Serum samples of 60 patients with cerebral microbleeds, including 32 males and 28 females, were collected. Furthermore, serum samples of 60 healthy subjects of similar age were collected, including 29 males and 31 females. Fasting blood samples were drawn in the morning from all subjects after an 8-h overnight fast. All subjects signed informed consent. Patients who met the following criteria (29) were eligible for the study: i) patients whose age was ≥18 years but <65 years; ii) patients who were diagnosed with brain microbleeds (CMB) by MRI in accordance with the diagnostic criteria of CMB; and iii) patients' family agreed to participate in the study and signed an informed consent form. Patients who met the following criteria were excluded from this study: i) patients who did not take an MRI exam; ii) patients who had intracerebral hemorrhage due to abnormal structures in the brain; iii) patients who were experiencing parenchymal hemorrhage due to intracranial aneurysm rupture; iv) patients who had cerebral bleeding due to traumatic brain injury; v) patients who had circulatory system diseases; vi) patients who had moyamoya disease; vii) patients who were pre-treated with anticoagulant therapy; and viii) patients who had severe respiratory diseases, advanced cancers, severe liver and kidney dysfunction, severe heart dysfunction, hyperthyroidism or severe endocrine system diseases. The study was approved by the Ethics Committee of The Third Affiliated Hospital of Qiqihar Medical University (Qiqihar, China).

Spontaneously hypertensive rats (SHR) were purchased from Shanghai Experimental Animal Center affiliated with Chinese Academy of Sciences. The animals were given a batch number of SYXK Black 2008004 and were raised and reproduced in our institution's laboratory animal center.

Materials and reagents used in this study were purchased from commercial sources: Human cystatin C kit (item no. ab179883) from Abcam; CysC drug from Enzo Life Sciences; RIPA lysis buffer from Beyotime Biotechnology Co., Ltd.; sucrose, glucose, KCl, NaHCO₃, NaH₂PO₄, CaCl₂ and MgCl₂ from Sigma-Aldrich; Merck KGaA; rat brain stereotaxic device from Anhui Zhenghua Biological Instrument Co., Ltd.; primary antibodies to cystatin C (item no. ab109508), p-extracellular signal-regulated kinase 1/2 (ERK1/2) (item no. ab223500), synapsin I (item no. ab8), and p-synapsinI-S549 (item no. ab119370) from Abcam; and β-actin primary antibody (item no. 66009-1-Ig) from Proteintech.

Venous blood was drawn from subjects into a 15 ml tube after more than 8 h fast. Immediately after collection, the blood was centrifuged at 1,006.2 × g at room temperature (25–28°C) for 20 min. The supernatant was carefully collected, aliquoted into 200 µl, and stored at −80°C. Multiple freeze-thaw cycles were avoided. In accordance with the CysC kit manual, absorbance of the sample was measured at 450 nm using a microplate reader, and the expression level of CysC in the serum was calculated.

Following conditions were provided in SHR housing. All the SHRs were housed in a specific pathogen-free room maintained at 18–26°C with a daily temperature difference of ≤3°C. The housing was kept at a relative humidity of 40–70%, a noise level of ≤60 dB, and a 12 h light (150–300 Lux)/12 h dark cycle. The animals were maintained with free access to food and water. All experimental procedures were consistent with experimental animal welfare and ethical principles.

The animals were randomly divided into 4 groups: sham surgery control group (sham), model group (CMB), model + empty vector control group (CMB + vehicle), and model + cystatin C overexpression group (CMB + CysC). Of each group, 12 animals were used for behavioral experiment, 6 for electrophysiological experiment, and the remaining 6 for molecular biology experiments.

Following protocol was for establishing the animal model. Forty-eight specific pathogen-free SHRs at 10 weeks of age were selected, of which 36 were randomly chosen for the rat model with brain microbleeds. A published protocol was followed (30). Rats were first restricted from food and water for 8 h prior to the experiment. The rats were anesthetized with 1% sodium pentobarbital via intraperitoneal injection at 0.6 ml/10 g. After induction of deep anesthesia, the rats were fixed on a stereotactic instrument. The skin area in the head was prepared and disinfected for surgery. Two holes were drilled 5 mm behind the anterior fontanelle, one at 2 mm left and the other at 2 mm right of the midline. To elicit microbleed formation, the rat brain was pierced perpendicularly to a depth of 4 mm from the dura mater using stainless steel needles of 474 and 159 µm in diameter, respectively, at both sides of the midline. Bone wax was used to patch the holes, and the skin over it was closed with sutures. The other 12 rats were used for sham surgery which was the same as the procedure described above except that there was no piercing of stainless-steel needles into the brain.

To elicit cystatin C overexpression, following steps were performed. The drug CysC was dissolved in physiological saline to a final concentration of 50 µg/ml (31). The lateral ventricle was injected with 10 µl of the drug solution. The same amount of saline was injected into the brain of rats in the negative control group. The lateral ventricle injection was at 0.8 mm behind the anterior fontanelle, 1.5 mm distal to the midline, and 4.5 mm deep.

