The study protocol was approved by the Institutional Review Board of Asan Medical Center, Seoul, Republic of Korea (2015–1123), and was in accordance with the declaration of Helsinki. After receiving their informed consent, bone marrow was obtained from healthy donors. All the manufacturing and product testing procedures for hBMSC generation were performed using good manufacturing practices (Pharmicell Co. Ltd., Seongnam, Gyeonggi, Republic of Korea). A previously described isolation protocol was used (30). Cells were cultured in Dulbecco’s modified Eagle’s medium-low glucose (Gibco, Grand Island, NY, US), containing 10% human plate lysate (PLUS^TM Cell Culture Supplement, Compass Biomedical, Inc. MA, US). Then, cells at passage 8 were cultured for 72 h in a hypoxia-conditioned incubator (at 37°C in 1% O₂, 5% CO₂ and 94% N₂, for 72 h). After hypoxic preconditioning, the HP-hBMSCs to be injected were harvested using trypsin, washed twice with PBS and once with Plasma Solution A inj. (Multiple Electrolytes Injection, Type 1, UPS; CJ Health Care) and resuspended to a final concentration of 1~4×10⁶ cells in Plasma Solution A inj. Criteria for release of HP-hBMSCs for preclinical use included the absence of microbial contamination (bacteria, fungus, mycoplasma, or endotoxin), a viability greater than 70% after assessment using a trypan blue exclusion assay, the proven expression of CD73 and CD105 surface molecules, and absence of CD14, CD34, and CD45, via immune phenotyping by flow cytometric analysis.

All aspects of animal care and treatment and the surgical procedures used conformed to the eighth edition of the Guide for the Care and Use of Laboratory Animals, published in 2011. The protocols for performing animal experiments were approved by the Institutional Animal Care and Use Committee of Asan Medical Center (2016-12-068). One-hundred-and-eighty 12-week-old male Sprague-Dawley rats were purchased from Orient Bio Inc. (Seongnam, Gyeonggi, Republic of Korea) and housed for 1 week for acclimatization. During the experiments, all rats were maintained under a 12 h:12 h light/dark cycle (lights on at 08:00, lights off at 20:00), a temperature of 22±2°C, and a humidity of 50~55%, and ad libitum access to food and water was provided.

Sixty 13-week-old rats were randomly and equally distributed into the following six groups (10 animals per group): sham, nephrectomy control, IRI control, renal arterial injection, renal parenchymal injection, and tail venous injection groups. Each rat was intramuscularly anesthetized with 0.3 ml of tiletamine (Zoletil; Virbac Laboratories, Carros, France) and xylazine hydrochloride (Rompun, Bayer, Germany) in a 4:1 mixture. Laparotomy was performed in the sham group and right nephrectomy was performed in the nephrectomy control group, while right nephrectomy was performed and IRI of the left kidney was induced in the other four groups. IRI induction was performed in the left kidney as described previously (17). Rats from the renal arterial injection group were injected immediately prior to the induction of renal IRI with 1×10⁶ HP-hBMSC/50 μl via a 33-gauge Hamilton syringe (Hamilton, Zurich, Swiss) that was inserted into the distal aorta after clamping the aorta above and below the left renal artery. The injection site was closed using TachoSil^® Sealant matrix (BBF Sterilisatoin Service GmbH, Kernen, Germany), and the clamps were removed to restore aortic and renal blood flow. Rats from the renal parenchymal injection group were injected using a Hamilton syringe needle, which was pierced from the lower pole to the upper pole of the kidney. While pulling out the needle, the cells were gradually released from the syringe. Rats from the tail venous injection group were injected with 1×10⁶ HP-hBMSC/50 μl through the tail vein using a 26-gauge Kovax-syringe (Koreavaccine, Gyeonggi, Republic of Korea). The renal function of all the rats was evaluated 4 days before and 1, 2, 3, 4, 7 and 14 days after IRI. They were sacrificed 14 days after IRI, and the kidney tissues were then collected for histological examination.

