Mice

C57BL/6J and C57BL/6-Tg (CAG-EGFP) female mice were purchased from Japan SLC (Shizuoka, Japan) and CLEA Japan (Tokyo, Japan). Young mice were 2–3 months-old, whereas those older than 24 months were regarded as aged. Mice with grossly visible tumors, hepatomegaly or splenomegaly were excluded from the study. Some but not all mice were assessed for body weight, liver weight, lung weight and serum ALT levels before the experiments began. For the quantification of serum ALT levels, blood was sampled from the heart and centrifuged to separate serum for ALT measurements using an ALT assay kit (#700,260, Cayman Chemical, Ann Arbor, Michigan) and a microplate reader (PowerScanHT; DS Pharma Biomedical, Osaka, Japan). All experiments were carried out following the guidelines of Osaka University Committee for animal and recombinant DNA experiments. Mice were handled and maintained according to the Osaka University guidelines for animal experimentation.

EC isolation and cell counting by flow cytometry

Isolation of ECs from murine liver and lung was performed as previously described [37]. Briefly, the organs were chopped up with scissors. Tissue fragments were then digested with several enzymes including dispase (#17105-041, Gibco, Waltham, MA) and collagenases (#034-22363, Wako, Osaka, Japan and #LS004176, Worthington Industries, Columbus, Ohio) followed by mechanical dissociation using a bioshaker (#BR-22FH, Taitec, Saitama, Japan) to prepare single-cell suspensions. Cells were labeled with fluorescent antibodies and ECs were isolated by FACS. Fluorescent antibodies used were as follows: Purified rat anti-mouse CD16/CD32 (#553,142, BD Biosciences, Franklin Lakes, NJ), Brilliant Violet 421 rat anti-mouse CD31 (#102,423, Biolegend, San Diego, CA), FITC rat anti-mouse CD31 (#11-0451-85, Thermo Fisher Scientific, Waltham, MA), APC-Cy7 anti-mouse CD45 (#103,116, BioLegend), APC anti-mouse CD157 (#140,208, BioLegend), PE rat anti-mouse CD200 (#123,808, BioLegend). Propidium Iodide (PI) (#P4170-10MG, Sigma-Aldrich, St. Louis, MO) was used to exclude dead cells. The gating strategy for the isolation of VESCs was as described previously [37]. Cell populations were selected based on forward scatter area (FSC-A) and side scatter area (FSC-A), and cell doublets were excluded using forward scatter width (FSC-W) and side scatter width (SSC-W). PI-negative cells were analyzed for CD31 and CD45 expression. Finally, CD31+CD45− ECs were divided into three fractions based on their expression of CD157 and CD200 antigens (CD157+CD200+, CD157−CD200+ and CD157−CD200−). Fluorescence minus one (FMO) controls were used to set the gates. We ran the samples at a low event rate (< 3,000 events/s) and used purity sorting mode (yield mask 32, purity mask 32, phase mask 0) to isolate ECs. In this setting, the sorting efficiency was usually 80–90% and the purity of sorted cells > 98%. Sorting efficiency was checked by counting the number of sorted cells with a cell counter (#WC2-100, Waken, Kyoto, Japan) for each experiment. Actual cell counts of sorted cells were calculated considering the sorting efficiency. In Fig. 1, the number of ECs counted by analyzing the whole cell suspension prepared from one liver by flow cytometry is illustrated.

