Hybrid Manganese Dioxide Nanoparticles Potentiate Radiation Therapy by Modulating Tumor Hypoxia

Radiotherapy remains the main line of treatment for solid tumors in more than 50% of patients with cancer with locally advanced solid tumors. However, poor tumor oxygenation (hypoxia) found in many types of solid tumors can lead to tumor resistance to radiation and thereby incomplete eradication and local tumor recurrence (1–10). Hypoxia has been consistently shown as a strong prognostic clinical indicator of radioresistance and decreased disease-free survival in many cancer types (8–10). Nearly 40% of locally advanced breast cancers, for example, exhibit tumor regions with oxygen concentrations below what is required for half-maximal radiosensitization, contributing to the ineffectiveness of radiotherapy alone, which has encouraged investigations involving the combination of radiotherapy with other therapies and treatment alternatives (7).

The efficacy of radiotherapy critically depends on the relative level of oxygen in the tumor at the time of irradiation as oxygen can enhance the formation of DNA double strand breaks (DSB) caused directly by ionization of DNA and indirectly by free radicals generated by the radiolysis of water (5). Various studies have shown that hypoxic cells can be two-to-three times more resistant to a single fraction of ionizing radiation than those irradiated in the presence of normal levels of oxygen (6). Hypoxia has also been linked to the development of aggressive phenotype in many different cancers mostly due to the overexpression of hypoxia inducible factors (HIF; refs. 1, 8).

Strong clinical evidence of the direct and indirect effects of tumor hypoxia, including hypoxia-related genes and their proteins on radiation response has motivated studies of various methods to modulate tumor hypoxia for enhancement of radiotherapy. For example, delivery of oxygen by hyperbaric oxygen therapy or via oxygen carriers has been attempted but results have lacked reproducibility or not achieved significant benefit (11, 12). A variety of antiangiogenic agents (e.g., inhibitors of HIF-1α and VEGF) have been developed to normalize the tumor vasculature network and subsequently reduce tumor hypoxia. This approach has been tested clinically in combination with radiotherapy (13–16); however, it yielded marginal therapeutic benefit, with increased cardiotoxicity and metastases in some cases (16–18). Tumors have also appeared to rapidly develop resistance to such treatments (19). Other approaches have also been attempted to mitigate hypoxia-associated radioresistance, including the use of drugs that preferentially kill or sensitize hypoxic cells to radiation. But these approaches were often limited in the clinical setting due to safety concern and inconsistent response (20–22).

While advanced image-guided radiotherapy (IGRT) is commonly used clinically to “dose-paint” solid tumors, a major side effect of treatment remains the unavoidable exposure of surrounding healthy tissues to radiation (23). Yet, to overcome the inherent radioresistance of hypoxic cells, higher radiation doses are often employed, resulting in a greater probability of collateral damage to healthy tissues and vital organs that are in the radiation field (24, 25). For example, chest irradiation with high cumulative radiation doses (40–60 Gy) for treatment of breast and lung cancers are associated with late presentation of cardiovascular complications (25, 26). Thus a means of achieving maximal tumor regression from radiotherapy while sparing surrounding healthy tissues is highly desirable, which may be attained by using tumor-specific radiosensitizers [including in situ oxygen (O₂)-generating nanoparticles (NP)] that potentiate radiotherapy.

To achieve this goal, we have recently proposed the use of biocompatible manganese dioxide (MnO₂) NPs (MDNPs) for sustained and localized generation of molecular O₂ in solid tumors to modulate the tumor microenvironment (TME) and enhance radiation efficacy (27, 28). In our initial work, we developed the first pharmaceutically acceptable albumin-based MnO₂ (AMD) formulation, and demonstrated its effects on the concomitant attenuation of hypoxia, acidosis, HIF-1α, and VEGF in a murine EMT6 tumor model. The proof-of-principle study showed that intratumoral (i.t.) injection of the AMD formulation followed by a single dose of 10 Gy irradiation resulted in higher relative tumor cell kill and short-term tumor growth delay (TGD) than radiation alone (27).

