Male CB-17 mice (C.B-17/Icr- + / + Jcl, Oriental Yeast, Tokyo, Japan) aged 5, 10, 52, and 104 weeks were used for this study (N = 5, each). Minimal variation of cerebrovascular structure is known in CB-17 mice (Taguchi et al., 2010) and CB-17 mice were used for the experiment of cerebrovascular disorders (Kasahara et al., 2013; Takeuchi et al., 2020). After deep anesthesia by pentobarbital, a blood sample was obtained by the puncture to left ventricle of heart. Total RNA was isolated using RNeasy Plus Universal Mini Kit (Qiagen, CA, United States) according to the manufacturer’s instructions. cDNA was synthesized from 1μg total RNA using PrimeScript™ II 1st strand cDNA Synthesis Kit (TAKARA, Kyoto, Japan). Transcription of mRNA was analyzed using PowerUp™ SYBR™ Green Master Mix (Applied Biosystems, CA, United States) and the Agilent AriaMx real time PCR System. 18S ribosomal RNA was used for the reference gene. Relative quantification of RNA was analyzed by phaffle method. For the comparison of RNA transcription of Connexin 43 (Cx43) between young and aged mice, 5 and 80 weeks old Male CB-17 mice were used (N = 6, each). The list of target genes, primer sequences, and amplification protocols of qPCR are shown in Table 1.

Brain tissue was harvested from young (5W) and aged (more than 80 weeks) mice (N = 7, each). After harvesting the brain, coronal sections (4 mm thick) of the forebrain between 2 mm and 6 mm from frontal pole were cut followed by immersion in RNAlater (Thermo Fisher, Waltham, MA, United States) to prevent RNA degradation. Under stereoscopic microscope (Olympus, Tokyo, Japan), the part of the hippocampus, was dissected by medical tweezers, as we described previously (Takeuchi et al., 2020). Total RNA was isolated and cDNA was synthesized from 1 μg total RNA. The list of target genes, primer sequences, and amplification protocols are shown in Table 1.

Human umbilical vein endothelial cells (HUVEC, Kurabo, Osaka, Japan) were cultured with medium, serum and growth factors (HuMedia-EB2, Kurabo). Total RNA of HUVEC (1 × 10⁵ cells) at passage 3 and 10 were was isolated and expression of Cx43 and Connexin 37 (Cx37) were evaluated by qPCR (N = 4). The list of target genes, primer sequences, and amplification protocols of qPCR are shown in Table 1.

Male CB-17 mice aged 5 and 80 weeks old were used for this study (N = 5, each). Blood sample was obtained by a puncture to the left ventricle of the heart and red blood cells were removed using lysing buffer (BD Biosciences, Woburn, MA, United States) (Muirhead et al., 1986). WBC were incubated with 5 μM of BCECF-AM (2′,7′-bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein, acetoxymethyl ester, Dojindo, Kumamoto Japan) for 30 min at 37°C as we described previously (Kikuchi-Taura et al., 2020). BCECF-AM was converted to BCECF (2′,7′-bis-(2-carboxyethyl)-5-(6)-carboxyfluorescein) in the cytoplasm and BCECF loaded WBC were washed twice with PBS before use. HUVEC in passage 6 were used in the subsequent experimental work. BCECF loaded WBC (1 × 10⁶ cells in 20 μl PBS) and HUVEC (1 × 10⁵ cells in 100 μl PBS) were co-cultured at 37°C for 3 h. After co-incubation, the cell mixture was washed twice with PBS, and stained with PE-conjugated anti-human CD31 antibody (BD Biosciences, Woburn, MA, United States), FITC-conjugated anti-mouse CD45 antibody (BD Biosciences, Woburn, MA, United States) and 7-AAD (7-Amino-Actinomycin D) (BD Biosciences, Woburn, MA, United States). The level of BCECF in HUVEC (CD31-positive, CD45-negative and 7AAD-negative) was evaluated using a FACS Calibur fluorescence activated cell sorter (BD Biosciences, Woburn, MA, United States). Cell populations that communicate with HUVEC via gap junction were characterized using FACS as described previously (Kikuchi-Taura et al., 2021).

HUVEC were treated with gap junction blocker (0.01 v/v 1-octanol. Wako, Osaka, Japan) or PBS for 10 min at 37°C. A blood sample was obtained by a puncture to the left ventricle of the heart of male CB-17 mice (5 weeks old) and co-cultured at 37°C for 3 h with HUVEC that were treated or non-treated by gap junction blocker. The change of RNA transcription was evaluated by qPCR with mouse specific primers.