In vitro culture of CD157-positive VESCs

One day before the isolation of primary murine VESCs by FACS, OP9 feeder cells (#RCB1124, RIKEN cell bank, Ibaraki, Japan) were seeded into a 24-well plate at a concentration of 2.0 × 104 cells/well. On the day of seeding the VESCs, the OP9 cells were almost confluent. CD157+ VESCs were sorted as described above and then seeded onto OP9 cells in the 24-well plates (1,000 cells/well for liver VESCs, 5,000 cells/well for lung VESCs). Cells were cultured in RPMI medium (#R8758-500ML, Sigma-Aldrich) supplemented with 10% fetal bovine serum (FBS) (#172012-500ML, Sigma-Aldrich) and 0.1% 2-mercaptoethanol (#21,985,023, Gibco). Cell culture plates were maintained under 5% CO2 in humidified air at 37℃. Ten ng/mL of VEGF165 (#100–20, PeproTech, Cranbury, NJ) was added to the culture medium every 3 days. Ten days after seeding, cells were fixed with 4% PFA in phosphate-buffered saline (PBS) for immunostaining. In Fig. 3c, VESCs were isolated from EGFP mice to visualize the process of the colony formation. EGFP fluorescence was observed every other day at the same position in the culture plate using a fluorescence microscope (Leica Microsystems, Wetzlar, Germany).

Immunostaining of cultured ECs

After PFA fixation, immunostaining with rat anti-mouse CD31 antibody (#553370, BD Biosciences) was performed as previously described [38]. After the primary antibody, cells were incubated with biotin-conjugated anti-rat IgG antibody (Dako, Santa Clara, CA). Then avidin-biotin complexes were formed employing VECTASTAIN Elite ABC-HRP Kits (#PK-6100, Vector Laboratories, Newark, CA). Finally, EC colonies were visualized with 3,3’-diaminobenzidine (DAB) and nickel chloride (NiCl2). Images were captured with a Canon EOS kiss X7. Areas and diameters of EC colonies were quantified with Fiji software [39].

Immunohistochemical staining

Tissue preparation and staining was as previously described [40]. Briefly, fixed liver specimens were embedded in OCT compound (Sakura Finetek, Torrance, CA) and frozen at -80℃. Frozen specimens were then sectioned at 40 μm thickness and washed with 0.1% Tween 20 in PBS (PBS-T). They were then incubated with blocking buffer (2% skimmed milk in PBS-T) for 1 h at room temperature. Primary antibodies were diluted with the blocking buffer at 1:200–1:400. Antibody was left on the sections overnight at 4℃. Primary antibodies used were as follows: rat anti-mouse CD31 (#553,370, BD), Armenian hamster anti-mouse CD31 (#MAB1398Z, Merck Millipore, Burlington, MA), rat anti-mouse CD157 PE (#140,204, Biolegend), rat anti-mouse CD45 PE (#12-0451-83, Biolegend). For primary antibodies that were not directly conjugated, sections were sequentially incubated with fluorophore-conjugated secondary antibodies diluted with the blocking buffer at 1:500 for 3 h at 4℃ on the next day. Second antibodies used were as follows: Alexa Fluor 488-conjugated goat anti-rat IgG (#A11006, Thermo Fisher), Alexa Fluor 647-conjugated goat anti-Armenian hamster IgG (#127-605-160, Jackson ImmunoResearch, West Grove, PA). Finally, sections were mounted with fluorescent mounting media (#S3023, Dako) and imaged using a Leica TCS SP5 confocal microscope. Images were processed with the Leica application suite (Leica Microsystems), and Adobe Photoshop CC software (Adobe Systems, San Jose, CA). For quantitative measurements of vascular area as well as CD157- and CD45-positive cells, the left lobe of the liver was sectioned at the middle. Sections were assessed at 20 μm thickness and > 3 images were acquired at periportal sinusoids or portal veins from each section. Vascular density was quantified using Angiotool software [41] for Fig. 2b. CD157- and CD45-positive areas were quantified with Fiji software [39] for Figs. 2d and 5f.

VESC transplantation assay

Transplantation experiments were performed as described previously [14]. A genotoxic pyrrolizidine alkaloid, monocrotaline (MCT) (#C2401, Sigma-Aldrich) was administered to the recipient mice at 500 mg/kg intraperitoneally 48–72 h before transplantation. On the day of transplantation, recipient mice additionally underwent whole body irradiation with a single dose of 6 Gy. Liver VESCs were isolated as described above from young (2–3 month-old) and aged (25–26 month-old) EGFP mice. Five thousand VESCs in 4% FBS/PBS were transplanted directly into the recipient’s liver through a 27-gage needle. Four weeks (28 days) after transplantation, engrafted EGFP-positive ECs were counted by analyzing the recipient’s whole liver by flow cytometry. In Fig. 4f and S3, the recipient’s liver and lung were observed using a stereoscopic fluorescence microscope (Leica Microsystems). The area of engrafted colonies was quantified with Fiji software [39].

