An mTOR and VEGFR inhibitor combination arrests a doxorubicin resistant lung metastatic osteosarcoma in a PDOX mouse model

In order to identify more effective therapy for recalcitrant osteosarcoma, we evaluated the efficacy of an mTOR-VEGFR inhibitor combination on tumor growth in a unique osteosarcoma patient-derived orthotopic xenograft (PDOX) mouse model derived from the lung metastasis of an osteosarcoma patient who failed doxorubicin therapy. We also determined the efficacy of this inhibitor combination on angiogenesis using an in vivo Gelfoam fluorescence angiogenesis mouse model implanted with osteosarcoma patient-derived cells (OS-PDCs). PDOX models were randomly divided into five groups of seven nude mice. Group 1, control; Group 2, doxorubicin (DOX); Group 3, everolimus (EVE, an mTOR and VEGF inhibitor); Group 4, pazopanib (PAZ, a VEGFR inhibitor); Group 5, EVE-PAZ combination. Tumor volume and body weight were monitored 2 times a week. The in vivo Gelfoam fluorescence angiogenesis assay was performed with implanted OS-PDCs. The nude mice with implanted Gelfoam and OSPDCs also were divided into the four therapeutic groups and vessel length was monitored once a week. The EVE-PAZ combination suppressed tumor growth in the osteosarcoma PDOX model and decreased the vessel length ratio in the in vivo Gelfoam fluorescent angiogenesis model, compared with all other groups (p < 0.05). There was no significant body-weight loss in any group. Only the EVE-PAZ combination caused tumor necrosis. The present study demonstrates that a combination of an mTOR-VEGF inhibitor and a VEGFR inhibitor was effective for a DOX-resistant lung-metastatic osteosarcoma PDOX mouse model, at least in part due to strong anti-angiogenesis efficacy of the combination.

We have also established a fluorescent-protein based angiogenesis mouse model that can image angiogenesis induced by cancer cells in Gelfoam (Pfizer, New York, NY, U.S.A.) implanted in mice [29][30][31][32] . A previous study demonstrated strong angiogenesis by an osteosarcoma cell line using this model 31 .
In the present study, we assessed the efficacy of the combination of EVE and PAZ on a unique osteosarcoma PDOX mouse model derived from the lung metastasis of an osteosarcoma patient who failed DOX therapy. In addition, we assessed the efficacy of the EVE-PAZ combination on vessel length ratio in the in vivo Gelfoam angiogenesis mouse model implanted with osteosarcoma-patient-derived cells (OS-PDCs).

Results
Efficacy of the combination of EVE and PAZ on osteosarcoma PDOX tumor growth. The therapeutic schedule for the DOX-resistant lung metastasis osteosarcoma PDOX mouse models is presented in Fig. 1. The treatment period was for 14 days.
Tumor volume ratios at the end of the study were as follows: 4.70 ± 0.58 in Group 1 (untreated control group); 3.97 ± 0.53 in Group 2 (DOX); 3.89 ± 0.80 in Group 3 (EVE); 4.16 ± 1.01 in Group 4 (PAZ); and 1.70 ± 0.30 in Group 5 (EVE and PAZ). DOX, EVE-alone, and PAZ-alone showed no significant difference in the tumor-volume ratio compared with the untreated control group (p = 0.24, 0.38, 0.94, respectively). However, the combination of EVE and PAZ arrested tumor growth and had a significant difference in relative tumor volume compared with all other groups (p = 0.018) (Fig. 2). Representative photographs of tumors in situ and resected tumors also showed strong efficacy of the combination of EVE and PAZ (Fig. 3).
Histopathology in the treated and untreated osteosarcoma PDOX mouse model. The untreated control-group tumor had a high density of pleomorphic and spindle-shaped osteosarcoma cells. Osteosarcoma cells in the DOX, EVE-alone, and PAZ-alone groups had a slightly lower density compared with the untreated-control group. However, the combination of EVE and PAZ resulted in a lower density of osteosarcoma cells and had more necrosis compared with all other groups (Fig. 4).
Body weight in the osteosarcoma PDOX mouse model. Animal deaths were not observed in any group. Body weight ratios at the end of the study were 1.05 ± 0.04 in Group 1 (untreated control); 1.04 ± 0.04 in Group 2 (DOX); 1.04 ± 0.06 in Group 3 (EVE); 1.04 ± 0.03 in Group 4 (PAZ); and 0.99 ± 0.05 in Group 5 (EVE and PAZ).
The combination of EVE and PAZ resulted in a slight body weight loss. However, no significant differences were observed among any group (Fig. 5).   Representative vessel images obtained with the FV1000 confocal microscope in each group are shown in Fig. 7. The time-dependent change of the vessel length ratio is presented in Fig. 8.

