Deletion of B-cell translocation gene 2 (BTG2) alters the responses of glial cells in white matter to chronic cerebral hypoperfusion

Subcortical ischemic vascular dementia, one of the major subtypes of vascular dementia, is characterized by lacunar infarcts and white matter lesions caused by chronic cerebral hypoperfusion. In this study, we used a mouse model of bilateral common carotid artery stenosis (BCAS) to investigate the role of B-cell translocation gene 2 (BTG2), an antiproliferation gene, in the white matter glial response to chronic cerebral hypoperfusion. Btg2−/− mice and littermate wild-type control mice underwent BCAS or sham operation. Behavior phenotypes were assessed by open-field test and Morris water maze test. Brain tissues were analyzed for the degree of white matter lesions and glial changes. To further confirm the effects of Btg2 deletion on proliferation of glial cells in vitro, BrdU incorporation was investigated in mixed glial cells derived from wild-type and Btg2−/− mice. Relative to wild-type mice with or without BCAS, BCAS-treated Btg2−/− mice exhibited elevated spontaneous locomotor activity and poorer spatial learning ability. Although the severities of white matter lesions did not significantly differ between wild-type and Btg2−/− mice after BCAS, the immunoreactivities of GFAP, a marker of astrocytes, and Mac2, a marker of activated microglia and macrophages, in the white matter of the optic tract were higher in BCAS-treated Btg2−/− mice than in BCAS-treated wild-type mice. The expression level of Gfap was also significantly elevated in BCAS-treated Btg2−/− mice. In vitro analysis showed that BrdU incorporation in mixed glial cells in response to inflammatory stimulation associated with cerebral hypoperfusion was higher in Btg2−/− mice than in wild-type mice. BTG2 negatively regulates glial cell proliferation in response to cerebral hypoperfusion, resulting in behavioral changes.


Background
Subcortical ischemic vascular dementia, one of the major subtypes of vascular dementia, is characterized by lacunar infarcts and white matter lesions caused by chronic cerebral hypoperfusion [1][2][3]. Chronic cerebral hypoperfusion is modeled in rodents by bilateral common carotid artery stenosis (BCAS), in which micro-coils are placed bilaterally on the common carotid arteries [4][5][6]. BCAS model mice exhibit apparent glial activation in addition to white matter lesions, providing important insight into the mechanism of subcortical ischemic vascular dementia [4]. Brain inflammation, defined by glial activation, has attracted a great deal of attention in stroke research [7][8][9][10]. Notably in this regard, proliferation of glial cells, including astrocytes, is observed after ischemic injury in animal models [11][12][13].
B-cell translocation gene 2 (BTG2, also called PC3/ Tis21) was identified as an anti-proliferative protein regulated by p53 activation [14] and was initially described as an immediate early gene induced by growth factors and tumor promoters in PC12 and Swiss 3T3 cells [15,16]. BTG2 belongs to the BTG/TOB family, whose members regulate cell-cycle progression, apoptosis, and differentiation [17,18]. In particular, BTG2 is involved in several biological processes including cell-cycle progression [14], induction of cellular senescence [19], cellular differentiation including neurogenesis [20], hematopoiesis [21][22][23][24], the genotoxic stress response [14,25,26], and integrated stress responses such as the response against oxidative stress [18,27]. Btg2 is broadly expressed in multiple organs, including the kidney, lung, prostate, pancreas, thymus, skeletal muscle, and central nervous system [18,[28][29][30]. In the mouse brain, Btg2 is highly expressed in astrocytes, microglia, and endothelial cells (https://www.brainrnaseq. org). Notably, Btg2 mRNA is induced in the brain response to hypoxia-ischemia [31] and in the brains of AD model mice with obesity-linked diabetes but not in those of AD model mice or mice with obesity-linked diabetes [32]. Moreover, Btg2 is also upregulated during chronic cerebral hypoperfusion in BCAS-treated mice [33]. However, the roles of BTG2 in chronic cerebral hypoperfusion remain unclear. To answer this question, in this study, we generated Btg2 −/− mice, performed BCAS surgery, and analyzed locomotor activity, cognitive behavior, white matter lesions, and glial cells in the white matter.

