdown aggravates the motor and cognitive dysfunction of Huntington’s disease by reducing the expression of morphine peptides in the striatum region, and that increasing A
2A receptor expression in peripheral blood mononuclear cells can improve the prognosis of mild cognitive function disorder patients by reducing inflammation [
14]. These data suggest that adenosine A
2AR plays two distinct roles in the recovery of cognitive function in different diseases. However, the role and mechanism of adenosine A
2AR in cognitive dysfunction induced by CHWMLs has not been reported, and it is worthy of further exploration.
Our previous research confirmed that BMDC A
2ARs can inhibit inflammation [
15], while non-BMDC A
2ARs activate microglia and promote inflammatory responses [
16]. Different sources of adenosine A
2AR play contradictory roles. In general, the anti-inflammatory effect mediated by BMDC A
2ARs is stronger than the pro-inflammatory effect of non-BMDC A
2ARs. Promoting overall adenosine A
2A expression can inhibit inflammation and reduce CHWMLs [
17]. However, in the persistent chronic hypoperfusion model, the role of different cell-derived adenosine A
2AR in cognitive function recovery and its mechanism need to be further studied. In consideration of the effect of the inflammation in the cognitive function recovery, we hypothesize that BMDC A
2ARs could improve the cognitive function of chronic white matter ischemia model and non-BMDC A
2ARs served the opposite effects.
Therefore, the aims of our study were to apply bone marrow transplantation after irradiating wild-type mice and A2A knockout mice to establish models with selective inactivation or reconstruction of BMDC A2ARs, to explore the effect and mechanism of adenosine A2AR from different tissue sources on cognitive function in CHWMLs, and to find corresponding targets to improve the quality of life of patients with chronic white matter ischemia.
Methods
Experimental Animals and Drugs
Global Adenosine A2AR gene knockout (A2AR gKO) C57BL/6 mice were a gift from Dr. Jiang-Fan Chen (Boston University School of Medicine, Boston, MA). All mice in this experiment were 8–9 weeks old and weighed 23–30 g. Female mice were donors, and male mice were recipients. The adenosine A2AR-specific agonist CGS21680 was purchased from TOCRIS, UK.
Adenosine A2AR KO mice and littermate WT mice were sacrificed by cervical dislocation, and the femur and tibia were separated on a clean platform. The bone marrow cells in the medullary cavity were washed out with RPMI 1640 medium. The cells were collected, filtered through a 200-mesh sieve, and centrifuged at 1500 rpm for 5 min, and the precipitated cells were washed twice with RPMI 1640 medium, centrifuged at 1500 rpm for 5 min, and pulverized in RPMI 1640 medium to form a cell suspension. The cell concentration was adjusted to 2 × 108/mL for use. The cell viability was checked by 0.2% trypan blue staining, and the number of viable cells was greater than 95% for bone marrow transplantation.
Establishment of a Chimeric Model of Bone Marrow Transplant Cells
The recipient mice (male) were subjected to whole body gamma irradiation with a total dose of 12.5 Gy from a
60Co source, which destroyed the bone marrow hematopoietic capacity. The collected bone marrow cell suspension was injected into the recipient mice via the tail vein (the concentration of bone marrow cells was 2 × 10
8/mL, 0.3 mL/rat), and the bone marrow cell transplantation was completed [
18]. Bone marrow cells of female adenosine A
2AR KO mice were transplanted into male C57BL/6 mice (KO → WT) to establish a selective inactivated model of BMDC A
2ARs; bone marrow cells of female WT mice were transplanted into male C57BL/6 mice as controls (WT → WT); bone marrow cells of female WT mice were transplanted into male adenosine A
2A receptor KO mice (WT → KO), which was used as a selective reconstruction model of BMDC A
2ARs; and bone marrow cells of female adenosine A
2A receptor KO mice were transplanted into male adenosine A
2A receptor KO mice (KO → KO), which served as controls. The experiment was carried out 8 weeks after transplantation.
