Antibodies anti-glyceraldehyde 3-phosphate dehydrogenase (GAPDH), anti-BDNF and horseradish peroxidase-conjugated goat anti-rabbit secondary antibody were purchased from Abcam. ECL reagent and Hyperfilm ECL were purchased from Amersham Biosciences (Amersham, UK). Pure standards of phenolics and phenolic acid standards, gallic acid (Cat. No. 7384), gentisic acid (Cat. No. G5129), (1)-catechin (Cat. No. C0567), chlorogenic acid (Cat. No. C3878), caffeic acid (Cat. No. C0625), (2)-epicatechin (Cat. No. 855235), p-coumaric acid (Cat. No. C9008), sinapic acid (Cat. No. D7927), benzoic acid (Cat. No. B3250), p-anisic acid (Cat. No. A0631), myricetin (Cat. No. M6760), 3,4,5-trimethoxycinnamic acid (Cat. No. T70408), and quercetin (Cat. No. Q4951) were purchased from Sigma. High-performance liquid chromatography (HPLC)-grade trifluoroacetic acid and methanol were purchased from Sigma.

Blueberry powder was prepared as follows: fresh blueberries were purchased from local supermarkets. Then, 100 g blueberry fruits were blended by using a blender (DS-1, Shanghai. Precision Instruments Co., Ltd, Shanghai, China). The fresh juice was filtered by using a Whatman #2 (Whatman Inc.; Florham, NJ, USA) under vacuum. Clear juice was freeze-dried and powdered by using a miller. A dried powder sample (5 g) was extracted with 50 ml 80% methanol at 30°C for 1 day in a THZ-Q shaking incubator (HZQ-F, Harbin Donglian Electronic Technology Development Co., LTD., Harbin, China) at 100 r/min. The extracts were collected via centrifugation at 10,000 rpm for half an hour and filtered via a Millipore filter (0.22 μm) at room temperature.

Standard stock solution (gallic acid, gentisic acid, [1]-catechin, chlorogenic acid, caffeic acid, [2]-epicatechin, p-coumaric acid, sinapic acid, benzoic acid, p-anisic acid, myricetin, 3,4,5-trimethoxycinnamic acid and quercetin) was prepared in methanol at concentrations of 0.2–1 mg/mL, and filtered through a membrane filter (0.45 mm). Phenolic compounds in blueberry extracts were analyzed using an reversed-phase HPLC with a Waters 600 pump controller equipped with an Eclipse XDB-C18 reversed-phase column (25 cm × 4.6 mm ID × 5 μm) (Supelco; Sigma-Aldrich Co.). Compounds were eluted with a mobile phase solvent A (ddH₂O, pH 2.5 with trifluoroacetic acid) and solvent B (methanol). Gradient elution step started with 100% solvent A at 0 min until 50% solvent A and 50% solvent B for 40 min. Flow rate was 1.0 mL/min. The detection wavelength was set at four different UV wavelengths, 280, 320, and 370 nm because the compound mixtures have different UV spectra.

Phenolic contents in blueberry extracts were identified by using corresponding standards. Standard cures were made as follows: standard concentration was chosen based on final testing of phenolic contents in blueberry juice within the linear range. Regression equation was calculated by using peak area as Y (vertical axis), and the concentration as X (horizontal axis). Each compound was calculated correspond to the standard curve after detecting HPLC peak area and substituting corresponding standard curve equation.

The following inclusion criteria were used: (1) patients diagnosed with CVT based mainly on clinical manifestations and neuroimaging method; (2) diagnosis of hypertension and or chronic hypertension; (3)depressive symptoms measured with a total score ≥ 16 on the 20-item Center for Epidemiologic Studies—Depression Scale (CES-D) (Macedo et al., 2017; Siddaway et al., 2017); (4) Clostridium difficile, Campylobacter spp., Aeromonas spp., Plesiomonas shigelloides, and Shigella spp., and mixture organisms isolated from gastrointestinal tract tissue or drainage fluid; and (5) endoscopic findings indicating infection.

The following exclusion criteria were used: (1) current psychiatric symptoms that necessitated treatment as determined by clinical interview; (2) presence of a major medical problem; (3) undergoing medication that can affect gastrointestinal infection, CVT, and depression symptoms; (4) current involvement in weight loss or psychotherapy; and (5) pregnancy.

Before experiments, all procedures were approved by the ethical committee of the First Hospital of Jilin University. Informed consent was obtained in writing from all patients. A total of 124 patients were selected after excluding the patients did not pass selection criteria. Final selected patients were randomly and evenly assigned into two groups: blueberry extract group (BG, n = 62, received 10 mg of blueberry extract daily) and placebo group (PG, n = 62, received placebo daily, Figure 1). Follow-up period was 6 weeks.

Depression symptoms were measured by a total score ≥ 16 on the 20-item CES-D. Effects of blueberry on depressant symptoms were measured between BG and PG groups. Severity of depression symptoms of CVT patients were compared by using CES-D scores.

