Introduction
Vascular dementia (VAD) has been the second common type of dementia and lack of effective treatments [
1,
2]. Chronic hypoperfusion leads to white matter lesions, which is the main pathogenic mechanism of VAD [
3]. It has been proposed that accumulated myelin debris and local inflammation cascade secondary to cerebral chronic ischemia hinders endogenous white matter repair. Resident microglia/macrophage phagocytose myelin debris, which resolves excessive inflammation in the damaged white matter and promotes white matter recovery [
4].
Pentoxifylline (PTX) is a non-specific phosphodiesterase inhibitor and a methylxanthine derivative which is extracted from cacao beans [
5,
6]. PTX was approved by the National Formulary of China for patients with ischemic stroke in 2010. In addition to ischemic stroke, A few controlled clinical studies suggested that PTX could be a therapeutic option for VAD with the tendency to improve cognitive impairment [
7‐
9], but there is no consensus on this point in the international community. Moreover, with the development of the research on cerebrovascular disease and dementia, the classification of vascular dementia is becoming more and more refined [
1,
10]. There is still no effective treatment for vascular dementia caused by white matter injury associated with cerebral small vessel disease. Extensive research has shown that PTX could reduce the secretion of inflammatory cytokines such as tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β), IL-6, IL-8 by inhibiting microglia/macrophage activation in neuropathic pain and ischemic stroke models [
11,
12]. Moreover, it was shown that PTX could regulate macrophage migration and myelin debris uptake in vitro [
13]. However, the effect of PTX on microglial phagocytosis of myelin debris has not been confirmed.
Mer receptor tyrosine kinase (Mertk) belongs to TAM (Tyro3, Axl and Mertk) family, which is a professional phagocytic receptor expressed on the surface of microglia and other immune cells. All phagocytic processes start with the exposure of “eat-me” signal from the apoptotic cells or debris. For instance, phospholipid phosphatidylserine on apoptotic cells/myelin debris could act as a strong eat-me signal initiating microglia rapid recognition and clearance by Mertk [
14,
15]. Efficient clearance of apoptotic cells by Mertk and Axl has fundamental roles facilitating adult neurogenesis [
16]. In addition to apoptotic cell clearance, Mertk was required for myelin engulfment and subsequent remyelination in multiple sclerosis [
17,
18].
A recently published study showed that activated peroxisome proliferator-activated receptors-γ (PPAR-γ) could augment the expression of Mertk in microglia and accelerate the clearance of hematoma after intracranial hemorrhage, suggesting that PPAR-γ may act as an upstream regulator of Mertk [
19]. However, the protective patterns of Mertk and its transcriptional regulation in VAD model are yet to be examined.
Considering the above factors, we aimed to explore the therapeutic effects of PTX on cognitive function and the white matter integrity in mouse bilateral common carotid artery stenosis (BCAS) models. We hypothesized that PTX activated PPAR-γ and promoted white matter repair by modulating Mertk-mediated myelin clearance in microglia. These results could further enhance our understanding on PTX and provide a novel therapeutic strategy to limit white matter impairment after vascular dementia.
Discussion
White matter injury caused by hypoperfusion or hypoxia participated in the development of vascular dementia. BCAS mouse model is widely accepted to mimic the cognitive impairment and white matter injury of clinical vascular dementia [
20‐
22]. In this study, we demonstrated that short-term working memory, characteristic exploratory feature, spatial learning and memory performance of mice were disrupted following BCAS, whereas this cognitive impairment was reversed by the administration of PTX. It is noteworthy that locomotor activity and behavioral despair could be affected in severe BCAS model. In such cases, the outcomes of the MWM cannot be attributed only to cognitive function [
23,
24]. Therefore, we performed open field test to exclude the impact of the locomotor activity and behavioral despair. After BCAS, white matter lesions were often observed in the corpus callosum, caudate putamen, internal capsule, and to a lesser extent in the optic tract [
25]. Consistently, our data also showed that myelin and axon integrity were damaged in CC and IC area 30 days after BCAS. With the application of MRI, immunofluorescence and electron-microscopy, we strongly confirmed that PTX increased white matter integrity after BCAS. Clinical trials suggested a trend toward improved cognitive function in VAD patients administrated with pentoxifylline [
26]. To our knowledge, this is the first study confirming its efficacy in VAD animal model with detailed behavior, imaging and pathological data. Using animal model of maternal LPS-induced white matter injury, Mustafa et al. demonstrated that PTX may minimize periventricular leukomalacia in the developing rat brain [
27]. This is the only evidence showing the protective potency of PTX in white matter injury. Moreover, mechanisms responsible for the neuroprotective actions of PTX were largely unknown.
