Background: Brain metastasis is becoming increasingly prevalent in breast cancer due to improved extra-cranial disease control. With emerging availability of modern image-guided radiation platforms, mouse models of brain metastases and small animal magnetic resonance imaging (MRI), we examined brain metastases’ responses from radiotherapy in the pre-clinical setting. In this study, we employed half brain irradiation to reduce inter-subject variability in metastases dose-response evaluations. Methods: Half brain irradiation was performed on a micro-CT/RT system in a human breast cancer (MDA-MB-231- BR) brain metastasis mouse model. Radiation induced DNA double stranded breaks in tumors and normal mouse brain tissue were quantified using γ-H2AX immunohistochemistry at 30 min (acute) and 11 days (longitudinal) after half-brain treatment for doses of 8, 16 and 24 Gy. In addition, tumor responses were assessed volumetrically with in-vivo longitudinal MRI and histologically for tumor cell density and nuclear size. Results: In the acute setting, γ-H2AX staining in tumors saturated at higher doses while normal mouse brain tissue continued to increase linearly in the phosphorylation of H2AX. While γ-H2AX fluorescence intensities returned to the background level in the brain 11 days after treatment, the residual γ-H2AX phosphorylation in the radiated tumors remained elevated compared to un-irradiated contralateral tumors. With radiation, MRI-derived relative tumor growth was significantly reduced compared to the un-irradiated side. While there was no difference in MRI tumor volume growth between 16 and 24 Gy, there was a significant reduction in tumor cell density from histology with increasing dose. In the longitudinal study, nuclear size in the residual tumor cells increased significantly as the radiation dose was increased. Conclusions: Radiation damages to the DNAs in the normal brain parenchyma are resolved over time, but remain unrepaired in the treated tumors. Furthermore, there is a radiation dose response in nuclear size of surviving tumor cells. Increase in nuclear size together with unrepaired DNA damage indicated that the surviving tumor cells post radiation had continued to progress in the cell cycle with DNA replication, but failed cytokinesis. Half brain irradiation provides efficient evaluation of dose-response for cancer cell lines, a pre-requisite to perform experiments to understand radio-resistance in brain metastases. Keywords: Breast cancer, Brain metastases, Small animal radiation therapy, Radiation dose-response, Magnetic resonance imaging, DNA double-strand breaks, γ-H2AX * Correspondence: firstname.lastname@example.org Department of Medical Biophysics, University of Western Ontario, London, Ontario, Canada London Regional Cancer Program, University of Western Ontario, London, Ontario, Canada Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Zarghami et al. Radiation Oncology (2018) 13:104 Page 2 of 11 Background radiation using half-brain irradiation to reduce inter- The parallel developments of modern image-guided pre- subject variability. We accomplished this using two animal clinical radiotherapy devices, small animal magnetic res- cohorts. In the first cohort, DNA DSBs within cancer cells onance imaging, and mouse model of brain metastasis and the brain was assessed via immunohistochemistry presents us with a unique opportunity to ask brain staining of γ-H2AX in the acute setting (30 min after half- metastasis-specific radiobiology questions. We, and brain treatment) at three radiation dose levels. Tumor others, have recently employed whole brain irradiation in dose-response over time was evaluated in the second co- mouse models of brain metastasis due to breast cancer to hort using longitudinal MRI (prior to and 11 days after study tumor response after different timing or fraction- half-brain treatment) as well as immunohistochemistry at ation regimens of radiotherapy [1–3]. Despite the use of a the endpoint using two radiation dose levels. MRI was tumor bearing animal model, inter-subject variability used to obtain tumor volumes. In addition to assessing remained the major contributor to experimental uncer- DNA DSB, 4′,6-diamidino-2-phenylindole (DAPI) immu- tainties requiring typically 6-12 animals per longitudinal nohistochemistry staining of the cell nuclei was used to study group each lasting approximately 30 days, making assess tumor cell density and nuclear size. By performing these studies challenging. half brain irradiations in conjunction with MRI and im- Examples of contributors to inter-subject variability in- munohistochemistry in the acute and longitudinal set- clude variations in the number of cells delivered to the tings, we were able to compare responses in the tumors brain from intra-cardiac injection, number of proliferat- versus normal mouse brain tissues, and radiated tumors ing metastases, and their subsequent growth . In versus un-irradiated tumors in the same animal at the addition, post-sacrifice immunohistochemistry (IHC) various dose levels. slide staining results can also vary despite following the same protocol . This led us  and others  to de- Methods velop and validate platforms for specifically half-brain ir- Table 1 provides an overview of the study experiments radiations , allowing us to reduce inter-animal and performed and analyzed. We will describe them in more inter-histological slide variability by using the contralat- details in this section. eral brain as the control. Due to these challenges, tumor radiation dose-response Cell culture is generally not well established in-vivo, and we expect For this study, the brain tropic clone of human triple- that the dose-response would depend