Serum lipid profile (triglycerides (TG), total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), and high-density lipoprotein cholesterol (HDL-C)), which is closely associated with CVT progression, was measured. Serum concentration of TG was measured by immunometric assay (Immulite 2000 Thyroglobulin; Diagnostic Products, Los Angeles, CA, USA). Serum concentration of TC was measured by using an automated clinical chemistry analyzer kit (Biosino Biotech, Beijing, China). HDL-C was measured by using an automated chemistry analyzer (Hitachi, Tokyo, Japan). Serum concentration of LDL-C was measured by using a LDL-cholesterol kit (Sekisui Medical Co., Ltd., Tokyo, Japan). Serum malondialdehyde (MDA) was analyzed by using a MDA kit in Nanjing Jiancheng Institute of Biotechnology (Nanjing, China).

Venous thrombi were detected with computed tomography (CT) venography. Brain CT was carried out in a CT 64-slice multi-detector row CT scan (Toshiba Medical Systems Corporation, Otawara, Tochigi, Japan). CT images were obtained with 2.5 mm thickness via the posterior fossa with 5 mm section thickness via supratentorial hemispheres.

Gastrointestinal infection was first observed by endoscopy. After determination, stool specimens were received by our laboratory and were submitted with orders by providers of stool culture. Genome from the specimen aliquots was extracted by using Genomic DNA purification kit (Promega, Madison, WI, USA) according to manufacture's instruction. Bacterial species were identified by using rRNA gene bidirectional sequencing.

Protein was extracted by using Protein Extraction Kit (Millipore, Bedford, MD, USA) according to manufacturer's protocol. Protein concentration was determined by using a BCA protein concentration determination Kit (Beyotime, Beijing, China). Samples were incubated for 10 min at 95°C in boiling buffer (50 mM Tris, pH 8.8, 1% sodium dodecyl sulfate (SDS); 5% 2-mercaptoethanol, 10% glycerol, and 0.002% bromophenol blue). Protein samples were run on12% SDS- polyacrylamide gel electrophoresis (PAGE) and then transferred to nitrocellulose membranes. Membranes were then blocked with 5% nonfat milk (w/v) and incubated with rabbit anti-human anti-BDNF (1:1,000) or anti-GAPDH (1:3,000); and a goat anti-rabbit secondary antibody. Bands were analyzed with software Image J. 1.42q (National Institutes of Health, United States). GAPDH was used to normalize BDNF protein levels.

HT22 hippocampal cells were purchased from American Type Culture Collection ATCC and were cultured in Dulbecco's Modified Eagle Medium (DMEM) with 10% fetal bovine serum (FBS), 100 μg/ml streptomycin, and 100 μg/ml penicillin in 5% CO₂ at 37°C. HT22 cells were transfected with miR-155 mimics and inhibitors (synthesized by Shanghai Sangon Biological Engineering Technology & Services Co., Ltd., Shanghai, China) by using Lipofectamine (Thermofisher, MA, USA). Cells were treated with 200 μg/ml filtered and sterilized blueberry juice.

For immunoblot data, statistical comparisons were carried out using to one-way ANOVA with the diet group as the main factor. Correlation coefficients were calculated using Pearson product-moment correlation coefficient. All the data were expressed as mean ±standard deviation (S.D.) and analyzed using SPSS version 20.0 (SPSS Inc., Chicago, IL, USA).

For HPLC analysis, the maximum absorption of phenolic compounds is at 280 nm for gallic acid, [1]-catechin, [2]-epicatechin, benzoic acid and p-anisic acid; 320 nm for p-coumaric acid, gentisic acid, chlorogenic acid, sinapic acid and 3,4,5-trimethoxycinnamic acid; and 370 nm for myricetin and quercetin (Figure 2A). HPLC analysis showed that blueberry extracts are the main phenolic acids (Figure 2B, gentisic acid, chlorogenic acid, [2]-epicatechin, p-coumaric acid, benzoic acid, p-anisic acid, and quercetin) compared with standard (Figure 2A). Table 1 showed 7 kinds of phenolic compounds detected by HPLC via standards. Detection response values of standard solution had a good linear relationship between phenolic concentration and HPLC peak area, and the limit of detection are low, suggesting that this method is very sensitive. There are 0.18, 0.85, 0.26, 0.72, 0.66, 0.41, and 1.92 mg/g of gentisic acid, chlorogenic acid, [2]-epicatechin, p-coumaric acid, benzoic acid, p-anisic acid, and quercetin in blueberry extracts, respectively (Table 1).

Table 2 showed baseline clinical characteristics among all CVT patients with gastrointestinal infection. No statistically significant difference was observed in body mass index (BMI), daily food intake, baseline depressive symptom severity, and demographic characteristics between BG and PG groups (P > 0.05).