Encouraged by these compelling results, we subsequently developed new formulations of hybrid MDNPs with tunable oxygen generation rates and optimized colloidal stability under physiologic in vivo conditions with an eye on translating our approach clinically (29). Two new MDNP formulations were then designed with distinct O₂-generating kinetics (28). A hydrophilic terpolymer-albumin–based formulation of MDNPs (TMD) was shown to generate O₂ quickly, which is practical for clinical use in treating anatomically accessible tumors via intratumoral injection followed thereafter by a single dose of radiation. In comparison, a lipid-polymer–based formulation of MDNPs exhibited excellent tumor accumulation and retention after intravenous injection but a lower reaction rate compared with TMD especially at the normal circulating blood pH range (28). These properties of the LMD formulation are ideal for a more gradual O₂ generation via either intratumoral or intravenous treatment alone or in combination with radiotherapy. TMD and LMD formulations are prepared using biocompatible starch-based terpolymer and myristic acid (a lipid derived from animals and vegetable fats), respectively. As both starch and myristic acid occur naturally, we expect MDNP formulations should be biocompatible and safe as MnO₂ carriers for clinical use. Hence, expanding on our previous studies, the current preclinical work investigates the new MDNP formulations (designed for clinical use), in two breast tumor experimental models to determine whether, and to what extent, they effectively overcome hypoxia-mediated radioresistance to achieve meaningful therapeutic benefit compared with radiotherapy alone. In this study, we used the established murine EMT6 breast tumor model in BALB/c mice as a clinically representative hypoxic tumor model. The EMT6 tumors grow quickly and aggressively with a relatively large portion of xenografted tumors being hypoxic and radioresistant (28, 30). To emphasize clinical relevance, we also used the well-characterized human MDA-MB-231 breast tumor xenograft model in SCID mice to investigate the possible effects of MDNPs on radiation response in vivo. The human MDA-MB-231 model is slow-growing with “triple negative” characteristics underscoring its clinical relevance (31, 32). Triple-negative tumor cells do not express estrogen receptor (ER), progesterone receptor (PR), or HER2. They are less responsive to treatments (including radiotherapy) and have a poor prognosis with relatively high risk of local and regional recurrence (33).

EMT6 breast tumor cell were originally provided by Dr. Ian Tannock (Ontario Cancer Institute, Ontario, Canada) and have been maintained in the Wu laboratory. MDA-MB-231 cells were obtained from ATCC. The EMT6 cells were authenticated using short tandem repeat analysis (Toronto Center for Applied Genomics, Toronto, Ontario, Canada), while the MDA-MB-231 cell line was not independently authenticated. Cells were grown in growth medium [α-minimal essential medium (α-MEM) supplemented with 10% FBS (Life Technologies)] at 37°C in a humidified incubator with 5% CO₂ atmosphere. Early passage cells were used for all the in vitro cell and in vivo tumor studies.

Two different MDNPs formulations were prepared by closely following our previously published protocol (28). The details of chemicals and synthesis protocols for MDNPs are also described in Supplementary Section S1. The average diameter of TMD and LMD NPs was 140 and 170 nm, respectively, determined using transmission electron microscope (TEM, Hitachi H7000, Hitachi Canada) and dynamic light scattering (DLS, Zetasizer; Supplementary Fig. S1; ref. 28).

All animal experiments were performed in accordance with the guidelines and regulation of the Animal Care Committee at University Health Network (AUP 3166). Animals were euthanized if any of the following endpoints were reached: rapid weight loss, ulcerated tumor, or the tumor volume exceeding 0.6 cm³.

Murine EMT6 and human MDA-MB-231 tumors were grown orthotopically by injecting 1 × 10⁶ EMT6 or MDA-MB-231 cells suspended in 30 μL of cell culture growth medium into the mammary fat pads of 6 to 8 weeks old female BALB/c (Jackson Laboratory) or female SCID mice (Ontario Cancer Institute, Ontario, Canada), respectively. EMT6 tumors were allowed to grow for one week before the treatment. MD-MB-231 tumors were allowed to grow for almost 2 weeks before treatment. Tumor-bearing mice were randomized at the same age and divided into four groups: (i) saline, (ii) saline + radiotherapy, (iii) MDNPs, and (iv) MDNPs + radiotherapy for both intratumoral and intravenous treatments. For the survival studies, 5 to 10 animals were in each group. For the IHC studies, 3 animals were used for the each treatment group.

