Thrombospondin-1 and CD47 limit cell and tissue survival of radiation injury

Abstract

Radiation, a primary mode of cancer therapy, acutely damages cellular macromolecules and DNA and elicits stress responses that lead to cell death. The known cytoprotective activity of nitric oxide (NO) is blocked by thrombospondin-1, a potent antagonist of NO/cGMP signaling in ischemic soft tissues, suggesting that thrombospondin-1 signaling via its receptor CD47 could correspondingly increase radiosensitivity. We show here that soft tissues in thrombospondin-1-null mice are remarkably resistant to radiation injury. Twelve hours after 25-Gy hindlimb irradiation, thrombospondin-1-null mice showed significantly less cell death in both muscle and bone marrow. Two months after irradiation, skin and muscle units in null mice showed minimal histological evidence of radiation injury and near full retention of mitochondrial function. Additionally, both tissue perfusion and acute vascular responses to NO were preserved in irradiated thrombospondin-1-null hindlimbs. The role of thrombospondin-1 in radiosensitization is specific because thrombospondin-2-null mice were not protected. However, mice lacking CD47 showed radioresistance similar to thrombospondin-1-null mice. Both thrombospondin-1- and CD47-dependent radiosensitization is cell autonomous because vascular cells isolated from the respective null mice showed dramatically increased survival and improved proliferative capacity after irradiation in vitro. Therefore, thrombospondin-1/CD47 antagonists may have selective radioprotective activity for normal tissues.

Figure 1

TSP1 and CD47 limit survival of irradiated soft tissue. A: Age- and sex-matched C57BL/6 wild-type, TSP1-null, and CD47-null mice received 25 Gy of irradiation to the right hindlimb. B: Tissue changes were assessed every week, and scores are presented for 2 months. Significance was determined using the independent two-tailed t-test. *P < 0.05 versus wild type. C: Representative H&E-stained sections of muscle and skin are shown from irradiated hindlimbs harvested after 2 months. Original magnifications, ×20. D: Mitochondrial viability of hindlimb muscle biopsies was assessed at 2 months after irradiation by the reduction of a tetrazolium salt to water insoluble formazan through mitochondrial oxidation as described. Significance was determined using the independent two-tailed t-test. *P < 0.05 versus wild-type nonradiated. Results were expressed as absorbance normalized to dry tissue weight. E: Copy numbers in control and irradiated muscle tissue harvested at 2 months were determined for two mitochondrial genes (NADH6 and 16s rRNA, mt-Rnr2) and two nuclear genes (the general transcription factor Gtf2ird1 and osteocalcin, Bglap2) by quantitative real-time PCR. For each sample, results were normalized to the nuclear gene Pkd1.

Figure 2

TSP2 does not modulate radiation-induced tissue damage. Age- and sex-matched B6129sf1/J wild-type and TSP2-null mice received 25 Gy to the right hindlimb. A: Tissue changes were assessed every week and scored throughout the next 2 months. B: Mitochondrial viability of hindlimb muscle biopsies was assessed at 2 months by the reduction of a tetrazolium salt to water insoluble formazan through mitochondrial oxidation as described. Results were expressed as absorbance normalized to dry tissue weight.

Figure 3

Limb flexibility is preserved in irradiated tissue in the absence of TSP1. Limb extension in age- and sex-matched C57BL/6 wild-type, TSP1-null, and CD47-null mice was measured as described 8 weeks after 25 Gy to the right hindlimb. Significance was determined using the independent two-sample t-test. *P < 0.05 versus wild type nonradiated.

Figure 4

TSP1 limits tissue vascular response after irradiation. A: Age- and sex-matched wild-type and TSP1-null mice received 25 Gy to the right hindlimb. Eight weeks later limb perfusion was assessed via BOLD MRI. Images were obtained for 30 minutes from T₂*-weighted gradient echo sequences. DEA/NO (100 nmol/g body weight) was administered 5 minutes after starting the scan. The percent change in integrated BOLD values as a function of time is presented as mean ± SE of eight pairs of wild-type and TSP1-null mice, respectively. B: Representative multislice multiecho, T₂ maps, and BOLD images show normal and irradiated hindlimbs of wild-type and TSP1-null animals. In all images, the irradiated (XRT) limb is on the left.