Animals
All mice shared the same genetic background (C57BL/ 6J). The mice were group housed without enrichment structures in a specific pathogen-free environment in ventilated cages and were used in the experiments according to the Guideline for the Care and Use of Laboratory Animals of our research facilities. Several strains of Btg2 −/− mice were produced by electroporation using the CRISPR/Cas9 system [34] at the Institute of Experimental Animal Sciences, Graduate School of Medicine, Osaka University. Briefly, the following reagents were purchased: Cas9 mRNA, T7-NLS-hCas9-polyA (RIKE N BRC #RDB13130), and guide RNA (gRNA). To design the gRNA sequence (5′-ACATGCTCCCGGAGATCGCC-3′), a software tool (http://crispr.mit.edu/) for predicting unique target sites in the mouse genome was used. The specificity of CRISPR/Cas9-mediated DNA cleavage is conferred by the 20-base sequence of the single-guide RNA (sgRNA) targeting the chromosomal mouse Btg2 genomic sequence. The CCG trinucleotide, the protospacer-adjacent motif (PAM), which is recognized by Cas9, resides 16 bases downstream from the start codon of the mouse Btg2 genomic sequence. Pronuclear-stage mouse embryos were prepared by thawing frozen embryos (CLEA Japan Inc., Japan). For electroporation, 50-100 embryos at 1 h after thawing were placed into a chamber with 40 μl of serum-free media (Opti-MEM, Thermo Fisher Scientific, USA) containing 400 ng/μl Cas9 mRNA and 200 ng/μl gRNA and then electroporated with a 5-mm gap electrode (CUY505P5 or CUY520P5 Nepa Gene, Chiba, Japan) in a NEPA21 Super Electroporator (Nepa Gene, Chiba, Japan). After introduction of Cas9 and gRNA, mouse embryos were transferred to the oviduct of a female surrogate anesthetized with sevoflurane (Mylan Pfizer Japan Inc.). We obtained a line of Btg2 −/− mice with an 11-bp deletion of the Btg2 genomic sequence. Genotype was confirmed by PCR using the following primers: forward: 5′-CTCCCCGAGTGGTATGAA AG-3′, reverse: 5′-TCAAGGTTTTCAGTAGGGCG-3′. When PCR products were separated by 6% acrylamide gel electrophoresis, two products were detected: 220 bp, indicating the Btg2 deletion, and 231 bp, indicating the wild type. Sanger sequencing of PCR products from wild-type and Btg2 −/− mice was performed using the same set of primers (Japan GENEWIZ, Saitama, Japan). To exclude offtarget effects, heterozygous Btg2 +/− mice were backcrossed for six generations to wild-type C57BL/6J mice.

BCAS surgery
Surgical procedures for BCAS were described previously [4]. Briefly, mice were habituated in the surgery room with free access to food and water for 3-5 days before surgery at 26°C under a 12-h/12-h light/dark cycle. Nine-to eleven-week-old wild-type mice and Btg2 −/− littermates (body weight, 22-28 g) underwent BCAS or sham operation. Before surgery, mice were intraperitoneally injected with a combination of three anesthetics: 0.3 mg/kg medetomidine hydrochloride (Nippon Zenyaku Kogyo, Fukushima, Japan), 4.0 mg/kg midazolam (Astellas, Tokyo, Japan), and 5.0 mg/kg butorphanol tartrate (Meiji Seika Pharma, Gifu, Japan). For sham-operated mice, the common carotid artery (CAA) was exposed and carefully freed from its sheath. For BCAS-treated mice, an 0.18mm inner diameter micro-coil (Samini, Shizuoka, Japan) was then attached bilaterally to the CAA. After surgery, 3.0 mg/kg atipamezole hydrochloride (Nippon Zenyaku Kogyo), an antagonist of medetomidine hydrochloride, was injected intraperitoneally to promote safe awakening. To maintain body temperature after surgery, the mice were warmed with a heating pad at 37°C.