Tumor growth was monitored by measuring the tumor size in two dimensions using a vernier caliper and tumor volume was calculated using the formula,

During the study, mice were allowed free access to food and water. Tumors were transplanted one week after receipt of animals.

Tumor irradiation was performed using XRAD 225Cx irradiator with photon energy of 225 KVp and a tube current of 13-mA and 0.3-mm Cu filter. For local MDNP treatment, animals were treated with single intratumoral injection of TMD or LMD NPs (50 μL 1 mmol/L, MnO₂) or saline as control, and irradiated locally with a single radiation dose of 10 Gy 30 minutes postinjection. For the intravenous treatment, animals received a single tail vein injection of LMD NPs (200 μL 1 mmol/L, MnO₂) or saline as control and irradiated with a single radiation dose of 10 Gy 4 hours postinjection to maximally benefit from oxygenation. We previously determined that the maximum peak of tumor oxygenation (and thus the lowest hypoxia level) occurred 30 minutes, and 4 hours post intratumoral and intravenous injection of MDNPs, respectively, (Supplementary Figs. S2 and S3; ref. 28). Tumors were localized by X-ray fluoroscopy and then irradiated locally using a collimator with 1-cm radius with a dose rate of 2.71 Gy/min. During the irradiation, animals were maintained at 1.8% isoflurane delivered through a nose cone.

Rabbit anti-VEGF antibody (ab46154, Abcam), and rabbit anti-Ki67 antibody (RM-9106-51, Thermo Scientific) were used for staining VEGF and cell proliferation, respectively. The TUNEL assay (11093070910, Roche) was used to measure tumor cell apoptosis. To stain tumor vasculature, rabbit anti-CD31 antibody (ab28364, Abcam) was used. For all tumor tissues, hematoxylin and eosin (H&E) staining was used to determine viable tumor areas for the image analysis. Tumor tissue preparation was performed by the Centre for Modeling Human Disease Pathology Core laboratory at Mount Sinai Hospital (Ontario, Canada). Tumor tissue was stained with γH2AX antibody for DNA DSBs. All slides were scanned with a NanoZoomer 2.0 RS whole slide scanner (Hamamatsu). Images were analyzed with Visiopharm 4.4.4.0 software. IHC image quantifications were conducted using ImageJ software (Supplementary Section S4 and Supplementary Fig. S4). For immunofluorescence staining, at predetermined times, animals received an intraperitoneal injection of pimonidazole HCl solution (60 mg/kg; Hypoxyprobe-1 Plus Kit, Hypoxyprobe Inc), and were sacrificed 30 minutes after pimonidazole administration. Tumor tissues were quickly harvested, formalin fixed, then embedded in paraffin and stained following the manufacturer's protocol. Separate sections were stained with H&E. Stained tumor slides were imaged using an Olympus upright microscope, and ImagePro software was used to quantify the pimonidazole signals. H&E sections were utilized as a guide to delineate viable tumor areas. Nontumor tissue (including lymph nodes), necrotic areas and artefacts were carefully excluded.

Data are presented as mean ± SD, Student t test, or ANOVA followed by Tukey t test (OriginPro8) was utilized to determine statistical significance between two or more groups, respectively. P < 0.05 was considered statistically significant.

We have previously demonstrated the ability of MDNPs to modulate hypoxia in TME by reacting with endogenous H₂O₂ (27, 28), known to be produced by cancer cells under oxidative stress (34, 29; for details see Supplementary Section S5 and Supplementary Table S1). In the current study, we observed the effect of MDNPs on radiotherapy. For this purpose, we first examined the short-term effects of various MDNP and radiotherapy treatment combinations on tumor growth rate and the TME characteristics, for example, VEGF expression, cell proliferation, tumor vasculature, and DNA DSBs. Having confirmed the positive effects of the MDNPs on TME and radiotherapy through short-term studies, we then investigated whether, and to what extent, pretreatment with MDNPs via local or systemic administration followed by single dose (10 Gy) radiotherapy effected long-term survival of both murine and human tumor-bearing mice. The detailed study arms and evaluations are presented in Supplementary Section S6 and Supplementary Table S2.