on cell lines and negative breast cancer cell line, MDA-MB-231-BR, sta- sublines with specific genes inserted or deleted. In this bly transfected with enhanced green fluorescent protein study, we present our dose-response findings from our (EGFP) was used . Cells were cultured and main- half brain irradiation of the brain metastasis mouse model tained in Dulbecco’s modified Eagle’s medium (DMEM) using a well published human triple negative cell line containing 10% fetal bovine serum and 1% penicillin/ MDA-MB-231-BR. Endpoints include both tumor metas- streptomycin. Cultured cells were kept in 5% CO at tases volumes from longitudinal magnetic resonance im- 37 °C. Trypan blue exclusion assay was done to deter- aging brain imaging and histological endpoints. mine cell viability. Ionizing radiation induced DNA double-strand breaks (DSBs) are known to be lethal lesions that are respon- Animal tumor model sible for cell’s mitotic death . In response to DSBs, a To deliver MDA-MB-231-BR cells into the brain, the histone H2A family member X, H2AX, is rapidly phos- intra-cardiac injection method was used to distribute phorylated to form γ-H2AX . Staining for γ-H2AX cells through arterial circulation. Female nu/nu mice (N are therefore be employed as a measurement of DNA = 19, 6–8 weeks old; Charles River Laboratories) were DSBs . It is known that tumors have higher amounts anesthetized with 1.5 to 2% vaporized inhaled isoflurane of “cryptogenic” γ-H2AX due to endogenous sources in O . A suspension containing 1.5 × 10 MDA-MB-231- such as replication stress, genomic instability, uncapped BR cells in 0.1 ml of Hanks balanced salt solution was telomeres and apoptosis compared to the healthy tissue slowly injected into the left ventricle of the beating heart [12–14]. Previous studies have investigated the residual of the mouse . Animals were housed in ventilated γ-H2AX of murine normal tissues from days to two cages with a 12-h light/dark cycle and controlled months after exposure to detect radiation-induced tox- temperature (20-22 °C), fed normal chow and given icity such as fibrosis and myelopathy [15–17]. To the water ad libitum. Animal’s appearance and behavior was best of our knowledge, tumors’ residual γ-H2AX after scored daily through the experiment and no profound in-vivo irradiation has not been previously reported. effect of pain and distress on behavior was observed. The aim of this study is to measure the radiation dose- This study followed animal care protocols approved by response of a breast cancer brain metastases model to the Animal Use Subcommittee of The University of Zarghami et al. Radiation Oncology (2018) 13:104 Page 3 of 11 Table 1 Summary of experiment: number of animals and MRI-identified irradiated metastases for the acute and longitudinal study Study Group Number of Mice Minimum number of tumors visualized on Dose (Gy) Dissection after radiation therapy MR Irradiated Shielded Acute (dissection after 30 min) A 3 90 90 8 30 minutes B 3 90 90 16 30 minutes C 4 120 120 24 30 minutes Number of tumors tracked on longitudinal MR Irradiated Shielded Longitudinal (dissection 11 days) A 3 68 85 16 11 days B 3 49 60 24 11 days Western Ontario and were consistent with the policies be 16 Gy +/− 0.6 Gy. This dose variation is minimal com- of the Canadian Council on Animal Care. Mice received pared to the dose levels of 8, 16 and 24 Gy. The dose re- half brain radiation 26 days after cell injection. ceived by the un-irradiated side of the brain and tumors were denoted as 0* and will be employed as the control of the irradiated side in the same mouse. After recovery from Mouse half-brain irradiation radiotherapy, mice were selected either for acute or longi- Mice received half brain radiation therapy on the modi- tudinal dose-response study. fied GE eXplore CT 120 (GE Healthcare, Milwaukee, WI) preclinical imaging system [20, 21]. They were anes- In-vivo MRI thetized using 1.5 to 2% vaporized inhaled isoflurane All mice were imaged on a 3 T GE clinical MR scanner and were immobilized using the customized 3D-printed (General Electric, Mississauga, Canada) with a custom- mouse head holder with a targeting accuracy of < 0. built gradient insert coil on day 26 after tumor injection 15 mm . Mice were set-up in a feet first prone position. and before receiving radiation. MRI was performed to ver- The longitudinal fissure (LF) was visually set as the ana- ify the presence of the tumors in the mouse brain, particu- tomical target for the radiation field. Setup lasers and CT larly in both brain hemispheres. Mice that had no images were used to verify the alignment of the animal’s identifiable brain metastases on MR did not proceed to RT head in the head holder. Once the mouse was immobi- and excluded from this study. Images were acquired using lized for treatment, online dorsal-ventral fluoroscopy was 3D balanced steady-state free precession (bSSFP) protocol acquired to identify the rim of the skull and to position (acquisition resolution = 100 × 100 × 200 μm, repetition the collimators. A small CT localization marker was time = 8 ms, echo time = 4 ms, flip angle = 35°, receive placed on the right side of the head holder to help with bandwidth = 19.23 kHz, signal averages = 2, radiofrequency animal orientation on CT and fluoroscopy. The right half phase cycles = 8, scan time = 29 min, along with ZIP2 and of the brain was irradiated with a single field (14 × 20 2 ZIP512 upscaling), a well-established imaging technique mm ) from the dorsal direction. Mice received doses of 8, for this model [25–27]. To evaluate the response of breast 16 or 24 Gy in a single fraction. These dose levels were cancer brain metastases to different radiation doses in- chosen because