Establishment of a Chronic Cerebral Blood Flow Hypoperfusion Model
A chronic cerebral blood flow hypoperfusion (CCH) model was established with reference to previously described methods [
19]. Anesthesia was achieved with isoflurane inhalation, and a supine position was applied. A median longitudinal incision was made after cervical anterior skin depilation and sterilization, the neck muscle was separated, and the bilateral common carotid artery was fully liberated. Two 3-0 nylon threads were used to loosely wrap the near and far end of the right common carotid artery, and two nylon threads were suspended with a hemostat and placed in a specially designed microspring with an inner diameter of 1.8 mm. Then, the spring coil was carefully placed over the common carotid artery; after 30 min, the spring loop was placed over the left common carotid artery with the same method. In the sham-operated control group, the bilateral common carotid arteries were exposed but not clipped. The skin was sutured and disinfected after surgery. The anal temperature of the mice was maintained at 36.5–37.5 °C during the operation. During the whole operation, the Doppler blood flow meter 414-1 probe was fixed at the junction of the scale and the sphenoid wing above the zygomatic arch to measure the cerebral blood flow and evaluate the modeling.
Morris Water Maze Task
The learning and memory functions of the mice were evaluated using a Morris Water Maze (MWM) task, which was started on the 12th day before modeling and the 21st day after modeling. A black pool with a diameter of 120 cm, a height of 40 cm, and a depth of 25 cm was selected, and warm water with a temperature of 20–23 °C was infused. The pool was divided into 4 quadrants (NW, NE, SE, and SW), and a platform with diameter of 5 cm was placed at the center of the NE quadrant, 1 cm below the horizontal plane. Visual cues of different shapes were given on the NW- and SE-oriented swimming pool walls. The first day was visible platform training (VP), and a marker was placed over the platform to allow the mouse to identify the platform. The mice were placed into the water in a clockwise direction (beginning from the NW quadrant), and the time required for the mouse to find the platform stay on the platform for 10s (escape latency) was recorded. If the mouse did not find the platform within 120s, the experimenter guided it onto the platform, and it stayed on it for 10s. In this situation, the escape latency was recorded as 120s. The second to fifth days were the acquisition phase (AP). At this stage, the mark was removed from the platform, and the above experiment was repeated to record the escape latency and swimming distance. A probe test (PT) was carried out on the sixth day. At this stage, the platform was removed, and the mice were allowed to enter the water from the NW quadrant. The time in the correct quadrant (SE) and the number of passages through the platform location within the 120s were recorded.
Immunohistochemistry
At the 2nd week and 4th week after CCH, 5 mice were randomly selected from each group, anesthetized with 1% sodium pentobarbital (60 mg/kg), and perfused with 4% paraformaldehyde. Their brain tissues were taken out after dehydration and then frozen (10 μm thick per section). After the sections were antigen-repaired, permeated, and blocked, they were incubated at 4 °C overnight with the ZO-1 antibody. The sections were then incubated with biotin-labeled secondary antibody and horseradish peroxidase label for 2h at room temperature, and finally, coloration was performed with a DAB kit. Then, the expression of ZO-1 was observed under a microscope. Semiquantitative analysis of staining results was performed using Image Pro Plus software. The Nishigaya method was used [
20]. The staining intensity of the WT mouse in the sham group (normal) was rated as 4. It was rated as 3 when it was slightly lower than normal, rated as 2 when moderately lower than normal, rated as 1 when significantly lower than normal, rated as 0 when the negative result was positive, rated as 5 when slightly stronger than normal, rated as 6 when it was moderately stronger than normal, and rated as 7 when it was significantly stronger than normal.
Measurement of Blood-Brain Barrier Permeability
To analyze the alterations in cerebral vascular permeability, Evans blue dye was used as a marker of albumin extravasation. Briefly, mice were injected with 2% EB (3 mL/kg) via the tail vein 4 weeks after CCH. After an hour, animals were then perfused transcardially with phosphate buffer saline (PBS) to purge the intravascular EB dye. Whole brains were isolated, and the dye was extracted with N,N-dimethylformamide overnight at 50 °C, followed by centrifugation at 5000 rpm for 20 min. The concentration was determined from the OD (610 nm) values according to the standard curve, and then the EB content in the brain tissue was calculated to evaluate blood-brain barrier (BBB) permeability.
Immunofluorescence Staining
At the 4th week after CCH, tissue sections were fixed in 4% paraformaldehyde for 30 min, followed by permeabilization with a 0.1% Triton X-100 solution for 30 min and antigen blocking with 5% goat serum at room temperature for 30 min. The sections were incubated with cystatin F (1:50, Santa Cruz Biotechnology, Inc., USA) and CD11b (1:100, Chemicon, Colorado, USA) overnight at 4 °C. Then, staining with a fluorescein-labeled secondary antibody and DAPI was performed in the dark. The cells were finally mounted on glass slides and observed under a laser confocal microscope.