Behavioral experiment was conducted 5 days after rat model creation. Y-maze test was performed using a Y-maze apparatus. The Y-maze as a labyrinth was composed of three arms of equal length, i.e. arm I, arm II and arm III, as well as their junction area. The bottom of the box was covered with an electric grid (0.2 cm in wire diameter, 14 cm in length and 1 cm in grid spacing). There were multiple buttons on the Y-maze control panel such as I, II, III and 0, as well as a voltage adjusting knob and a time control knob. When button I, II or III was pressed, a signal light corresponding to that arm became on, indicating the arm was not energized and therefore was a safe zone, while the other two arms and the junction area were dark and energized and therefore were non-safe zones. When the button 0 was pressed, the junction area was energized while the three arms were not. Five seconds after the signal light was on, electricity (60 v) was on in the grid. Each rat was tested 20 times a day for 2 days. Day 1 test was for study, and day 2 test was for memory retention. Safe zones were changed in a random and alternating way. In the beginning of the test, two arms including the one where the rat hid were energized. The animal would escape to the light area after receiving an electric shock. A test was completed when the rat stayed in the safe zone for 30 sec. Then next test started from where the rat was. It would be a correct reaction if the rat fled directly to the safe zone after an electric shock or ran to the safe zone within 10 sec after the shock. It would be an incorrect reaction if the rat escaped to any other arm without lights.

Testing indicators are listed below: i) Error number (EN) was defined as the number of icorrect reactions in all reactions. ii) Total reaction time (TRT) was defined as sum of the time required to complete correct reactions and incorrect reactions. Reaction time referred to the time taken from when the signal light was on to when the rat escaped to the safe zone. iii) Active avoidance rate (AAR) was defined as the percentage of times a rat completed its escape reaction within 5 sec after the light was on but the arm was not energized.

Artificial cerebrospinal fluid (ACSF) for anatomical use was made containing 210 mM sucrose, 12 mM glucose, 2 mM KCl, 24 mM NaHCO₃, 1 mM NaH₂PO₄, 0.5 mM CaCl₂ and 7 mM MgCl₂. A mixed gas of 5% CO₂/95% O₂ was bubbled through the ACSF for at least half an hour at an osmotic pressure of 310–320 Osm in order to saturate it with oxygen. The pH was adjusted to 7.4. A rat was rapidly decapitated, and the head was placed in ice-cold artificial cerebrospinal fluid where the brain tissue was dissected out. The well-extracted brain tissue was transferred to fresh ice-cold and oxygen-saturated artificial cerebrospinal fluid. The brain tissue was trimmed with a scalpel while it was completely submerged in ACSF. The trimmed brain tissue was fixed on a sample tray, and a 2% agar block was used to hold the ventral side of the brain to prevent the slice from tilting. Ice-cold artificial cerebrospinal fluid for anatomical use was poured into the sample tray to completely submerge the brain tissue. The sample tray together with the brain tissue was mounted on a microtome. A mixed gas of 5% CO₂/95% O₂ was bubbled through the ACSF. After a blade was installed and parameters such as slicing speed and oscillation frequency were set, the brain tissue was cut into coronal slices of 400 µm thickness.

MED64 planar microelectrode array recording system was purchased from Alpha Med Science, Japan and used for electrophysiological recordings (32). ACSF for recording use was prepared containing 120 mM NaCl, 27 mM NaHCO₃, 20 mM glucose, 1 mM NaH₂PO₄, 3 mM KCl, 2.6 mM CaCl₂ and 1 mM MgCl₂. A mixed gas of 5% CO₂/95% O₂ was bubbled through the ACSF for half an hour at an osmotic pressure of 310–320 Osm, followed by pH adjustment to 7.4. The well-cut brain slices were incubated in the ACSF for recording use at room temperature for at least one hour to restore brain slice activity. A slice was selected, and after it was placed on the MED64 probe, the slice was infused continuously. At this stage it was ready for recording. A stimulation site in the hippocampal CA3 region was selected. All the sites in the CA1 region were for recording. An input-output (I-O) curve was constructed. The current corresponding to 30–50% of the maximum excitatory post-synaptic potential amplitude was chosen as the subsequent stimulation intensity. High-frequency stimulation was delivered to the brain slice after baseline recording. The change in amplitude was recorded during the 60 min period after stimulation.

Cells were lysed with ultrasound in RIPA lysis buffer. Protein concentration was measured using the BCA kit. Samples were subjected to electrophoresis, membrane transfer, blocking, incubation with antibodies and image development. Incubation with primary antibodies, i.e. anti-cystatin C (dilution factor, 1:20,000), p-ERK1/2 (dilution factor, 1:400), synapsin I (dilution factor, 1:1,000), p-synapsinI-S549 (dilution factor, 1:1,000), and β-actin (dilution factor, 1:10,000), was maintained at 4°C overnight. After washing, the membrane was incubated with corresponding secondary antibody (dilution factor, 1:4,000), followed by exposure and image development. Quantity One software was used to analyze the image in Grayscale format.

Experimental data were expressed as mean ± standard deviation (mean ± SEM). Statistical analysis was performed using SPSS 16.6 statistical software. The results were analyzed using one-way ANOVA. A Bonferroni test was used for comparison between groups. A difference was statistically significant at P<0.05, P<0.01 and P<0.001. GraphPad Prism 5 software was used for graph drawing.