Research ArticleCancer

Tumor cells, but not endothelial cells, mediate eradication of primary sarcomas by stereotactic body radiation therapy

See allHide authors and affiliations

Science Translational Medicine  11 Mar 2015:
Vol. 7, Issue 278, pp. 278ra34
DOI: 10.1126/scitranslmed.aaa4214

Not all cells are eradicated equally

Personalized cancer therapies dominate the news. But radiation therapy continues to be an essential part of the treatment regimens of nearly half of all cancer patients—sometimes achieving complete tumor regression through the safe delivery of high doses of radiation. Previous research with transplanted tumor models in mice has suggest that radiation targets, not only the tumor cells themselves, but also components of the surrounding milieu, which comprises blood vessels and various cell types that influence tumor growth. Now Moding et al. challenge the earlier findings in studies conducted with primary sarcomas in mice that carried, in either tumor or endothelial cells, genetic mutations that modulate radiation sensitivity. The authors found that it was the tumor, rather than endothelial, cells that mediate primary sarcoma shrinkage by radiation therapy and that selective small-molecule inhibition of a DNA-damage response enzyme can enhance radiosensitization of some tumors.

Abstract

Cancer clinics currently use high-dose stereotactic body radiation therapy as a curative treatment for several kinds of cancers. However, the contribution of vascular endothelial cells to tumor response to radiation remains controversial. Using dual recombinase technology, we generated primary sarcomas in mice with targeted genetic mutations specifically in tumor cells or endothelial cells. We selectively mutated the proapoptotic gene Bax or the DNA damage response gene Atm to genetically manipulate the radiosensitivity of endothelial cells in primary soft tissue sarcomas. Bax deletion from endothelial cells did not affect radiation-induced cell death in tumor endothelial cells or sarcoma response to radiation therapy. Although Atm deletion increased endothelial cell death after radiation therapy, deletion of Atm from endothelial cells failed to enhance sarcoma eradication. In contrast, deletion of Atm from tumor cells increased sarcoma eradication by radiation therapy. These results demonstrate that tumor cells, rather than endothelial cells, are critical targets that regulate sarcoma eradication by radiation therapy. Treatment with BEZ235, a small-molecule protein kinase inhibitor, radiosensitized primary sarcomas more than the heart. These results suggest that inhibiting ATM kinase during radiation therapy is a viable strategy for radiosensitization of some tumors.

INTRODUCTION

About half of all cancer patients are treated with radiation therapy (1), which may be given with palliative intent in cases where a delay in tumor regrowth (growth delay) can be clinically meaningful. However, the majority of cancers treated with radiation therapy are treated with the intent to cure, where the goal of radiation therapy is to achieve complete and permanent tumor regression (local control). When patients with cancer are treated with radiation, they usually receive relatively small (1.8 to 2.0 Gy) daily fractions for 1 to 2 months. Recently, advances in radiation treatment planning and delivery have made it possible to safely deliver a small number of high radiation doses (15 to 24 Gy), termed stereotactic body radiation therapy (SBRT) or radiosurgery, to improve the local control of some tumors (2).

The tumor microenvironment of human cancers consists of blood vessels, fibroblasts, and immune cells that modulate cancer development, progression, and response to therapy (3). However, whether or not stromal cells, such as endothelial cells, are critical targets of radiation therapy remains controversial. Indeed, experiments using transplanted tumors in mice with radiosensitive stroma have suggested that tumor stromal cells do not contribute to local control of cancer by radiation therapy (4).

Recently, endothelial cell apoptosis and microvascular collapse were reported to contribute to the radiation response of transplanted melanoma and fibrosarcoma cell lines (5). Endothelial cell apoptosis can occur because of membrane damage, which triggers rapid ceramide-mediated apoptosis after high doses of radiation exposure (6, 7). As a result, transplanted tumors with acid sphingomyelinase– or Bcl-2–associated X protein (Bax)–deficient stroma, which have defective radiation-induced endothelial cell apoptosis, grow 200 to 400% faster than transplanted tumors with wild-type stroma and display a decreased growth delay after radiation doses of up to 20 Gy (5). Notably, endothelial cell apoptosis has been proposed to occur at a threshold of 8 to 10 Gy and to increase up to 20 to 25 Gy (7), suggesting that endothelial apoptosis may be contributing to tumor cure by SBRT (8). However, the conclusion that microvascular damage regulates tumor response to radiation has been challenged (9, 10), and additional experiments using transplanted model systems have failed to resolve the controversy (1114).

Unlike transplanted tumor models, which may not fully recapitulate the vasculature and immune surveillance of autochthonous tumors (15, 16), genetically engineered mouse models (GEMMs) develop tumors within the native microenvironment in immunocompetent mice (17) and may more faithfully recapitulate the tumor stroma and microenvironment of human cancers (18). In addition, the response of these primary mouse cancer models to therapeutics might mimic the response of human cancers in clinical trials (19, 20).