Bone marrow was obtained from 5-week-old syngeneic male CB-17 mice and BM-MNC were prepared by Ficoll-Paque (GE-Healthcare, Little Chalfont, United Kingdom) density-gradient centrifugation as described elsewhere (Kikuchi-Taura et al., 2020). Aged (more than 80 weeks) mice received 1 × 10⁵ BM-MNC in 100 μl PBS or PBS alone through the tail vein (5 times in total), and young (5 weeks) mice received PBS (N = 5, each). Each injection was performed without anesthesia and injection was finished in 10 s. Mouse behavior was evaluated using wire hang test and passive avoidance test before blood sample collection. The experimental design is shown in Supplementary Figure 1A. The Passive avoidance test is a fear-aggravated test used to evaluate learning and memory in rodent models. The apparatus (TMS-2; Melquest, Toyama, Japan) was divided into two compartments, an illuminated compartment (120 mm × 120 mm × 135 mm) and a dark compartment (120 mm × 120 mm × 135 mm), that were connected via a guillotine door. The dark compartment was equipped with a grid floor (6 mm in diameter spaced at 10 mm) through which a footshock (0.2 mA, 3 s) could be delivered. In the conditioning and acquisition phases, each mouse was placed in an illuminated compartment and habituated for 10 s before door opening. In the conditioning phase, a mild electric foot shock was administered 10 s after each mouse crossed the adjacent dark compartment and door closing. On the next day, each mouse was again placed in the illuminated compartment and habituated for 10 s, and the time to cross over to the dark compartment after door opening, up to a maximum of 180 s, was recorded as the test phase (Ogawa et al., 2020). The wire hang test evaluates muscular strength or motor function. In this test, each mouse was placed on the center of wire mesh plate (450 mm × 300 mm; mesh wire was 1.3 mm in diameter and spaced at 12 mm intervals, KK23-8331; Kohnan, Osaka, Japan) and allowed to accommodate to this environment for 5 s. The wire mesh plate was inverted and secured to the top of a cubic transparent open-topped glass box (250 mm × 250 mm × 250 mm). Latency to fall was recorded, up to a maximum of 180 s. This trial was repeated five times with an interval of 1 min and the average time to fall was calculated (Ogawa et al., 2020).

For immune historical analysis, aged mice (more than 80 weeks) received 1 × 10⁵ BM-MNC in 100 μl PBS or PBS alone intravenously totally five times every 2 days from tail vein, and young (5 weeks) mice received PBS (N = 6,7,8, for aged + PBS, aged + BM-MNC and young + PBS, respectively). Twenty-four hours after the last injection, the brain was removed, fixed with 2% paraformaldehyde (PFA; Thermo Fisher, Waltham, MA, United States), and cut into coronal sections (20 μm) using a vibratome (Leica, Wetzlar, Germany). Sections were immunostained with primary antibodies against Nestin (Merck Millipore,1:200), Doublecortin (DCX; Merck Millipore, 1:350) and DAPI (Thermo Fisher, 1:1,000). Confocal images were obtained with fluorescence microscope (BZ-X810: Keyence, Osaka Japan). The number of Nestin positive fibers and DCX positive cells at the granular cell layer and sub granular zone of dentate gyrus, respectively, in the hippocampus of aged mice were counted by blinded investigator. The number of positive cells per 1 mm of each zone was evaluated.

HUVEC (1 × 10⁵ cells in 100 μl PBS) and BM-MNC (1 × 10⁶ cells in 20 μl PBS) were co-cultured at 37°C for 3 h, and the change in expression of Cx43 in HUVEC was evaluated with PE-conjugated anti-human CD31 antibody, FITC-conjugated anti-human Cx43 antibody (Santa Cruz Biotechnology, Santa Cruz, CA, United States) and 7-AAD (7-Amino-Actinomycin D) by FACS Calibur fluorescent cell sorter.

Statistical comparisons among groups were performed using one-way ANOVA (analysis of variance) followed by post hoc analysis using Steel-Dwass test or Dunnett’s test. Correlation between RNA transcription of metabolism related genes in circulating WBC and neurological functions was evaluated with linear regression analysis. Individual comparisons were performed using Student’s t-test. All data are shown as mean ± SD.