Discussion and conclusion
In the present study, the combination of EVE and PAZ demonstrated anti-tumor efficacy in the DOX-resistant lung-metastatic osteosarcoma PDOX mouse model compared with the untreated control, DOX, EVE-alone, and PAZ-alone. Moreover, the combination of EVE and PAZ had significant anti-angiogenesis efficacy in the in vivo Gelfoam angiogenesis OS-PDC mouse model compared to the untreated control and all other treatment.
EVE, an mTOR inhibitor, has anti-proliferative activity through inhibition of the PIK3/AKT/mTOR pathway and also has anti-angiogenesis activity by reducing production of hypoxia-inducible factor 1α and VEGF [5][6][7][8] . EVE has been used in the treatment of advanced breast and renal-cell cancer 3,4 . However, mTOR inhibitors have not shown efficacy against osteosarcoma clinically 40,41 . PAZ, a VEGFR inhibitor, inhibits the VEGF/VEGFR system, which diminishes new blood vessel formation [11][12][13] . Previous reports have shown that high expression of VEGF/  www.nature.com/scientificreports/ VEGFR in osteosarcoma patients is a poor-outcome predictor [42][43][44] . Therefore, it is thought that both VEGF and VEGFR contribute to poor prognosis of osteosarcoma patients. PAZ had efficacy in some advanced cases of sarcoma 45,46 . However, there has not been a report that a VEGFR inhibitor could prolong overall survival and progression-free survival of osteosarcoma patients. Blockade of angiogenesis in two different pathways with an mTOR inhibitor and a VEGFR inhibitor has been demonstrated to increase anti-tumor efficacy against several cancers in preclinical and clinical studies 15,[47][48][49] . The combination of an mTOR and VEGFR inhibitor demonstrated 3.6 months progression free survival and 9.1 months overall survival in a urothelial carcinoma clinical trial 15 . In the present study, the efficacy of the mTOR and VEGFR inhibitor combination on osteosarcoma is the first report to the best of our knowledge.
We previously developed the in vivo Gelfoam angiogenesis mouse model that can image nascent vessel formation with GFP [29][30][31][32] . Our previous studies demonstrated strong induction of angiogenesis with an osteosarcoma cell line and anti-angiogenesis efficacy of tumor-targeting Salmonella typhimurium A1-R 31,32 . The present study indicated that combination of an mTOR and a VEGFR inhibitor reduced angiogenesis induced by the OS-PDC mouse model. The result of the present study also suggested that the anti-angiogenesis activity of the combination of mTOR and VEGFR inhibitors contributed to the efficacy on osteosarcoma tumor-growth inhibition.  www.nature.com/scientificreports/ Future gene-profiling studies will determine how the level of mTOR and VEGFR affect the response to the EVE-PAZ combination in the present and other PDOX models of osteosarcoma. Future experiments will also study the efficacy of the EVE-PAZ combination on the tumor micro-environment. The EVE-PAZ combination will also be studied for anti-metastatic efficacy and survival on the present PDOX model in future long-term experiments. Future experiments will also study the osteosarcoma cells that resisted DOX in the PDOX model to determine if they have increased DOX resistance and we will establish PDOX models from these super-DOX resistant osteosarcoma cells.
The results of the present study demonstrates that the combination of mTOR and VEGFR inhibitors is a promising therapeutic strategy for DOX-resistant recalcitrant metastatic osteosarcoma, in part due to strong anti-angiogenesis activity. Long-term treatment studies in additional osteosarcoma PDOX models will be necessary before clinical translation of the novel treatment strategy described in the present report. The osteosarcoma PDOX model mimiced the DOX resistant of the tumor in the patient.  www.nature.com/scientificreports/

Materials and methods
Mice. 4-6 weeks old athymic nu/nu nude mice and nestin-regulatory-element-driven green fluorescent protein (ND-GFP) transgenic nude mice (AntiCancer Inc., San Diego, CA) were used in the present study for the PDOX and in vivo Gelfoam angiogenesis assay study, respectively. Mouse housing, imaging, and surgical procedures were performed as described in our previous publications [33][34][35][36] . The mice were humanely sacrificed as described in our previous publications [33][34][35][36]  Patient-derived tumor. Informed consent from the patient and her parents was previously obtained under an Institutional Review Board-approved protocol of University of California, Los Angels (UCLA) (IRB #10-001857) 38 . A 16-year-old female with high-grade osteosarcoma of left-distal-femur received neoadjuvant chemotherapy with DOX and had surgery with replacement of the distal femur. One year after surgery, three bilateral metachronous pulmonary metastases recurred, which were resected at UCLA. The resected tumor was previously established as an osteosarcoma PDOX mouse model 38 . All methods performed with the patient-derived tumor were performed with the relevant guidelines and regulations.

