Unlocking the potential of ultra-high dose fractionated radiation for effective treatment of glioblastoma

Background: Conventional radiation therapy for glioblastoma (GBM) has limited efficacy. Regenerative medicine brings hope for repairing damaged tissue, opening opportunities for elevating the maximum acceptable radiation dose. In this study, we explored the effect of ultra-high dose fractionated radiation on brain injury and tumor responses in immunocompetent mice. We also evaluated the role of the HIF-1α under radiation. Methods: Naïve and hypoxia-inducible factor-1 alpha (HIF-1α)+/− heterozygous mice received a fractionated daily dose of 20 Gy for three or five consecutive days. Magnetic resonance imaging (MRI) and histology were performed to assess brain injury post-radiation. The 2×105 human GBM1 luciferase-expressing cells were transplanted with tolerance induction protocol. Fractionated radiotherapy was performed during the exponential phase of tumor growth. BLI, MRI, and immunohistochemistry staining were performed to evaluate tumor growth dynamics and radiotherapy responses. Additionally, animal lifespan was recorded. Results: Fractionated radiation of 5×20 Gy induced severe brain damage, starting 3 weeks after radiation. All animals from this group died within 12 weeks. In contrast, later onset and less severe brain injury were observed starting 12 weeks after radiation of 3×20 Gy. It resulted in complete GBM eradication and survival of all treated animals. Furthermore, HIF-1α+/− mice exhibited more obvious vascular damage 63 weeks after fractionated radiation of 3×20 Gy. Conclusion: Ultra-high dose fractionated 3×20 Gy radiation can eradicate the GBM cells at the cost of only mild brain injury. The HIF-1α gene is a promising target for ameliorating vascular impairment post-radiation, encouraging the implementation of neurorestorative strategies.


Background
Glioblastoma (GBM) is the most common primary central nervous system malignancy in adults; 1 the three-year survival rate of patients is only 10.3%. 2 This dismal clinical outcome highlights that treating GBM remains one of the greatest challenges in medicine.
Radiotherapy is an essential adjuvant strategy used in conjunction with other treatments. 3However, the negative impact of side effects, particularly on neurological and cognitive function, 4 can signi cantly reduce the quality of life for patients after treatment.Furthermore, these side effects often set the threshold for the maximum tolerable levels of anti-tumor radiation therapies that can be administered.
Several targeted therapeutic techniques have been developed to address these issues, including threedimensional conformal radiation therapy, 5 intensity-modulated radiation therapy, 6 hyper-fractionated radiotherapy, 7 and functional imaging-guided dose-escalated radiation therapy. 8These techniques can selectively destroy the bulk of the tumor.Unfortunately, the early invasive growth of GBM, with satellite cancer cells found within the brain far from the tumor mass, makes conformal radiation therapy prone to a high recurrence rate.As a result, there is a dire need for a breakthrough treatment.
The eld of regenerative medicine has advanced tremendously in recent years, paralleling the progress in oncology.However, these elds have not yet been su ciently integrated, leaving many untapped opportunities.Regenerative medicine techniques now enable the repair and replacement of damaged white matter and vasculature, the tissue components most vulnerable to radiation damage. 9It presents an exciting opportunity for elevating the maximum acceptable dose of radiation.Combining ultra-high dose fractionated radiation therapy with brain repair techniques may be feasible to achieve effective brain tumor eradication.In addition, we sought to explore the mechanism of radiation-induced vascular damage.To do so, we studied hypoxia-inducible factor-1 alpha (HIF-1α) signaling, a well-established system that is a master regulator of endothelial homeostasis and angiogenesis. 10HIF-1α signaling has not been systematically studied in the context of responses to radiation injury, and the modulation of HIF-1α signaling has not been exploited as a therapeutic strategy for brain radiation injury.In this study, we explored the e ciency of ultra-high dose fractionated radiation therapy for human GBM in immunocompetent mice using the described immunological tolerance induction protocol. 11We also investigated radiation-induced brain injury in both wild-type mice and HIF-1α +/− heterozygote mice to better understand the role of HIF-1α signaling in radiation-induced vascular damage.