the biological effective dose (BED, vivo, the longitudinal group was imaged again 11 days after assuming α/β = 10 Gy) of 16 Gy and 24 Gy in a single receiving half brain radiotherapy (37 days after tumor in- fraction are meant to represent doses prescribed for whole jection) with the same imaging protocol. brain radiation therapy (30 Gy in 10 fractions) [22, 23] and stereotactic radiosurgery respectively (18-24 Gy in one fraction) . Figure 1 shows a representative dose MRI analysis distribution in the mouse brain for 16 Gy. The 16 Gy iso- Brain metastases were segmented manually on pre and dose line (magenta color) in Fig. 1 shows homogenous ra- post-radiotherapy images by a single observer using diation dose for the hemisphere away from the field edge open-source OsiriX image software version 6.0. Tumors near the midline of the brain. We have measured the dose in the midline of the brain (±200 μm of the longitudinal drop off to be 7.5% per 5 mm . We prescribed the fissure) were excluded from the study as only part of dose to the midplane of the brain, and expected then the these tumors may have been irradiated. Figure 2a variation to be +/− 3.75%. That is, when we prescribed showed an example of the manual segmentation of the 16 Gy to the midplane, the variation across the brain will tumors performed on an MR acquired on day 11 after Zarghami et al. Radiation Oncology (2018) 13:104 Page 4 of 11 Fig. 1 a Calculated dose distribution on coronal CT plane of the mouse brain for a 16 Gy (magenta isodose line) half brain irradiation. b Whole brain image of γ-H2AX stained section (red), imaged at 10X. DAPI counterstaining of DNA is shown in blue. Stable EGFP labeled tumors are in green. γ-H2AX stain shows the sharp edge of the beam in the middle of the brain along the longitudinal fissure RT. Mean fractional volume changes of the tumors were were harvested and post-fixed in 4% PFA and transferred calculated by dividing the post-treatment tumor volume to 30% sucrose solution until the specimen sank to the by the volume of the same tumor before treatment and bottom. Brain samples were embedded in Tissue-Tek averaged for all brain metastasis for mice in each group. OCT Compound (Sakura, Torrance, CA) and frozen. One mouse in the 24 Gy longitudinal cohort had to be Cyrosectioning of coronal slices was performed with 10- sacrificed at 7 days due to its deteriorating condition. μm slice thickness. Tissue sections were stained with hematoxylin and eosin (H&E) to assess the morphology of Immunohistochemistry the tumors. At the two post-irradiation time-points (30 min or 11 days) Immunostaining was performed with the primary mouse brain samples were collected and processed for im- monoclonal antibody against γ-H2AX using a protocol munohistochemistry staining. Mice were perfused with 0. published by Ford et al. . Staining of sections con- 9% saline followed by 4% paraformaldehyde (PFA). Brains sisted of antigen retrieval with sodium citrate, 1 h Fig. 2 a Manual segmentation of tumors on an MR scan acquired 11 days after RT of an animal treated to 24 Gy to the right brain. Original MR image is on the left panel and segmented MR image is on the right. Tumors segmented in green are in the right (irradiated) half of the brain, and tumors segmented in orange are in the left (shielded) brain. b An example of our segmentation of DAPI-stained tumor nuclei. Original DAPI image of a tumor cluster is shown on the left panel. Segmented tumor nuclei are shown on the right which we employed in our analyses Zarghami et al. Radiation Oncology (2018) 13:104 Page 5 of 11 incubation in blocking serum (10% goat serum with 0. fluorescence intensity for each field of view was nor- 1% Triton X-100 for membrane permeabilization), over- malized to the total area of segmented nuclei for the night incubation at 4 °C in mouse anti-γ-H2AX antibody same field (Eq. 1). (anti-phospho-histone H2AX, Ser139, clone JBW301; Total γ−H2AX intensity in segmented nuclei γ−H2AX intensity density¼ Millipore, Billerica, MA, USA) at the dilution of 1:700, Total area of segmented nuclei 1 h incubation in secondary antibody (1:500 goat anti- ð1Þ mouse Alexa Fluor 594 conjugated, Life Technologies, Carlsbad, CA, USA.), DAPI counterstain 5 min, and Mean γ-H2AX intensity per unit area was determined mount with anti-fade mounting medium Vectashield for each treatment condition in the acute and longitu- (Vector Laboratories, Inc. Burlington, ON). This proto- dinal settings. The total number of nuclei analyzed for col was used consistently to stain sections from the two each dose level varied from 350 to 950. time-points. For quantification, images were acquired We observed that MDA-MB-231-BR tumors grew in with 100X (oil immersion) objective lens on a fluores- clusters surrounded by edema. We obtained the number cence microscope (Carl Zeiss Canada Ltd). Imaging pa- of tumor nuclei per cluster area. This index gave us the rameters such as intensity, exposure time and gain were density of tumor nuclei/cells in each cluster (Eq. 2). kept consistent during the experiment. We collected a Number of tumor nuclei in cluster total of ten to thirteen images of different tumors for Tumor cell density¼ ð2Þ Area of segmented cluster each mouse. We quantified both the tumor cell density and size of Histological quantification tumor nucleus for all radiation doses at the two time- To evaluate the DNA damage response, γ-H2AX points. Figure 3 shows the flow chart of the processes in- stained sections of tumors were analyzed for each volved in these histological quantifications. IHC staining radiation dose level. The amount of damage was also was repeated three times for the acute study and twice quantified in neighboring normal