Lewis lung carcinoma (LL/2) and human embryonic kidney 293 cell lines (purchased from American Type Culture Collection, ATCC, Rockville, MD) were maintained in Dulbecco’s modified Eagle’s medium (Gibco-BRL, Grand Island, NY) supplemented with 10% heat-inactivated fetal bovine serum (Gibco, Auckland, NZ), 2 mM L-glutamine and 100 μg/ml of Amikacin. Human umbilical vein endothelial cells (HUVECs) were isolated from human umbilical cord veins as previously described (22), and grown in EBM-2 medium with SingleQuots (Lonza) containing VEGF and other growth factors. HUVECs at passages 2 to 6 were used for all experiments.

Female C57BL/6 mice 4- to 6-week-old were sacrificed, femurs and tibias were isolated, and whole bone marrow was retrieved by flushing these bones with low-glucose DMEM (Gibco), supplemented with 10% FBS and 50 U/ml penicillin, 50 μg/ml streptomycin. All of these bone marrow cells were then cultured at 37°C in a 5% CO₂ and a 95% humidified atmosphere. Two days later, the nonadherent hematopoietic cells were discarded and the adherent MSCs were preserved for further expansion. The medium was changed twice/week and the MSCs were cultured for 14-days before first passage. Cells of 4–5 passages were used in the experiments (23). MSCs have been proved to express several cell lineage-specific antigens (24). Therefore, the plate-adhering cells were further sorted by flow cytometry using CD34, CD44, CD45, CD73, CD90 and CD105 (BD Biosciences) for phenotype characteristics of MSCs. Isotypic control analyses were conducted in parallel. Flow cytometric analysis was performed on a BD Biosciences FACSCalibur flow cytometer, and data were analyzed with the CellQuest Pro software (BD Immunocytometry Systems).

The adenoviruses were created using the AdEasy system. The viruses were amplified in HEK293 cells and purified on CsCl gradients according to standard methods (25). When the confluence of proliferating MSCs reached or surpassed ∼90%, the cells were transduced with recombinant adenovirus at a multiplicity of infection (MOI) of 1,500 for 4 h. MSCs were also infected with adenovirus-LacZ (Ad-LacZ) at an MOI of 1,500 as a control.

Western blot analysis was performed as described previously (26). Briefly, MSCs were infected with adenoviruses for 4 h and then the virus-containing medium was removed and additionally incubated in low-serum medium (low-glucose DMEM containing 2% FBS). After a further 48 h of incubation, the conditioned media was collected. The secreted proteins in the supernatants of the culture were precipitated by TCA-DOC/acetone and the western blotting assay was performed using a standard method with a mouse anti-human PEDF monoclonal antibody (R&D Systems, Boston, MA, USA) and a biotinylated secondary antibody, and the bands were detected using an enhanced chemiluminescence detection system (Pierce, Rockford, IL, USA). The concentration of the PEDF secreted in the culture supernatants was detected using a sandwich enzyme-linked immunosorbent assay (ELISA) kit for the human PEDF protein (Groundwork Biotechnology Diagnosticate Ltd., San Diego, CA, USA) following the manufacturer’s protocol.

The tube formation assay was conducted as described previously (27) to evaluate the antiangiogenic activity of PEDF produced by MSCs-PEDF. Each well of a 96-well plate was coated with 50 μl Matrigel (BD Biosciences). After polymerization at 37°C for 30 min, HUVECs (2×10⁴/well) were suspended in 200 μl of the conditioned media derived from MSCs, MSCs-LacZ, MSCs-PEDF respectively and seeded onto the Matrigel. Six hours later, cells were viewed and photographed under an inverted microscope with a digital camera. The tubes of five fields were counted and the data were averaged.

To determine whether PEDF secreted by MSCs-PEDF can inhibit HUVEC invasion, a Transwell invasion assay was performed as previously described (28). Briefly, the filter of the Transwell chamber (Millipore) was coated with 50 μl Matrigel (BD Biosciences). After Matrigel polymerization, HUVEC (2×10⁴/well) were suspended in 200 μl of the conditioned media derived from MSCs, MSCs-LacZ, MSCs-PEDF respectively and loaded in the upper chamber. The lower well of the Transwell plate was filled with 600 μl of EBM-2 medium containing various growth factors. Cells were allowed to migrate for 24 h at 37°C. Non-migrated cells were scraped with a cotton swab, and migrated cells were fixed with 100% methanol and stained with 0.05% crystal violet. The number of cells that had migrated to the lower side of the filter was quantified by manual counting under a light microscope with five fields (x100). All assays were conducted in triplicate.