Animal irradiation
All procedures involving live animals were approved and performed in accordance with the Guide for the Care and Use of Laboratory Animals at the local Animal Care and Use Committee.Ten naïve C57BL/6J mice (8-10 weeks (w), 20-25 g, Jackson Laboratory), as well as 12 transgenic HIF-1α +/− mice (which are heterozygous for a null (knockout) allele at the locus encoding HIF-1α) and 12 wild-type mice of the same litter (6-8 w, both provided by Prof. Gregg L. Semenza), received fractionated daily dose of 20 Gy radiation for consecutive three (Fr 3×20 Gy) or ve (Fr 5×20 Gy) days.Brie y, the mice were anesthetized with iso urane (4% for induction, 1%-2% for maintenance) and placed in the prone position.A custombuilt small animal radiation research platform (SARRP) equipped with onboard computed tomography (CT)-guidance was used for radiotherapy.Following three-dimensional CT acquisition, a target spot (2.5 mm deep relative to the skull) was irradiated with a single beam (60° angle), using a 5 × 5 mm 2 collimator at a daily dose of 20 Gy for three consecutive or ve days.

Magnetic resonance imaging
Magnetic resonance imaging (MRI) was performed under general anesthesia (iso urane) after radiation 1 w, 3 ws, 5 ws, and then monthly until the mice deteriorated or reached the end of our observation time.

Tumor implantation
Fifteen male C57BL/6J mice (6-8 w, 20-25 g, Jackson Laboratory) were used for tumor transplantation.Animals were randomly assigned to three groups: GBM1 without radiation (GBM1, n = 5), GBM1 with a fractionated daily dose of 2 Gy (GBM1 + Fr 3×2 Gy, n = 5) or 20 Gy (GBM1 + Fr 3×20 Gy, n = 5) for consecutive three days.Before transplantation, GBM1 oncospheres were harvested, dissociated into single cells, and suspended in PBS at a nal concentration of 1×10 5 /µL.Then animals were anesthetized with iso urane.A total number of 2×10 5 cells were injected at a rate of 1 µL/min into the right striatum (anteroposterior = 0.5 mm; mediolateral = 2.0 mm; dorsoventral = 2.5 mm) using a 10 µL Hamilton syringe with an attached 31-gauge needle.After injection, the needle was kept in place for 2 minutes (min) to avoid back ow of the injected cells through the needle tract and then withdrawn.
Bioluminescence imaging (BLI) was used to monitor tumor growth rate and was initiated a day after tumor implantation.Fractionated radiotherapy (Fr 3×2 Gy or Fr 3×20 Gy) was performed after tumor inoculation when BLI showed accelerating tumor growth.Then BLI and MRI were performed to assess the tumor response and brain injury.

Bioluminescence imaging
Animals were anesthetized with iso urane and injected i.p. with 150 µl luciferin (30 mg/mL, Gold Biotechnology).Images were acquired 5-15 min after substrate injection at the peak of the bioluminescence signal by IVIS Spectrum In Vivo Imaging System (PerkinElmer).For BLI analysis, images were quanti ed by drawing regions of interest (ROIs).The data were expressed as photon ux (p/sec).
Imaging began on day 1 post-implantation and was then done weekly within the rst month, biweekly within the second month, and monthly after that until the mice deteriorated or reached the end of our observation time.

Behavior Assessment
Open-eld test: Open-eld test was performed using the SDI Open Field System (San Diego Instruments, San Diego, CA).The open eld arena was a 20×40 cm Lexan shoebox cage equipped with a camera to monitor overall locomotor activity.For each open-eld test session, mice were placed in the center of the eld, and activity was monitored for 5 min.In addition, the travel distance and average speed were measured by the investigator blind to the study design.
Y-maze test: Mice were placed at the end of one of the three arms and allowed to explore freely for 5 min.
Spatial recognition memory was assessed.The testing was video recorded, and the number of times the mouse entered three different arms consecutively divided by the total visits were measured by an observer blind to the condition.