Behavioral tests
To assess the effect of Btg2 deletion on the behavior of the BCAS model, we performed the open-field and Morris water maze tests. One month after the surgery, the open-field test was performed: mice were placed in a cubical plastic box (30 × 30 × 30 cm 3 , width × length × height) and allowed to roam freely for 15 min. One and a half months after the surgery, the Morris water maze test was performed to test spatial learning and memory, as previously described [35,36]. The test was conducted in a circular pool filled with water adjusted to ambient temperature. For hidden platform training, a transparent platform was submerged 1.5 cm below water level. The pool was located in a test room in which there were many external cues. On day 1, a preparative session was carried out in which mice were given 90 s for free swimming without the platform, followed by two training trials (1 session) with the platform. From day 2 to day 5, four training trials (two sessions) were performed each day. During each trial, the mice were released from four pseudo-randomly assigned starting points and allowed to swim for 90 s. The inter-trial interval was approximately 10 min, and the intersession interval was 2 h. After mounting the platform, the mice were allowed to remain there for 10 s and were then placed in a holding cage with a heating lamp until the start of the next trial. If a mouse failed to find the platform, it was guided to the platform and allowed to rest on the platform for 10 s. The probe test was performed 24 h after the last hidden platform training. In the probe test, the hidden platform was removed, and the mouse was released from the opposite quadrant and allowed to swim freely for 60 s. In the visible platform test performed after the last probe test on the same day, the platform was elevated above the water surface. The mice underwent three trials lasting 90 s. All experiments were conducted at approximately the same time each day. An overhead camera was used to track the movement of mice in both the open-field and Morris water maze tests, and the video was analyzed using the ANY-maze software (Muromachi Kikai, Tokyo, Japan).

Tissue collection and preparation
Seven weeks after surgery, mice were anesthetized with isoflurane (Fujifilm Wako Pure Chemicals, Osaka, Japan) and perfused transcardially with phosphate-buffered saline (PBS) plus complete protease inhibitor cocktail (Roche Diagnostics, Basel, Switzerland). After perfusion, the brain was removed from the skull and divided along the sagittal plane. The right hemisphere was divided into the cortex plus hippocampal areas and other areas, frozen in liquid nitrogen, and stored at − 80°C until the biochemical analysis. The left hemisphere was fixed with 4% paraformaldehyde/PBS overnight and embedded in paraffin for histological analysis. For biochemical analysis, tissues were pulverized in a BioPulverizer (BioSpec Products, Bartlesville, OK, USA) that had been prechilled on dry ice, and then stored at − 80°C.

ELISA assay
Protein levels of glial fibrillary acidic protein (GFAP) and CD11b in the sham/BCAS-treated wild-type and Btg2 −/− mice brain were determined by ELISA as previously described [38]. In brief, pulverized brain samples were homogenized on ice using a Polytron homogenizer (Kinematica, Luzon, Switzerland) in ice-cold RIPA lysis buffer (10 ml/g of wet weight; MilliporeSigma, Burlington, MA, USA) containing 0.1% SDS and complete protease inhibitor cocktail (Roche). After centrifugation at 100,000g for 1 h at 4°C, the supernatant was aliquoted and analyzed by ELISA with the appropriate sets of antibodies. Colorimetric quantification was performed on an iMark plate reader (Bio-Rad) after incubation with horseradish peroxidase (HRP)-linked Avidin-D (Vector Laboratories, Burlingame, CA, USA) or anti-rabbit IgG (H+L), HRP conjugate (Promega, Madison, WI, USA) and 3,3′,5,5′-tetramethylbenzidine substrate (Sigma Aldrich, St. Louis, MO, USA). Protein concentrations in each sample were performed using the Protein Assay BCA kit (Fujifilm Wako Pure Chemical).

Histological preparation
Paraffin-embedded brain samples were sectioned at 5μm thickness by microtome (SM2000R, Leica Microsystems, Wetzlar, Germany). Four non-adjacent sections including the hippocampal region were mounted per slide on MAS (Matsunami Adhesive Slide)-coated glass slides (Matsunami Glass Ind., Osaka, Japan). These tissue samples were kept at room temperature prior to histological examination.

