Protons show greater relative biological effectiveness for mammary tumorigenesis with higher ERα and HER2 positive tumors relative to γ-rays in APC^Min/+ mice

Abstract

Purpose.

Exposure to ionizing radiation increases risk of breast cancer. Although proton radiation is encountered in outer space and in medicine, we do not fully understand breast cancer risks from protons due to limited in vivo data. The purpose of this study was to comparatively assess the effects of γ-rays and protons on mammary tumorigenesis in APC^Min/+ mice.

Methods and Materials.

Female APC^Min/+ mice were exposed to 1 GeV protons (1.88 or 4.71 Gy) and ¹³⁷Cs γ-rays (2 or 5 Gy). Mice were euthanized 100 to 110 days after irradiation, mammary tumors scored, tumor grades assessed, and relative biological effectiveness (RBE) calculated. Molecular phenotypes were determined by assessing estrogen receptor α (ERα) and human epidermal growth factor receptor 2 (HER2) status. ERα downstream signaling was assessed by immunohistochemistry.

Results.

Exposure to proton radiation led to increased mammary tumor frequency at both proton radiation doses compared to γ-rays. The calculated RBE for proton radiation-induced mammary tumorigenesis was 3.11 for all tumors and >5 for malignant tumors relative to γ-rays. Tumor frequency per unit of radiation was higher at the lower dose suggesting a saturation effect at the higher dose. Protons induced more adenocarcinomas relative to γ-rays, and proton-induced tumors show greater ERα and HER2 positivity and higher activation of the ERα-downstream PI3K/Akt and cyclin D1 pathways relative to γ-rays.

Conclusions.

Our data demonstrate that protons pose a higher risk of mammary tumorigenesis relative to γ-rays. We also show that proton radiation-induced tumors in APC^Min/+ mice are ERα and HER2 positive, which is consistent with our previous data on radiation-induced estrogenic response in wild-type mice. While this study establishes APCMin/+ as a model with adequate signal-to-noise ratio for space radiation-induced mammary tumorigenesis, further studies will be required for addressing the uncertainties in space radiation-induced breast cancer risk estimation.

Introduction

Epidemiological studies have demonstrated a link between radiation exposure and increased breast cancer risk (1–5). However, since most of these studies involved photon radiation such as γ-rays and x-rays, the issue of breast cancer risk from proton radiation remains unresolved. The use of proton radiation, due to its higher precision dose delivery and greater damaging effects relative to conventional photon-based radiotherapy, has been gaining ground in radiation oncology (6–8). Proton radiation deposits its energy at a depth in the tissue with maximum deposition at the Bragg peak. In contrast, photons deposit most of the energy close to the entrance into the body and decrease exponentially with increasing tissue depth. While normal tissue exposure during proton radiotherapy can be minimized with the adoption of appropriate beam design techniques, it cannot be completely avoided in the entrance plateau region of the Bragg curve as well in the peripheral region of the spread-out Bragg peak (SOBP) (9). Unlike the situation with photon radiation, we have limited long-term follow up data on proton radiation due to its recent incorporation into mainstream radiotherapy.

Small number of female astronauts is also a constraining factor in our understanding of breast cancer risk from proton radiation in outer space where protons contribute up to 90% of the dose equivalent to the space radiation spectrum (10–12). While 85% of the galactic cosmic radiation (GCR) spectrum is comprised of protons of various energies, 90% of a solar particle event (SPE) consists of protons and during an active SPE, astronauts could receive up to 2 Gy of acute proton radiation dose (10, 12). While low energy protons might be blocked, high energy protons collide with the spacecraft materials to produce a mixed secondary radiation environment inside the spaceship (13) which is a major risk for superficial breast tissues. Given that breast tissue is sensitive to even low doses of radiation (1, 14) and astronauts could be exposed up to 2 Gy of protons during an SPE along with continuous low dose proton exposure during long duration space missions (12, 15), female astronauts are predicted to have a cumulative proton dose that will increase risk of breast cancer during and after the mission (16, 17). Currently, there is much uncertainty in the prediction of breast cancer risk from protons due mostly to lack of qualitative and quantitative in vivo human or animal data.