We previously observed that transcription of metabolism related genes, including energy source transport genes, in the brain are significantly decreased with aging with impaired neurological functions. We expected similar changes in WBC and investigated the sequential change in RNA transcription of these genes in circulating WBC. Contrary to our expectation, the energy source transport genes, including Glut1, Glut3, and monocarboxylate transporter 4 (MCT4), in WBC significantly increased with aging (Figure 1A). Hif1-α is one of the major transcriptional factors that regulates the expression of energy source transporters (Masoud and Li, 2015; Zorzano et al., 2005) and has been shown to decrease with aging in the brain (Takeuchi et al., 2020). The transcription of Hif1-α and its down-stream genes, prolyl hydroxylase 3 (PHD3) and pyruvate dehydrogenase kinase 1 (PDK1), in circulating WBC were investigated and it was found that transcription of PHD3 and PDK1 increased with aging (Figure 1B).

We have shown low molecular weight molecules are transferred from BM-MNC to endothelial cells via gap junction mediated direct cell-cell interaction (Kikuchi-Taura et al., 2020; Takeuchi et al., 2020). To investigate gap junction mediated cell-cell interaction between WBC and endothelial cells, low molecular weight fluorescence compound, namely BCECF, was loaded to WBC and co-cultured with HUVEC. Similar to BM-MNC, transfer of BCECF from WBC to HUVEC was observed and the level of BCECF transfer significantly decreased in WBC of aged mice, relative to that of young mice (Figure 2A). The change of RNA transcriptions with aging was evaluated and found the level of Cx43 in WBC was significantly decreased in aged mice, compared with young mice (Figure 2B). To confirm the significance of gap junction mediated cell-cell interaction on RNA transcription in WBC, WBC were co-cultured with HUVEC with or without blockade of gap junction. A significant increase of RNA transcription of Hif1-α and PDK1 was observed following blockade of gap junction (Figure 2C). These findings show the importance of gap junction mediated cell-cell interaction in the decrease of metabolism related gene expression in WBC, mainly monocytes and lymphocytes. The change in RNA transcription with passage number was evaluated in HUVEC and it was found that the level of Cx43 significantly decreased during passaging (Figure 2D). The change in RNA transcription with aging at hippocampus was evaluated and it was found that the level of Cx43 significantly decreased in aged mice, compared with young mice (Figure 2E). The change of RNA transcription of Cx37 was evaluated and found its expression decreased in HUVEC during passaging (Figure 2F) and in hippocampus with aging (Figure 2G).

Next, we evaluated the correlation between RNA transcription of metabolism related genes in circulating WBC and neurological function in mice, including aged mice that received BM-MNC transplantation. Figures 3A-F shows the correlation between increased RNA transcription of metabolism related genes and shortening of step through latency in passive avoidance test. A strong and significant correlation (| r| > 0.7 and p < 0.05) was observed for Glut1, MCT4, PHD3, and PDK1. Supplementary Figure 2 shows the correlation between increased RNA transcription of metabolism related genes and shortening of latency to fall in wire hang test. A strong and significant correlation (| r| > 0.7 and p < 0.05) was observed for MCT4, PHD3, and PDK1. Supplementary Figures 3, 4 show the correlation in aged mice, including those that received PBS or BM-MNC but not young mice, in the passive avoidance test (Supplementary Figure 3) and in the wire hang test (Supplementary Figure 4). A strong and significant correlation was observed for RNA transcription of MCT4 and PHD3. It should be noted that, a statistically significant reduction of the latency in passive avoidance test and wire hang test in aged mice with PBS, compared with young mice, was reported in our previous report. In contrast, BM-MNC injection to aged mice improved the scores of both tests and there had been no significant difference in the latencies between in aged mice with BM-MNC and young mice, though no statistically significant difference had been observed between in the scores of aged mice that received PBS or BM-MNC (Takeuchi et al., 2020).

Mice that received BM-MNC transplantation showed decreased RNA expression of Glut3, MCT4, PHD3, and PDK1 in circulating WBC (Figures 4A,B). To investigate the possible cause of decreased level of these genes by BM-MNC transplantation, endothelial cells were co-cultured with BM-MNC. We observed increased expression of Cx43 in HUVEC when co-cultured with BM-MNC (Figure 4C). These findings indicated that BM-MNC transplantation has a potential to improve cell-cell interaction between circulating WBC and endothelial cells through enhancing expression of Connexin on endothelial cells.

To further investigate the link between BM-MNC transplantation and increased memory function, the level of neurogenesis at hippocampus was evaluated using anti-Nestin and anti-DCX antibody, and it was found that the level of neurogenesis significantly increases with BM-MNC transplantation in aged mice (Figures 5A-D). These findings are consistent with our previous report that BM-MNC transplantation ameliorates memory disorder in aged mice (Takeuchi et al., 2020).