Establishment of the osteosarcoma PDOX mouse model by surgical orthotopic implantation (SOI).
The osteosarcoma tumor from the lung metastasis was previously initially implanted on the subcutaneous back space of nude mice to establish the tumor 38 . The subcutaneously-grown tumors were excised and 9 mm 3 small fragments were made from excised tumor. After mice were put under anesthesia using a ketamine solution, a 10-mm skin incision was made on the lateral side of right knee. The vastus lateralis muscle was cut and divided to expose the lateral condyle of the distal femur. The lateral condyle of distal femur was cut with a scissors to make a cavity for implantation of the tumor fragment. A single 9 mm 3 tumor fragment was orthotopically implanted the resected cavity of the lateral condyle in order to establish a PDOX mouse model. The wounds in the muscle and skin were sutured as described in our previous publication 38 . Therapeutic study design in the osteosarcoma PDOX mouse model. Five groups of seven mice were randomly assigned from the osteosarcoma PDOX mouse models as follows: Group 1, control with PBS, i.p. (i.p., n = 7); Group 2, DOX alone (3 mg/kg, i.p., weekly for 2 weeks, n = 7); Group 3, EVE alone (5 mg/kg, oral gavage, daily for 2 weeks, n = 7); Group 4, PAZ alone (50 mg/kg, oral gavage, daily for 2 weeks, n = 7); Group 5, combination of EVE and PAZ (EVE: 5 mg/kg, PAZ: 50 mg/kg, n = 7) (Fig. 1). Treatment was started 21 days after orthotopic implantation of the tumor. All mice were sacrificed on day 15 after treatment initiation. www.nature.com/scientificreports/ The long and short diameters of tumors were monitored with calipers and body weight was monitored with a digital weight scale twice a week. The tumor volume was calculated using the following formula: Tumor volume (mm 3 ) = long diameter (mm) x short diameter (mm) x short diameter (mm) × 1/2. The tumor volume ratio was equal to the tumor volume at each time point divided to the initial time point.
Histopathology. Resected tumor specimens were placed in 10% formalin for fixation. The tumor specimens were embedded in melted paraffin before sectioning. Tissue sections (5 μm) were rehydrated using ethanol and deparaffinized using xylene. Hematoxylin and eosin (H&E) staining was conducted with standard protocols 27,32 . Implantation of Gelfoam. ND-GFP transgenic nude mice were put under anesthesia with the ketamine solution. Basic fibroblast growth factor (βFGF; Millipore, Billerica, MA, U.S.A., 300 ng) and RPMI-1640 medium (75 μl) were added to Gelfoam blocks (5 × 5 mm). The treated Gelfoam was then implanted in the subcutaneous space of ND-GFP transgenic nude mice 29-31 . Primary cell culture of OS-PDCs. OS-PDCs were established as follows: tumors were minced mechanically with scissors and seeded in DMEM (Sigma-Aldrich, St. Louis, MO) with 10% fetal bovine serum (GIBCO, Grand Island, NY), 100 U penicillin G, and 100 μg/ml streptomycin (Gibco) in 100 mm plastic dishes. The minced tumors with medium were cultured at 37 °C, in a 5% CO 2 incubator. 24 h after seeding, medium was changed to remove loose debris and unattached cells.

Injection of OS-PDCs into mouse-implanted Gelfoam. Seven days after implantation of Gelfoam,
OS-PDCs (5 × 10 5 cells) were injected into the previously implanted Gelfoam in ND-GFP transgenic nude mice which express GFP in nascent blood vessels. ND-GFP transgenic nude mice were put under anesthesia with the ketamine solution. A skin incision was made on the back from the shoulder to the buttock. The subcutaneous tissue was exfoliated to make a skin flap without damaging the vessels. The skin flap was unfolded and immobilized on a flat stand. A 6-0 nylon suture was used to suture the skin after imaging 30,31,39 . Imaging and measurement of fluorescent vessel length. Skin flaps were made for imaging at 22, 29, and 36 days after implantation of Gelfoam. The FV1000 confocal microscope with an XLUMPLFLx20x (0.95 numerical aperture (NA)) water immersion objective (Olympus Corp., Tokyo, Japan) was used to visualize GFPexpressing blood vessels. GFP had a peak excitation at 488 nm with an argon laser. FV10-ASW Fluoview software (Olympus) was used to display image. Measurement of the vessel length was conducted with FV10-ASW Fluoview software. A 6-0 nylon suture was used to suture the skin after imaging 31 . Therapeutic study design in the in vivo Gelfoam fluorescence angiogenesis mouse model with OS-PDCs. In vivo Gelfoam fluorescence angiogenesis mouse models were randomly assigned into the control and four-therapeutic groups the same as the PDOX model groups described above (Fig. 6). Imaging of vessel length measurements started on day-22, which is 14 days after implantation of OS-PDCs into the Gelfoam (Fig. 2). The vessel-length ratio was determined on the vessel length at any given time point relative to the vessel length on day 22. For each mouse, eight fields were selected to measure vessel length. Statistical analysis. JMP version 13.0 (SAS Institute Japan, Tokyo, Japan) was employed for statistical analyses. The Steel-Dwass test was used to analyze the tumor volume ratio and vessel length ratio, and the Tukey-Kramer test was used to analyze body weight ratio. Line graphs or bar graphs show the mean data, and error bars show ± standard deviation. A probability value of p < 0.05 was determined to have significant difference. www.nature.com/scientificreports/