Qualitatively, elevated level of estrogen is linked to increased risk of breast cancer (18, 19). Most of the biological effects of estrogens are mediated through the estrogen receptor alpha (ERα) (20) and ERα is expressed in the majority of the breast tumors (21). Estrogen, through interaction with its receptor, exerts a potent stimulus on breast cell proliferation through direct and/or indirect activation of proliferative pathways (22) and the production of pro-proliferative factors including insulin-like growth factor 1 (IGF1) (23), cyclin D1 (24), and STAT5 (25). Activation of the MAPK and PI3K/Akt signaling pathway by estrogen have been extensively studied in several cell types including breast cancer (26). ERá has also been reported to interact directly with ErbB2 (HER-2/neu) (27). Notably, ERá mediated activation of the epidermal growth factor (EGF) receptor occurs through activation of G proteins, Src kinase, and matrix metalloproteinases leading activation of the MAPK and Akt pathways (28). In the current study, mammary tumorigenesis after exposure to two different doses of high-energy protons were assessed and compared to two equitoxic γ-ray doses.

Methods and Materials.

Mice and radiation exposure.

APC^Min/+ mice were bred according to a protocol described previously (12) and female mice (6 to 8 weeks) were irradiated whole-body without anesthesia to either γ-rays (energy: 0.662 MeV; LET: 0.8 keV/μm) or protons (energy: 1 GeV; LET: 0.22 keV/μm at the entrance plateau region of the Bragg curve) and control mice were sham irradiated. Mice (n=25 mice/study group) were exposed to either 2 or 5 Gy of γ-rays and for proton radiation exposure, we used either 1.88 or 4.711 Gy. Radiation exposures were at a dose rate of 1 Gy/min and exposures were lateral from horizontal radiation sources. The proton and γ radiation doses are equitoxic, which were calculated using an RBE of 1.06 determined previously from a survival (LD_50/30) study (29). The dosimetry for the proton irradiation was calculated by the physics team at NASA Space Radiation Laboratory at Brookhaven National Laboratory using parallel-plate ion chambers and a thimble chamber, and has been discussed previously in detail (30–33). Mice were exposed at the entrance plateau region of the Bragg curve so that they are exposed at a constant LET. For γ-irradiation dosimetry, we used pre-calibrated dosimetry table and mice were exposed on a turntable at the third position of a J.L. Shepherd Mark I Model ¹³⁷Cs γ-irradiator. For mice transportation between Georgetown University and Brookhaven National Laboratory, radiation exposure, and radiation dosimetry, we followed procedures established and described previously (12, 29, 34). We followed the Guide for the Care and Use of Laboratory Animals for our studies.

Mammary tumor count.

Mice were observed daily and euthanized between 100 and 110 days post-irradiation according to procedures established previously (12). Mice were dissected to expose the mammary fat pads and mammary tumors were scored. Total mammary tumor data along with malignant mammary tumor data derived from total tumor frequency data are presented as average tumor/malignant tumor frequency per mouse. Since radiation doses are different due to the use of equitoxic doses, we converted the data into tumor frequency per unit (cGy) of radiation (tumor frequency / radiation dose in cGy). Tumor samples were fixed in 10% buffered formalin for 24 h and placed in 70% ethanol prior to paraffin embedding and sectioning.

Radiation dose response modeling and relative biological effectiveness (RBE) calculation.

Tumor data were also used to calculate the RBE of proton radiation-induced mammary tumorigenesis relative to γ-rays (35). We modeled the mammary tumor dose responses for protons and γ-rays by a linear dependence. RBE is defined as the ratio of doses for different radiation types, which result in the same magnitude of the studied effect. In this case, the radiations being compared are protons and γ-rays. For linear dose responses, such as those used here, RBE can be calculated using the ratio of dose response slopes for protons vs γ-rays. This approach was used for total tumors and for malignant tumors as separate endpoints. This approach was justified because two non-zero dose points are available for this initial study. More complex dose response modeling may be implemented in the future when more data are gathered. The fitting of the linear dose response model was performed in R 3.5. 2 software for total tumors as well as for malignant tumors per mouse.