Histological analysis
Mice were deeply anesthetized with iso urane and perfused intracardially with 5% sucrose, followed by 4% paraformaldehyde (PFA).Brains were dissected, post-xed in 4% PFA overnight at 4 ℃, then successively cryopreserved in 20% and 30% sucrose until the tissue sank.Brains were cryosectioned into 30-µm-thick coronal sections.Hematoxylin and eosin (H&E) staining was performed.For eriochrome cyanin staining, slides were oven-dried and dehydrated in 95% and 70% ethanol, then put into the eriochrome cyanine solution (0.2% eriochrome cyanine, 0.4% FeCl 3, and 0.5% H 2 SO 4 ).Staining was developed by alternating exposure to 0.1% NH 4 OH for 3-7 s and rinsing in distilled water for 30 s until the blue background was reduced, and the cells turned faintly pink but still had blue shading.After that, sections were put in two changes of 70%, 95%, and 100% ethanol and three changes of xylene for 10 min each.

Statistical analysis
The overall mouse survival time was calculated and expressed by the Kaplan-Meier curve, and the statistical difference was analyzed by log-rank tests.The Shapiro-Wilk test was used to test if the distribution was normal.Those results were presented as mean ± standard deviation (SD).The intergroup differences between more than 2 groups were assessed using ANOVA followed by a post-hoc the least signi cance difference (LSD) test.The comparison of 2 groups was evaluated using an independent ttest.Data without normal distribution were presented as median with interquartile range (IQR).The Mann-Whitney test was used when comparing two groups, and the Kruskal-Wallis test was used when comparing more than 2 groups.All statistical analysis was performed using the SPSS software package (Version 22.0, USA).P < 0.05 was considered to be statistically signi cant.

Ultra-high dose fractionated radiation leads to late-phase brain injury
We studied the effects of ultra-high dose radiation on the brain of immunocompetent mice.The mice were exposed to two radiation doses, 60 Gy divided into 3 fractions and 100 Gy divided into 5 fractions delivered daily.Longitudinal MRI showed that Fr 5×20 Gy radiation-induced multiple microhemorrhages starting from 3 weeks after radiation (Fig. 1A-B) and that was accompanied by blood-brain barrier (BBB) breakdown (Fig. 1C-D).Additionally, we observed white matter damage and brain edema as evidenced by the ipsilateral lateral ventricle compression 8 weeks after Fr 5×20 Gy radiation (Fig. 1E-G).Most of the Fr 5×20 Gy animals showed sudden deterioration, and all mice died within 12 weeks (Fig. 1H).
Fractionated radiation of 3 × 20 Gy also caused brain damage, but the damage was observed much later, after three months.Animals developed vascular abnormalities manifested as microbleeds and BBB breakdown, rst observed 12 weeks after radiation (Fig. 1A-D).At 79 weeks after radiation, we observed compression of the lateral ventricle on the ipsilateral side (Fig. 1E-G).We conducted histological staining in the very late stage (80 weeks) after radiation to further assess brain damage.As shown in Fig. 2, only mild demyelination (Eriochrome, MBP) and vascular abnormalities (Collagen IV) were observed.Minor neuroin ammation, including astrocyte (GFAP) or microglia (Iba1)/macrophage (CD68) activation, was detected at 80 weeks.However, we found no neuronal loss (NeuN) even at that late time.All mice from the Fr 3×20 Gy group survived as long as 80 weeks (the end of our observation time) post-irradiation.Therefore, this dose was selected for further GBM1 eradication studies.The remarkably long window encourages further study to identify targets and opportunities to implement therapeutic intervention addressing the observed damage.