brain tissues under for the longitudinal study. the same conditions as the tumors. Initially, we We also observed an increase in tumor nuclei size and employed an inverted confocal microscope (Olympus we quantified the size of tumor nuclei by computing the Fluoview FV1000 Confocal Imaging System) for high average area of each nucleus from DAPI images (Eq. 3). resolution 3D images of γ-H2AX foci within the nu- clei . We observed in the acute setting γ-H2AX Total area of segmented nuclei Average area of tumor nucleus¼ foci were over-lapping, which made detection of in- Number of segmented nuclei dividual foci impossible. Similarly, foci saturation ð3Þ was observed in the irradiated tumors in the longitu- dinal experiment. Unable to count individual foci, we quantified γ-H2AX based on the fluorescent stain in- Statistics tensity, which is a more reliable method for high ra- Statistical analyses were performed using SPSS (Armonk, diation doses [30, 31]. NY: IBM Corp) and confirmed by GraphPad Prism soft- All IHC analyses were performed on images taken ware (La Jolla, CA, USA). The normality of the measured from the fluorescence microscope using 100X oil variables was tested using the Shapiro-Wilk test and the p immersion objective. The γ-H2AX intensity was mea- < 0.05 was used as the significance threshold. For normally sured for both normal mouse brain and tumor tis- distributed variables, between-groups analysis of variance sues. Tumor nuclei were visually distinguished from (ANOVA) followed by Tukey post-hoc test was conducted mouse nuclei based on the characteristic punctuate to determine whether the response was statistically signifi- pattern of mouse DAPI staining . To quantify γ- cant (p < 0.05). Nonparametric Kruskal-Wallis analysis H2AX intensity, DAPI-stained nuclei were used to followed by Mann-Whitney U test was used for variables generate nuclear outlines in which the γ-H2AX inten- that were not normally distributed. sity would be measured. Nuclear segmentations were used to eliminate signal from background fluores- Results cence. Nuclei on DAPI images were manually seg- γ-H2AX radiation dose-response mented using Adobe Photoshop CC. For each field of In the acute radiation dose-response study, mice re- view, total γ-H2AX fluorescence intensity was ob- ceived half brain radiation of 8, 16 and 24 Gy (minimum tained by summing the intensity values of all pixels N = 3 per dose) and were sacrificed approximately within the segmented boundary using an in-house 30 min after treatment. Tissue sections were stained for code developed and validated in MATLAB (Math- γ-H2AX to quantify the initial damage induced in both Works, Natick, MA, USA). The total γ-H2AX normal mouse brain and tumors. Figure 1b displays a Zarghami et al. Radiation Oncology (2018) 13:104 Page 6 of 11 Fig. 3 Flow chart of the processes involved in the quantification of γ-H2AX intensity, tumor nucleus size and tumor cell density. DAPI and γ-H2AX images were overlaid and nuclei were segmented based on DAPI. The intensity of γ-H2AX from segmented nuclei was acquired. From the seg- mented DAPI images, number and total area of segmented nuclei were quantified. For tumor cell density analysis, tumor clusters were seg- mented based on DAPI and the area of the cluster was computed mouse whole brain coronal section, which received half In-vivo dose-response brain radiation of 16 Gy. To assess the changes in the volume of tumors in re- Figure 4a shows the tissue sections of tumors and nor- sponse to radiation doses in-vivo, MR images were taken mal mouse brain stained with DAPI and γ-H2AX at the before and 11 days after half brain radiation therapy. acute time point. Figure 4b shows our quantification of Representative images of brain metastases at two dif- γ-H2AX based on fluorescence intensity density in the ferent time-points for doses of 16 and 24 Gy are nuclei of normal brain and tumor tissues evaluated at shown (Fig. 5a). The mean fractional growth of the the acute time point. In normal brain, the amount of tumors was calculated for each group (Fig. 5b). There γ-H2AX intensity density increased linearly (R = 0.78, was a statistically significant difference (Mann-Whit- p < 0.001) with increasing radiation dose. However, in ney U p ≤ 0.05) between the growth of un-irradiated tumors, this trend stopped at 16 Gy; the level of γ-H2AX and irradiated brain metastases for both doses of 16 intensity density dropped at the dose of 24 Gy compared and 24 Gy. A second observer segmented tumors on to 16 Gy. The γ-H2AX intensity density in both tumors MRI on two animals treated at 24 Gy and confirmed and normal brain of the irradiated side were significantly this finding. The fractional reduction in tumor vol- increased (p < 0.0001) compared to the respective un- ume growth as assessed by MRI was not statistically irradiated side (8 versus 0*(8), 16 versus 0*(16) and 24 ver- different between 16 and 24 Gy in the longitudinal sus 0*(24) Gy). setting. Tumor Cell Density. To investigate how much of the initial damage is retained We observed on H&E samples from the longitudinal in both tumors and normal brain tissues, γ-H2AX intensity cohort that irradiated tumors are less compacted with density was measured for the longitudinal group 11 days cells, and surrounded by a more substantial amount of after hemi brain radiation (Figs. 4c, d). We observed edema compared to tumors on the un-irradiated side that γ-H2AX intensity density in irradiated normal (Fig. 6a). We quantified this by calculating tumor cell brain nuclei returned to background levels when density based on DAPI staining for tumors in both the compared to un-irradiated side of the brain 11 days acute and longitudinal settings. The acute