Tumor Response to Radiotherapy Regulated by Endothelial Cell Apoptosis

See allHide authors and affiliations

Science  16 May 2003:
Vol. 300, Issue 5622, pp. 1155-1159
DOI: 10.1126/science.1082504


About 50% of cancer patients receive radiation therapy. Here we investigated the hypothesis that tumor response to radiation is determined not only by tumor cell phenotype but also by microvascular sensitivity. MCA/129 fibrosarcomas and B16F1 melanomas grown in apoptosis-resistant acid sphingomyelinase (asmase)–deficient or Bax-deficient mice displayed markedly reduced baseline microvascular endothelial apoptosis and grew 200 to 400% faster than tumors on wild-type microvasculature. Thus, endothelial apoptosis is a homeostatic factor regulating angiogenesis-dependent tumor growth. Moreover, these tumors exhibited reduced endothelial apoptosis upon irradiation and, unlike tumors in wild-type mice, they were resistant to single-dose radiation up to 20 grays (Gy). These studies indicate that microvascular damage regulates tumor cell response to radiation at the clinically relevant dose range.

Ionizing radiation is a widely used therapy for solid tumors and is thought to act by directly targeting tumor clonogens, also known as stem cells (1, 2). Tumor curability is believed to be determined by the most resistant clonogen, because one surviving stem cell appears sufficient for reconstituting tumor growth (3, 4). This model appears relevant to several normal tissues, particularly those classified as rapid-turnover systems. For example, gastrointestinal (GI) damage is believed to result from direct interaction of radiation with the clonogenic compartment at the crypt of Lieberkühn base (5, 6). However, we recently reported that microvascular endothelial apoptosis is required for clonogenic cell dysfunction (7). GI damage was prevented when endothelial cell apoptosis was inhibited genetically by asmase–/– depletion or pharmacologically by intravenous basic fibroblast growth factor (bFGF) (79). Because endothelial cells, but not crypt cells, expressed bFGF receptors, GI protection afforded by bFGF appeared to result from prevention of microvascular dysfunction. The notion that the microvasculature, rather than the clonogenic component, might be the primary target for radiation provides a basis for exploring new mechanisms of radiation effects on tumors.

The important role of the microvasculature in controlling tumor growth is well established. Tumors are considered to be growing near their maximal growth rate (2) because they purportedly limit their own growth by regulating neoangiogenesis (10). To evaluate whether the asmase genotype of the host angiogenic compartment affects tumor growth, we implanted MCA/129 fibrosarcoma and B16F1 melanoma cells into asmase+/+ and asmase–/– littermates (11, 12). Whereas MCA/129 fibrosarcomas grew at a rate of 10.8 ± 1.2 mm3/day (mean ± SEM) in asmase+/+ mice from day 8 to day 27 after implantation (Fig. 1A, upper panel), their growth rate was more than doubled in asmase–/– littermates (26.2 ± 2.8 mm3/day; P < 0.05). A similar increase in growth rate, by a factor of 3 to 4, occurred with B16F10 melanomas (fig. S1) and B16F1 melanomas implanted into asmase–/– mice (Fig. 1A, lower panel). To evaluate whether these differences correlated with the apoptotic competence of the host endothelium, we first determined (using hematoxylin, anti-CD34 immunohistochemical staining for endothelium, and morphological criteria) that MCA/129 tumors at 8 days after implantation contained 35 ± 2 endothelial cells and 1200 ± 38 fibrosarcoma cells per 400× field. Double staining with anti-CD34 and terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling (TUNEL), as described (7, 13), revealed that when fibrosarcomas were implanted into asmase+/+ mice, baseline endothelial apoptosis was increased by a factor of 2.4 relative to asmase–/– mice (6.1 ± 0.6 versus 2.5 ± 0.3 apoptotic endothelial cells per 400× field, respectively; P < 0.01) (Fig. 2B).

Fig. 1.

(A) Growth patterns of MCA/129 fibrosarcomas (upper panel) and B16F1 melanomas (lower panel) in sv129 × C57BL/6 asmase+/+ and asmase–/– mice. Cells (106, resuspended in phosphate-buffered saline) were injected subcutaneously into the right hindlimb, and tumor volume (based on caliper measurements) was calculated daily according to the formula of Kim et al. (30). (B) MCA/129 fibrosarcomas implanted as in (A) were treated with 15 Gy at a volume of 100 to 150 mm3. Mice, lightly sedated with ketamine (0.1 mg/g) and xylazine (0.02 mg/g), were irradiated using a Philips MG-324 x-ray unit at 105.5 cGy/min (source-to-skindistance, 50 cm). Only tumor, surrounding skin, and subcutaneous tissues were exposed; the rest of the mouse was shielded with a specialized lead jig. Values are means ± SEM. Numbers of mice are in parentheses. Arrow indicates day of irradiation.