Sixty 13 week-old rats were randomly and equally distributed into the following six groups (10 animals per group): sham, nephrectomy control, IRI control, low-dose (1×10⁶ HP-hBMSC/100 μl), moderate-dose (2×10⁶ HPhBMSC/100 μl) and high-dose (4×10⁶ HP-hBMSC/100 μl) groups. HP-hBMSC was injected through effective routes to protect the renal function during the optimal route study. IRI induction and the renal function evaluation method were similar to those for the optimal injection route study.

Sixty 13-week-old rats were randomly and equally distributed into the following six groups (10 animals per group): sham, nephrectomy control, and IRI control, pre-IRI (2 days before IRI), concurrent-IRI, and post-IRI (2 days after IRI) groups. An effective dose of HP-hBMSC was injected through the certain route to protect the renal function in studies to determine the optimal injection route and dose. The IRI induction process was similar to that followed in the above two studies, and all rats underwent kidney function evaluation 1 and 4 days before IRI, and 1, 3, 4, 7, and 14 days after IRI.

Serum levels of creatinine and blood urea nitrogen (BUN) and urine levels of creatinine were measured at these time points in all animals. Urine was collected from animals in a metabolic cage at 24 h, to estimate the daily urine volume. The glomerular filtration rate (GFR) was calculated using the following equation for the creatinine clearance rate (Ccr): Ccr (ml/min/100 g)=(urine creatinine×24 h urine volume×100)/(serum creatinine× 1440×body weight). Quantification of BUN and creatinine levels was performed using standard laboratory equipment at our hospital.

After rats were sacrificed, the kidney tissues of each rat were collected. Half of each kidney was cryopreserved in liquid nitrogen for protein extraction or immunofluorescence staining. The remaining half was fixed in 4% paraformaldehyde and embedded in paraffin prior to sectioning and immunohistochemistry analysis.

Four μm kidney specimen sections were prepared and stained with hematoxylin and eosin for light microscopic analysis. Histopathological scoring was carried out based on a previous study in a blind fashion (31). The tubular injury scores estimated by deriving the percentage of destroyed tubules in the outer medulla that showed a loss of brush border, tubular necrosis, tubular dilatation, and cast formation in 10 randomly chosen, non-overlapping fields (200×) were as follows: 0 (normal histology), 1 (≤ 24%), 2 (25~49%), 3 (50~74%) and 4 (75~100%). Histopathological scoring was described in a previous study (18).

The slides were dipped in Xylene I and Xylene II, treated with Xylene III, and dipped in ethanol solutions in the following order: 100% to 90% and 80% to 70%. After dipping slides in distilled water thrice, we dipped them in the Harris hematoxylin solution (Sigma, St. Louis cat no: HHS-16, MO, US). Sirius red reagent (Abcam, Cambridge, MA., US) was dropped on the slides and they were incubated. After dipping slides twice in the acid solution, we washed the slides thrice using distilled water, and then dipped them again in ethanol solutions 70% to 80%, 90%, and 100% in 3 steps for each solution. We repeatedly dipped the slides in Xylene I and Xylene II, treated them with Xylene III, and covered the slides with coverslips using permanent mounting media.

A TUNEL assay kit (Roche Molecular Biochemicals, Mannheim, Germany) was used. Each sectioned specimen was deparaffinized in xylene I, II, and III, and rehydrated in increasingly diluted ethanol solutions (diluted 100%, followed by 90%, and so on upto 70%). After washing in distilled water, the slides were placed into a pretreatment solution (heat-induced epitope retrieval solution/ target retrieval solution). After washing, the protein blocking serum-free ready-to-use (Dako, Glostrup, Denmark) reagent containing 20% FBS was added into the sections, and they were incubated in a humid chamber. The tunnel reaction mixture was dispensed onto the slides, which were incubated in a 37°C humid chamber. Converter-PODs (Converter-Peroxidase) were dispensed and slides were incubated in a 37°C humid chamber. After washing, DAB-starting working solution was added to the specimens and they were incubated.