The short-term effects of MDNPs with or without radiotherapy (MDNPs+/−RT) were assessed by measuring tumor growth up to 5 days after treatments. The detail results are presented in Supplementary Section S7. Strikingly, significant regression up to 87% was observed in both EMT6 and MDA-MB-231 tumor models treated with MDNPs + radiotherapy at day 5 after treatment when compared with tumor volume at day zero, while radiotherapy alone only resulted in TGD without regression (Supplementary Fig. S5). Consistently, a decrease in tumor weight up to 85% was observed at day 5 for the mice treated with MDNPs + radiotherapy combination, compared with radiotherapy alone (Supplementary Fig. S5). Interestingly, MDNP treatment alone also led to TGD up to 94% as compared with saline group (Supplementary Table S3), probably due to the effects of MDNPs on TME such as attenuation of hypoxia (Supplementary Figs. S2 and S3) and reduction of HIF-1α (28). Here we examined the connection between the TME modulation and MDNPs + radiotherapy treatment–induced tumor regression and TGD to gain insights into the potential mechanisms of action for the tumor response to the treatment. The experimental design for the intratumoral treatment studies is shown schematically in Fig. 1A. Tumor tissues were harvested at 24 hours after treatment, sectioned, and stained for the antigen Ki67 and apoptosis [determined using deoxynucleotidyl transferase (TdT)–mediated deoxyuridine triphosphate (dUTP) nick end labeling (TUNEL)]. Radiation-induced DNA DSBs were evaluated in the EMT6 tumor model using γH2AX tumor tissue staining. Representative images of IHC-stained tissue and corresponding staining quantification are presented in Fig. 1 for both tumor models. As seen in Fig. 1B and C, cell proliferation, measured by Ki67 expression, decreased by 30% (TMD + RT) and 37% (LMD + RT) in EMT6 tumors, and 16% (TMD + RT) and 22% (LMD + RT) in MDA-MB-231 tumors as compared with saline + RT (P < 0.001). MDNPs treatment alone showed a slight decrease in cell proliferation in both EMT6 and MDA-MB-231 tumors. As expected, tumor cell apoptosis exhibited an opposite trend (Fig. 1). Both tumor models showed an approximately 1.3-fold increase in TUNEL-positive cells in the MDNP + RT groups compared with saline + RT, with 230/mm² (TMD + RT; P < 0.001), 229/mm² (TMD + RT; P < 0.05) versus 181/mm² for saline + RT in EMT6 tumors, and 128/mm² (TMD + RT; P < 0.05), 132/mm² (LMD + RT; P < 0.05) versus 99/mm² (saline + RT) in MDA-MB-231 tumors.

As DNA DSBs are a main mechanism of radiation-induced cell death, we stained EMT6 tumors 24 hours after treatments using γH2AX antibody. As shown in Fig. 1B, bottom right plot, the combination of MDNPs with radiotherapy resulted in 1.7- to 1.9-fold increase in DNA DSBs with 29% and 25% positive cells in the TMD + RT and LMD + RT treatment group, respectively, compared with 15% observed in the saline + RT group (P < 0.001). These results attest that the enhanced short-term (e.g., up to 5 days) antitumor effect of radiotherapy by MDNP treatment is largely attributable to an increase in tumor cell DNA DSBs.

After the initial assessment of TGD using the EMT6 tumor model (Supplementary Fig. S6), we studied the influence of intratumorally injected MDNPs + RT on survival in the human MDA-MB-231 breast tumor model by measuring the tumor volume once a week until the humane endpoint was reached (n = 5/group). The time course of change in tumor volumes for individual mice in control and treatment groups is plotted in Fig. 2A, from which Kaplan–Meier survival curves were obtained and median survival times for each group were determined (Fig. 2B). Consistent with results in EMT6 tumors (Supplementary Fig. S6), the majority of MDA-MB-231 tumors regressed over a relatively short period (5–10 days) after treatment with MDNPs and a single 10 Gy dose of focal radiation and then regrew until the humane end point, albeit at a slower rate compared with saline + RT group (Fig. 2A). Compared with a 25% increase in the median survival time (from 20 days for saline to 25 days for saline + RT group), the median survival time increased to 35 days and 38 days following TMD + RT and LMD + RT treatments, respectively. These results correspond to a 3- and 3.6-fold improvement in radiotherapy efficacy for the TMD + RT and LMD + RT groups, respectively, compared with the saline + RT group with a single 10 Gy dose. Note that an intratumoral injected single doses of TMD or LMD alone did not significantly alter the survival time of the human breast tumor-bearing mice.

