UMR-106 osteosarcoma cells (UMRs), which are a clonal derivative of a transplantable rat osteosarcoma induced by the injection of radiophosphorus, were purchased from the American Type Culture Collection (ATCC; Manassas, VA, USA). The cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% fetal bovine serum (FBS) and 100 IU/ml penicillin, 100 IU/ml streptomycin and 0.25 μg/ml amphotericin at 37°C in an atmosphere of 5% CO₂, and were passaged by standard methods of trypsinization. Only cells at passages 2 and 3 were used in the experiments.

All experiments were performed following animal protocols approved by the Institutional Animal Care and Use Committee at Stanford University. ADMSCs with permanent expression of a reporter gene signal during follow-up BLI were extracted from the adipose tissues of transgenic mice expressing the β-actin promoter and double reporter genes of GFP and firefly luciferase. This was performed in order to minimize errors in BLI data caused by the viral transfection of reporter genes into ADMSCs. The transgenic mice used for the harvesting of ADMSCs were kindly supplied by Dr Contag’s laboratory at Stanford University. Subcutaneous fat tissues were collected from the lower anterior abdominal wall to the inguinal area of the transgenic mice. The tissue fragments were rinsed 3 times in phosphate-buffered saline (PBS) before being finely minced for 5 min. These tissues were then digested using 0.075% collagenase (Sigma-Aldrich, St. Louis, MO, USA) at 37°C for 1 h. Neutralized cells were centrifuged and the mature adipocytes and fibrovascular fraction were selected and removed. Pelleted stromal cells were passed through a 100-μm cell strainer before plating. Cells were cultured at 37°C in an atmosphere of 5% CO₂, in DMEM containing 10% FBS, 100 IU/ml penicillin, 100 IU/ml streptomycin and 0.25 μg/ml amphotericin. Only the cells that adhered to plastic were used in this experiment, and cells floating during culture were discarded. Preliminary in vitro BLI of cultured ADMSCs was performed to validate the performance of luciferase gene expression before implantation in animals.

Twenty nude mice (NU/NU) aged 8–10 weeks were purchased from Charles River Laboratories (Wilmington, MA, USA) and categorized into 4 groups. Each group was injected with ADMSC-UMR mixtures containing 5, 10, 15, or 25% ADMSCs, defined as G1–G4, respectively. This experiment was performed twice separately due to limitations in stem cell harvest and culture, the preparation of UMRs and BLI and US imaging capacity.

Different numbers of ADMSCs and UMRs were mixed with 30 μl of solubilized basement membrane preparation without growth factor (Matrigel; BD Biosciences, San Jose, CA, USA) and subcutaneously injected into the right flank of the mice. The tumor xenograft was performed in 4 groups at 2 different time points. The first 2 groups, G1 and G3, were injected with mixtures of 5×10⁴ and 1.5×10⁵ ADMSCs in 1×10⁶ UMRs. The second two groups, G2 and G4, were injected with 2×10⁵ and 5×10⁵ ADMSCs in 2×10⁶ UMRs. As a control, 1×10⁶ and 2×10⁶ UMRs in basement membrane preparation were xenografted at the contralateral area of mice in the first (G1 and G3) and second groups (G2 and G4), respectively. The numbers of ADMSCs and UMRs and proportions of ADMSCs in each group are summarized in Table I. These procedures were carried out under general anesthesia with isoflurane inhalation.

A specially designed high-resolution micro-US system for small animal with a 40 MHz transducer (Vevo 770; Visual Sonics, Toronto, Ontario, Canada) was used to evaluate the characteristics of the growing tumor implants with and without ADMSCs. The spatial resolution and penetration depth of the 40 MHz transducer used in this study was 30 μm and 6 mm, respectively. US imaging was performed at 2- or 3-day intervals before sacrificing the animals 16 days after the xenograft. In order to obtain the exact tumor volume taking into account various contours, we assumed each tumor mass was an elliptical ball and obtained two tangential US images along the longitudinal and transverse planes to calculate the exact volume. The equation for the calculation of this mass is as follows: Volume = 4/3πABC, where A, B and C are the measured radii of each elliptical ball, and the unit of the calculated volume was mm³ (Fig. 1).

The changes in the tumor volumes of the ADMSC-treated and control sides were monitored and compared in each group. The ratio of tumor volume (volume of ADMSC-treated site/ volume of control side) was also measured for each mouse, and these results were compared between the groups to reduce the bias from individual variance. The tumor-suppressive or tumor-promoting effect of ADMSCs was determined by the comparison of the ratios of tumor volume on the final day as well as the mean volumes.