Histology and immunohistochemistry.

A board-certified pathologist evaluated the H&E stained tumor sections for histologic grading of the tumors. Tumor sections were immunostained for ERα, HER2, p85, Akt, and cyclin D1 as per procedure described previously (19). While ERα and cyclin D1 positive nuclei were scored and the results are expressed as percent of the total cells counted, the staining results for HER2, p85, and Akt were expressed in arbitrary pixel units.

Statistical analysis.

We used IBM SPSS Statistics for Macintosh, v25.0 (IBM Corp., Armonk, NY) to determine normality, equality of variance, and statistical significance in our data set as described previously (34). Briefly, since Shapiro-Wilk test (36) to determine normality of data distribution show non-normal distribution, we used a non-parametric Levene’s test to determine equality of variance (37). Inequality of variance was observed in the mammary tumor dataset because the p value was <0.05. Although tests on the mammary tumor frequency data show non-normal distribution and inequality of variance, we have equal sample size and accordingly we used Welch’s 1-way analysis of variance (ANOVA) with Games-Howell post hoc test (38).

Tumor data are presented graphically as mean ± standard error of the mean (SEM) and the y-axis scales in low and high dose tumor frequency graphs are kept same for comparison.

Immunohistochemistry staining was analyzed as described previously (19). Briefly, 10 to 15 fields of vison at the microscopic magnification specified in each figure were randomly chosen and were either scored for ERα and cyclin D1 positive nuclei or pixel intensity measured for p85, Akt, and HER2 using Image J v1.51. We used color deconvolution and Image-based Tool for Counting Nuclei (ITCN) plug-ins of the ImageJ software as per protocol described earlier (19, 39, 40). We used student’s t-test to determine statistical significance between two groups with statistical significance set at ≤0.05. The error bars represent mean ± SEM.

Results

Increased mammary tumorigenesis in APC^Min/+ mice after proton radiation.

Mammary tumor frequency (± SEM) shows significantly higher tumorigenesis after 1.88 (0.90 ± 0.10; Figure 1A) and 4.71 (1.36 ± 1.60; Figure 1B) Gy of proton radiation relative to equitoxic doses of 2 (0.30 ± 0.07) and 5 (0.52 ± 0.13) Gy γ-rays respectively. Control mice developed few spontaneous tumors (0.04 ± 0.04). Analyses for total tumor counts per mouse showed a slope of 0.316 (SE 0.047, p-value 3.7×10⁻¹⁴) Gy⁻¹ for protons and 0.102 (SE 0.028, p-value 3.8×10⁻⁵) Gy⁻¹ for γ-rays. The ratio of slopes for protons / γ-rays, which represents the RBE, was 3.11 (SE 0.96; Table S1 and Supplemental Methods in Supplemental Information). Since not all the mice develop tumors, we determined the percent of mice with tumor. Our data show a higher percent of mice with tumor after 1.88 (Figure 1C) and 4.71 (Figure 1D) Gy of protons relative to γ-ray doses. Tumor frequency per unit (cGy) of radiation shows that lower doses induced higher tumor number per cGy of radiation relative to higher doses of γ and proton radiation (Figure 1E).

Figure 1. Higher mammary tumorigenesis in APC^Min/+ mice after proton radiation relative to γ-rays.