To investigate the contribution of endothelial cells to the radiation response of primary sarcomas, we developed the technology to contemporaneously mutate different genes specifically in the tumor cells versus the endothelial cells of primary sarcomas (21). In this system, an adenovirus expressing FlpO (adeno-FlpO) activates a conditional allele of oncogenic Kras (FSF- KrasG12D) and deletes both copies of a conditional allele of p53 (p53FRT), whereas a tissue-specific Cre driver mutates floxed alleles specifically in endothelial cells. We recently used this dual recombinase technology to demonstrate that selectively sensitizing endothelial cells to mitotic cell death by deleting the DNA damage response gene ataxia telangiectasia mutated (Atm) (22) prolongs sarcoma growth delay after SBRT (23). However, an increase in growth delay does not necessarily translate into improved local control (4, 24, 25).

Here, we irradiated sarcomas with Atm deleted in endothelial cells with a curative dose of radiation. We also used dual recombinase technology to selectively protect endothelial cells from apoptosis by deleting the proapoptotic gene Bax (26). We found that tumor endothelial cells in primary sarcomas did not die via apoptosis within four hours of SBRT. In addition, endothelial cell death did not contribute to sarcoma eradication by radiation therapy. In contrast, radiosensitizing tumor cells by deleting Atm increased local control of primary sarcomas after radiation therapy. These results demonstrate that tumor cells, but not endothelial cells, are critical targets of curative radiation therapy in primary sarcomas. Also, to test whether ATM inhibition can improve the therapeutic ratio during SBRT, we compared the radiation response of primary sarcomas and hearts after Atm inhibition with BEZ235 to demonstrate that targeting ATM during radiation therapy might be a viable approach for radiosensitization of tumors at certain anatomic sites.

RESULTS

Bax and Bak do not regulate endothelial cell death after SBRT in primary sarcomas

Because apoptosis of tumor endothelial cells is dependent on Bax in transplanted tumor models (5), we used (i) dual recombinase technology to initiate primary sarcomas in conditional FSF-KrasG12D; p53FRT/FRT (KPFRT) mice with adeno-FlpO (21) and (ii) VE-Cadherin-Cre (27) to delete Bax floxed alleles (BaxFL) (28) specifically in endothelial cells. We recently demonstrated that this approach efficiently deletes floxed alleles in endothelial cells of primary sarcomas (23). We confirmed by quantitative polymerase chain reaction (qPCR) that VE-Cadherin-Cre efficiently deleted Bax in endothelial cells (fig. S1A).

We next investigated the effect of Bax deletion specifically in endothelial cells on sarcoma initiation and growth by generating primary sarcomas in KPFRT; VE-Cadherin-Cre; BaxFL/+ (KPFRTVBaxFL/+) and KPFRT; VE-Cadherin-Cre; BaxFL/FL (KPFRTVBaxFL/FL) mice. In contrast to previous reports with tumor cells transplanted into Bax-null mice (5), we observed no change in primary sarcoma initiation or growth in mice with deletion of Bax specifically in endothelial cells (fig. S1, B to D). To determine whether Bax deletion from endothelial cells protected tumor endothelial cells from radiation in an autochthonous model system, we irradiated sarcomas in KPFRTVBaxFL/+ and KPFRTVBaxFL/FL mice with a single dose of 20 Gy using fluoroscopy-guided radiation therapy and examined endothelial cell death through terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick-end labeling (TUNEL) staining at various time points after irradiation. In transplanted tumor models, tumor endothelial cell apoptosis peaks 4 to 6 hours after radiation exposure (5). Consistent with our previous results (23), we did not observe a significant change in apoptotic endothelial cell death 4 hours after irradiation of primary sarcomas, but endothelial cell death did increase 48 hours after radiation, likely as a result of mitotic catastrophe (Fig. 1, A and B). Furthermore, deletion of Bax did not affect endothelial cell death, suggesting that endothelial cells in primary sarcomas do not undergo Bax-mediated apoptosis after irradiation.

Fig. 1. Deletion of Bax from mouse endothelial cells does not affect primary sarcoma response to radiation therapy.

(A) Representative immunofluorescence for CD31 and TUNEL in sarcomas from KPFRTVBaxFL/+ and KPFRTVBaxFL/FL mice 4 and 48 hours after irradiation with 20 Gy. Areas enclosed by dashed lines are shown at higher magnification in the insets. Scale bar, 100 μm. (B) Quantification of CD31 and TUNEL double-positive cells in sarcomas from KPFRTVBaxFL/+ and KPFRTVBaxFL/FL mice at the indicated time points after irradiation with 20 Gy (n = 4 mice per group). Two-way analysis of variance (ANOVA) for genotype and time interaction followed by Bonferroni’s post hoc tests for pairwise comparisons between genotypes showed that the differences were not statistically significant. (C and D) Tumor growth curves (C) and time to volume tripling for sarcomas (D) in KPFRTVBaxFL/+ and KPFRTVBaxFL/FL mice after irradiation with 20 Gy (n = 8 mice per group). Two-tailed Student’s t test showed that the differences were not statistically significant. (E) Kaplan-Meier plot of local sarcoma control defined as the absence of tumor volume tripling for sarcomas in KPFRTVBaxFL/+ and KPFRTVBaxFL/FL mice after irradiation with 50 Gy (n = 18 and 20 mice per group). Log-rank test showed that the differences were not statistically significant.