Our previous study demonstrated superior tumor accumulation and longer retention time of hydrophobic LMD NPs compared to hydrophilic TMD NPs after intravenous administration in EMT6 tumor model (28). In this work, we further confirmed that LMD NPs also exhibited good systemic circulation and tumor accumulation in the MDA-MB-231 tumor model by in vivo and ex vivo fluorescence imaging (Fig. 5A and Supplementary Section S8).

We first assessed tumor response to intravenously injected MDNPs and radiotherapy treatments in both EMT6 and MDA-MB-231 models by measuring percent change in tumor volumes from time zero to day 5 posttreatment, and measuring the tumor weights of various groups (Supplementary Section S7). Similar to the intratumoral administration, the combination of intravenous LMD NPs and radiotherapy treatment also led to regression in tumor volume (up to 32%), when compared with tumor volume at day zero, and lower weight of resected tumors (up to 46%) compared with radiotherapy alone, at day 5 (Supplementary Fig. S7). Immunohistochemical analysis was performed on EMT6 and MDA-MB-231 tumor tissues. The experimental design for the intravenous treatment studies is shown schematically in Fig. 3A. In agreement with the results of intratumoral treatments, the LMD NPs intravenous treatment alone also inhibited cell proliferation (Ki67 staining) and enhanced apoptosis (TUNEL staining) in both tumor models (Fig. 3). This effect was more prominent when LMD NPs were combined with radiotherapy, Ki67-positive areas decreased by 27% (P < 0.01) and 24% (P < 0.001) in EMT6 and MDA-MB-231 tumors, respectively, compared with saline + RT group, while TUNEL-positive areas increased by approximately 1.6-fold in EMT6 tumors [273/mm² vs. 174/mm² for saline + RT (P < 0.05)] and by approximately 1.2-fold in MDA-MB-231 tumor model [139/mm² vs. 109/mm² for saline + RT (P < 0.001)]. The combination treatment resulted in 43% increase in DNA DSBs in EMT6 tumors compared with saline + RT group (P < 0.001; Fig. 3B). These results confirm that systemic administration of LMD NPs could also produce antitumor effects.

For long-term survival studies, mice bearing EMT6 tumors or MDA-MB-231 tumors were treated intravenously with LMD NPs (200 μL 1 mmol/L, MnO₂) or the same volume of saline as control and irradiated with a single radiation dose of 10 Gy at 4 hours after injection (n = 5–10/group). Tumor volumes were monitored until the humane endpoint was reached. To confirm the reduction in tumor hypoxia following systemic administration of LMD NPs, we performed H&E and immunofluorescence staining and quantification for EMT6 tumors (n = 10). At 4 hours after LMD intravenous treatment before application of irradiation, a significant decrease in tumor hypoxia was observed as compared with saline control group (n = 5) evidenced by dimmed green color and reduced positive area (46%, P < 0.001; Fig. 4A). This result is consistent with our previous observation (28). Figure 4B depicts the tumor growth curves of individual EMT6 tumor-bearing mice from different treatment groups. The LMD + RT treatment extended median survival time from 15 days (saline) to 40 days, whereas saline + RT increased the median survival time to 20 days (Fig. 4C). In other words, a 5-fold improvement in median survival time was obtained by combining LMD NPs with radiotherapy as compared with saline + RT. More strikingly, curative treatment was achieved in approximately 40% of mice treated with LMD + RT. To confirm the complete regression of tumors, we stained ex vivo mammary fat pad tissues of these long-term surviving animals (n = 4) excised at 120 days after treatment using H&E and compared them with H&E-stained mammary fat pad tissues from healthy (untreated) animals (n = 3). From the H&E staining, we did not observe any signs of remaining tumor cells in the mammary fat pad of the surviving animals (Supplementary Fig. S8).