A) Mammary tumor frequency was higher after 1.88 Gy of protons relative to control and equitoxic 2 Gy γ-rays (p<0.001 compared to control; p<0.03 compared to γ-rays). Higher mammary tumor frequency was also noted after 2 Gy γ radiation relative to control (p<0.02). B) Higher mammary tumor frequency was observed after 4.71 Gy of protons relative to control and equitoxic 5 Gy of γ-rays (p<0.0001 compared to control and γ-rays). Mammary tumor frequency was also increased after 5 Gy γ radiation relative to control (p<001). C) Percent of tumor-bearing mice was higher after 1.88 Gy protons relative to equitoxic 2 Gy γ-rays. D) Percent of tumor bearing mice was higher after 4.71 Gy of protons relative to 5 Gy of γ-rays. E) Average number of mammary tumors per unit of proton radiation was higher at the lower dose relative to the higher dose (p<0.03 when compared between 1.88 and 4.71 Gy of protons). Comparison between low and high dose γ-rays was not statistically significant. Statistical significance for mammary tumorigenesis data is set at p ≤ 0.05 and the error bars represent mean ± standard error of the mean (SEM). *Significant compared to control. **Significant compared to γ-rays.

Proton radiation shows greater RBE and higher malignant tumor frequency relative to γ-rays.

Histologic analysis of mammary tumor grade shows higher malignant tumor frequency after 1.88 (0.56 ± 0.11; Figure 2A) and 4.71 (1.02 ± 0.11; Figure 2B) Gy proton radiation relative to equitoxic 2 (0.07 ± 0.03) and 5 (0.19 ± 0.04) Gy doses of γ-rays respectively. Analyses for malignant tumor counts per mouse showed a slope of 0.215 (SE 0.021, p-value 3.5×10⁻¹⁷) Gy⁻¹ for protons and 0.039 (SE 0.007, p-value 5.0×10⁻⁷) Gy⁻¹ for γ-rays. The ratio of slopes for protons / γ-rays, which represents the RBE, was 5.52 for malignant tumors (SE 1.16; Table S1 in Supplemental Information). Our calculated RBE for malignant tumors was higher at ~8 if the lowest tested doses only (1.88 Gy protons and 2 Gy gamma rays, respectively) were considered. Calculated proportion of benign and malignant mammary tumors show that greater percent of tumors were malignant after 1.88 (Figure 2C; 62%) and 4.71 (Figure 2D; 75%) Gy of proton radiation relative to equitoxic doses of γ-rays (25% after 2 Gy and 37% after 5 Gy). While the majority of the benign tumors were epidermoid cystadenoma, malignant tumors were predominantly acinar adenocarcinoma along with some squamous cell carcinoma and basaloid squamous cell carcinoma. Representative images of an epidermoid cystadenoma from 2 Gy γ-rays (Figure 2E, upper panel) and an acinar adenocarcinoma (Figure 2E, lower panel) from equitoxic 1.88 Gy protons are shown.

Figure 2. Proton exposure was associated with higher frequency of carcinomas relative to γ-rays.

A) Malignant tumor frequency was higher after 1.88 Gy of proton radiation relative to control and 2 Gy γ-rays (p<0.0001 compared to control and γ-rays). Higher malignant mammary tumor frequency was also scored after 2 Gy γ radiation relative to control (p<0.03). B) Higher malignant mammary tumor frequency was observed after 4.71 Gy of proton radiation relative to control and γ-rays. (p<0.0001 compared to control and γ-rays). Exposure to 5 Gy γ radiation also increased malignant mammary tumor frequency (p<0.04). C) Higher percent of carcinomas was noted after 1.88 Gy protons relative to 2 Gy γ-rays. D) Higher percent of carcinomas was noted after 4.71 Gy protons relative to 5 Gy γ-rays. E) Representative image of an epidermoid cystadenoma and an acinar adenocarcinoma.

Proton radiation-induced mammary tumors show increased ERα and HER2 positivity.

Mammary tumor showed a greater number of positive nuclear ERα staining in proton irradiated samples relative to γ-ray samples (Figure 3A). Quantification show significantly higher ERα positive cells after proton radiation relative to γ-rays and the number of ERα positive cells in γ-ray exposed samples were higher relative to controls (Figure 3B). Immunohistochemical stained for HER2 show higher staining in proton tumors relative to γ-ray tumors (Figure 3C). Quantification of HER2 show significantly increased HER2 expression in proton radiation-induced tumors (Figure 3D).

Figure 3. Proton radiation-induced mammary tumors showed higher ERα and HER2 positivity relative to γ-rays.

