Establishment of a canine hemangiosarcoma (HSA) xenograft model. A biopsy was taken from a subcutaneous HSA of the right shoulder of a 9-year-old German shepherd dog. Several small pieces of the HSA biopsy were engrafted subcutaneously in the dorsal cervical area of a bg/nu/XID mouse and the skin wound was sutured closed. The mouse was monitored for 120 days, at which time a 1-cm-diameter tumor was present and the animal was sacrificed. The tumor was excised and a portion was snap-frozen with liquid nitrogen in OCT compound for immunohistochemical staining and the remainder was used for establishing the primary canine HSA cell line. After 14 passages, the cells were sorted and cells labeling positive for expression of α_v3 integrin (with a canine-selective antibody) and CD146 (P1H12) were collected and expanded. About 5 x 10⁶ cells were injected subcutaneously into the dorsum of a second bg/nu/XID mouse and allowed to grow for 30 days until a 1-cm³ tumor had developed. The tumor was then excised. A portion was snap-frozen in OCT compound for immunohistochemical staining, and the remaining tumor was used to establish a single-cell suspension that was then expanded. These cells, called the SB-HSA cell line developed from the second passage in a bg/nu/XID mouse, were used for the remainder of the experiments to induce tumors in NOD-SCID mice.

SB-HSA tumors in NOD-SCID mice are histologically identical to canine hemangiosarcoma. The histological appearance of SB-HSA tumors grown in NOD-SCID mice, shown above, is indistinguishable from naturally occurring hemangiosarcoma in the dog. At low (x 40) power (A), there are sheets of tumor cells with blood-filled spaces and areas of necrosis (light pink, lower right). At high (x 400) power (B), plump atypical endothelial cells (black arrows) can be seen forming tortuous vascular spaces (white arrows).

Expression of endothelial growth factors and receptors by SB-HSA cells. (A) RT-PCR was used to detect messenger RNA for endothelial cell growth factors (VEGF, bFGF, and IL-8) and some of their receptors (VEGFR-1 and VEGFR-2). Other molecules of interest included PTEN, MMPs (MMP-2 and MMP-13), and c-kit. Primers were designed using canine sequences and β-actin was used as a control. Expression of VEGFR-1 and VEGFR-2 is considered selective for cells of the endothelial lineage. (B) Flow cytometry using an anti-c-kit antibody demonstrated bright expression of c-kit by SB-HSA cells confirming that the protein was expressed for one of the transcripts amplified by RT-PCR.

IL-12 suppresses growth of canine hemangiosarcoma. Cohorts (n = 7) of NOD-SCID mice were injected subcutaneously with 5 x 10⁶ SB-HSA cells over the dorsal thorax. Two days later, mice were anesthesized and small osmotic pumps containing PBS or murine recombinant IL-12 (IL-12) were inserted subcutaneously in the mice from respective cohorts. The pumps were designed to deliver continuous subcutaneous infusion of 6 µl/day for 4 weeks. The mice in the IL-12 group each received 1 µg/day IL-12 during that time period. At the conclusion of the study, the pumps were examined to verify that the contents had been delivered. Compared to the control (PBS) group, tumors in the mice that received IL-12 were significantly smaller (P = .016). Bars are standard errors.

SB-HSA cells produce tumors in immunodeficient mice. (A) Injection of 5 x 10⁶ SB-HSA cells subcutaneously into NOD-SCID mice routinely produced large tumors within 3 to 4 weeks. Such a tumor has grown on the dorsal midline of the mouse in this photo. Inset: The gross appearance of SB-HSA tumors from NOD-SCID mice exhibits features indistinguishable from hemangiosarcoma in the dog. There is a characteristic bright red/maroon color to the tumor and the tumor bled excessively when cut. The tumor was highly vascular. Although implanted in the subcutaneous tissue, the tumors were invasive into the underlying musculature. (B) NOD-SCID mice were injected intravenously with 70,000 SB-HSA cells through the tail vein. After 2 months, mice were sacrificed and found to have multifocal maroon tumor nodules (5 mm diameter) disseminated throughout the lungs. This route of administration routinely produced pulmonary metastases.

Immunohistochemical staining confirms endothelial lineage of SB-HSA cells. Immunohistochemical staining was performed on fresh frozen and formalin-fixed, paraffin-embedded sections from SB-HSA tumors obtained from bg/nu/XID and NOD-SCID mice. Anti-factor VIII polyclonal antibody (A and B) and anti-CD31 (C and D) were used on fixed tissue sections following antigen retrieval, and P1H12 (E and F) was used on fresh frozen sections. The final product was developed using alkaline phosphatase with New Fuchsin as the chromogen. Positive staining is appreciated as a red color. Expression of these markers supports an endothelial lineage for the cells comprising the SB-HSA tumor. Original magnification, x40 (panels A, C, and E) and x 100 (panels B, D, and F).

SB-HSA cells are potent inducers of angiogenesis and IL-12 inhibits this effect. (A) A polyvinyl sponge containing 70,000 SB-HSA cells was placed into a surgically created corneal pocket in a BALB/c mouse. After 12 days, a robust angiogenic response is visible in the cornea, with neovascularization arising from the limbus and proceeding toward the sponge implant (white structure in the center of the cornea). Implanted sponges that did contain SB-HSA cells failed to provoke an angiogenic response. Experiment shown is representative of eight mice, all exhibiting a similar response. (B) Matrigel, an artificial connective tissue matrix, was injected subcutaneously in the flank of a BALB/c mouse. After solidifying, a nick was made into the Matrigel through a skin incision, and a polyvinyl sponge (arrow) containing 5 x 10⁶ SB-HSA cells was inserted into the Matrigel. After 2 weeks and 3 to 5 minutes prior to sacrifice, 200 µl of FITC-dextran was injected into the tail vein and the Matrigel plug was removed. Under fluorescence microscopy, a pronounced angiogenic response was observed (arborized fluorescent green vessels), which progressed toward the proliferating SB-HSA cells in the sponge. These experiments confirm the proangiogenic properties of SB-HSA cells. The effect of IL-12 on SB-HSA-induced corneal angiogenesis was compared in BALB/c control mice given PBS (C) by continuous subcutaneous infusion versus BALB/c that received 1 µg/day mrIL-12 for 7 days by continuous subcutaneous infusion (D). Mice treated with IL-12 had significantly less corneal neovascularization compared to the control mice (P < .05). Mice were injected with FITC-dextran just prior to euthanasia and corneal vessels were evaluated with fluorescence microscopy. In this example, there was a 56% reduction in corneal neovascularization in the IL-12-treated mouse. The fluorescent vessel in the ventral aspect of the cornea in (D) is the normal limbic vessel and the green squares in the centers of (C) and (D) are the SB-HSA sponges that have absorbed fluorescein. Note that vessels have nearly reached the sponge in (C) whereas the cornea in (D) remains relatively avascular.