Myelin debris is a major obstacle to white matter repair both as a physical impediment [
28] and through expressing axon growth inhibitory signals [
29]. Microglia could be highly efficient in clearing myelin debris from demyelination which is critical for oligodendrocyte precursor cell recruitment and their maturation into oligodendrocytes [
30]. Modulating microglia-mediated myelin debris-clearance could be a therapeutic target to improve white matter injury [
31]. In Wallerian degeneration, PTX may regulate different effector functions of macrophages such as migration and myelin phagocytosis [
32]. Therefore, we hypothesis that PTX might promote myelin debris scavenge by microglia. In this study, we found that microglia in PTX group localized around damaged white matter bundles, showing higher capacity of phagocytosing myelin debris after BCAS.
The phagocytic receptor Mertk belongs to TAM receptors, which sense eat-me signals (for instance, phosphatidylserine of lipid-rich myelin debris) from the extracellular space and initiates debris clearance [
33]. The functional roles of Mertk in white matter injury have been addressed in multiple sclerosis (MS) [
18,
34]. Peripheral blood monocyte-derived macrophages of MS patients displayed a decreased Mertk expression and reduced ability to phagocytose human myelin [
18,
34]; Likewise, Mertk knockout mice showed impaired clearance of myelin debris and remyelination in cuprizone-mediated demyelination [
17]. In addition to central nervous system (CNS), Mertk expressed on Schwann cells is critical for myelin clearance in a mouse model of peripheral nerve crush injury [
35]. Generally, no studies concerned functional profile of Mertk in the pathogenesis of ischemic white matter injury. In this present study, microglia-specific Mertk knockout mice demonstrated equal memory compared with their wild littermates in the intact status. After BCAS, Mertk knockout mice showed cognitive decline and white matter damage much more pronounced than their age matched controls, indicating that Mertk gene is related to an increased susceptibility of ischemic white matter injury.
Interestingly, in this study, PTX treatment upregulated Mertk expression in microglia after BCAS and in myelin debris-stimulated microglia. This regulation is dependent on the activation of PPAR-γ, since PPAR-γ antagonist could abolish Mertk upregulation by PTX. PPAR-γ is a type of ligand-activated transcription factor under the class of nuclear receptor superfamily [
36]. Studies have demonstrated that PPAR-γ activation could reduce pro-inflammatory cytokines such as TNF-α and promote white matter integrity after ischemic stroke [
37,
38]. In mouse models of intracerebral hemorrhage, activated PPAR-γ augmented the expression of Mertk and accelerated the phagocytosis of microglia and the clearance of hematoma [
19]. PTX is a phosphodiesterase inhibitor that modulates inflammatory response through increasing intracellular cyclic adenosine monophosphate (AMP) concentration and reducing TNF-α gene transcription [
26]. In this study, we found that PTX treatment facilitated PPAR-γ translocation into the nucleus after BCAS in mice and in primary cultured microglia, which is a critical for myelin clearance.
There are several limitations to the present study. BCAS model has some drawbacks. Firstly, BCAS model could not reflect actual small vessel dysfunction of cerebral small vessel disease (CSVD) in clinic. Secondly, other vascular risk factors that induce white matter lesions, such as hypertension, diabetes, aging and gender could not be reproduced by the BCAS model [
21]. Thirdly, although BCAS model could mimic chronic hypoperfusion, which is the main pathogenic mechanism of white matter lesion and vascular cognitive impairment, it is not identical to the pathogenetic mechanism of the patient. Cerebral blood flow in BCAS model decreased dramatically within the first postoperative day and then slowly gradually returned to a more stable hypoperfusion level [
39], but cerebral hypoperfusion in patients with vascular dementia progressed slowly and lasted for a long time. Thus, in the future, novel and more reliable animal models are needed to study the pathogenesis of vascular dementia as well as the mechanism of PTX. Moreover, BCAS lacks an in vitro model, and we used myelin debris stimulation to mimic the demyelination process in vivo, which could not reflect the complicated pathogenesis observed in vivo. In summary, we demonstrated that PTX may represent an effective therapeutic strategy for patients with VAD, via promoting microglial phagocytosis of myelin debris depending on Mertk.