Figure 5

The absence of TSP1 preserves blood oxygenation responses to NO in irradiated hindlimbs. Positive (top) and negative (bottom) BOLD MRI signal curves are shown after NO treatment for irradiated (open symbols) and normal hindlimbs (filled symbols) in age- and sex-matched wild-type and TSP1-null 8 weeks after 25 Gy to the right hindlimb.

Figure 6

TSP1 and CD47 modulate irradiated hindlimb responses to vasoactive challenge. Age- and sex-matched C57BL/6 wild-type, TSP1-null, and CD47-null mice received 25 Gy irradiation to the right hindlimb. Eight weeks after irradiation, limb perfusion was assessed by laser Doppler imaging for 45 minutes after NO vasoactive challenge (1 μl/g weight 100 mmol/L DEA/NO stock via rectal bolus). A: Animals were maintained at a core temperature of 35.5°C and 1.5% isoflurane anesthesia. Significance was determined using the independent two-tailed t-test. *P < 0.05 versus wild type. B: Cutaneous skin sections 1 × 2 cm in dimension were harvested from age- and sex-matched mice, stained for H&E, and vessels counted.

Figure 7

Endogenous TSP1 and CD47 sensitize muscle and bone marrow to radiation-induced apoptosis. A: The presence of apoptosis within mouse hindlimb tissue was examined by immunohistochemistry using a peroxidase in situ apoptosis detection kit. Nonirradiated control hindlimbs (top) and hindlimbs irradiated for a total of 25 Gy from the indicated strains were prepared for paraffin-embedded tissue sections 12 hours after radiation and were examined for the presence of apoptotic nuclei. Apoptosis is inferred by intranuclear staining of muscle and bone marrow cells. Images were acquired using a ×20 objective. B: Lungs were obtained from 12-week-old wild-type, TSP1-null, and CD47-null mice and stained for H&E. Original magnifications, ×100.

Figure 8

TSP1 and CD47 limit vascular cell survival and proliferation after radiation. A: Primary vascular endothelial cells were harvested as previously described from wild-type, TSP1-null, and CD47-null mice and plated in growth medium on 96-well culture plates (5 × 10³ cells/200 μl per well), treated with the indicated doses of radiation, and cell viability determined by MTT assay at 72 hours. Significance was determined using the independent two-tailed t-test. *P < 0.05 versus wild type. B: Primary vascular endothelial cells were harvested from wild-type and TSP2-null mice and plated in growth medium on 96-well culture plates (5 × 10³ cells/200 μl per well), treated with the indicated doses of radiation and cell viability determined by MTT assay after 72 hours of incubation. Primary murine wild-type and TSP1 vascular endothelial cells were seeded into 96-well plates at 5 × 10⁴ cells/well density. After 24 hours cells were then either irradiated at 40 Gy or left to incubate at 37°C as a control. C: At the indicated time points after irradiation, proliferation was assessed by quantifying incorporation of BrdU into new synthesized DNA using a BrdU cell proliferation assay (Calbiochem, La Jolla, CA). Results are presented as relative fluorescence units (RFUs), and significance was determined using the one-way analysis of variance test. *P < 0.05 versus wild type. C57BL/6 wild-type (D) and TSP1-null 12-week-old female mice (E) were inoculated with B16F10 melanoma cells (2.5 × 10⁶ cells/animal) to the right lateral thigh. On tumors reaching 200 mm³ half of each group underwent 10 Gy of radiation to the tumor-bearing limb. Tumor size was measured bi-weekly. Animals were sacrificed if tumors exceeded 2 cm³. Results are from 16 animals, 8 of each strain.