Klüver-Barrera staining
Paraffin-embedded sections were deparaffinized, hydrated, and immersed in Luxol Fast Blue solution (Muto Chemical, Tokyo, Japan) at 65°C overnight. Sections were chilled at room temperature, rinsed in 95% ethanol, treated with 0.1% lithium carbonate, rinsed in 70% ethanol, counterstained with 0.1% Cresyl Violet Acetate (Muto Chemical) containing 10% acetic acid, and mounted with Entellan New (Merck Millipore). Sections were analyzed on a BZ-X810 microscope (Keyence, Osaka, Japan). Using images of the white matter area (optic tract and corpus callosum), the severity of white matter lesions was classified into four grades as described by Wakita et al. [40]: normal (grade 0), disarrangement of nerve fibers (grade 1), formation of marked vacuoles (grade 2), and disappearance of myelinated fibers (grade 3). Grading was performed by three individuals in a blinded manner.

Statistical analyses
All statistical analyses were performed using the JMP Pro software (version 14.2, SAS Institute, Cary, USA). In the animal experiments, the effects of each genotype with or without BCAS were analyzed through (1) linear regression models adjusting for sex, followed by Tukey's HSD test (when both sexes were included), or (2) oneway ANOVA followed by Tukey-Kramer test or Dunnett's multiple comparison test (when only one sex was included). Morris water maze test results were analyzed by repeated-measures one-way ANOVA followed by Tukey-Kramer test. For gene expression analyses of the BTG/TOB family, Student's t-test was used to compare the two genotypes. In cellular experiments, the effects of genotype with or without treatment were analyzed by one-way ANOVA followed by Tukey-Kramer test. We used standard error of the mean (SEM) for adjusted values by sex in Figs. 3, 4, and 5, and Supplementary  Figures 1, 2, and 3, and standard deviation (SD) for nonadjusted values in Figs. 2, 7, and 8, and Supplementary  Figures 4 and 5. p-values less than 0.05 were considered significant. The individual effect of Btg2 deletion or BCAS treatment was analyzed by two-or three-way ANOVA (Supplementary Table 1).

Development of Btg2 −/− mice
To assess the role of Btg2 in chronic cerebral hypoperfusion, we generated Btg2 −/− mice by genome editing (Fig. 1a-c). RT-PCR revealed that Btg2 mRNA was absent from Btg2 −/− mice (Fig. 1d). Because BTG2 belongs to the BTG/TOB family, we asked whether the expression of other members of the BTG/TOB family was upregulated by a compensatory mechanism in Btg2 −/− mice. However, expression levels of other members of the BTG/TOB family were not upregulated in Btg2 −/− mice (Fig. 2), suggesting deletion of Btg2 is not compensated by upregulation of other family members.
Behavioral characteristics of BCAS-treated Btg2 −/− mice We treated Btg2 −/− mice and littermate wild-type mice with or without BCAS. One month after the surgery, behavioral changes were initially assessed by the open-field test. Total distance traveled (Fig. 3a) and average speed (Fig. 3b) were significantly higher in BCAS-treated Btg2 −/− mice (p < 0.01). The ratio of travel distance in the outer or inner part of the open field to total field, which is an index of anxiety or exploratory behavior, respectively, did not significantly differ among groups (Supplementary Fig. 1).
We then examined spatial learning ability using the Morris water maze test; the experiment consisted of 5day training trials followed by one probe test. We observed no significant difference in the results of the visible platform test (data not shown). Sham-operated wild-type, BCAS-treated wild-type, and sham-operated Btg2 −/− groups exhibited a significant reduction in escape latency on the fifth day relative to the first day (p < 0.01, p < 0.01, and p < 0.001, respectively, Fig. 3c). By contrast, in the BCAS-treated Btg2 −/− group, we observed no such reduction in escape latency (p = 0.3306, Fig. 3c), suggesting that the spatial learning ability of BCAS-treated Btg2 −/− mice was impaired relative to other mice. In the probe test, we observed no significant difference among groups in the time spent in the target quadrant ( Supplementary Fig. 2).

Histological observations of white matter in BCAS-treated
Btg2 −/− mice To assess the effect of Btg2 deletion on white matter lesions after BCAS treatment, we performed histological examinations of white matter by Klüver-Barrera staining following behavioral assessments. The appearance of myelin fibers in the white matter (corpus callosum and optic tract) was graded for the severities of lesions as described by Wakita et al. [40]. Although the severities of lesions in the corpus callosum and optic tract were significantly increased by BCAS treatment, we observed no significant difference between wild-type and Btg2 −/− mice with or without BCAS (Fig. 4).