A) Increased ERα positive nuclei were observed in proton-induced mammary tumors. B) Quantification shows higher percent of ERα positive nuclei after protons relative to control and γ-rays (p<0.0001compared to control and γ-rays). Increased ERα was also observed after γ radiation relative to control (p<0.0001). C) Increased HER2 staining was observed in proton-induced mammary tumors. D) Quantification shows higher staining of HER2 after protons relative to control and γ-rays (p<0.0002 compared to control; p<0.0007 compared to γ-rays). Increased HER2 was also observed after γ radiation relative to control (p<0.0004). Statistical significance for immunohistochemistry quantification data is set at p ≤ 0.05 and the error bars represent mean ± standard error of the mean (SEM). *Significant compared to control. **Significant compared to γ-rays.

Increased activation of ERα downstream proliferative pathways in proton radiation-induced tumors.

Both PI3K and Akt are downstream of membrane ERα (mERα) and expression of p85, the regulatory subunit of the PI3K, as well as that of Akt were assessed by immunohistochemistry. Higher expression of p85 (Figure 4A and B) and Akt (Figure 4C and D) was observed in proton radiation- relative to γ-ray-induced mammary tumors. We also immunohistochemically assessed the expression of cyclin D1, which is downstream of nuclear ERα (nERα) signaling. The data show increased cyclin D1 positive cells in proton radiation-induced mammary tumors relative to γ-ray tumors (Figure 4E and F).

Figure 4. Increased expression of ERα downstream molecules after proton radiation.

A) Higher p85 expression in proton radiation-induced mammary tumors relative to γ-rays. B) Quantification shows increased p85 expression in proton radiation-induced tumors relative to control and γ-rays (p<0.001 compared to control; p<0.002 compared to γ-rays). Exposure to γ-rays also increased p85 level (p<0.002). C) Akt expression was higher in proton radiation-induced tumors. D) Quantification shows increased Akt expression in proton radiation-induced tumors relative to control and γ-rays (p<0.0009 compared to control; p<0.006 compared to γ-rays). Increased Akt was also observed after γ radiation (p<0001). E) Higher number of cyclin D1 positive nuclei in proton radiation-induced mammary tumors. F) Quantification shows increased cyclin D1 expression in proton radiation-induced tumors relative to control and γ-rays (p<0.0001 compared to control and γ-rays). Cyclin D1 was also increased after γ radiation (p<0.0001). Statistical significance for immunohistochemistry quantification data is set at p ≤ 0.05 and the error bars represent mean ± standard error of the mean (SEM). *Significant compared to control. **Significant compared to γ-rays.

Discussion.

Studies in the cohort of atomic bomb survivors implicated radiation as a breast cancer risk factor and showed radiation exposures were associated with increased estrogenic response (1, 41). Our previous data showed radiation-induced increased systemic and local estrogenic responses in wild type mice even twelve months after radiation exposure (18, 19). However, none of these studies involved proton radiation and therefore, a knowledge gap exists in our understanding of proton radiation-induced breast cancer risk. The current study demonstrates proton radiation induced a higher frequency and grade of mammary tumors in APC^Min/+ mice relative to γ-rays with an RBE of ~3. Molecular phenotyping demonstrate that mammary tumors are ERα and HER2 positive and that proton radiation was associated with increased activation of PI3K/Akt and cyclin D1, which are ERα downstream proliferative signaling pathways relative to γ-rays.