Kidney sections (4 μm thick) were prepared for immunohistochemistry analysis from rats of each group. Tissue sections were placed on microslides and deparaffinized in xylene, hydrated in increasingly diluted serial ethanol solutions, and immersed in 3% H₂O₂ to quench endogenous peroxidase activity. To permit antigen retrieval, all sections were microwaved for 15 min in Tris-EDTA buffer (pH 9.0), prepared in distilled water. Next, the sections were incubated for 2 h at an ambient temperature with anti-Ki67 (1:200, Abcam, Cambridge, MA, US), anti-Glutathione reductase (GR; 1:2000, Abcam), and anti-Glutathione peroxidase (GPx; 1:2000, Abcam) antibodies. After washing, the slides were incubated using a biotin-free polymeric horseradish peroxidase–linker antibody conjugate system (Dako, Glostrup, Denmark) for 30 min at an ambient temperature. The slides were washed, and chromogen development was performed for 10 min. The slides were counterstained with Meyer’s hematoxylin and mounted using Immu-mount (Fisher Scientific, Geel, Belgium). For image analysis, four randomly selected kidney fields per animal in both the cortex and the medulla (40 kidney fields per group) were photographed. Images were recorded at 40× magnification via a Panoramic viewer program software (3D Histech, Budapest, Hungary). All images were analyzed by using Adobe Photoshop CS2 to quantify signals. The results of proliferation are presented as the percentage of Ki67-positive areas to the total area in each field (cortex and medulla). A semiquantitative analysis was used to express GR and GPx as the percentage of the positive area in a blind fashion.

All results are expressed as means±standard error values. Multiple groups were compared using 1-way analysis of variance followed by the Tukey’s HSD (honest significant difference) test for post hoc comparison. All statistical tests were 2-sided, and statistical significance was defined as p<0.05, p<0.01, or p<0.001. p≥0.05 was not considered significant. The data were analyzed by using IBM SPSS Statistics Version 21 (IBM Corporation, Armonk, NY, US).

The viability of HP-hBMSCs as assessed by trypan blue exclusion assay, was 89.30±7.41%. Flow cytometric analysis showed that HP-hBMSCs were positive for MSC markers CD105 (99.41%), CD73 (99.96%), and HLA-ABC (99.65%), and negative for hematopoietic markers CD45 (0.36%), CD14 (1.42%), CD80 (0.52%), CD86 (0.47%), CD40 (0.23%), and HLA-DR (0.64%) (Fig. 1A). The multilineage differentiation capacity was confirmed after observing the adipogenic and osteogenic differentiation (Fig. 1B).

There were no significant differences in body weight and food intake among the groups (Supplementary Fig. S1A, S1B). The weight of the kidneys increased about 0.6±0.3 g, as compared to the sham group after nephrectomy, but no significant difference was observed between nephrectomy groups (Supplementary Fig. S1C). No gross differences in kidney cross-sections were observed (Supplementary Fig. S1D).

Two and 1 cases of mortality were observed in the renal parenchymal and tail venous injection groups, respectively. Renal arterial injection significantly reduced the extent of elevation in serum BUN values, as compared with those for the IRI control group, 1 and 14 days after IRI (Fig. 2A). Renal arterial injection significantly reduced the extent of elevation in serum creatinine, as compared with the IRI control group, 1, 7, and 14 days after IRI (Fig. 2B). Renal arterial injection significantly reduced the extent of decrease in the GFR, as compared with the GFR for the IRI control group, 2 and 4 days after IRI (Fig. 2C).

Histopathology scores were significantly different from those for the sham, nephrectomy control, and IRI control groups, but there was no significant difference in values for the IRI control and HP-hBMSC injection groups (Fig. 2D, 2E). Sirius red staining for collagen ratio was significantly increased in the IRI control group, as compared to that for the sham group. Sirius red staining in the cortex and medulla indicated that the renal arterial and parenchymal injection groups were significantly less fibrotic than the IRI control group (Fig. 2F, 2G). The TUNEL assay showed that apoptosis was more significantly decreased in the renal arterial injection group than in the IRI control group in both the cortex and medulla (Supplementary Fig. S2A, S2B). We used the Ki67 marker to determine proliferation; it was observed that the renal arterial injection group underwent increased proliferation in the cortex and medulla, as compared to the IRI control group (Supplementary Fig. S2C, S2D). A greater increase in GR and GPx was observed in HP-hBMSC injection groups, especially for the arterial injection group in comparison with the IRI control group, regardless of the cortex or medulla regions (Supplementary Fig. S2F, S2G).