The LMD + RT treatment also enhanced the antitumor efficacy of radiotherapy and extended survival time of MDA-MB-231 tumor-bearing mice (Fig. 5B). The median survival time increased from 20 days (saline) to 27 days (saline + RT), and from 22 days (LMD) to 38 days (LMD + RT), demonstrating a 2.6-fold improvement in median survival with LMD + RT treatment compared with radiotherapy alone (Fig. 5C).

To test whether the observed enhancement of radiation efficacy was induced mainly by MDNP pretreatment that produced oxygen in situ, we treated EMT6 tumor-bearing mice with intravenous injection of LMD NPs (at the same dose used for the efficacy experiment) 10 minutes after the radiotherapy. Tumor growth was monitored until the humane endpoint was reached. Supplementary Figure S9 shows the median survival time was 20 days for mice that received radiation before LMD NP treatment, or saline + RT. This agreement between median survival times supports our working hypothesis that the radiotherapy enhancement by LMD NPs is due largely to the reoxygenation of the TME by the action of the MDNPs prior to radiotherapy treatment.

VEGF plays an important role in tumor angiogenesis by stimulating endothelial cell promotion and proliferation (35), and has a great effect on tumor growth (36, 37). We investigated the effect of MDNPs + RT on VEGF expression in both EMT6 and MDA-MB-231 tumors at 24 hours post intratumoral and intravenous treatments using IHC staining. As seen in Fig. 6A, the intratumoral treatment of TMD and LMD alone caused a decrease in EMT6 tumor VEGF by 23% and 29%, respectively, compared with saline (top; P < 0.001). These results were consistent with our previous findings for EMT6 tumor using AMD NPs (27). Furthermore, when MDNPs were combined with radiotherapy, VEGF expression in EMT6 tumors decreased by 58% and 61%, that is, with only 4.5% and 4.2% of VEGF-positive area in TMD + RT and LMD + RT groups, respectively, compared with 11% positive area in saline + RT group (P < 0.001). A similar trend was observed in the MDA-MB-231 tumor model but with a less reduction (Fig. 6A, bottom), perhaps due to markedly high VEGF expression at the basal level as observed in this work and in literature (38). For the human breast tumor model, MDNP treatment alone decreased VEGF expression over 5% compared with the saline control, while TMD + RT and LMD + RT respectively reduced VEGF expression by 26% and 27% (with 44% and 43% of VEGF-positive area, respectively), compared with 59% for the saline + RT (P < 0.001).

Consistent with the results from intratumoral treatment, the intravenous treatment with LMD NPs decreased VEGF levels by 14% in murine EMT6 tumors and by 10% in MDA-MB-231 tumors compared with the saline groups (Fig. 6B). The LMD + RT combination reduced VEGF expression by 56% (P < 0.001) in EMT6 tumors and by 24% in the MDA-MB-231 tumors (P < 0.001) compared with saline + RT. In contrast, the VEGF expression considerably increased after saline + RT treatment in both EMT6 and MDA-MB-231 tumors.

As VEGF is a major driver of tumor angiogenesis, we examined possible changes in tumor vasculature using cluster of differentiation 31 (CD31) staining of tumor slices collected at day 5 after treatment. Representative immunohistochemical staining images of tumor sections and corresponding quantification of CD31 marker in tumors after intratumoral treatments are presented in Fig. 7A and those for intravenously treated tumors in Fig. 7B. As seen in Fig. 7A (top), EMT6 tumors treated with intratumoral injection of MDNPs showed reduced CD31-positive area by 18% (TMD) and 20% (LMD), respectively, compared with the saline group. When combined with radiotherapy, the percentage of CD31-positive area decreased by 42% and 46% for TMD and LMD groups, respectively, compared with the saline + RT groups (P < 0.01). In MDA-MB-231 tumors, the CD31-positive areas decreased by 24% and 29% for TMD + RT and LMD + RT, respectively, compared with saline + RT groups (P < 0.001).