Assessment of glial cell markers in BCAS-treated
Btg2 −/− mice To assess the effects of Btg2 deletion on the response of glial cells to BCAS, we performed immunohistochemical analysis using anti-GFAP, anti-Iba1, and anti-Mac2 antibodies (Fig. 5). The GFAP-immunoreactive area in the optic tract was significantly increased by BCAS. Moreover, BCAS treatment also increased the GFAP-and Mac2-immunoreactive areas in Btg2 −/− mice relative to wild-type mice (Fig. 5b). Iba1-immunoreactive areas were larger in BCAS-treated Btg2 −/− mice than in shamoperated wild-type mice. In the corpus callosum, we observed similar trends in GFAP and Iba1 immunoreactivities in the optic tract following BCAS treatment, but no significant difference in the GFAP-and Mac2immunoreactive areas in mice harboring the Btg2 deletion ( Supplementary Fig. 3). To investigate the relationship between Mac2-and GFAP-or Iba1-positive cells, we performed double-staining (Fig. 6). Although most Mac2-positive cells did not contain GFAP or Iba1 immunoreactivities, co-localization of Mac2 and GFAP or Iba1 immunoreactivities was observed in some cells ( Fig. 6a and b, respectively) (white arrowheads).
To analyze the changes in these glial cells by other means, we performed real-time PCR to quantify mRNA levels of the following representative markers: Gfap (astrocytes), Cd11b (microglia, monocytes, and macrophages), Trem2 and Dap12 (disease-associated microglia [DAM]), Cd45 (pan-hematopoietic), Cd68 and F4/80 (macrophages), and Cd14 (monocytes and macrophages). We analyzed both the cortex and hippocampal areas. Although we analyzed fewer mice than in the histochemical analyses, we could still see that the level of Gfap mRNA was significantly increased by BCAS in Btg2 −/− mice (p = 0.0484, vs. sham-treated wild-type mice; p = 0.0362, vs. BCAS-treated wild-type mice; p = 0.0406, vs. sham-treated Btg2 −/− mice; Dunnett's multiple comparison test). Cd45 mRNA was also significantly higher in BCAS-treated Btg2 −/− mice than in sham-treated wild-type mice (p = 0.0490). On the other openfield test. Data are adjusted means ± SEM. **p < 0.01, as determined by Tukey's HSD test. c Average escape latency in each day during the 5-day training trial for the Morris water maze test. Data are adjusted means ± SEM. **p < 0.01 and ***p < 0.001 compared with the scores from day 1 and day 5, as determined by repeated-measures one-way ANOVA followed by Tukey's HSD test. n.s., not significant hand, mRNA levels of Cd11b, Trem2, Dap12, Cd68, F4/ 80, and Cd14 (p = 0.1606, p = 0.340, p = 0.307, p = 0.1634, p = 0.7505, and p = 0.0923, respectively; one-way ANOVA) exhibited no significant difference, although BCAS-treated Btg2 −/− mice generally expressed higher levels of these mRNAs than the other groups (Fig. 7).
We also analyzed the protein levels of these glial cell markers in the same area, using ELISAs to confirm realtime PCR results. Although no significant difference was observed, likely due to the small number of mice examined, the level of GFAP tended to be higher in the BCAS-treated Btg2 −/− mice than in other groups, consistent with the real-time PCR results. The level of the microglial marker CD11b did not significantly differ among groups (Supplementary Fig. 4).