Enzyme-Linked Immune Sorbent Assay
At the 4th week after CCH, 5 mice were randomly selected from each group, and peripheral blood was taken for an enzyme-linked immune sorbent assay (ELISA) to evaluate systemic inflammatory responses. The eBioscience ELISA kit was used, and the reagent instructions were as follows. After preparing the standard, the sample was added and incubated at 37 °C for 30 min. It was then washed repeatedly with PBS buffer, the enzyme standard reagent was added, and the sample then was incubated again at 37 °C for 30 min. Finally, the developer was added, and the absorbance of each well was measured at 450 nm with a microplate reader.
Bielschowsky Silver Staining
Bielschowsky silver staining was carried out according to the instructions. Paraffin sections were dewaxed, 20% silver nitrate solution was dip-coated for 24 h at room temperature, the sections were washed with double distilled water, 10% formaldehyde was reduced, and then 20% silver nitrate solution + ammonia water was applied again for coloration. The sections were then fixed with 5% sodium thiosulfate and sealed after anhydrous ethanol dehydration. The judgment criteria were graded for the degree of damage of white matter nerve fibers [
21]: grade 0, normal; grade 1, disordered arrangement of nerve fibers; grade 2, significant vacuolation (nucleus damage); and grade 3, loss of myelinated nerve fibers (neural fiber axon damage).
Statistical Methods
All data are presented as the mean ± standard error. Data analysis was performed using SPSS 22.0. A one-way analysis of variance with the appropriate LSD post hoc test was used for comparison of experimental groups. The MWM test results were analyzed with multivariate analysis of variance with the appropriate LSD post hoc test. P < 0.05 indicated statistical significance.
Discussion
Our results confirmed that BMDC A2ARs promoted cognitive function recovery in a CCH mouse model, while non-BMDC A2ARs aggravated cognitive impairment. In general, the function of BMDC A2ARs to promote cognitive recovery plays a leading role, and its mechanism is mainly achieved by reducing inflammation and white matter lesions.
Dementia caused by vascular cognitive impairment accounts for more than 20% of the number of dementia patients, which is the second leading cause of dementia, after Alzheimer’s disease (AD) [
22,
23]. Chronic hypoperfusion is an important factor leading to vascular cognitive impairment, and it has received increasing attention in recent years. The phosphodiesterase inhibitor roflumilast [
24] and the transient receptor potential melastatin 2 (TRPM2) inhibitor [
25] can suppress microglial activity after CCH injury and promote cognitive function recovery through anti-inflammatory effects. IL-1β receptor inhibitors and IL-1β knockout mice block the IL-1β signaling pathway, thereby inducing oligodendrocyte premature cell migration to the periphery of the corpus callosum, and avoiding white matter damage and reversing chronic hypoperfusion-induced cognitive dysfunction by improving the local inflammatory environment [
26]. MiR-181c [
27] and MiR-96 [
28] can ameliorate cognitive dysfunction by regulating autophagy in CCH [
29]. In general, demyelination, inflammatory responses, and mitochondrial dysfunction are common pathways leading to cognitive dysfunction [
30,
31], which can ameliorate the prognosis of chronic cortical hypoperfusion by reducing white matter lesions.