setting was after radiotherapy. However, irradiated tumors had employed to provide a baseline verification. As expected, higher levels of γ-H2AX intensity density compared to tu- no significant difference was detected in the density be- mors in the contralateral un-irradiated sides (0*(16) and tween treated and un-treated tumors and for different 0*(24) Gy). There was no significant difference in the radiation doses 30 min after radiation. amount of residual γ-H2AX between irradiated tumors On the other hand, there was a significant difference (16 Gy vs. 24 Gy). in tumor cell density between treated and un-treated Zarghami et al. Radiation Oncology (2018) 13:104 Page 7 of 11 Fig. 4 a Acute DNA damage response 30 min post-irradiation. Histology sections of fluorescent γ-H2AX and corresponding DAPI (nuclei) stained for tumor (MDA-MB-231-BR) and normal brain are shown. Images were taken with a fluorescence microscope (100X objective). Scale bar = 50 μm. b Quantification of the intensity of γ-H2AX staining versus radiation dose 30 min after radiotherapy. Tumors are plotted in green and normal brain tissue are plotted in blue. In irradiated normal brain tissue, the γ-H2AX intensity had a linear trend (R = 0.78, p < 0.001). In tumors, γ-H2AX did not continue to increase at the dose of 24 Gy even though the γ-H2AX intensity is significantly different between irradiated and un-irradiated sides (p < 0.0001). Error bar indicates standard error of the mean. c Residual DNA damage response 11 days post-irradiation. Scale bar = 50 μm. (d) Quantification of the intensity of γ-H2AX staining for the various radiation dose 11 days after radiotherapy. In normal brain, γ-H2AX intensities returned to the background level. In irradiated tumors, γ-H2AX intensity was higher than both the background level and tumors in the irradiated side. ** = p ≤ 0.01, *** = p ≤ 0.001, and error bar indicates standard error of the mean tumors in the longitudinal experiment (Fig. 6b). Further- second observer repeated this DAPI nuclei segmentation more, there was a significantly lower density in those on tumors that were treated at 24 Gy and their contralat- treated with 24 Gy compared to 16 Gy. eral control and confirmed the manual segmentation re- sults. However, in the longitudinal cohort, there was a Tumor cell nuclear size significant difference in the size of the nuclei between DAPI is used as a counterstain for the nucleus of the cell treated and un-treated sides of the same mice. Radiation and we used this stain to investigate the size of tumor dose at 24 Gy resulted in a significantly larger nuclei size nuclei for both acute and longitudinal studies. We ob- than 16 Gy in the longitudinal setting (Fig. 6d). served that the nuclei of treated tumors were signifi- cantly larger than the un-treated nuclei 11 days after Discussion radiotherapy. Figure 6c shows the different morpho- In this study, we used both in-vivo and ex-vivo methods logical appearances of irradiated versus un-irradiated to evaluate the response of MDA-MB-231-BR brain me- tumor nuclei stained with DAPI. The size of tumor nuclei tastases and normal brain to different radiation doses at was quantified for both acute and longitudinal studies. two time-points after treatment. In the longitudinal The acute setting quantification was employed to establish study, the normal brain’s response contrasted with the a baseline and no significant differences was found in the tumors’ after delivering 16 or 24 Gy half brain irradi- average size of tumor nuclei 30 min after treatment. A ation: γ-H2AX levels returned to normal in brain nuclei Zarghami et al. Radiation Oncology (2018) 13:104 Page 8 of 11 Fig. 5 a MR images (bSSFP) of the mouse brain at two-time points. Metastases appear as hyper-intense (bright) regions compared to brain parenchyma. Pre-treatment images are on day 26 and images on day 37 are for the same mouse 11 days after radiation therapy. Right half of the brain was irradiated. One mouse per radiation group is shown. Red arrows indicate the brain metastases in the irradiated side while green arrows show brain metastases in the un-irradiated side. b Mean fractional growth of brain metastases measured on MR images for the radiation doses normalized to that of the un-irradiated halves. Tumors irradiated with 16 and 24 Gy grew with significantly different growth rates than their respective un-irradiated sides (Kruskal-Wallis followed by Mann-Whitney U test). No difference was observed between irradiated tumors of 16 and 24 Gy. *** = p ≤ 0.001, error bar indicates standard error of the mean 11 days after radiation, while tumors retained signifi- multinucleated cells [37, 38]. Cancer cells are known to cantly higher density of phosphorylated γ-H2AX com- exhibit aneuploidy, and here, we showed radiation further pared to un-irradiated tumors. This higher amount of exacerbate this problem in cells that survived radiation in phosphorylated γ-H2AX is independent of the increase a dose-dependent manner. in the size of the tumor nuclei that we also observed be- Finally, we evaluated the response of treated and cause we have quantified γ-H2AX intensity per unit nu- un-treated breast cancer brain metastases with MRI. clei area. It has been shown that tumors that retain the In the bSSFP sequence, MDA-MB-231-BR brain me- induced γ-H2AX in the first 24 h after radiotherapy are tastases appear as hyperintense regions compared to more likely to die . This is supported by our imaging normal mousebrain duetotumor-associatededema finding that tumors in the half brain treated with radio- [1, 39, 40]. We found that treated tumors grew sig- therapy had significantly slower growth than tumors in nificantly less over 11 days compared to