Glial cell proliferation in Btg2 −/− mice in response to inflammatory stimuli
The results obtained with BCAS models suggested that Btg2 deletion could increase the proliferation of glial cells during cerebral hypoperfusion. To further confirm the effects of Btg2 deletion on glial cell proliferation in vitro, we performed BrdU incorporation assays in mixed glial cells derived from wild-type and Btg2 −/− mice. To mimic the pro-inflammatory conditions during cerebral hypoperfusion, mixed glial cells were treated Fig. 4 Histological observations of white matter in sham/BCAS-treated wild-type and Btg2 −/− mice. a, b Corpus callosum (a) and optic tract (b) subjected to Klüver-Barrera staining. Black arrowheads show vacuolation in myelin fibers. Scale bar = 20 μm. c, d Severity of white matter lesion (WML) in the corpus callosum (c) and optic tract (d). All mice were male. Data are adjusted means ± SEM. **p < 0.01 and ***p < 0.001, as determined by Tukey's HSD test with or without inflammatory stimuli (LPS and IFNγ), as reported [43] (Fig. 8). Cell proliferation rate was calculated as the ratio of BrdU-incorporating cells to total cells stained with DAPI. The ratio of BrdU-positive cells to total cells was significantly lower in the LPS, IFNγ, and LPS + IFNγ groups than in the control group. This result is consistent with a previous observation that inflammatory stimuli like LPS and IFNγ induce differentiation of proliferating primary glial cells into activated glial cells [44]. Importantly, the ratio of BrdU-positive cells to total cells was significantly higher in Btg2 −/− glial cells than in wild-type glial cells following LPS and LPS + IFNγ treatment, but no significant difference was observed after treatment with IFNγ alone (Fig. 8a, b). Coimmunostaining of the mixed glial cells with anti-BrdU and either anti-GFAP or anti-Iba1 antibody revealed that Btg2 deletion altered responses to inflammatory stimuli in astrocytes, but not in microglia ( Supplementary Fig. 5). OT is inside the white dotted lines. b Immunoreactive (IR) areas were compared among groups after adjusting for sex. Data are adjusted means ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001, as determined by Tukey's HSD test