The intravenous treatment with LMD NPs alone also decreased vascular density in EMT6 tumors by 9% compared with the control group. After treatment with intravenous LMD + RT, the CD31-positive area decreased by 31% compared with RT alone (Fig. 7B, top, 2.9% for LMD + RT vs. 4.2% for saline + RT; P < 0.05). Similarly, the MDA-MB-231 tumors treated with intravenous LMD + RT showed a 30% decrease in CD31-positive area, compared with the saline + RT group (Fig. 7B, bottom, 4.7% for LMD + RT vs. 6.7% for saline + RT, P < 0.01).

In this study, two newly developed, clinically suitable MDNP formulations (TMD and LMD), were evaluated in two different preclinical breast tumor models for their ability to improve radiotherapy efficacy. Short-term changes in TME in response to intratumorally or intravenously administered MDNPs alone or with single dose (10 Gy) radiation were also investigated to assess the cellular, molecular, and genetic mechanisms of the antitumor efficacy of the treatments. Supplementary Table S2 summarizes the relative changes in tumor growth, tumor cell proliferation, apoptosis, VEGF expression, and tumor vascular density after treatments with MDNPs alone or in combination with radiotherapy when compared with saline controls. Our in vivo studies indicated that MDNP + RT treatment prolonged the median survival time in both murine and human breast tumor models by 3- to 5-fold compared with radiotherapy alone. Hydrophobic, LMD NPs exerted an antitumor effect especially when combined with radiotherapy irrespective of the route of administration (i.e., intratumoral or intravenous injection). Remarkably, intravenous application of LMD NPs with radiotherapy resulted in 40% cure of EMT6 tumor-bearing mice with a 1 × 10 Gy radiation dose (Fig. 4; Supplementary Table S2). Recently a similar (41%) cure rate in the same tumor model was reported using a copper-containing porphyrin compound in combination with 1 × 25 Gy radiotherapy (39), 2.5 times the dose used in the current study. Collectively, these results indicate that MDNPs not only enhance the efficacy but also the potency of radiotherapy markedly in these preclinical breast tumor models. Our data support the notion that MDNP + RT treatment of solid breast tumors could allow for equivalent tumor reduction using a lower single dose of irradiation compared with that of a single high radiotherapy dose alone, thereby potentially reducing the collateral healthy tissue damage of the latter while achieving therapeutic benefit. Certainly, this would need to be validated in a randomized controlled clinical trial.

Ideally 100% curative treatment is desirable for all therapeutic modalities including the combination therapy of MDNPs + RT studied here. However, in reality, there always are poor-responding individuals in clinic and preclinical studies due to a variety of factors. In this work, we believe that heterogeneity in tumors such as differences in tumor size, shape, structure, hypoxia degree and distribution, as depicted in Fig. 4A, plays a role in the nonuniform response to the treatment. To answer how these factors and other biological factors influence the treatment outcomes requires extensive future studies.

The observation of increased DNA DSBs observed in the MDNP + RT–treated EMT6 tumors is consistent with the hypothesis that tumor growth inhibition following the combination treatment is mainly a result of localized tumor reoxygenation, which potentiates radiotherapy. The negligible improvement in TGD and survival in tumor-bearing animals receiving intravenous LMD NPs after radiotherapy strengthens the argument that intratumoral reoxygenation prior to the radiotherapy is critical for enhancing the antitumor effect of irradiation (Supplementary Fig. S6). TUNEL staining showed an increase in tumor cell apoptosis while Ki67 staining indicated significant decreases in tumor cell proliferation at 24 hours after MDNP + RT administration in both tumor models (Figs. 1 and 3; Supplementary Table S2). These results also confirmed the antitumor effect of MDNPs when combined with radiotherapy.

Consistent with our previous findings (27), the new MDNP formulations also reduced tumor VEGF expression. This effect was more pronounced in the groups treated with MDNPs + RT. The downregulation of tumor VEGF is not surprising given the potent effect of MDNPs on attenuation of tumor hypoxia and the transcription factor HIF-1α as reported previously (28), and the well-documented effect of HIF-1α inhibition on reducing tumor VEGF expression (40, 41). Our results suggest that in addition to sustainably generating intratumoral oxygen, a potent radiosensitizer, MDNPs can also function as potent an anti-HIF-1α and anti-VEGF agents.