We next evaluated whether Bax in endothelial cells contributed to the growth delay of primary sarcomas. To this end, we irradiated sarcomas in KPFRTVBaxFL/+ and KPFRTVBaxFL/FL mice with a single dose of 20 Gy and monitored tumor growth until the sarcomas tripled in size. In contrast to transplanted tumors in Bax-null mice (5), primary sarcomas in KPFRTVBaxFL/FL mice displayed a growth delay after irradiation similar to sarcomas in KPFRTVBaxFL/+mice (Fig. 1, C and D). To investigate whether there was a difference in local control in this model, we irradiated primary sarcomas in KPFRTVBaxFL/+ and KPFRTVBaxFL/FL mice with 50 Gy, which we determined to be the maximally tolerated dose in our system. Although 50 Gy was able to eradicate a small percentage of primary sarcomas, there was no difference in local control or growth delay between the two genotypes (Fig. 1E and fig. S2).

Because the proapoptotic genes Bax and Bcl-2 homologous antagonist/killer (Bak) can be redundant for executing programmed cell death in some settings (29), we also deleted the Bax gene from endothelial cells of Bak-null (Bak−/−) mice (30). Primary sarcomas in KPFRT; VE-Cadherin-Cre; Bak−/−; BaxFL/+ (KPFRTVBak−/−BaxFL/+) and KPFRT; VE- Cadherin-Cre; Bak−/−; BaxFL/FL (KPFRTVBak−/−BaxFL/FL) mice developed at the same time after injection of adeno-FlpO and grew at the same rate in the absence of radiation (fig. S3). After irradiation with 20 Gy, we observed no difference in endothelial cell death or growth delay for primary sarcomas in KPFRTVBak−/−BaxFL/+ and KPFRTVBak−/−BaxFL/FL mice (fig. S4), suggesting that endothelial cell apoptosis does not contribute to the radiation response of primary sarcomas. Because previous studies have demonstrated differences between the vasculature of transplanted and primary cancers (15, 31, 32), our discordant results on the contribution of endothelial cell death might reflect differences in the vasculatures of tumors derived from transplanted cell lines versus the autochthonous tumors studied herein. Together, these findings illustrate the importance of studying the tumor microenvironment using multiple complementary models, including GEMMs.

Endothelial cell death does not contribute to eradication of primary sarcomas by SBRT

We showed recently that deleting Atm specifically in endothelial cells radiosensitizes these cells in autochthonous sarcomas, causing them to undergo a delayed mitotic cell death that increases tumor growth delay after radiation therapy with 20 Gy (23). To determine whether Atm deletion increased endothelial cell death after a curative dose of radiation, we treated primary sarcomas in KPFRT; VE-Cadherin-Cre; AtmFL/+ (KPFRTVAtmFL/+) and KPFRT; VE-Cadherin-Cre; AtmFL/FL (KPFRTVAtmFL/FL) mice with 50 Gy. Consistent with our previous results, deletion of Atm significantly increased endothelial cell death 24 hours after irradiation (Fig. 2, A and B).

Fig. 2. Sensitizing sarcoma endothelial cells to radiation does not affect local control of primary sarcomas.

(A) Representative immunofluorescence for CD31 and TUNEL in sarcomas from KPFRTVAtmFL/+ and KPFRTVAtmFL/FL mice 24 hours after irradiation with 50 Gy. Areas enclosed by dashed lines are shown at higher magnification in the insets. Scale bar, 100 μm. (B) Quantification of CD31 and TUNEL double-positive cells in sarcomas from KPFRTVAtmFL/+ and KPFRTVAtmFL/FL mice 24 hours after irradiation with 50 Gy (n = 6 mice per group). Two-tailed Student’s t test, P < 0.05. (C and D) Kaplan-Meier plot of local sarcoma control (defined as the absence of tumor volume tripling) for sarcomas in KPFRTVAtmFL/+ and KPFRTVAtmFL/FL mice after irradiation with 50 Gy (n = 22 and 23 mice per group, respectively) (C) or four daily fractions of 20 Gy (n = 23 and 25 mice per group, respectively) (D). One KPFRTVAtmFL/FL mouse developed an abdominal metastasis 8 weeks after irradiation with 50 Gy, and two KPFRTVAtmFL/+ mice died of unknown causes at 7 and 10 weeks after irradiation with four daily fractions of 20 Gy before sarcoma tripling. Thus, these mice were scored as locally controlled until the time points at which they either developed metastasis or died. Log-rank tests showed that the differences were not statistically significant. Asterisks represent statistically significant difference between the indicated groups.