RNA-sequencing

Liver ECs (CD31+45−) were isolated from young (10-week-old) and aged (30-month-old) mouse livers by FACS as described above. Three independent samples were collected from each group and mRNA was extracted from ECs using RNeasy Plus Micro Kits (#74,034, Qiagen, Hilden, Germany) according to the manufacturer’s protocol. RNA libraries were prepared using the TruSeq Sample Prep v2 kit and sequenced on a HiSeq 2500 (Illumina, San Diego, CA) in 75-base single-end mode. CASAVA 1.8.2 software (Illumina) was used for base calling. Sequenced reads were mapped to the mouse reference genome sequence (mm9) using TopHat v2.1.0. Fragments per kilobase of exon per million mapped fragments (FPKM) values were determined with Cuffnorm v.2.2.1. For PCA and t-SNE visualization, read counts were uploaded and analyzed using the integrated Differential Expression and Pathway (iDEP.96) (http://bioinformatics.sdstate.edu/idep96/) program [42]. Genes with low expression values (CPM < 0.5) were filtered out with default settings for normalization (Constant c for started log: log (x + c) = 1). To analyze Gene Ontology (GO) terms of differentially expressed genes, the web-based Database for Annotation, Visualization and Integrated Discovery (DAVID) 6.8 was used (https://david.ncifcrf.gov/) [43]. GSEA v4.3.1 Mac App was used to perform gene set enrichment analysis [44]. The lists of differentially expressed genes containing log-fold changes and p-values are shown in Supplementary Tables S2 and S3.

Quantitative PCR

ECs (CD31+45−), CD157-positive ECs and CD157-negative ECs were isolated from young and aged livers by FACS. Total mRNA was extracted from ECs using RNeasy Plus Micro Kits (#74,034, Qiagen) according to the manufacturer’s protocol. mRNA was reverse-transcribed to cDNA using PrimeScript RT reagent Kits (#RR037A, Takara Bio, Shiga, Japan). Real-time PCR was performed with the LightCycler 96 System (Roche, Basel, Switzerland). Primers are listed in Supplementary Table S1. Threshold value was determined automatically within the exponential growth region and threshold cycle (Ct) value was defined as cycle number at which fluorescence passed the threshold. ΔCt value was determined by subtracting the Ct value of Gapdh from that of the target gene. ΔΔCt value was calculated by subtracting the ΔCt value of young CD31+45− ECs from the respective ΔCt value of each cell group. 2−ΔΔCt is presented as a fold-increase relative to young CD31+45− ECs in Fig. 5d.

Single-cell RNA-seq analysis using the Tabula Muris Senis data

The filtered h5ad file for FACS subsets was downloaded from the official Tabula Muris Senis repository (https://figshare.com/projects/Tabula_Muris_Senis/64982). Raw count data of liver were transferred into the R environment (v4.2.2). Next, we extracted 3- and 24-month samples and ran the standard Seurat (v4.3.0) pipeline with default parameters (log normalization, 15 PCs, and UMAP for dimensionality reduction). The Tabula Muris Consortium cell type designations were used for the clustering. The Seurat package was used for the analysis of differentially expressed genes and UMAP visualization.

Statistical analysis

Statistical analysis was performed with Prism 9 software. Data are presented as mean ± standard deviation (SD). Paired data were evaluated with two-tailed unpaired Student’s t tests and comparison of multiple groups was performed using two-way analysis of variance (ANOVA). P < 0.05 was considered to be statistically significant. P values are indicated as *(P < 0.05), ** P < 0.01) or ***(P < 0.001).