Fig. 2.

MCA/129 fibrosarcomas implanted into asmase–/– mice display reduced baseline and radiation-induced endothelial cell apoptosis. (A) Representative 5-μm histologic tumor sections obtained 4 hours after exposure to 15 Gy, stained for apoptosis by TUNEL (13, 31) and for the endothelial cell surface marker CD-34 (32). Apoptotic endothelium manifests a red-brown TUNEL-positive nuclear signal surrounded by dark blue plasma membrane signal of CD-34 staining. (B) TUNEL-CD34 doubly stained cells were counted manually in a single-blinded fashion. Data (means ± SD) were compiled from 20 different fields (400× magnification) at each time point. (C) Genetic evidence that endothelial apoptosis regulates tumor response: Growth of B16F1 melanomas implanted subcutaneously into C57BL/6 Bax+/+ and Bax–/– mice. Tumor volumes were measured daily and data analyzed as in Fig. 1.

To assess whether the radiosensitivity of MCA/129 fibrosarcomas and B16F1 melanomas also depends on host asmase, we irradiated tumors with doses from 10 to 20 Gy. Exposure of fibrosarcomas grown in wild-type mice to 10 Gy caused a growth delay of 9 ± 1 days and 10% tumor control [TCD10, established as in (14)] but did not affect tumors implanted into asmase–/– mice (fig. S2; P < 0.01 versus asmase+/+). At 15 Gy (Fig. 1B), fibrosarcomas in wild-type mice continued to grow for 2 to 3 days after irradiation, but thereafter displayed a 69 ± 5% reduction in size (P < 0.001 versus day 9). Analysis of individual mice for up to 120 days showed that this dose conferred 50% tumor control (TCD50). In contrast, these tumors appeared completely resistant to 15 Gy when implanted in asmase–/– mice (P < 0.01 versus asmase+/+), continuing to grow at a rate not significantly different from that of unirradiated controls. Exposure to 20 Gy, however, appeared to overcome the radioresistance of tumors growing in asmase–/– mice, as tumors in both genetic backgrounds manifested a similar 80% volume reduction by 1 week after irradiation (fig. S3). B16F1 melanomas were also sensitive to 15 Gy when implanted in asmase+/+ mice, showing a 40% reduction in volume after 1 week (P < 0.01 versus unirradiated), but were resistant when grown in asmase–/– mice (15).

To determine whether differences in tumor response to radiation correlated with induction of endothelial apoptosis, we stained tumor tissue specimens with TUNEL and anti-CD34 after exposure to 15 Gy (13). MCA/129 fibrosarcomas in asmase+/+ mice displayed markedly increased endothelial apoptosis beginning at 1 hour and peaking at 4 to 6 hours (17.5 ± 1.3 apoptotic endothelial cells per 400× field at 6 hours; P < 0.01 versus unirradiated) (Fig. 2B). Typical histologic fields at 4 hours after irradiation are shown in the upper panel of Fig. 2A. Minimal tumor cell apoptosis was detected subsequent to endothelial cell apoptosis, increasing to less than twice the baseline value of 4.2 ± 0.7 apoptotic tumor cells per 400× field at 6 to 10 hours after irradiation. Thus, 50% of the endothelial cells but only 0.35% of tumor cells underwent apoptosis within this time frame. In contrast, endothelium of tumors in asmase–/– mice was resistant to radiation-induced apoptotic death (Fig. 2A, lower panel, and Fig. 2B; 5.5 ± 0.6 apoptotic endothelial cells per 400× field at 6 hours after irradiation; P < 0.01 versus asmase+/+), and tumor cell apoptosis (2.3 ± 0.4 apoptotic tumor cells per 400× field) was not increased. Similar results were obtained with B16F1 melanomas (fig. S4). At 20 Gy, enhanced endothelial apoptosis was not detected in tumors growing in asmase–/– mice even though this dose overcame tumor radioresistance in these mice. This indicates activation of an alternative tumor response mechanism at the higher dose range.