To investigate whether endothelial cell death contributes to the local control of primary sarcomas, we irradiated sarcomas in KPFRTVAtmFL/+ and KPFRTVAtmFL/FL mice with 50 Gy. Although 50 Gy was able to eradicate primary sarcomas in both KPFRTVAtmFL/+ and KPFRTVAtmFL/FL mice, there was no difference in local control between the two genotypes (Fig. 2C). Moreover, at the curative dose of 50 Gy, there was no difference in growth delay for primary sarcomas in KPFRTVAtmFL/+ and KPFRTVAtmFL/FL mice (fig. S5, A and B). Because sarcomas in KPFRTVAtmFL/FL mice have a prolonged growth delay after 20 Gy compared to sarcomas in KPFRTVAtmFL/+ mice (23), we repeated this experiment with four daily fractions of 20 Gy. Similar to a single dose of 50 Gy, we observed no difference in local control or growth delay for sarcomas in KPFRTVAtmFL/+ and KPFRTVAtmFL/FL mice after 80 Gy of radiation was delivered in four 20 Gy fractions (Fig. 2D and fig. S5, C and D). These results demonstrated that although endothelial cell death can delay the regrowth of primary sarcomas at noncurative doses of radiation, endothelial cell death is not rate-limiting for sarcoma regrowth after curative doses of radiation and does not contribute to local control of primary soft tissue sarcomas.

Tumor cells in primary sarcomas are critical targets that mediate local control by SBRT

For comparison, we also deleted Atm in tumor cells. First, we injected the muscle of LSL-KrasG12D; p53FL/FL; AtmFL/+ (KPloxPAtmFL/+) and LSL-KrasG12D; p53FL/FL; AtmFL/FL (KPloxPAtmFL/FL) mice with adeno-Cre. However, the sarcomas that developed did not reliably delete both alleles of Atm (fig. S6A). Therefore, we crossed KPloxPAtmFL/+ and KPloxPAtmFL/FL mice to Pax7-CreER (P7) mice, which express a tamoxifen-inducible Cre recombinase in muscle satellite cells (33). P7KPloxP mice develop sarcomas after administration of systemic tamoxifen (34). In contrast to KPloxPAtmFL/FL mice injected with adeno-Cre, intramuscular injection of P7KPloxPAtmFL/FL mice with 4-hydroxytamoxifen generated primary sarcomas at the site of injection with efficient deletion of both Atm alleles (fig. S6B).

To determine whether Atm deletion sensitized sarcoma cells to radiation, we irradiated cell lines isolated from sarcomas in P7KPloxPAtmFL/+ and P7KPloxPAtmFL/FL mice. Sarcoma cells from P7KPloxPAtmFL/FL mice were more sensitive to radiation in clonogenic survival assays compared to sarcoma cells from P7KPloxPAtmFL/+ control mice that retained one allele of Atm (Fig. 3A). Next, we irradiated primary sarcomas in P7KPloxPAtmFL/+ and P7KPloxPAtmFL/FL mice with 50 Gy and monitored the sarcomas for local control. Deletion of Atm in tumor cells significantly improved local control of primary sarcomas (Fig. 3, B and C), suggesting that direct tumor cell death can mediate tumor eradication by SBRT. Before local recurrence, several mice developed second sarcomas at other locations in the body, presumably from systemic tamoxifen–mediated Cre recombination. When we analyzed the histology from the site of irradiated primary sarcomas in nine KPloxPAtmFL/FL mice that developed second-site sarcomas, there were no tumor cells present in five mice and a few scattered histologically intact tumor cells in three mice (fig. S7).

Fig. 3. Sensitizing tumor cells to radiation increases local control of primary sarcomas after radiation therapy.

(A) Clonogenic survival of primary sarcoma cell lines from P7KPloxPAtmFL/+ and P7KPloxPAtmFL/FL mice (n = 3 independent cells lines per genotype). Two-way ANOVA for genotype and dose interaction (P < 0.05) followed by Bonferroni’s post hoc tests for pairwise comparisons between genotypes after 2 and 4 Gy (P < 0.05). (B and C) Sarcoma growth curves (B) and Kaplan-Meier plot of local sarcoma control (C) defined as the absence of tumor volume tripling for primary sarcomas generated by intramuscular 4-hydroxytamoxifen injection into P7KPloxPAtmFL/+ and P7KPloxPAtmFL/FL mice after irradiation with 50 Gy (n = 13 mice per group). Several mice were euthanized before sarcoma tripling because of the development of second sarcomas at other locations in the body, presumably from systemic tamoxifen–mediated Cre recombination. These mice were scored as locally controlled until the time point of second tumor formation. Log-rank test, P < 0.05. Asterisks represent statistically significant difference between the indicated groups.