An ~30 mg/ml solution of luciferin was made by dissolving 1 g D-luciferin firefly potassium salt (Biosynth International, IL, USA) in 33.3 ml of PBS. Approximately 4.5 mg/150 μl of luciferin was injected into the intraperitoneal space of each mouse. BLI was obtained 10 min after the administration of luciferin using a cooled charge-coupled device camera (IVIS; Xenogen). Mice were imaged at high resolution for 5 min, and five mice were simultaneously imaged in the prone position. BLIs were performed for 16 days with 2-day intervals for G1 and G3, and 3-day intervals for G2 and G4. The acquired BLI data were analyzed by a specialized program (Living Image Software; Xenogen USA). A round region of interest with a 3.2-cm diameter was located in the ADMSC-inoculated right flank to measure the luciferase activity from the ADMSCs. The measured bioluminescence activity was expressed as average radiance (photon/cm²/sec/ steradian).

The rabbit anti-firefly luciferase antibody was obtained from Abcam (cat. no. ab21176, Cambridge, UK). Xenografts were embedded in optimum cutting temperature compound (Sakura Finetek, Torrance, CA, USA) using liquid nitrogen. Sections of 10-μm thickness were fixed in 2% paraformaldehyde for 10 min, and then washed using PBS. Prior to incubation with the antibodies, the sections were blocked using PBS containing 0.1% Triton X-100 and 5% normal goat serum. The primary antibody was diluted (1:50) in blocking solution and applied overnight at 4°C. Alexa 488-conjugated goat anti-rabbit IgG secondary antibody (Invitrogen, Carlsbad, CA, USA) was then added at a 1:100 dilution for 30 min at room temperature. Stained sections were mounted using antifade reagent containing 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI; Vecta Shield, Burlingame, CA, USA) to visualize cell nuclei. Fluorescent microscopic images were captured using the Nikon Eclipse E800 (Nikon Imaging).

The mean volumes of the xenografted tumors on the ADMSC-treated UMR side (right) and the UMR-only control side were compared using the Wilcoxon matched-pairs signed ranks test to determine the differences in tumor volumes in the same subjects. Then the volume ratio of the ADMSC-treated UMR mass to the control tumor was calculated in each mouse from the same group, and the means of these ratios were compared across groups using the Mann-Whitney U test.

The Pearson correlation coefficient was used to evaluate the correlation between BLI activity and tumor volume in G2 and G3. P<0.05 was taken to indicate a statistically significant difference. All statistical evaluations were performed using SAS statistical software (ver. 9.2; SAS Institute Inc., Cary, NC, USA).

The initial volume of injection on both sides of each mouse was ~30 mm³ in all 20 mice in the 4 groups. On post-operation date (POD)2, the volume of xenografts was slightly decreased compared with the initial measurement time due to the absorption of water from the basement membrane preparation used. Thereafter, the volume of the tumors continuously increased on both the ADMSC-treated and control sides, but the increment pattern was different among all 4 groups. In the group containing the lowest proportion of ADMSCs, G1 (5%), the ADMSC-treated xenograft was smaller than the contralateral tumor-only side in 2/5 mice. The remaining 3 mice had a larger tumor volume on the ADMSC-treated side. The injection of mice with higher proportions of ADMSCs (G2 and G3, 10 and 15%, respectively) led to the development of smaller tumors in 4/5 mice (Fig. 2). However, all mice in the group with the highest proportion of ADMSCs, group G4 (25%), showed larger tumor volumes. These results are summarized in Table II. The Wilcoxon matched-pairs signed-ranks test revealed no statistically significant difference in the mean volume of xenografted tumors on the right and left sides from all 4 groups because of the wide range of standard deviations. Thus, the tumor volume ratio of the AMDSC-treated side to the control side was introduced to reduce the bias from individual variance and this value showed significant differences between G2 and G4, and G3 and G4, by the Mann-Whitney U test (Fig. 3). However, the Pearson correlation coefficient between the last tumor volume and initial BLI activity in G2 and G3 was 0.453, which showed no significant correlation between these two factors (P=0.1915).

Initial BLI revealed strong luciferase activity from the ADMSCs in all xenografts, but these activities rapidly decreased at an early stage, and only minimal residual luciferase activities were detected during the later stages. There was no definite increment in BLI to suggest significant proliferation of ADMSCs during the follow-up period and no BLI activity was detected in contralateral tumors to suggest distant migration of the stem cells (Fig. 4)

The DAPI staining of both the control side and ADMSC-treated side revealed well stained nuclei (Fig. 5A and B) but immunohistochemical staining for luciferase in the control side showed no positive finding to suggest any migration of AMDSCs from the primary administration site (Fig. 5C). Fluorescent microscopic images of the ADMSC-treated side displayed a few bright cells stained by the luciferase antibody at the periphery of the tumor mass (Fig. 5D). However, these cells were not detected at the center of the mass, and there was no visible distribution of stained cells along the vascular structures or interstitial tissues, which might have suggested the contribution of ADMSCs to these components.