To confirm that apoptosis resistance of the endothelium confers enhanced tumor growth and reduced radiation responses, we performed studies with mice null for Bax, a proapoptotic Bcl-2 family member (16). Previous studies in asmase–/– and Bax–/– mice indicate that these genes function in the same antiapoptotic pathway in oocytes (17). B16F1 melanomas implanted in Bax–/– mice grew faster than in Bax+/+ littermates by a factor of 4 to 5 (Fig. 2C). Further, as occurred with asmase–/– mice, the tumors manifested marked reductions in baseline and radiation-induced endothelial cell apoptosis as well as in tumor radiation response (fig. S5).

We next tested whether asmase–/– bone marrow (BM) transplantation confers increased tumor proliferation and radioresistance. Recent studies showed that BM-derived VEGFR2-positive endothelial precursor cells contribute to rapid tumor neoangiogenesis (18). To assess whether the microvasculature of MCA/129 fibrosarcomas is derived from circulating BM precursors, we treated wild-type mice with 12 Gy whole-body irradiation (WBR) to sterilize the BM (7) and reconstituted with 107 BM cells from Rosa-26 mice. Tumors were implanted 4 weeks later and excised 2 weeks thereafter. β-Galactosidase staining revealed blue cells throughout the tumor (fig. S6). About 50% of tumor neovessels were derived from Rosa-26 BM, a finding confirmed by double staining with von Willebrand factor.

This technique enabled us to populate tumors with asmase–/– or asmase+/+ endothelium irrespective of the host genotype. MCA/129 tumors were implanted subcutaneously into wild-type mice 4 weeks after reconstitution with 107 asmase+/+ or asmase–/– BM cells. In mice reconstituted with asmase–/– BM (Fig. 3A), tumors grew nearly twice as fast as in mice reconstituted with asmase+/+ BM (12.3 ± 1.0 mm3/day versus 6.7 ± 0.7 mm3/day; P < 0.01). Note that tumors established in BM-transplanted mice grew somewhat slower than in naïve mice regardless of the genotype of BM transplanted (compare Figs. 3A and 1A), consistent with the known reduction in stem cell capacity that accompanies BM transplantation (19). Tumors grown in wild-type mice reconstituted with asmase–/– BM became resistant to 15 Gy (Fig. 3B; P < 0.01 versus tumors in mice reconstituted with asmase+/+ BM). Similar data were obtained with B16F1 melanoma (15). Additionally, radioresistance was observed with any combination in which transplanted BM was asmase–/–, irrespective of the genotype of the host, and vice versa for asmase+/+ BM (15). Furthermore, transplantation of rag–/–; asmase+/+ BM had no effect on tumor growth or radiation responsiveness, which suggests that the effects of loss of asmase were not due to defective immune responses (15).

Fig. 3.

Effects of BM transplantation on growth of MCA/129 fibrosarcomas implanted into asmase+/+ and asmase–/– mice. (A) Growth pattern of tumors transplanted into asmase+/+ mice harboring BM from asmase+/+ and asmase–/– donor mice. For autologous BM transplantation, 107 BM cells harvested from femur and tibia of donor mice by flushing medullary cavities with Hank's balanced salt solution were injected into the tail vein of recipient mice 16 hours after 12 Gy WBR. Data were analyzed as in Fig. 1. (B) BM transplantation from asmase–/– into asmase+/+ littermates conferred radiation resistance. Tumors were irradiated with 15 Gy (indicated by arrow) and data analyzed as in Fig. 1.

We next investigated whether resistance of asmase–/– endothelium to radiation-induced apoptosis in tumor tissue was due to direct loss of asmase within endothelium rather than to a systemic factor. MCA/129 fibrosarcomas were excised from mice, placed in Dulbecco's modified Eagle's medium (DMEM) supplemented with 0.2% fetal bovine serum at 37°C, and irradiated ex vivo with 15 Gy. Exposure of tumors containing asmase+/+ vasculature to 15 Gy ex vivo resulted in time-dependent endothelial cell apoptosis (Fig. 4A) similar to that observed in tumors irradiated in vivo (P < 0.001 versus unirradiated asmase+/+ controls at 2 to 4 hours). In contrast, tumors populated by asmase–/– endothelium irradiated ex vivo were resistant to radiation-induced apoptosis (P < 0.001 versus tumors excised from asmase+/+ mice at 2 to 4 hours after irradiation). Because it remained possible that another asmase–/– cell type derived from the host and infiltrating the tumor might regulate the endothelial response even in tumors irradiated ex vivo, we purified tumor endothelial cell populations of greater than 95% purity by a modification of the approach of Kinzler and co-workers (fig. S7) (20) and subjected them to irradiation. At 15 Gy, isolated asmase+/+ tumor endothelial cells underwent time-dependent apoptosis peaking at 8 hours (fig. S8; P < 0.01 versus control). In contrast, asmase–/– tumor endothelium was largely radioresistant (P < 0.001 versus irradiated asmase+/+ at 4 to 8 hours). Figure 4B shows full dose-response profiles of this effect 8 hours after irradiation. These studies show a maximal 63 ± 3% apoptosis at 25 Gy in asmase+/+ tumor endothelial cells (P < 0.001 versus unirradiated). In contrast, asmase–/– tumor endothelium was almost entirely resistant to ionizing radiation–induced apoptotic death (P < 0.001 versus asmase+/+ from 11 to 25 Gy).