The protein kinase inhibitor BEZ235 preferentially radiosensitizes primary sarcomas

Deletion of the Atm gene from tumor cells increased the probability of tumor eradication after SBRT, suggesting that therapeutic targeting of the ATM protein could improve tumor response to radiation therapy. However, systemic targeting of ATM in normal tissues might also increase radiation toxicity. To begin to address whether targeting ATM during radiation therapy can improve the response of tumors relative to some normal tissues (that is, to enhance the therapeutic ratio), we showed recently that deleting Atm preferentially radiosensitizes proliferating tumor endothelial cells compared with quiescent heart endothelial cells (23). Here, we used a pharmacological approach to investigate whether a therapeutic window exists for inhibiting ATM during radiation therapy. To this end, we treated KPloxP mice that had primary sarcomas with the phosphoinositide 3-kinase (PI3K)–like kinase (PI3KK) inhibitor BEZ235, which potently inhibits human ATM (35). Pretreatment of the mice with BEZ235 significantly decreased autophosphorylation of mouse Atm and phosphorylation of the Atm target Kap1 (Krüppel-associated box–associated protein 1) in primary sarcomas after irradiation with 20 Gy (fig. S8, A to D).

To investigate whether BEZ235 can selectively radiosensitize sarcomas, we collected hearts and primary sarcomas from vehicle- and BEZ235-treated KPloxP mice 24 hours after whole-body irradiation with 20 Gy. Treatment with BEZ235 significantly increased cell death in sarcomas but not in hearts (Fig. 4, A and B). Next, we monitored tumor growth in KPloxP mice treated with BEZ235 alone or in combination with focal 20 Gy irradiation. Although a single dose of BEZ235 alone did not delay tumor growth (fig. S8, E and F), BEZ235 treatment significantly delayed primary sarcoma regrowth after 20 Gy (Fig. 4, C and D).

Fig. 4. The PI3KK inhibitor BEZ235 preferentially radiosensitizes primary sarcomas compared to mouse heart tissue.

(A and B) Immunofluorescence (A) and quantification of TUNEL-positive cells (B) in hearts and sarcomas from vehicle- or BEZ235-treated KPloxP mice 24 hours after irradiation with 20 Gy (n = 5 mice per group). Two-way ANOVA for tissue and treatment interaction followed by Bonferroni’s post hoc tests for pairwise comparison between treatments showed that the differences were statistically significant in sarcomas but not in hearts (P < 0.05). Scale bar, 100 μm. (C and D) Tumor growth curves (C) and time to volume tripling (D) of primary sarcomas in KPloxP mice treated on day 0 with vehicle or a single dose of BEZ235 (50 mg/kg) 2 hours before irradiation with 20 Gy (n = 8 mice per group). Two-tailed Student’s t test, P < 0.05. (E) Kaplan-Meier plots of myocardial necrosis–free survival for mice that lacked p53 in endothelial cells (VPFL/FL) or mice that retained p53 in endothelial cells (VPFL/+), both of which were treated with vehicle or BEZ235 (50 mg/kg) 2 hours before whole-heart irradiation with 12 Gy (n = 7 to 9 mice per group). Two VPFL/FL mice treated with vehicle were censored because they developed tumors before developing myocardial necrosis. Log-rank test for VPFL/+ Vehicle versus VPFL/FL Vehicle, VPFL/+ Vehicle versus VPFL/FL BEZ235, VPFL/+ BEZ235 versus VPFL/FL Vehicle, and VPFL/+ BEZ235 versus VPFL/FL BEZ235 (P < 0.05). Asterisks represent statistically significant difference between the indicated groups.

We next investigated the effect of BEZ235 treatment on the development of radiation-induced heart disease. We showed previously that deletion of Atm radiosensitizes p53-null cardiac endothelial cells but not cardiac endothelial cells with intact p53 (23). Therefore, we treated VE-Cadherin-Cre; p53FL/+(VPFL/+) mice with one allele of p53 deleted in endothelial cells and VE-Cadherin-Cre; p53FL/FL (VPFL/FL) mice with two alleles of p53 deleted in endothelial cells with either vehicle or BEZ235 2 hours before whole-heart irradiation with 12 Gy and monitored the mice for the development of radiation-induced heart disease. BEZ235 shifted the myocardial necrosis–free survival curve to the left in VPFL/FL mice but did not promote the development of radiation-induced myocardial necrosis in VPFL/+ mice (Fig. 4E). Because we used BEZ235 with a single dose of radiation, we cannot exclude the possibility that ATM inhibition has some effect on radiation-induced heart disease with other radiation doses. However, taken together with the genetic studies, these experiments with a pharmacological inhibitor of ATM suggest that inhibiting ATM may preferentially radiosensitize sarcomas compared with quiescent normal tissues, such as the heart.