control, but the untreated side. Higher cryptogenic level of γ-H2AX not in a dose dependent manner. In contrast, hist- in tumor cells  is attributed to dysfunctional telo- ology sections of these tumors showed tumor cell meres that drives genomic instability . Sustained ele- density decreased with increasing radiation dose. It vation of γ-H2AX here could be predictive of an is expected that higher doses will lead to increased unstable genome, and may allow the acquisition of more cell kill, but edema must set in to achieve a lower aggressive characteristics  if the higher level of re- tumor cell density. One interpretation is that there sidual DSBs do not keep these cells from going through exists a dose-response relationship of radiation in- mitosis. Smart et al.  have successfully retrieved the duced edema, particularly in this cell line, and such surviving tumor cells after radiotherapy using the same edema masked the tumor volume response as animal model, and showed that they are more radiosen- assessed by bSSFP MRI. Diffusion MRI has the abil- sitive than before. Our results are consistent with this ity to detect such changes in tumor cell density and finding as we showed that remaining tumor cells after should be employed for future studies. radiation has a higher sustained level of DNA damage This study was limited by the exponential tumor with an elevated γ-H2AX. growth in the MDA-MB-231-BR model which left a We found that the tumor nuclear size increased at 16 short interval (maximum of about 11 days) between and 24 Gy compared to contralateral controls (Fig. 6c, d). MRI-visible metastasis and the need to sacrifice. This This suggests that while DNA replication had continued, left us with a limited opportunity to observe longer cells failed to undergo cytokinesis. When cell division is term changes in gross tumor volume beyond what we not possible, this leads to aneuploidy, polyploidy , or have reported. Moreover, while half brain irradiation Zarghami et al. Radiation Oncology (2018) 13:104 Page 9 of 11 Fig. 6 a H&E stained sections of shielded and irradiated tumors from the same section of a mouse brain 11 days after radiotherapy at 16 and 24 Gy (10X magnification). Scale bar = 1 mm. b Quantification of tumor cell density 11 days after radiotherapy. The densities of tumor cells treated with 16 and 24 Gy were significantly lower than their corresponding un-treated side. There was also a significant difference between treated tumors at 16 and 24 Gy. c DAPI staining of shielded and irradiated tumor nuclei from the same section of a mouse brain 11 days after radiotherapy at 16 and 24 Gy. Scale bar = 50 μm. d Average size of tumor nuclei 11 days after radiotherapy normalized by that of the respective un-irradiated halves. There was a significant difference between the sizes of tumor nuclei treated with 16 and 24 Gy compared to the contralateral side. The size of tumor nuclei was also significantly different between 16 and 24 Gy. ** = p ≤ 0.01, *** = p ≤ 0.001, **** = p ≤ 0.0001, error bar indicates standard error of mean allowed us to reduce inter-animal and inter-slide stain- become saturated at the higher dose levels. In the longi- ing variability, this technique can potentially introduce tudinal setting 11 days after treatment, we showed that radiation-induced bystander effect . We assumed in the response of irradiated tumors (at both 16 and 24 Gy) this work that the bystander effect is small in this brain differed from un-irradiated counterparts in γ-H2AX metastasis model due to the use of nude mice that fluorescence intensity, MRI-assessed tumor growth, lacked adapative immune T-cells. tumor cell density, tumor cell nuclear size, and the frac- tion of tumor cell proliferation. Decreased tumor cell Conclusions density and increased nuclear size were seen when we Brain metastasis is a growing problem in breast cancer increased the dose from 16 to 24 Gy, but not in γ- patients and new treatment strategies for brain metasta- H2AX intensities or MRI tumor volume. We con- sis are necessary. Radiotherapy is an established treat- clude that surviving MDA-MB-231-BR cells in the ir- ment that is currently used to treat the majority of brain radiated tumors must have continued DNA metastasis patients. Understanding the properties of can- replication but failed cyctokinesis in a dose-dependent cer cells surviving radiotherapy can provide evidence for manner, leading to increased nuclear size. Further- further improvements (e.g. molecularly targeted adjuvant more, lower tumor cell density implied the presence therapies) and optimization in the clinics. As a first step of radiation induced edema for this cell line. Add- toward this goal, we evaluated the radiation dose- itional pre-clinical research is warranted to further response of MDA-MB-231-BR breast cancer brain me- understand these responses, their generalizability, and tastases in the present study. We found in the acute set- ultimately to capitalize on such information to im- ting that γ-H2AX in tumors, unlike normal tissues, prove brain metastasis radiotherapy. Zarghami et al. Radiation Oncology (2018) 13:104 Page 10 of 11 Abbreviations 4. Perera M, Ribot EJ, Percy DB, Mcfadden C, Simedrea C, Palmieri D, et al. In ANOVA: Analysis of variance; BED: Biological effective dose; bSSFP: Balanced vivo magnetic resonance imaging for investigating the development and steady-state free precession; DAPI: 4′,6-diamidino-2-phenylindole fluorescent distribution of experimental brain metastases due to breast Cancer. Transl nuclear stain; DMEM: Dulbecco’s modified Eagle’s medium; DSB: Double Oncol [Internet]. 