Fig. 4.

Radiation effects on microvascular endothelial apoptosis. Radiation induces microvascular endothelial apoptosis in tissue explants from asmase+/+ but not asmase–/– mice harboring MCA/129 fibrosarcomas. (A) Tumors were excised, sliced into 0.5-cm fragments, washed in DMEM, irradiated with 15 Gy, incubated for the indicated times, fixed in 4% formaldehyde, embedded in paraffin, and TUNEL-CD34 double-stained as in Fig. 2A. TUNEL-CD34–positive cells were counted manually in 20 fields (400× magnification) in a single-blinded fashion. Total positive cells ± 95% confidence limits are presented. (B) Pure endothelial cells isolated by a modification (13) of the protocol of St. Croix et al. (20) (see fig. S7) from MCA/129 fibrosarcomas implanted in asmase+/+ mice are more radiosensitive than those isolated from asmase–/– mice. A minimum of 500 cells, stained with bis-benzimide, were scored for apoptosis at each dose 8 hours after irradiation, as described (33). Data (means ± SEM) were collated from three independent experiments.

Our results add to previous genetic evidence that tumors are angiogenesis dependent (21), defining the occurrence of endothelial apoptosis in the tumor microvasculature as a critical homeostatic factor regulating the rate of tumor growth (13, 22, 23). The striking apoptotic sensitivity of asmase+/+ endothelium is consistent with endothelium synthesizing 20 times as much ASMase as any other cell in the body, mostly in a nonlysosomal secretory form (24). The large excess of ASMase in endothelium may be involved in tissue remodeling and wound repair (25), events normally requiring waves of tissue neoangiogenesis and endothelial apoptosis (26). This suggests that a normal function of endothelial ASMase is to delimit rapid growth associated with tumor development.

Our studies define a similar pattern of tissue response in the GI tract and tumors— that is, rapid endothelial apoptosis followed over a period of days by tissue regression— which suggests that sustained microvascular dysfunction regulates tumor stem cell dysfunction in response to low-dose irradiation. This paradigm diverges from the current consensus that DNA damage and mitotic death, resulting from direct interaction of radiation with tumor cells, is an autonomous mechanism by which radiation confers its antitumor effect. That the radioprotective effect likely resulted from direct protection of tumor endothelium was confirmed here by isolating pure asmase–/– endothelial cell populations resistant to as much as 25 Gy.

If the microvasculature is a target for therapy, why has it been so difficult to establish this principle clinically? We believe the answer resides in the notion that endothelial apoptosis in tissue (and cell culture) is subject to transmodulation. Different tumors likely secrete growth factors, cytokines, and mitogens that differ in type and amount according to genetic factors (e.g., tumor cell type, mutations) as well as epigenetic factors (e.g., hypoxia, nutritional status), and these secretions would tend to modify the radiosensitivity of their endothelial cells (27). Thus, tumors may regulate the baseline apoptotic level and the radioresponsiveness of their own microvascular endothelium.

Our data define microvascular damage as a key mechanism in tumor response to radiation at the low dose range. This dose range is equivalent in its biological effects to the higher dose levels delivered clinically with daily fractionated radiotherapy schemes (28). In this context, antibodies to VEGFR2 sensitize human small cell lung carcinoma and glioblastoma xenografts in nude mice to fractionated radiotherapy (14), which suggests microvasculature involvement in this response (13, 29). Our studies also identify the sphingomyelin pathway as a possible target for improving cancer treatment, perhaps by manipulating asmase levels in tumor endothelial cells or in endothelial precursors.

Supporting Online Material

Materials and Methods

Figs. S1 to S8

  • * These authors contributed equally to this work.

References and Notes

Stay Connected to Science

Navigate This Article