Materials and methods
Mice
Male C57BL/6J mice weighing 25–28 g were provided by the Animal Center of Nanjing Drum Tower Hospital. Mice were randomly divided into four groups: the sham-operated (sham) group; BCAS group; the saline-treated BCAS (BCAS + Vehicle) group; and the 50 mg/kg PTX-treated BCAS (BCAS + PTX) group. Microglia-specific Mertk knockout mice (Cx3cr1-Cre: Mertk
fl/fl) and wild-type littermate (Mertk
fl/fl) were purchased from the GemPharmatech Co., Ltd. (Nanjing, Jiangsu, China). The mice that met both Mertk flox homozygosity and Cx3cr1-Cre heterozygosity were the microglia-specific Mertk knockout mice (Cx3cr1-Cre: Mertk
fl/fl) (Additional file
3: Fig. S3A, B). All animal experimental protocols were approved by the Standard Medical Laboratory Animals Care. Microglia-specific Mertk knockout mice and wild-type littermate were simultaneously randomly assigned to two groups: the saline-treated BCAS (BCAS + Vehicle) and the 50 mg/kg PTX-treated BCAS (BCAS + PTX) group.
BCAS surgery
The BCAS model was performed according to previous experiments [
25]. Briefly, mice were anaesthetized with 4% chloral hydrate (10 ml/kg) by intraperitoneal injection. After exposing both common carotid arteries (CCAs) through a midline cervical incision, a micro-coil (Inner diameter 0.18 mm, pitch 0.50 mm, total length 2.5 mm, purchased from Sawane Spring Co.) was rotated around the right CCA in a twining way. After 30 min, another micro-coil of the same size was twined around the left CCA.
Cell culture
Primary microglia were isolated and purified from 1-day-old C57BL/ 6J mice as described previously [
21]. In brief, brain membrane was removed and cerebral cortex was digested with 0.25% tripsin EDTA for 10 min. Then, we added the same amount of DMEM medium containing 10% FBS to terminate the digestion. The cells were centrifuged at 800 rpm at 37 ℃ for 10 min. The supernatant was aspirated and cells were cultured in 75 cm
2 flasks for 10–12 days, and then the microglia were separated from mixed primary glia by shaking the flasks and the floating microglia were replanted onto plates for about 48 h. The primary microglia were cultured in 90% DMEM (Invitrogen, Frederick, MD, USA), 10% fetal bovine serum (FBS, Hyclone, Logan, UT, USA) at 37 ℃ in a humidified atmosphere of 5% CO
2. The murine microglia cell line BV2 was prepared from the China Infrastructure of Cell Line Resources (Beijing, China) and its culture medium was the same as primary microglia.
Myelin debris preparation and stimulation
Mice were euthanized according to CO2 asphyxiation guidelines, followed by cervical dislocation. Brains were isolated in 0.32 M sucrose solution and cut into pieces approximately 5 × 5 × 5 mm3. 90 ml homogenized brain solution were gently added the top of the 0.83 M sucrose solution. After centrifuging at 100,000 rpm for 45 min at 4 °C, the crude myelin debris were collected from the interface of the two sucrose densities, dissolved in 35 ml using Tris·Cl buffer solution and homogenized for 30–60 s. After centrifuging at 100,000 rpm for 45 min at 4 °C, the pellets were resuspended in 10–15 ml of Tris·Cl buffer solution and then centrifuged at 100,000 rpm for 45 min. The pellets were resuspended in 5–6 ml of sterile phosphate buffer solution (PBS) and centrifuged at 22,000 rpm for 10 min t 4 °C. The pellets were myelin debris we need. Finally, we resuspend the pellets in PBS to a final concentration of 100 mg/ml for future use. We used 0.01 mg/ml myelin debris to stimulate primary microglia for different time and detected the protein level of Mertk. Moreover, pre-treated with Mertk inhibitor UNC2250 (100 nM), primary microglia were incubated with myelin debris (0.01 mg/ml) for 6 h and then used for flow cytometry. Finally, primary microglia were incubated with myelin (0.01 mg/ml), myelin (0.01 mg/ml) + GW9962 (10 μM), myelin(0.01 mg/ml) + PTX (25 μM), myelin(0.01 mg/ml) + PTX(25 μM) + GW9962 (10 μM) for 6 h.
Open field test
To determine motor function and anxiety-like behavior, a 50 × 50 × 50cm3 box was placed in a room with sound insulation, appropriate light intensity, temperature and humidity. The bottom of the box is divided into 16 small squares. The recording started immediately after the mouse is placed in the center of the box and continued for ten minutes. Total distance traveled and time spent in the center area and corner area were measured and recorded by ANY-maze software (Stoelting, USA).