DISCUSSION

Here, we used a new dual recombinase technology to manipulate the radiosensitivity of endothelial cells in primary sarcomas. In contrast to results with transplanted tumor models (5), we found that tumor endothelial cells in primary sarcomas did not die via apoptosis within 4 hours after irradiation. Although endothelial cells can contribute to primary sarcoma growth delay after noncurative radiation therapy (23), radiosensitizing endothelial cells did not increase local control. In contrast, radiosensitization of tumor cells increased the probability of sarcoma eradication by SBRT. Therefore, tumor cells, but not endothelial cells, are critical targets of SBRT that mediate sarcoma eradication.

These results have important clinical implications, suggesting that increased endothelial cell death does not contribute to the improved local control of some tumors after SBRT. Instead, the increased efficacy of SBRT is likely the result of delivering larger, biologically effective doses than are delivered with standard radiation therapy (36). Our results are consistent with the previous observation that local control is not affected when tumors are transplanted into radiosensitive mice (4) and extend these findings to primary sarcomas. Despite the extensive endothelial cell death observed after a dose of 50 Gy (Fig. 2, A and B), most of the sarcomas in KPFRTVAtmFL/FL mice recurred (Fig. 2C). These results raise the possibility that the vasculature in recurrent sarcomas may have come from outside of the irradiation field (37). Although the source of these newly formed blood vessels remains an area of ongoing debate (38), targeting vascular recruitment could improve local control.

Our results do not exclude a contribution from other stromal cells to tumor eradication with SBRT. Indeed, recent studies suggest that the recruitment of macrophages from outside the radiation field through a stromal cell–derived factor 1 (SDF-1)–C-X-C chemokine receptor 4 (CXCR4) axis might participate in tumor cell repopulation after radiotherapy (37). In addition, a growing body of data suggests that the immune system can contribute to the killing of tumor cells after radiation therapy (39, 40), and this response may be greater after SBRT than after standard radiation therapy. Furthermore, our experiments focused on soft tissue sarcomas, and additional experiments are needed to determine whether endothelial cell death contributes to the radiation response of other tumor types. In the future, dual recombinase technology can be applied to other primary tumor model systems and stromal cell populations to further explore how the tumor microenvironment contributes to curative radiation therapy.

Although most patients are treated with curative intent, for patients receiving palliative radiation therapy, an increased growth delay can be clinically meaningful. For example, radiation therapy is the standard of care for children with diffuse intrinsic brainstem gliomas and provides many months of relief from severe neurological symptoms (41). Targeting tumor endothelial cells during radiation therapy could prolong the time that these patients remain neurologically intact. However, our results suggest that this approach may not improve tumor eradication. Instead, targeting tumor cells with radiosensitizers, such as inhibitors of ATM, might be a more promising treatment approach to achieve local control.

MATERIALS AND METHODS

Study design

The goal of this controlled laboratory experiment was to determine the relative contribution of endothelial cells and tumor cells to primary sarcoma response to radiation therapy. We used genetically engineered mice and observed histological, growth delay, and local control endpoints. Sample sizes were selected before initiating the study on the basis of prior primary sarcoma radiation response data from the laboratory and power calculations performed as described previously (42); data collection was stopped if a smaller sample size achieved statistical significance. Outliers were defined before initiating the study as falling greater than 2 SD from the mean. No outliers were excluded from this study. Sarcoma eradication was assumed if a sarcoma failed to triple in size 18 weeks after radiation treatment. Assuming a constant sarcoma doubling time of 4 days and tumor cell diameter of 25 μm (volume = 8.18 × 10−6 mm3), it would take a single tumor cell 100 days (~14 weeks) to reach 300 mm3. Mice that died before sarcoma tripling as a result of metastasis or sarcoma development at a distant site were censored at the time of death and were scored as locally controlled until this point. The mice in this study were not randomized to their treatments and were selected based on availability. The investigators were not blinded when performing sarcoma measurements. Histological quantification was performed by an observer blinded to treatment and genotype. Histological assessment of residual sarcoma cells in the legs of irradiated mice was assessed by a sarcoma pathologist (D.M.C.) who was blinded to treatment and genotype.

Mouse strains and sarcoma induction

All animal studies were performed in accordance with protocols approved by the Duke University Institutional Animal Care and Use Committee. All of the mouse strains used in this study have been described previously, including LSL-KrasG12D, p53FL, FSF-KrasG12D, p53FRT, VE-Cadherin-Cre, AtmFL, BaxFL, Bak−/−, and Pax7-CreER (21, 27, 28, 30, 33, 4346). LSL-KrasG12D and FSF-KrasG12D mice were provided by T. Jacks, AtmFL mice were provided by S. Zha and F. Alt, BaxFL and Bak−/− mice were provided by S. Korsmeyer, p53FL mice were provided by A. Berns, and Pax7-CreER mice were provided by C.-M. Fan. VE-Cadherin-Cre mice were obtained from The Jackson Laboratory. Primary sarcomas were generated in the right hind leg of KPloxP or KPFRT mice between 6 and 10 weeks of age as described previously (21, 47). Tamoxifen-induced primary sarcomas were generated by injecting 50 μl of 4-hydroxytamoxifen (5 mg/ml; Sigma-Aldrich) in dimethyl sulfoxide into the right hind leg of P7KPloxP mice (48). All mice were on a mixed genetic background. To minimize the effect of genetic background, age-matched littermate controls were used for every experiment so that potential genetic modifiers would be randomly distributed between the experimental and control groups.