Discussion
Although BTG2 is an anti-proliferative protein, and its expression is increased after ischemic stress, it has not been clearly determined whether deletion of this protein has impacts on the increase in glial cell number after cerebral hypoperfusion. In this study, we showed that deletion of Btg2 increased the abundances of astrocytes and Mac2-positive cells in the white matter of optic tracts after BCAS, even though the severities of white matter lesions did not differ significantly between wild-type and Btg2 −/− mice.
Previous studies observed no difference between sham-operated and BCAS-treated wild-type mice in locomotor activity (as determined by the open-field test) or spatial learning ability (as determined by the Morris water maze) [5,20]. Our findings were consistent with those studies. In addition, we observed behavioral changes in BCAS-treated Btg2 −/− mice relative to BCAStreated wild-type mice: their locomotor activity increased in the open-field test, and their spatial learning ability was weak in the Morris water maze test. The ratio of travel distance in the outer or inner part of the open field to the total field did not significantly differ between wild-type and Btg2 −/− after BCAS, suggesting that the increase in locomotor activity in BCAS-treated Btg2 −/− mice was not related to anxiety or exploratory behavior. Because BTG2 is an anti-proliferative protein that is highly expressed in glial cells, these behavioral differences might be due to alteration of these cells caused by Btg2 deletion.
BCAS treatment worsened white matter lesions and increased GFAP immunoreactivities irrespective of Btg2 deletion, consistent with previous reports [4]. Moreover, immunoreactivities of GFAP and Mac2 were significantly higher in BCAS-treated Btg2 −/− mice than in BCAStreated wild-type mice (Fig. 5), despite the lack of a difference in the white matter lesions (Fig. 4). Consistent with this, Gfap mRNA levels were higher in BCAS-treated Btg2 −/− mice than in sham-treated Btg2 −/− mice, whereas we observed no difference between wild-type mice with or without BCAS (Fig. 7). Brain ischemia induces the upregulation of Gfap mRNA levels for several weeks after the ischemic stress [45,46]. Although Gfap mRNA levels were comparable between sham-treated and BCAS-treated wild-type mice 7 weeks after the surgery, the Gfap mRNA level in BCAS-treated Btg2 −/− mice might still be elevated at this time point. These results suggest that BTG2 negatively regulates astrocyte proliferation.
Mac2, also called galectin-3, was first reported as a macrophage-specific antigen defined by monoclonal antibodies [47]. It is expressed not only in activated microglia and macrophage/monocyte but also in astrocytes [48]. We examined the co-localization of Mac2 and To determine whether BTG2 negatively regulates glial cell proliferation in response to inflammatory stimuli, we performed in vitro experiments using mixed glial cells from wild-type and Btg2 −/− mice. Incorporation of BrdU in response to inflammatory stimuli associated with brain ischemic conditions was greater in mixed glial cells from Btg2 −/− mice than in those from wild-type mice [49][50][51]. This result is consistent with multiple studies reporting that BTG2 plays an anti-proliferative role in breast cancer [52,53], pancreatic cancer [54], and vertebral patterning by modulating bone morphogenesis [55], homeotic transformation of the axial skeleton [56,57], and stress response [18]. In addition, our study shows for the first time that BTG2 regulates astrocyte proliferation in chronic cerebral hypoperfusion models.
This study had several limitations. First, we examined BrdU incorporation in vitro using mixed glial cells, but not in separated astrocytes or microglia. By coimmunostaining mixed glial cells using both anti-BrdU and anti-GFAP or anti-Iba1 antibody, we observed that responses to inflammatory stimuli were altered by Btg2 deletion in astrocytes, but not in microglia. However, it  Fig. 7 Relative mRNA levels of cell-specific markers in brains of sham/BCAS-treated wild-type and Btg2 −/− mice. mRNA levels of a Gfap, an astrocyte marker; b Cd11b, a microglia, monocyte, and macrophage marker; c Trem2 and d Dap12, disease-associated microglia (DAM) markers; e Cd45, a pan-hematopoietic marker; f Cd68, a macrophage marker; g F4/80, a mature macrophage marker; and h Cd14, a monocyte and macrophage marker, were compared among groups. All mice were male. Data are means ± SD. *p < 0.05, as determined by Dunnett's multiple comparison test. n.s., not significant remains unclear whether these effects in astrocytes with Btg2 deletion are cell-autonomous or non-cell-autonomous. This in vitro culture system also did not include macrophages/ monocytes, although it is possible that the Mac2-positive cells we observed were macrophages/monocytes. In a future study, we will examine the effect of BTG2 on the proliferation of each cell type using conditional cell type-specific deletions of Btg2. Second, we could not determine why the increased abundance of astrocytes and Mac2-positive cells in BCAS-treated Btg2 −/− mice vs. BCAS-treated wild-type mice was observed in the optic tract but not the corpus callosum. This observation might be related to regional differences in vulnerabilities to ischemic stress or that of the expression level of Btg2. Indeed, previous studies have reported regional differences in the distribution of neuropeptides [58] and nitric oxide synthase [59] in cerebral blood vessels. Finally, although we have tried several anti-BTG2 antibodies, including commercial ones as well as one developed in our laboratory, none of them worked well in western blotting or immunohistochemistry of samples from wild-type and Btg2 −/− mice. Further studies using BTG2-specific antibodies are required to fully characterize when and where BTG2 functions to regulate glial cell proliferation during chronic cerebral hypoperfusion.

Conclusions
Deletion of Btg2 caused behavioral changes associated with altered responses of glial cells, especially those of astrocytes and Mac2-positive cells. Moreover, our in vitro experiments supported the idea that BTG2 negatively regulates proliferation of glial cells. These results suggest that BTG2 might serve as an anti-proliferative protein in glial cells during cerebral hypoperfusion, thereby protecting against vascular dementia.
Additional file 1: Supplementary Figure 1. Anxiety/exploratory parameters in the open-field test. The ratio of distance travelled in the inner zone (a) or outer zone (b) to the distance traveled overall was compared among groups by Tukey's HSD test. Data are adjusted means ± SEM. n.s., not significant. Supplementary Figure 2. Probe test results of the Morris water maze test. NE: northeast quadrant area; NW: northwest; SE: southeast; SW: southwest. The platform was in the NE quadrant area (target quadrant). Time spent in each non-target quadrant was compared with the duration spent in the target quadrant in each group. Data are adjusted means ± SEM. ***p < 0001; Tukey-Kramer test. The following numbers of mice were used: wild-type + Sham, n = 13; wild-type + BCAS, n = 15; Btg2 -/-+ Sham, n = 11; Btg2 -/-+ BCAS, n = 16. Supplementary  Table 1. Two-or three-way ANOVA Results