Novel object recognition test
In the adaptive trail, the mouse was placed in a box without any objects for 15 min on three consecutive days. Twenty-four hours after the last adaption, the training trail started. Two identical objects were placed in the box, and the mice were allowed to freely explore in the box for 10 min. An hour later, one randomly selected familiar object was replaced with a novel one, and the mouse was left to explore for another five minutes. The olfactory trace was removed by 70% alcohol after each trial. Climbing and sitting on an object is not designated as an exploration, but sniffing with the nose or touching with the forepaws is designated as an exploration. The percentage of exploratory preference is as follows: time to explore a novel object/total time to explore both two objects.
Y-maze test
Y-maze test was used to test short-term working memory as previously described [
40]. The Y-maze made of dark polyvinyl plastic consists of three arms (marked as A, B, C). Mice were individually placed on the junction of the arms and move freely over an 8-min period. Spontaneous alternations were manually recorded when mice consecutive entered into all three different arms (i.e., ABC, ACB, CAB, BCA, CBA, and BAC). The percentage of spontaneous alternations (%) was defined as the number of spontaneous alternations in behavior/(the total number of arm entries − 2) × 100.
Morris water maze test
Morris water maze test was performed to evaluate spatial learning and memory in the mice. The apparatus is a 1.6-m-diameter pool surrounded by four curtains. The pool was divided equally into four quadrants, with a hidden platform in the fourth quadrant. Before the experiment, the pool was filled with water 1–2 cm above the platform. During the acquisition phase (Day 1–5), the mice were trained once a day and placed in four different quadrants to search for a platform for one minute at a time. A period of searching time is known as latency time. Mice that have not found the platform in 1 min will be guided by the experimenter to stay on the platform for 30 s. In the probe test (Day 6), the platform was removed and a 60-s exploration experiment was performed to investigate the memory of the mouse on the platform. Latency time, the number of platform crossing and time in target quadrant were recorded by ANY-maze software (Stoelting, USA) automatically.
Western blot analysis
The cells and tissues were homogenized in RIPA buffer containing protease inhibitor. Protein concentrations were determined by the BCA Assay (Thermo Fisher Scientific), equal amounts of which were loaded on gels and transferred to PVDF membrane (EMD Millipore). Membranes blocked with 5% milks were probed with the indicated primary antibodies: Mertk (Abcam, ab95925, 1:1000), PPAR-γ (Abcam, ab45036, 1:500), Histone H3 (Cell Signaling Technology, 4499S, 1:2000), GAPDH (BioWorld, AP0066, 1:5000).
Real-time PCR
Total RNA was extracted from the brains and cells using the TRIzol reagent (Invitrogen) according to the manufacturer’s protocol. Then cDNA was reverse transcribed with a PrimeScript RT Reagent Kit (Takara). Finally, ABI StepOne Plus PCR instrument (Applied Biosystems) with a SYBR green kit (Applied Biosystems) was used to perform quantitative PCR. The expression of target gene levels was normalized to the endogenous control GADPH. The primer sequences were as follows:
TNF-α | Forward | AGTCTGCACAGTTCCCCAAC |
| Reverse | TTAGGAAGACACGGGTTCCA) |
Mertk | Forward | AAGGTCCCCGTCTGTCCTAA |
| Reverse | GCGGGGAGGGGATTACTTTG |
GAPDH | Forward | GCCAAGGCTGTGGGCAAGGT |
| Reverse | TCTCCAGGCGGCACGTCAGA |
Immunofluorescence staining
Briefly, 20-μm-thick coronal sections was fixed with 4% paraformaldehyde and then blocked with 5% BSA and 0.25% Triton X-100. Subsequently, brain sections were incubated with primary antibodies overnight at 4 °C. The primary antibodies were as follows: Rat anti-MBP (Abcam, ab7349, 1:500), Rabbit anti-Iba-1 (Abcam, ab178847, 1:500), Goat-anti-Mertk (R&D systems, AF591, 1:300), Rabbit anti-PPAR-γ (Abcam, ab45036, 1:500). After incubating with appropriate secondary antibodies (Invitrogen) at room temperature for 2 h, sections were observed by a fluorescence microscope (Olympus IX73) or confocal laser-scanning microscope (Olympus FV3000).