There were no significant differences in body weight and food intake among groups (Supplementary Fig. S3A, S3B). The weight of the kidneys increased about 0.5±0.4 g after nephrectomy, as compared to the sham group, but no significant difference was observed between nephrectomy groups (Supplementary Fig. S3C). No gross differences in kidney cross-sections were observed (Supplementary Fig. S3D).

One case of mortality was observed in the IRI control group. High-dose (4×10⁶) injection significantly reduced the extent of elevation in the serum BUN level, as compared to that for the IRI control group, 1 and 2 days after IRI (Fig. 3A). High-dose injection significantly reduced the extent of elevation in serum creatinine, as compared to the IRI control group, 1 day after IRI (Fig. 3B). High-dose injection significantly reduced the extent of decrease in the GFR, as compared to the IRI control group, 3 days after IRI (Fig. 3C).

Histopathology scores were significantly different from those for the sham, nephrectomy control, and IRI control groups, but there was no significant difference between those for the IRI control and HP-hBMSC injection (Fig. 3D, 3E) groups. Results of Sirius red staining in the cortex and medulla indicated that all HP-hBMSC injection groups were significantly less fibrotic than the IRI control group (Fig. 3F, 3G). The TUNEL assay showed that apoptosis was more significantly decreased in all HP-hBMSC injection groups than in the IRI control group in both the cortex and medulla (Supplementary Fig. S4A, S4B). Ki67 showed that moderate-dose and high-dose injection groups underwent significantly increased proliferation, as compared to the IRI control group in both the cortex and medulla (Supplementary Fig. S4C, S4D). A greater increase in GR and GPx was observed in moderate-dose and high-dose injection groups, as compared to the IRI control groups, regardless of whether the cortex or medulla was analyzed (Supplementary Fig. S4F, S4G).

There were no significant differences in body weight and food intake among groups (Supplementary Fig. S5A, S5B). The weight of the kidneys increased about 0.4±0.4 g after nephrectomy, as compared to that of the sham group, but no significant difference was observed between nephrectomy groups (Supplementary Fig. S5C). No gross differences in kidney cross-sections were observed (Supplementary Fig. S5D).

One case of mortality was observed in the IRI control group. Pre-IRI injection significantly reduced the extent of elevation in serum BUN levels, as compared with those of the IRI control group, 1 day after IRI (Fig. 4A). Concurrent-IRI injection reduced the extent of elevation in serum creatinine levels, as compared with those of the IRI control group, 1 day (p=0.054) after IRI (Fig. 4B). Pre-IRI injection significantly reduced the extent of decrease in GFRs, as compared with GFRs for the IRI control group, 1 day after IRI (Fig. 4C).

Histopathology scores were significantly different from those for the sham, nephrectomy control, and IRI control groups, but there was no significant difference in values for IRI control and HP-hBMSC injection groups (Fig. 4D and 4E). Sirius red staining in the cortex and medulla indicated that all HP-hBMSC injection groups were significantly less fibrotic than the IRI control group (Fig. 4F and 4G). The TUNEL assay showed that the apoptosis was more significantly decreased in all HP-hBMSC injection groups than in the medulla of the IRI control group (Supplementary Fig. S6A, S6B). Ki67 showed that all HP-hBMSC injection groups underwent significantly increased proliferation, as compared to the IRI control group in both the cortex and medulla (Supplementary Fig. S6C and S6D). A greater increase in GR and GPx levels was observed in all HP-hBMSC injection groups, as compared with those in the IRI control group, regardless of the analysis of the cortex or medulla (Supplementary Fig. S6F, S6G).