An alginate-encapsulated tumor cell assay was conducted as previously described (26). Alginate beads were formed with about 1×10⁵ LLC cells per bead and implanted s.c. into both dorsal sides of C57BL/6 mice. Then the mice were injected with PBS, MSCs-LacZ, MSCs-PEDF or free Ad-PEDF via the tail vein on Day 2 for once. Twelve days after the beads were implanted, the mice were injected intravenously with a dose at 100 mg/kg FITC-dextran solution (Sigma) in 100 μl. Alginate beads were removed and photographed with a digital camera within 20 min after FITC-dextran administration. The uptake of FITC-dextran was quantified against a calibration curve of FITC-dextran.

Female syngeneic C57BL/6 mice 6–8 weeks of age were purchased from the West China Experimental Animal Center of Sichuan University (Sichuan, China) and maintained in a dedicated pathogen-free environment. All animal procedures were approved by the Institute’s Animal Care and Use Committee. Mice were implanted s.c. in the right flank with LLC cells at 5×10⁵/mouse. When tumor diameters reached 3 mm, mice were randomly divided into four groups and i.v. injected with PBS, 10⁸ plaque forming units (PFU) of Ad-PEDF, 5×10⁵ MSCs-LacZ or 5×10⁵ MSCs-PEDF at 4 days intervals for two times. Tumor growth and survival rate were monitored every 3 days by caliper. Tumor volumes were determined at various time points using the formula: width² × length × 0.52. Cohorts of mice (n=8) from each group were sacrificed at Day-30 to determine the effect of therapy and remaining mice (n=10) were monitored for long-term survival. The duration of survival was recorded when the mouse died or had to be sacrificed secondary to tumor diameter >20 mm, tumor ulceration or bleeding. The difference in survival was determined by a log-rank test.

To determine the degree of tumor-induced angiogenesis, frozen sections were prepared from the tumor tissue 30 days after inoculation. The procedure of immunofluorescence staining was done as described previously (29). Briefly, sections were probed with a monoclonal anti-CD31 antibody (BD Biosciences) and nuclei were stained with DAPI. The sections were viewed at low magnification (x40) to identify the most vascular-rich area (hot spot) and 5 non-overlapping fields were selected. Microvessel counting was performed at a high-power field (×200) under fluorescence microscopy and data were expressed as the mean number of CD31-positive vessels per field from three sections in each tumor.

To determine the degree of apoptosis, terminal dUTP nick-end labeling (TUNEL) staining was performed using an in situ cell death detection kit (Promega) following the manufacturer’s protocol. Tumor species were prepared as described above. Three tumors from each group were analyzed. The number of apoptotic cells was quantified in 5 randomly selected fields at ×200 magnifications using a fluorescence microscope.

Serum samples and fresh tumor tissues were harvested and detected using ELISA (Groundwork Biotechnology Diagnosticate, Ltd.) in triplicate for the tendencies of PEDF levels at 7, 14 and 21 days after i.v. administration of treatment. Non-necrotic fresh tumor tissues were cut into small pieces, lysed in RIPA buffer on ice, passed through a sieve (BellCo Glass), and evaluated for the local production of PEDF as ng/mg.

All values are presented as means ± SEM (standard error of the mean). SPSS 17.0 was used for statistical analysis. Significance was evaluated using one-way ANOVA, and P<0.05 was considered significant differences. The survival rates were compared by means of the log-rank test.

The adherent cells have a typical spindle-like morphology (Fig. 1B). Analysis of immunophenotype showed that isolated BM-derived cells shared classical immunophenotype of MSCs, including positivity for CD44, CD73, CD90, CD105, but negativity for CD34, CD45 (Fig. 1A). Thus, as the data showed, we successfully obtained abundant MSCs from mouse bone marrow.

After BMSCs were cultured to reach 90% confluence and incubated with adenoviruses at a MOI of 1,500 for 4 h, the secreted PEDF in the supernatants of conditioned media was verified by western blotting. As shown in Fig. 1C, a specific 50 kDa band was observed only in the supernatants from Ad-PEDF-transduced MSCs, which indicated that PEDF was appropriately synthesized and processed. PEDF levels in the collected media (n=3) were determined utilizing a sandwich enzyme-linked immunosorbent assay (ELISA) kit for the human PEDF protein (Groundwork Biotechnology Diagnosticate Ltd.). It showed that PEDF was produced in vitro in great amounts for 48 h, with a concentration as high as 78.8±4.8 ng/ml (Fig. 1D).