Fractionated 3×20 Gy radiotherapy eradicated GBM in immunocompetent mice
The GBM1 tumor xenograft tolerance model was successfully established in the brain of immunocompetent recipients, and its growth was monitored with BLI and MRI.Tumors grew rapidly, as shown by both imaging modalities (Fig. 3).Radiotherapy (Fr 3×2 Gy (a lower dose used as clinicallyrelevant therapy control) or Fr 3×20 Gy) was performed when BLI showed exponential tumor growth and T2 MRI showed a visible tumor mass around 25 days after tumor inoculation.All the untreated tumorbearing mice died within 9 weeks.A lower dose of Fractionated 3×2 Gy irradiation slowed the tumor growth and slightly prolonged animal survival; however, all mice died within 14 weeks.In animals treated with ultra-high dose fractionated 3×20 Gy radiation, the bioluminescence signal of the tumor gradually disappeared and overlapped with background radiance (Fig. 3A-B).MR images one month later showed regression of cancer with some hypointensity at the tumor site (Fig. 3C).No tumor regrowth was observed even at 46 weeks after radiation, as evidenced by MRI (Fig. 3C).Negative HuNu (a marker of human cell nucleus) and Stem121 (a marker of human cell cytoplasm) staining also con rmed the disappearance of the GBM1 tumor (Fig. 3E-F), and there was a dramatic improvement in survival for up to 322 days (the end of our observation time, Fig. 3D).

HIF-1α de ciency exacerbates local vascular damage after radiation
We further evaluated the role of the HIF-1α gene in radiation-induced brain injury.Fractionated 5×20 Gy radiation-induced severe brain injuries in HIF-1α+/-heterozygote and wild-type mice.In T2*, hypointensity appeared starting at week 3 after irradiation (Supplementary Fig. 1A-B), indicating microhemorrhages.In addition, there was widespread BBB breakdown starting 3 weeks after radiation, as evidenced by Gd enhancement on T1 images (Supplementary Fig. 1C-D).The ipsilateral/contralateral hemisphere and lateral ventricle areas were measured on T2 images, which showed a volume increase at 8 weeks, probably due to hemorrhage and edema, and then reduced at 12 weeks, indicating brain atrophy (Supplementary Fig. 1E-F).However, we did not observe a signi cant difference in the severity of the above brain damage measures between HIF-1α +/− and wild-type mice under this dose treatment.
There was no apparent brain injury before 16 weeks post-fractionated 3×20 Gy irradiation.However, animals developed vascular abnormalities manifested as microbleeds and BBB breakdown at 63 weeks after radiation in both HIF-1α +/− heterozygote and wild-type mice (Fig. 4A-D).In addition, at 63 weeks, we observed compression on the ipsilateral side, indicating microhemorrhages and/or brain edema (Fig. 4E-F).Neither neurobehavioral manifestation (open-eld and Y-maze tests) nor survival time showed the difference between HIF-1α +/− heterozygote mice and wild-type mice under any radiation dose (Supplementary Fig. 2).Yet it is worth noting that, compared with wild-type mice, there was a clear trend towards more microhemorrhages, more extensive BBB breakdown, and more ipsilateral lateral ventricle compression in HIF-1α +/− heterozygote mice (Fig. 4).Furthermore, the HIF-1α+/-mice exhibited higher levels of IgG leakage, which is considered one of the markers of vascular disruption (Fig. 5).Additionally, we observed malformation of blood vessels, as characterized by increased vessel diameter in the brain tissue of HIF-1α +/− mice (Fig. 5).These results indicate that HIF-1α gene de ciency exacerbates vascular damage after radiation.