The risk of space radiation-induced breast cancer in women astronauts is expected to increase during and after undertaking prolonged space missions such as missions to Mars (17). Indeed, a new breast mass may be diagnosed during a mission (16, 17). However, there is uncertainty in breast cancer risk prediction and risk management protocols especially during and after space radiation exposures due to the lack of sufficient in vivo data. Furthermore, although breast cancer risk in the general population due to radiation is well understood based on epidemiological studies, breast cancer risk after proton exposure remains unexplored. Our data in APC^Min/+ mice demonstrate higher mammary tumorigenesis after proton radiation relative to γ-rays. We also show that a higher percentage of proton-irradiated mice developed mammary tumors relative to γ-rays suggesting the higher tumorigenic potentials of protons. The importance of the current study lies in the fact that the APC tumor suppressor gene is non-functional in up to 70% of human breast cancers and APC^Min/+ mouse strain is an established model for mammary tumorigenesis studies (42–44). The APC^Min/+ model has also been demonstrated to be an effective model for radiation-induced mammary tumorigenesis studies (45, 46). Importantly, since there is almost no spontaneous tumor (<5% frequency) in control, the APC^Min/+ mouse model demonstrates a high signal-to-noise ratio for proton radiation-induced mammary tumorigenesis. The molecular phenotypic similarities between tumors in APC^Min/+ mice and in human indicate that the proposed mouse model is a valid representation of human breast cancer. Furthermore, the chosen (APC^Min/+) mouse model with its single genetic change that produces an organ-specific, consistent, and discrete tumor phenotype provides unique opportunities to gain insights into the molecular pathogenesis of proton radiation-induced breast cancer development.

Although protons are considered low-LET, the biological effects of protons are considerably different from those of photon radiation (8). Compared to photons, protons show higher damaging effects due to their physical characteristics, energy deposition pattern, and secondary radiation, which are exploited in radiotherapy to inflict damage to tumor cells (47, 48). However, these characteristics, while useful in eliminating cancer cells, are harmful to normal cells: especially the neutrons that are generated due to interaction between proton and beam instrumentation (and the metal surface of spacecraft) and increase the risk of cancer (49, 50). While the current study used entrance plateau region of the proton (energy: 1 GeV) Bragg curve to expose mice, the clinical proton beam therapy uses Spread-Out Bragg Peak (SOBP) with energy ranging from 70 to 250 MeV (51, 52). The LET at the entrance plateau region of both the clinical and the current study protons is lower relative to Bragg peak (7, 8, 53). However, for our study, the entrance plateau region provided a uniform LET and thus reduced variability in experimental outcome. In contrast, proton energy in clinical radiotherapy is varied for SOBP and the peak LET could reach between 8 and 12 keV/μm in the SOBP with higher LET in the later relative to the initial part of the SOBP (54). Although protons in our study are high-energy and low-LET relative to clinical proton beams, reports from others’ and our studies demonstrate that high-energy protons have distinct biological effects compared to γ-rays (8, 9). Since normal tissue exposure in the entrance plateau region of the clinical proton beams also occurs at lower LET ranging from 0.39 keV/μm (energy: 250 MeV) to 1 keV/μm (energy: 70 MeV) relative to SOBP, our data has implications for understanding breast cancer risk from clinical proton radiation (47, 55, 56). Studies with photon radiation show that breast cancer risk increases linearly with increasing dose (5, 50). Our in vivo data in APC^Min/+ mice support the notion that protons have a higher risk for breast cancer relative to photon radiation. Indeed, calculation of tumorigenesis per unit (cGy) of radiation demonstrate that tumorigenesis is higher at the lower dose relative to the higher dose. This could be due to a higher rate of multiple hits to the same cell along with increased apoptosis of the potential carcinogenic cells expected at higher doses relative to lower doses and is consistent with our previous intestinal tumorigenesis data (12). Additionally, the relatively higher non-targeted effects at lower doses, which have been predicted to contribute higher tumorigenic effects at lower doses, could have contributed to higher mammary tumorigenesis at the lower dose of protons (35). We acknowledge that while the lower doses used in the current study have implications for clinical and space radiation exposures, the higher doses have limited applicability for the clinics as well as for the space travel.