Adenosine is an intermediate product of energy metabolism, which is widely distributed in various tissues and organs. It is also an important neuromodulator in the nervous system that plays an important role in regulating inflammation. Our previous studies reported that the anti-inflammatory effects of BMDC A
2ARs are stronger than the pro-inflammatory effects of non-BMDC A
2ARs [
15], but it is not clear whether its role in cognitive impairment is caused by CHWMLs. MWM showed that the cognitive function-related indexes in the CCH (WT) group were significantly different from those in the CCH (KO) group. These data suggested that adenosine A
2AR involved in cognitive impairment after CHWMLs. To further evaluate the role of adenosine A
2AR in cognitive impairment in different tissue sources, we used irradiation and bone marrow transplantation to establish a selective inactivated or reconstruction model of BMDC A
2ARs. The results of the MWM test suggested that after selective deletion of BMDC A
2ARs, the escape latency was prolonged, the time in the target quadrant was shortened, the number of passages through the platform location was reduced, and the trend was consistent with the trend of A
2AR whole gene knockout. The selective reconstruction of BMDC A
2ARs can reverse the above results, further indicating that BMDC A
2ARs have protective effects on cognitive function. In the course of the experiment, we also found an interesting phenomenon; that is, the cognitive function-related index in the KO → WT group had a significant difference in comparison with the WT → KO group, which suggested that non-BMDC A
2ARs may aggravate cognitive function damage. To further confirm this hypothesis, we used the adenosine A
2AR agonist CGS21680 to interfere in different bone marrow transplantation groups. The experimental results suggested that only activated non-BMDC A
2ARs prolong the escape latency, reduce the time in the target quadrant, and decrease the number of passages through the platforms. Only activated BMDC A
2ARs will shorten the escape latency and increase the time in the target quadrant and the number of passages through the platform location. These data confirmed our hypothesis that BMDC A
2ARs promote cognitive function recovery and that non-BMDC A
2ARs aggravate cognitive impairment. However, the overall trend is that adenosine A
2AR provides a protective effect on cognitive impairment in CHWMLs.
As we previously mentioned, the inflammatory response is the basic pathway for cognitive impairment caused by CHWMLs. We therefore further verified whether anti-inflammation is the mechanism by which adenosine A
2AR promotes the recovery of neurological function. Selective inactivated BMDC A
2ARs or activated non-BMDC A
2ARs led to systemic inflammatory response, and increased the expression of CF, accordingly activated microglia, whereas selective reconstitution reversed the above phenomenon. CF, a potent endogenous cysteine protease inhibitor, was substantially up-regulated in regions of white matter rarefaction that occurred in various demyelinating diseases of the CNS [
32,
33]. In this study, we found that BMDC A
2ARs suppressed CF-mediated microglia activated and non-BMDC A
2ARs promoted CF-mediated microglia activated, which was consistent with our previous research [
16]. The immunohistochemistry suggested BBB alterations was not the key point for cognitive function recovery with adenosine A
2ARs. However, elevated BBB permeability was the pathological basis of systemic inflammatory cytokines acting on brain parenchyma cells. Silver staining results suggested that activation of BMDC A
2ARs can alleviate corpus callosum white matter damage. These data confirmed our hypothesis that adenosine A
2AR promotes the recovery of cognitive function in mice by inhibiting inflammatory responses and reducing white matter lesions. In this process, BMDC A
2ARs play a leading role.
In the AD and PD disease models, adenosine A
2AR aggravates cognitive dysfunction, whereas adenosine A
2AR can improve cognitive function in the HD and schizophrenia disease models. It is worth further consideration to determine the reason for the exact opposite effect of A
2AR. Literature reports have suggested that adenosine A
2AR had two main pathways for regulating cognitive function: one is via integrating dopamine, glutamate, N-methyl-D-aspartate (NMDA), and brain-derived neurotrophic factor signaling pathways, and the other is regulating inflammatory responses [
34]. Adenosine A
2AR mediates cognitive dysfunction in AD and PD disease models mainly through the first pathway [
9,
35,
36], while in the HD and schizophrenia disease models, it mainly protects through the second pathway [
13,
14,
37,
38]. This study showed that adenosine A
2AR also improved cognitive function through anti-inflammatory effects in CHWMLs, and this mechanism of action is consistent with the above conjecture. However, at the same time, we must also realize that in other disease models [
12,
39‐
42], the role of adenosine A
2AR in cognitive function cannot be explained by the above mechanism, indicating that there may be other functional ways, which requires further study.
Although the FDA has approved acetylcholinesterase inhibitors and an NMDA receptor antagonist, memantine, as a drug for the treatment of cognitive dysfunction since 1993, these drugs currently only play a role in delaying the course of cognitive dysfunction and cannot effectively stop or reverse the course of the disease [
43]. Moreover, specific drugs for cognitive impairment of chronic hypoperfusion leukoencephalopathy are lacking. Adenosine A
2AR has multiple biological effects and can play different roles in different disease models. This study found that BMDC A
2ARs inhibit cognitive function after CHWMLs through anti-inflammatory effects. Non-BMDC A
2ARs can aggravate cognitive impairment through inflammation. Overall, BMDC A
2ARs play a leading role. This study initially explored the role and mechanism of adenosine A2AR in the pathogenesis of cognitive impairment in CHWMLs and provides a new treatment strategy for CHWMLs.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.