Black-gold staining
After preheating 0.3% black-gold and 1% sodium thiosulfate solution to 60 °C, the sections were incubated in black-gold solution at 60 °C for about 12 min until the finest myelin fibers were stained dark red. Then the sections were incubated in sodium thiosulfate solution at 60 ℃ for 3 min. The sections were incubated in cresyl violet solution for 3 min at room temperature. Finally, after alcohol gradient dehydration, xylene clearing for 2 min, and resin mounting, the images were screened by BX Series.
Electron microscopy
After the mice were euthanized, a 1 × 1 × 1 mm
3 piece of tissue from the corpus callosum was taken and fixed in 2.5% glutaraldehyde. The following processing procedure was performed as previously reported protocol [
41]. In brief, the blocks of tissue were incubated in 2% paraformaldehyde in PBS at 4℃ overnight and then fixed by 2% osmium tetroxide in PBS. After alcohol gradient dehydration, samples were embedded in epoxy resin. Subsequently, semi-thin sections which were cut with an ultramicrotome were stained with 1% toluidine blue, uranyl acetate and lead citrate. Then, the images were screened by transmission electron microscopy (Hitachi, HT7800). G-ratio, which is referred to the diameter of the axon/the diameter of the entire myelinated fiber, is used for the indicator.
MRI
Magnetic resonance images were acquired on a 9.4T Bruker MR system (BioSpec 94/20 USR, Bruker) using a 440-mT/m gradient set with an 86-mm volume transit RF coil and a single channel surface head coil. Mice were anesthetized using isoflurane inhalation (2.5–3%) and monitored to maintain constant physiological parameters. Tooth bar and ear bars were used to restrain mice on a mouse holder for data acquisition. T2-weighted images were acquired using the 2D RARE (rapid acquisition with relaxation enhancement) sequence with the following parameters: repetition time (TR): 2500 ms, echo time (TE): 33 ms, field of view (FOV): 20 mm × 20 mm, matrix: 256 × 256 and 22 adjacent slices of 0.7 mm slice thickness. Diffusion-weighted images were acquired with spin-echo echo-planar imaging (SE-EPI) sequence with the following parameters: Two b-values (b = 0 and 1000 s/mm2) along with 30 non-collinear directions, δ = 4.1 ms, Δ = 10.3 ms; TR: 1500 ms, TE: 23.27 ms, FOV: 20 mm × 20 mm, matrix: 128 × 128, and 22 adjacent slices of 0.7 mm slice thickness. Images were converted into NIFTI format using MRIcron. Diffusion data were post-processing using FSL (v.5.0.9) pipeline including corrections for eddy currents and movement artifacts (eddy_correct), rotations of gradient directions according to eddy currents corrections (fdt_rotate_bvecs), brain mask extractions based on b0 images (bet) and FA maps calculations by fitting a diffusion tensor model at each voxel (dtifit). EC and IC were drawn using the itk-SNAP to extract the FA values.
Flow cytometry and microglia isolation
Mice were trans-cardial perfused with 50 ml cold PBS containing 5 IU/ml heparin. Ischemic brains were isolated in 1 × HBSS with 25% glucose and HEPES. After thorough grinding, the total mixture passed through a 70-μm pore filter. Harvested single-cell suspensions were stratified on a 30–70% Percoll gradient (GE Healthcare BioSciences). After centrifuging at 2500 rpm at slow acceleration and deceleration for 20 min, cells at the interface were collected and stained with anti-mouse CD45 (BioLegend, 103114, 1:500), CD11b (Invitrogen, 11-0112-82, 1:500), CD68 (BioLegend, 137007, 1:300), Mertk (Invitrogen, 12-5751-82, 1:500). Among viable cells (FVS-780low), microglia (CD45intCD11b+) were isolated and analyzed by Fluorescence activated cell sorter (FACS) (BD Biosciences, Carlsbad, CA, USA).
Statistical analysis
Statistical analysis was performed using SPSS 18.0 software (IBM Corp, Armonk, NY, USA) and data were expressed as the mean ± standard error of the mean (SEM). The normality of data distribution was analyzed by the Shapiro–Wilk test. For two group comparison, we used Student's t test to analyze if normally distributed continuous variables, otherwise use Mann–Whitney test. For multiple comparison, the data were analyzed by one-way analysis of variance (ANOVA) followed by Bonferroni’s post hoc test if data were normally distributed or by the Kruskal–Wallis test if the data were non-normally distributed. A statistically significant difference was established at p < 0.05.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.