While the literature suggests an RBE of 1.1 to 1.3 for protons (57), the RBE for mammary tumorigenesis calculated from our data is about 3 and an even higher RBE for the development of carcinoma, calculated to 5.52 was associated with increased frequency of mammary carcinoma demonstrating that protons have higher breast cancer risk relative to γ-rays, which induced mostly epidermoid cystadenoma. The higher RBE could be due to the high energy protons used in the current study that may have triggered a higher level of secondary particle generation and resultant non-uniform energy deposition (57). It has been proposed that differences in ionization pattern between γ-rays and protons produce distinct DNA damage patterns and unique DNA damage response and repair signaling, which we believe may be promoting differential adenoma/carcinoma frequency after the two different types of radiation exposures (57). Proton radiation-induced increased mammary carcinogenic potential could also be attributed to its differential oxidative, inflammatory, and oncogenic stress responses with higher transformation potential relative to γ-rays (52). However, currently, we do not have evidence to support either of these propositions and further studies will be required to determine the molecular underpinning of our observations.

A key oncogenic stress inducing factor in breast cancer is estrogen (58), which acts predominantly through ERα to activate downstream proliferative pathways (20). More than 50% of breast cancer patients show overexpression of ERα (21) and normal breast tissue as well as benign breast lesions showing increased ERα expression tend to have increased risk of breast cancer development (59, 60). Indeed, immunohistochemical analysis of proton radiation-induced mammary tumors show higher expression of ERα relative to γ-ray-induced and spontaneous tumors. Increased ERα has been implicated in breast cancer progression with ~70% of breast cancer expressing ERα (61). The current study shows that a greater proportion of the proton radiation-induced mammary tumors progressed to carcinomas and we believe ERα-induced upregulation of proliferative pathways could in part be driving the development of more carcinomas. ERα regulates target genes’ transcription either via binding to estrogen response elements (EREs) on promoters or via protein-protein interactions with other transcription factors (62, 63). Interaction between ERα and the transcription factor AP-1 triggers upregulation of cyclin D1, which has a consensus AP-1 binding site on its promoter and is a major downstream oncogenic effector of ERα (62, 64). Increased carcinoma along with increased cyclin D1 expression is suggestive of activation of genomic signaling component of the estrogen/ERα axis after exposure to proton radiation. However, it has been recognized that ERα also engages in protein-protein interaction(s) at the cell membrane to mediate what is known as the non-genomic pro-growth effects of estrogen (62, 63). Interaction of membrane ERα with human epidermal growth factor 2 (HER2), which is overexpressed in ~30% of breast cancer patients and is a major oncogenic receptor tyrosine kinase, is known to activate HER2 and related proliferative pathways leading to breast cancer progression (65, 66). HER2 can also be activated independent of ERα interaction due to ligands binding to heterodimers of HER2 with other members of the EGFR family, and spontaneous activation of HER2 homodimers at high expression levels, leading to upregulation of proliferative pathways such as PI3K/Akt (66). Data presented in the current study not only demonstrate increased expression of HER2, but also upregulation of downstream PI3K and Akt. The protein-protein interaction of ERα at the membrane has also been demonstrated to activate the PI3K/Akt axis directly independent of HER2 (62, 63). We speculate that ERα is acting in tandem with HER2 to promote higher mammary tumorigenesis after proton radiation exposure.

In summary, our quantitative and qualitative data provide insight into the effects of proton radiation quality factors on breast cancer risk that has implications for space exploration and proton radiotherapy. Our data also demonstrate that the APC^Min/+ mouse model with its low spontaneous mammary tumorigenesis, established utility in radiation-induced mammary tumorigenesis, and dysregulated APC in a considerable proportion of human breast cancer patients is of advantage for studying risk proton radiation-induced breast cancer development. Since ERα is actively involved in promoting breast cancer, ERα collaborate with HER2 upregulation, and mammary tumors in our study is ERα and HER2 positive, developing strategies to block the proton radiation-induced estrogenic response could benefit astronauts as well as radiotherapy patients. However, further studies using lower doses of proton radiation of varying energies simulating space environment along with interventions studies using countermeasure agents will be required for reducing the uncertainty in breast cancer risk prediction as well as for developing a comprehensive strategy for risk reduction. To this end, the APC^Min/+ model supported by encouraging data from this study could be a valuable animal system towards molecular mechanism-based understanding of proton radiation-induced breast cancer risks in astronauts and radiotherapy patients.