Radiation treatment

Sarcoma and whole-heart irradiations were performed using the X-RAD 225Cx small animal image-guided irradiator (Precision X-Ray). The irradiation field was centered on the target via fluoroscopy with 40 kilovolt peak (kVp), 2.5 mA x-rays using a 2-mm aluminum filter. Sarcomas were irradiated with parallel-opposed anterior and posterior fields with an average dose rate of 300 cGy/min prescribed to midplane with 225 kVp, 13 mA x-rays using a 0.3-mm copper filter and a collimator with a 40 × 40 mm2 radiation field at treatment isocenter.

Whole-heart irradiation was performed using a collimator to produce a 15-mm circular radiation field at treatment isocenter. Dose rates were measured with an ion chamber by the Radiation Safety Division at Duke University. Sarcomas were irradiated at ~250 mm3 by caliper measurement for growth delay and histological endpoints and ~100 mm3 for local control endpoints. After sarcoma irradiation, sarcomas were measured three times per week for growth delay and once per week for local control experiments until they tripled in size from the initial volume measured at the time of radiation. After whole-heart irradiation, mice were monitored daily for symptoms of heart disease. Myocardial necrosis was confirmed histologically by hematoxylin and eosin (H&E) staining of heart sections.

Histological analysis

H&E was performed on paraffin-embedded tissue sections. Tissue specimens were fixed in 10% neutralized formalin overnight and preserved in 70% ethanol until paraffin embedding. Five micron sections were deparaffinized with xylene and rehydrated with a graded series of ethanol and water washes before performing H&E staining. Sarcoma and heart immunohistochemistry was performed on frozen tissue sections. Specimens were embedded directly in optimal cutting temperature compound (Sakura Fintek) by snap freezing in a dry ice/isopentane slurry and stored at −80°C until sectioning. Ten-micrometer sections were fixed in 4% paraformaldehyde before immunofluorescence staining. The primary antibodies were rat anti-mouse CD31 (1:250; BD Pharmingen, #553370), rabbit anti-mouse Ser824-phosporylated KAP1 (1:250; Bethyl Laboratories, #A300-767A), and polyclonal rabbit anti-mouse Ser1987-phosphorylated ATM [diluted 1:500; provided by M. Kastan (49)]. The secondary antibodies were Alexa Fluor 488–conjugated donkey anti-rat immunoglobulin G (1:500; Invitrogen, #A21208) and Alexa Fluor 555–conjugated goat anti-rabbit immunoglobulin G (1:250; Invitrogen, #A21429). Nuclear staining was performed using Hoechst 33342 (10 μM; Sigma-Aldrich). TUNEL staining was performed with the In Situ Cell Death Detection Kit, TMR red (Roche) according to the manufacturer’s instructions. Pictures were acquired with a Leica DFC340 FX fluorescence microscope (Leica Microsystems) using Leica Suite software (Leica Microsystems). Quantification was performed using ImageJ [National Institutes of Health (NIH)]. Each data point represents the average of eight randomly selected 200× fields per sample.

Flow sorting of endothelial cells

Lungs were dissected, washed in phosphate-buffered saline, homogenized, and digested in type I collagenase (0.8 mg/ml; Worthington Biochemical Corp.) for 1 hour at 37°C. Digested tissues were filtered, and red blood cells were lysed with ACK lysing buffer (Lonza Group). Total number of cells was counted by Coulter counter (Beckman Coulter Inc.). Three million cells were stained with phycoerythrin (PE)–conjugated anti-mouse CD31 (BioLegend, #102407), PE-Cy5–conjugated anti-mouse CD45 (eBioscience, #15-0451), and eFluor 660–conjugated anti-mouse CD34 (eBioscience, #50-0341) antibodies. Dead cells were excluded by staining with 7-aminoactinomycin D (BD Pharmingen). Viable CD45-negative and CD31 and CD34 double-positive cells were sorted by FACSVantage (BD Pharmingen) and used for RNA isolation.

Quantitative reverse transcription PCR analysis

Total RNA was extracted from sorted lung and tumor endothelial cells with the RNAqueous-Micro Kit (Ambion), and reverse transcription was performed with the iScript complementary DNA (cDNA) Synthesis Kit (Bio-Rad). Quantitative reverse transcription PCR was performed using TaqMan Universal PCR Master Mix (Applied Biosystems) and TaqMan Gene Expression Assay Mix (Applied Biosystems) for Bax (Mm00432050_m1) or Hprt (Mm0446968_m1). Hprt was used as an internal control to correct for the concentration of cDNA in different samples. Each experiment was performed with three replicates from each sample, and the results were averaged.