As shown in Figure 3, all outcomes improved significantly between baseline and subsequent follow-ups in the BG group (P < 0.05) but the PG group (P > 0.05). Improvements from baseline to 3 months included the following: CES-D from mean (S.D.) of 36.58 (6.54) to 18.76 (9.28) (P < 0.001) in the BG group. Comparatively, the PG group showed no improvement from baseline to 3 months: CES-D from mean (S.D.) of 37.29 (7.13) to 36.54 (7.98) (P > 0.05). All patients did not use antidepressants or psychotherapy during these 3 months. After 3-month therapy, CES-D scores for depressant symptoms were reduced 48.7% when compared with before therapy in the BG group. Comparatively, there was no change for the CES-D scores for depressant symptoms in the PG group.

CT venography is a rapid, simple, and accurate technique for detecting CVT. The technique provides detailed information for CVT measurement. All patients were diagnosed with CVT as indicated by arrows in Figure 4. One disadvantage of CT venography is difficulty in reconstructing maximum intensity projection image from image data, which require the subtraction of bone adjacent to the venous sinus.

Before therapy, no statistically significant difference was noted between BG and PG groups (P > 0.05). Gastrointestinal symptoms were mainly caused by infection with Clostridium difficile, Campylobacter spp., Aeromonas spp., Plesiomonas shigelloides, and Shigella spp., and mixture infection. After 3 months of blueberry therapy, the BG group presented lower cases of gastrointestinal infection than the PG group (P < 0.05, Table 3). After 1-month therapy, gastrointestinal infection rates were reduced by 33.9% in the BG group while the rates were only reduced by 8.1% in the PG group. After 2-month therapy, gastrointestinal infection rates were reduced by 74.2% in the BG group while the rates were reduced by 16.1% in the PG group. After 3-month therapy, gastrointestinal infection rates were reduced by 83.9% in the BG group while the rates were reduced by 21.0% in the PG group.

Blueberry may inhibit gastrointestinal infection via inhibition of pathogens. To explore this possibility, the effects of blueberry on the growth of pathogens isolated from CVT patients were measured here. However, blueberry failed to inhibit the growth of bacterial pathogens even after using 10 mg/ml blueberry juice was used. Thus, blueberry extract cannot inhibit pathogens Clostridium difficile, Campylobacter spp., Aeromonas spp., Plesiomonas shigelloides, and Shigella spp., and mixture infection directly even high concentration was used (data were not shown). However, gastrointestinal infection can be inhibited by blueberry via improvement of autoimmune activity.

Long-term blueberry consumption improved lipid profile in CVT patients from the BG group when compared with PG group. In the BG group, serum levels of TG, TC, LDL-C, and MDA decreased, whereas serum level of HDL-C changed after 1 month (Table 4). However, no statistically significant difference between the two groups (P > 0.05, Table 4). Serum lipid profile improved further in the BG group after 2 months, and statistically significant differences were observed between the two groups (P < 0.05, Table 4). Furthermore, the lipid profile remarkably improved in the BG group after 3 months, and the two groups featured statistically significant differences (P < 0.05, Table 4). These findings demonstrate that blueberry improved lipid profile in CVT patients. Therefore, blueberry treatment changed the serum levels of lipid profiles in CVT patients. Blueberry dietary supplementation not only performs antidepressant-like function but also improves lipid profile of CVT patients. Present findings demonstrate that blueberry treatment exerts antidepressant effect and regulates lipid profile simultaneously.

Real-time quantitative polymerase chain reaction (qRT-PCR) analysis showed t no statistically significant difference for serum miR-155 levels between BG and PG groups before therapy (P > 0.05). After 3 months of blueberry therapy, miR-155 level was higher in the BG than PG group (P < 0.05, Figure 5A). Relative level of miR-155 was lowest level when it was silenced whereas the level was highest when it was overexpressed (Figure 5B). Meanwhile, no statistically significant difference in BDNF mRNA levels between BG and PG groups before the therapy (P > 0.05, Figures 5C,E). After 3-month blueberry therapy, the BG group exhibited higher BDNF mRNA levels than the PG group (P < 0.05, Figure 5C). Cellular level test showed that miR-155 overexpression or inhibition can increase or reduce mRNA levels of BDNF (Figure 5D). A significant positive correlation was observed between miR-155 or BDNF levels and blueberry treatment (P < 0.05, Figure 5D). Blueberry treatment cannot affect the mRNA levels of BDNF when miR-155 was overexpressed or inhibited (P > 0.05, Figure 5D). Western Blot analysis showed no statistically significant difference in BDNF levels between BG and PG groups before the therapy (P > 0.05, Figure 5E). After 3 months of blueberry therapy, BDNF level was higher in BG than the PG group (P < 0.05, Figures 5E).Additionally, miR-155 overexpression or inhibition increased or reduced protein levels of BDNF (P > 0.05, Figures 5F). Blueberry treatment cannot affect t protein levels of BDNF when miR-155 was overexpressed or inhibited. All these results suggest that blueberry increases BDNF level via miR-155.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.