2012;5:217–25. Available from: https://www.ncbi.nlm.nih. stranded break; EGFP: Enhanced green fluorescent protein; H&E: Hematoxylin gov/pmc/articles/PMC3384276/. and eosin; H2AX: H2A histone family, member X; IHC: Immunohistochemistry; 5. Brown CM. Fluorescence microscopy - avoiding the pitfalls. J Cell Sci [Internet]. LF: Longitudinal fissure; MRI: Magnetic Resonance Imaging; 2007;120:1703–5. Available from: https://doi.org/10.1242/jcs.03433%5Cnpapers2:// PFA: Paraformaldehyde publication/uuid/759DE9EB-603A-4B32-B29C-24BD6059A59C 6. Zarghami N, Jensen MD, Talluri S, Foster PJ, Chambers AF, Dick FA, et al. Acknowledgments Technical note: Immunohistochemical evaluation of mouse brain irradiation We would like to thank Carmen Simedrea for assistance in cardiac injection targeting accuracy with 3D-printed immobilization device. Med Phys and Yuhua Chen for the animal perfusion. We would like to thank Suzanne [Internet]. 2015;42:6507–13. Available from: http://scitation.aip.org/content/ Wong and Samveg Shah for assistance in performing confirmation tumor aapm/journal/medphys/42/11/10.1118/1.4933200 segmentations on the MR and histology images. 7. Grams M, Wilson Z, Sio T, Beltran C, Tryggestad E, Gupta S, et al. Design and characterization of an economical 192Ir hemi-brain small animal irradiator. Int J Radiat Biol. 2014;90:936–42. Funding 8. Ford E, Deye J. Current instrumentation and technologies in modern This work is funded by a Natural Sciences and Engineering Research Council radiobiology research – opportunities and challenges. Semin. Radiat. Oncol. of Canada Grant, and a London Regional Cancer Program Catalyst Grant. NZ [internet]. Elsevier. 2016;26:349–55. Available from: http://linkinghub.elsevier. was supported by a Translational Breast Cancer Studentship award funded in com/retrieve/pii/S1053429616300170 part by the Breast Cancer Society of Canada. 9. Sedelnikova OA, Pilch DR, Redon C, Bonner WM. Histone H2AX in DNA damage and repair. Cancer Biol Ther. 2003;2(3):233-5. Review. PubMed Availability of data and materials PMID: 12878854. Data can be made available upon acceptance. 10. Rogakou EP, Pilch DR, Orr AH, Ivanova VS, Bonner WM. DNA double- stranded breaks induce histone H2AX phosphorylation on serine 139. J Biol Authors’ contributions Chem [Internet]. 1998 [cited 2014 Jul 16];273:5858–5868. Available from: AFC (brain metastasis mouse model), PJF (small animal MR imaging), and EW http://www.jbc.org/content/273/10/5858.long. (small animal image guided irradiation) conceived the idea and planned the 11. Rothkamm K, Horn S. γ-H2AX as protein biomarker for radiation exposure. experiment. NZ, DHM, MJ acquired and analyzed the data. All histology Ann Ist Super Sanita. 2009;45:265–71. staining was performed by NZ in the lab of FAD and reviewed by FAD. NZ 12. Bonner WM, Redon CE, Dickey JS, Nakamura AJ, Olga A, Solier S, et al. drafted the manuscript and all authors have reviewed and edited the γH2AX and cancer. Cancer. 2011;8:957–67. manuscript. 13. Olive PL. Endogenous DNA breaks: gammaH2AX and the role of telomeres. Aging (Albany NY). 2009;1(2):154-6. PubMed PMID: 20157507; PubMed Ethics approval and consent to participate Central PMCID: PMC2806006. This study followed animal care protocols approved by the Animal Use 14. Sedelnikova OA, Bonner WM. GammaH2AX in cancer cells: a potential Subcommittee of The University of Western Ontario and were consistent biomarker for cancer diagnostics, prediction and recurrence. Cell Cycle. with the policies of the Canadian Council on Animal Care.. 2006;5(24):2909-13. Epub 2006 Dec 15. PubMed PMID: 17172873. 15. Kunwar A, Haston CK. DNA damage at respiratory distress, but not acute Competing interests time-points, correlates with tissue fibrosis following thoracic radiation The authors declare that they have no competing interests. exposure in mice. Int J Radiat Biol [Internet]. 2015;91:360–7. Available from: https://www.tandfonline.com/doi/full/10.3109/09553002.2015.997897 16. Bhogal N, Kaspler P, Jalali F, Hyrien O, Chen R, Hill RP, et al. Late residual Publisher’sNote gamma-H2AX foci in murine skin are dose responsive and predict Springer Nature remains neutral with regard to jurisdictional claims in radiosensitivity in vivo. Radiat Res. 2010;173:1–9. published maps and institutional affiliations. 17. Andratschke N, Blau T, Schill S, Nieder C. Late residual γ-H2AX foci in murine spinal cord might facilitate development of response-modifying strategies: a Author details research hypothesis. Anticancer Res. 2011;31:561–4. Department of Medical Biophysics, University of Western Ontario, London, 18. Yoneda T, Williams PJ, Hiraga T, Niewolna M, Nishimura RA. Bone-seeking Ontario, Canada. Department of Biochemistry, University of Western Ontario, clone exhibits different biological properties from the MDA-MB-231 parental London, Ontario, Canada. London Regional Cancer Program, University of human breast cancer cells and a brain-seeking clone in vivo and in vitro. Western Ontario, London, Ontario, Canada. Imaging Research Laboratories, J Bone Miner Res. 2001;16:1486–95. Robarts Research Institute, London, Ontario, Canada. Department of 19. Conley F. Development of a metastatic brain tumor model in mice. Cancer Oncology, University of Western Ontario, London, Ontario, Canada. Res [Internet]. 1979;39:1001–7. Available from: http://cancerres.aacrjournals. Department of Physics and Astronomy, University of Western Ontario, org/content/39/3/1001.short London, Ontario, Canada. 20. Jensen MD, Hrinivich WT, Jung J a, Holdsworth DW, Drangova M, Chen J, et al. Received: 27 December 2017 Accepted: 13 April 2018 Implementation and commissioning of an integrated micro-CT∕RT system with computerized independent jaw collimation. Med. Phys. [Internet]. 2013; 40(8):081706. Available from: https://doi.org/10.1118/1.4812422. 