Cell line experiments and clonogenic survival assays

Cell lines were isolated by digestion of primary sarcomas with trypsin, type IV collagenase, and dispase (Invitrogen). Cells were cultured in Dulbecco’s modified Eagle’s medium with high glucose and pyruvate (Gibco) supplemented with 10% fetal bovine serum. Cells were passaged five times to deplete stromal cells before isolation of genomic DNA for PCR or clonogenic survival assays. Deletion of Atm was verified by PCR using primers flanking the 3′ loxP site (sense, 5′- GGGCTACGAAATGAGACACACAC-3′; antisense 5′-CTTCCCCTGTTCAAAAGCCACTC-3′) and primers flanking the recombined loxP site (sense, 5′-TGAGTTCAAATCCCAGGAGCCAG-3′; antisense, 5′- CTTCCCCTGTTCAAAAGCCACTC-3′). For clonogenic survival assays, cells were plated in triplicate and allowed to adhere overnight before irradiation with an X-RAD 320 Biological Irradiator (Precision X-Ray). Cells were placed 50 cm from the radiation source and were irradiated with a dose rate of 161 cGy/min using 320 kVp, 10 mA x-rays and a 2-mm aluminum filter. After development of colonies, cells were fixed with 70% ethanol, stained with Coomassie Brilliant Blue (Bio-Rad), rinsed with deionized water, and dried. A population of more than 50 cells was counted as one colony, and surviving fractions were calculated relative to unirradiated controls.

BEZ235 treatment

Mice were treated by oral gavage with a single dose of BEZ235 (50 mg/kg; Novartis) dissolved in 10% N-methyl-2-pyrrolidone/90% polyethylene glycol 300 (Sigma-Aldrich) 2 hours before radiation therapy.

Statistics

Results are presented as means ± SEM. Two-tailed Student’s t test was performed to compare the means of two groups. Two-way ANOVA was performed to examine the interaction between genotypes and treatments followed by Bonferroni’s post hoc tests for pairwise comparisons of individual treatments or genotypes. Non-normally distributed data were log-transformed before applying statistical tests. For survival studies, Kaplan-Meier analysis was performed followed by log-rank test for statistical significance. Significance was assumed at P < 0.05. All calculations were performed using Prism 5 (GraphPad).

SUPPLEMENTARY MATERIALS

www.sciencetranslationalmedicine.org/cgi/content/full/7/278/278ra34/DC1

Fig. S1. Deletion of Bax in primary tumor endothelial cells does not affect sarcoma initiation or growth.

Fig. S2. Deletion of Bax in endothelial cells does not affect sarcoma growth delay after a curative dose of irradiation.

Fig. S3. Loss of Bak and Bax in endothelial cells does not affect primary soft tissue sarcoma initiation or growth.

Fig. S4. Deletion of Bak and Bax in endothelial cells does not affect primary sarcoma response to radiation therapy.

Fig. S5. Deletion of Atm in endothelial cells does not affect sarcoma growth delay after curative doses of irradiation.

Fig. S6. Pax7-CreER, but not adeno-Cre, efficiently recombines both AtmFL alleles in primary sarcomas in vivo.

Fig. S7. Spectrum of histological responses for primary sarcomas in KPloxPAtmFL/FL mice after irradiation with 50 Gy.

Fig. S8. BEZ235 inhibits Atm in primary sarcomas.

REFERENCES AND NOTES

Acknowledgments: We thank S. Zha and F. Alt for providing the AtmFL mice, T. Jacks for providing the LSL-KrasG12D and FSF-KrasG12D mice, A. Berns for providing the p53FL mice, S. Korsmeyer for providing the BaxFL and Bak−/− mice, and C.-M. Fan for providing the Pax7-CreER mice. We also thank Novartis for providing BEZ235 to complete this work, M. Kastan for providing the antibody against pATM, L. Woodlief and L. Luo for their help in caring for the mice, and C.-Y. Li and J. Chute for critical reading of the manuscript. Funding: This work was supported by the National Cancer Institute of the U.S. NIH under award numbers F30 CA177220 (E.J.M.) and R21 CA175839 (D.G.K.) and by Susan G. Komen under award number IIR13263571 (D.G.K.). Author contributions: E.J.M., C.-L.L., and D.G.K. conceived and designed the experiments. E.J.M, K.D.C., B.A.P., P.O., and C.-L.L. performed the experiments. E.J.M. and D.G.K. analyzed the data. H.D.M. and H.N. irradiated the mice and measured the sarcomas. Y.M. performed the histology. D.M.C. reviewed the histopathology. E.J.M. and D.G.K. wrote the manuscript. Competing interests: The authors declare that they have no competing interests.
View Abstract

Navigate This Article