Radiation induces upregulation of VEGF expression and angiogenesis (42, 43); however, it also causes damage to endothelial cells following combination treatment with VEGF inhibitors (44, 45). The formation of new blood vessels helps to deliver oxygen, nutrients, and growth factors to cancer cells while removing waste products, resulting in tumor development and metastasis (46–48). Thus, blocking VEGF can not only inhibit new vessel growth, but can also cause a regression in tumor blood vessels that could lead to enhanced TGD (49). Radiotherapy alone induced an increase in tumor CD31 expression as found in this work. Interestingly, treatment with MDNPs alone resulted in a decrease in tumor CD31 expression in both tumor models, which correlates with the observed anti-VEGF activity of MDNPs. Immunohistochemical analysis also confirmed the reduction in tumor vessel density at day 5 after MDNP or MDNP + RT treatments, which could have affected TGD in the treated mice. The decrease in tumor vascular density in MDNP + RT groups is similar to published observation for tumors treated with a combination of RT and anti-VEGF antibody-producing virus (50).

Overall, the murine EMT6 tumors displayed better response to the treatments compared with human MDA-MB-231 breast tumors. The lower radiosensitivity of MDA-MB-231 cells might be expected due to their triple-negative nature. Tumors generated from this cell largely consist of epithelial-to-mesenchymal transition and stem cell-like phenotypes, which respond slowly to treatment (51). In addition, radiotherapy is often more effective during G₂–M phase (52). Therefore, a rapidly growing EMT6 tumor might consist of more cells in a radiation sensitive phase of the cell cycle, and thereby exhibit higher radiation response.

We have shown that hydrophilic TMD NPs can generate radiotherapy response following local intratumoral treatment, while the slow-reacting hydrophobic LMD NPs showed radiotherapy enhancement independently of their route of administration (intratumoral or intravenous). We anticipate in the future that local administration of MDNPs formulations may become useful for the treatment of locally accessible tumors eligible for radiation treatment. The local intratumoral injection could deliver the maximum MDNPs to the target site with a lower administered dose compared with intravenous administration and less nonspecific tissue deposition. Intratumoral administration could also minimize the overall treatment time for accessible tumors where, for example, the patient could receive MDNPs shortly before receiving radiotherapy. On the other hand, intravenous-injectable LMD NPs could be a promising candidate for the treatment of many difficult-to-access locally advanced tumors such as those in the lungs, liver, and pancreas where radiotherapy is a common treatment option. Moreover, the clinically compatible MDNP formulations presented here have the potential to improve the antitumor efficacy of radiotherapy at lower radiation doses, thereby effectively reducing the risk of normal tissue injury that remains an important concern to the radiation oncologist.

No potential conflicts of interest were disclosed.

Conception and design: A.Z. Abbasi, C.R. Gordijo, A.M. Rauth, R.S. DaCosta, X.Y. Wu

Development of methodology: A.Z. Abbasi, C.R. Gordijo, M.A. Amini, A.M. Rauth, R.S. DaCosta, X.Y. Wu

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A.Z. Abbasi, C.R. Gordijo, M.A. Amini, A. Maeda, R.S. DaCosta, X.Y. Wu

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A.Z. Abbasi, C.R. Gordijo, M.A. Amini, A. Maeda, A.M. Rauth, R.S. DaCosta, X.Y. Wu

Writing, review, and/or revision of the manuscript: A.Z. Abbasi, C.R. Gordijo, M.A. Amini, A. Maeda, A.M. Rauth, R.S. DaCosta, X.Y. Wu

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): A.Z. Abbasi, M.A. Amini, A.M. Rauth, R.S. DaCosta, X.Y. Wu

Study supervision: A.M. Rauth, R.S. DaCosta, X.Y. Wu

We acknowledge the University of Toronto for the scholarship to M.A. Amini and Drs. P. Cai and Dr. C. He for their technical support.

This work was supported by Canadian Cancer Society Research Institute Innovation grant and Natural Sciences and Engineering Research Council of Canada Discovery grant.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.