HUVECs when seeded on Matrigel become elongated and form capillary-like structures mimicking the in vivo neoangiogenesis process (30). We used this assay to examine the antiangiogenic activity of PEDF expressed by MSCs-PEDF. As shown in Fig. 2A, HUVECs plated on the surface of Matrigel formed capillary-like structures in the MSCs and MSCs-LacZ groups after 6 h, however, treatment with the conditioned media (CM) from MSCs-PEDF dramatically blocked the tube formation (P<0.01). We also conducted Transwell invasion assays to evaluate the ability of HUVECs to pass through the Matrigel and membrane barrier of the Transwell in the presence of the CM derived from MSCs, MSCs-LacZ and MSCs-PEDF. Data showed that the CM from the MSCs-PEDF significantly inhibited the invasion properties of endothelial cells (Fig. 2B) (P<0.01).

To learn if the PEDF secreted in vivo by PEDF gene-modified MSCs can engender a beneficial effect against LLC cells in vivo, we implanted isogenic C57BL/6 mice with LLC cells s.c. and treated with PBS, 10⁸ PFU of Ad-PEDF, 5×10⁵ MSCs-LacZ or 5×10⁵ MSCs-PEDF at 4 days intervals for two times i.v. Cohorts of mice (n=8) from each group were sacrificed at Day-30 to determine the effect of therapy and remaining mice (n=10) were monitored for long-term survival. As shown in Fig. 3A, the tumor volume in the MSCs-PEDF group was dramatically smaller compared with that of the PBS group (1038.8±139.0 vs. 2897.7±274.9 mm³, ANOVA; P<0.01). By contrast, the growth rate of tumors was not affected by administration of free Ad-PEDF (2622.8±359.1 mm³) and MSCs transduced with Ad-LacZ (3019.2±360.9 mm³). These data suggest that intravenous administration of MSC-PEDF cells inhibits the growth of tumors, whereas systemically injected free Ad-PEDF or MSC-LacZ cells does not.

Moreover, MSCs producing PEDF was found to significantly increase the survival of mice. All animals in the control groups died by Day 42, whereas 40% of mice treated with MSC-PEDF survived beyond that time (P<0.01). There was no significant increase in survival of mice treated with MSC transduced with Ad-LacZ or free Ad-PEDF (Fig. 3B).

Immunofluorescence anti-CD31 staining was performed to evaluate the consequence of anti-angiogenesis therapy. The most highly vascularized area of each tumor section was identified on low power and five high-powered fields were counted in this area of greatest vessel density. The tumor tissue from MSCs-PEDF-treated mice showed apparently decreased microvessel density compared with the other three control groups (Fig. 4A). The inhibition of angiogenesis was confirmed in the alginate-encapsulated tumor cell assay. The treated tumor exhibited relatively little vascularity (Fig. 4B). Vascularization of beads over 12 days can then be measured by uptake of FITC-dextran into beads. Vascularization of alginate beads was apparently reduced, and FITC-dextran uptake was decreased in MSCs-PEDF-treated mice compared with that in controls (ANOVA; P<0.01).

The TUNEL assay was performed to detect tumor cell apoptosis to further investigate the role of MSCs-PEDF treatment in tumors in vivo. As shown in Fig. 5, MSCs-PEDF-treated tumors showed significantly more apoptotic cells (with green nuclei) than tumors from the PBS, Ad-PEDF or MSCs-LacZ-treated groups. The apoptosis index was also significantly higher in the MSCs-PEDF-treated group compared with the controls (ANOVA; P<0.01). The inhibition of angiogenesis and the consequent induction of apoptosis underlie the mechanism of tumor volume decrease and the lifespan prolongation in tumor-bearing mice.

The systemic and local (serum and intratumoral) expression tendencies of PEDF at indicated points by ELISA are shown in Fig. 6. In the MSCs-PEDF group, the serum levels of PEDF were ∼38 ng/ml on Day 7, and declined to ∼26 ng/ml after that; meanwhile, the intratumoral levels of PEDF rose from ∼168 ng/mg (7 days) to ∼253 ng/mg (14 days), and then slightly declined to ∼238 ng/mg in the 21 days, ∼7-fold vs. the intratumoral level in the free Ad-PEDF group (32 ng/mg). In free Ad-PEDF group, the levels of PEDF, diminished quickly from ∼60 to 25 ng/ml in serum and from ∼50 to ∼32 ng/mg intratumorally. Other controls including PBS, MSCs-LacZ hardly led to any elevation of serous or intratumoral levels of PEDF.