21. Thind K, Jensen MD, Hegarty E, Chen AP, Lim H, Martinez-Santiesteban F, References et al. Mapping metabolic changes associated with early radiation induced 1. Murrell DH, Zarghami N, Jensen MD, Chambers AF, Wong E, Foster PJ. lung injury post conformal radiotherapy using hyperpolarized 13C-pyruvate Evaluating changes to blood-brain barrier integrity in brain metastasis over magnetic resonance spectroscopic imaging. Radiother Oncol [Internet] time and after radiation treatment. Transl. Oncol. [Internet]. The Authors. Elsevier Ireland Ltd. 2014;110:317–22. Available from: https://doi.org/10.1016/ 2016;9:219–27. Available from: https://doi.org/10.1016/j.tranon.2016.04.006 j.radonc.2013.11.016 2. Murrell DH, Zarghami N, Jensen MD, Dickson F, Chambers AF, Wong E, et al. 22. Perez CA, Brady LW. In: Halperin Edward C, Wazer DE, editors. Perez and MRI surveillance of cancer cell fate in a brain metastasis model after early Brady’s principles and practice of radiation oncology. 6th ed. Philadelphia: radiotherapy. Magn Reson Med [Internet]. 2016;0:1–7. Available from: http:// Wolters Kluwer Health/Lippincott Williams & Wilkins; 2013. doi.wiley.com/10.1002/mrm.26541 3. Smart D, Garcia-Glaessner A, Palmieri D, Wong-Goodrich SJ, Kramp T, Gril B, 23. McTyre E, Scott J, Chinnaiyan P. Whole brain radiotherapy for brain et al. Analysis of radiation therapy in a model of triple-negative breast cancer metastasis. [Internet] Surg Neurol Int. 2013:S236–44. Available from: http:// brain metastasis. Clin Exp Metastasis [Internet] Springer Netherlands. 2015;32: surgicalneurologyint.com/surgicalint-articles/whole-brain-radiotherapy-for- 717–27. Available from: http://link.springer.com/10.1007/s10585-015-9739-9 brain-metastasis/. Zarghami et al. Radiation Oncology (2018) 13:104 Page 11 of 11 24. Ranjan T, Abrey LE. Current management of metastatic brain disease. Neurotherapeutics. 2009;6:598–603. 25. Perera M, Ribot EJ, Percy DB, McFadden C, Simedrea C, Palmieri D, et al. In vivo magnetic resonance imaging for investigating the development and distribution of experimental brain metastases due to breast Cancer. Transl Oncol [Internet]. 2012 [cited 2014 Jul 22];5:217–225. Available from: http:// linkinghub.elsevier.com/retrieve/pii/S1936523312800106 26. Percy DB, Ribot EJ, Chen Y, McFadden C, Simedrea C, Steeg PS, et al. In vivo characterization of changing blood-tumor barrier permeability in a mouse model of breast Cancer metastasis. Investig Radiol. 2011;46:718–25. 27. Scheffler K, Lehnhardt S. Principles and applications of balanced SSFP techniques. Eur Radiol. 2003;13:2409–18. 28. Ford EC, Achantac P, Purgerc D, Armoura M, Reyesa JJ, Fonga L, et al. Localized CT-guided irradiation inhibits neurogenesis in specific regions of the adult mouse brain. Radiat Res. 2011;175:774–83. 29. Ivashkevich A, Redon CE, Nakamura AJ, Martin RF, Martin OA. Use of the γ- H2AX assay to monitor DNA damage and repair in translational cancer research. Cancer Lett. 2013;487:109–13. 30. Anderson D, Andrais B, Mirzayans R, Siegbahn E a, Fallone BG, Warkentin B. Comparison of two methods for measuring γ-H2AX nuclear fluorescence as a marker of DNA damage in cultured human cells: applications for microbeam radiation therapy. J Instrum [Internet]. 2013;8:6008–6016. Available from: http://iopscience.iop.org/1748-0221/8/06/C06008 31. Hernández L, Terradas M, Martín M, Tusell L, Genescà A. Highly sensitive automated method for DNA damage assessment: gamma-H2AX foci counting and cell cycle sorting. Int J Mol Sci. 2013;14:15810–26. 32. Pereira CF, Terranova R, Ryan NK, Santos J, Morris KJ, Cui W, et al. Heterokaryon-based reprogramming of human B lymphocytes for pluripotency requires Oct4 but not Sox2. PLoS Genet. 2008;4:1–14. 33. Banáth JP, Klokov D, MacPhail SH, Banuelos CA, Olive PL. Residual gammaH2AX foci as an indication of lethal DNA lesions. BMC Cancer. 2010;10:1–12. 34. Nakamura AJ, Redon CE, Bonner WM, Sedelnikova OA. Telomere-dependent and telomere-independent origins of endogenous DNA damage in tumor cells. Aging (Albany NY). 2009;1:212–8. 35. Yu T, MacPhail SH, Banáth JP, Klokov D, Olive PL. Endogenous expression of phosphorylated histone H2AX in tumors in relation to DNA double-strand breaks and genomic instability. DNA Repair (Amst). 2006;5:935–46. 36. Coward J, Harding A. Size does matter: why Polyploid tumor cells are critical drug targets in the war on Cancer. Front Oncol [Internet]. 2014;4:123. Available from: https://www.frontiersin.org/articles/10.3389/fonc.2014.00123/full 37. Sato N, Mizumoto K, Nakamura M, Ueno H, Minamishima Y a, Farber JL, et al. A possible role for centrosome overduplication in radiation-induced cell death. Oncogene [Internet]. 2000;19:5281–90. Available from: http:// www.nature.com/doifinder/10.1038/sj.onc.1203902 38. Zhang XR, Liu YA, Sun F, Li H, Lei SW, Wang JF. p21 is responsible for ionizing radiation-induced bypass of mitosis. Biomed EnvironSci. 2016;29: 484–93. Available from: http://linkinghub.elsevier.com/retrieve/pii/ S193652331500025X. 39. Heyn C, Bowen CV, Rutt BK, Foster PJ. Detection threshold of single SPIO- labeled cells with FIESTA. Magn Reson Med. 2005;53:312–20. 40. Murrell DH, Hamilton AM, Mallett CL, van Gorkum R, Chambers AF, Foster PJ. Understanding heterogeneity and permeability of brain metastases in murine models of HER2-positive breast Cancer through magnetic resonance imaging: implications for detection and therapy. Transl Oncol [Internet] The Authors. 2015;8:176–84. Available from: https://www.transonc.com/article/ S1936-5233(15)00025-X/fulltext 41. Azzam EI, de Toledo SM, Little JB. Stress signaling from irradiated to non- irradiated cells. Curr Cancer Drug Targets. 2004;4:53–64.
– Springer Journals
Published: Jun 1, 2018