Discussion
3][14] Unfortunately, patients often face a high rate of recurrence and dismal prognosis.While increasing the radiotherapy dose may improve the success rate, it inevitably leads to brain damage.Actively dividing cells, including oligodendrocytes, vascular endothelium, and various precursor cells, are particularly vulnerable to this damage.Therefore, severe side effects preclude higher doses of radiotherapy. 15This gloomy reality has dominated for decades, but there is hope for a long-awaited breakthrough, thanks to new developments in regenerative medicine.While cell replacement strategies for brain repair have brought mixed results and have yet to prove successful in replacing neurons, replacing glial cells 16,17 or vascular components 18 is highly feasible.It creates an opportunity for innovation in treating GBM by escalating the radiotherapy dose followed by restorative therapy.
Our initial studies have shown that elevating the single dose of radiotherapy to as high as 80 Gy in mice is well tolerated over several weeks (data not shown), which is encouraging as this time window would allow regenerative intervention.However, despite an initial tumor response, the tumor could not be entirely eradicated, and recurrence was observed within two weeks.Therefore, we explored the effect of ultra-high dose fractionated radiation for brain tumor eradication and found that a fractionated 3×20 Gy radiation dose can fully eradicate the tumor and signi cantly extend the survival time of mice while only resulting in mild brain injury over 12 weeks.
A variety of glial cells are extremely vulnerable to irradiation.These include astrocytes and microglia, 19 but, most importantly, the oligodendrocytes and oligodendrocyte progenitor cells. 20Oligodendrocytes are post-mitotic cells that provide a myelin sheath for neurons, and it has been shown that oligodendrocytes undergo continuous turnover and are replaced by local tissue progenitors. 21,22Depletion of these cells has been reported to occur as early as 3 days after irradiation, 23 ultimately resulting in demyelination and white matter loss, which are features typically associated with post-irradiation brain atrophy.Astrocytes are another important glial phenotype to consider.Given their recently elevated functionality, including maintaining potassium and neurotransmitter homeostasis, calcium signaling, and the production of cytokines, as well as playing a critical role in the formation of the BBB, 24 it is warranted to assess the contribution of astrocytes to radiation injury, and consider them as a therapeutic target.
The vascular consequences of radiotherapy have been relatively well characterized and include early and chronic effects in the brain. 25,26Early effects manifest as denudating endothelial cells, parenchymal cell damage, and exhaustion of the cell renewal system.It is followed by abnormal proliferation of the endothelial cells, which are critical components of vascular remodeling, occurring months after the initial exposure.It has been shown that clinically relevant doses of ionizing radiation cause changes in the permeability of the BBB due to the abnormal synthesis of multiple proteins. 27Vascular changes tend to predominate and can range from thrombosis, hemorrhage, and hyalinization to brinoid necrosis, which can further exacerbate the hypoxic/ischemic necrosis in the area.Critical to the feasibility of our study is the slow development of vascular abnormalities, which, while preceding necrosis, still require a long time to evolve. 26adequate repair of damaged endothelial cells and BBB disruption after radiation exposure increase hypoxia in the local microenvironment, leading to upregulation of hypoxia-inducible factors HIF-1α and HIF-2α are master transcription factors for the cellular response to hypoxia. 28Inhibition of HIF-1 activity is believed to have therapeutic bene ts in tumor treatment. 29Oligodendrocyte precursor cells (OPCs) are particularly susceptible to oxidative stress and radicals due to their low levels of anti-oxidants and free radical scavengers. 30,31It has been shown that oligodendrocytes and OPCs, play a central role in cerebral angiogenesis via HIF signaling.Yuen et al. found that constitutive HIF-1/2α stabilization led to OPC maturation arrest, induced excessive postnatal angiogenesis in vivo, and directly stimulated (in a paracrine mechanism) endothelial cell proliferation in vitro.
Conversely, OPC-speci c HIF-1/2α loss of function led to insu cient angiogenesis in the corpus callosum and extensive axonal loss. 32These ndings point to critical interactions between OPCs, angiogenesis, and axonal integrity, with HIF signaling being a master regulator.Dysregulation of this system during radiation therapy may be the focal point in the pathomechanism of radiation-induced brain injury.Indeed, this is highly feasible, as the early elimination of OPCs after radiotherapy removes their pro-angiogenic role and leads to the deterioration of endothelial cells.The exact mechanism could also contribute to the failed recruitment of new pro-angiogenic precursors despite their presence in circulating blood.Our study evaluated the status of HIF signaling after ultra-high dose radiotherapy and showed that HIF-1α de ciency aggravated local vascular damage after radiation.This important result provides initial evidence for developing a strategy to prevent radiation injury via intervening with HIF-1α signaling.
Regenerative strategies offer the potential to alleviate or even eliminate the side effects of anti-tumor treatment. 17An important implication is an opportunity to elevate radiation doses beyond the currently acceptable maximum dose.There is a precedent for this approach in hematology, where treating malignancies involves delivering a high dose of chemo/radiotherapy to the whole body, thereby completely destroying the tumor cells and hematopoietic stem cells; however, these stem cells can be restored via bone marrow transplantation. 33This method is now a standard of care, with spectacular outcomes and, frequently, a complete cure for some malignancies. 34Similarly, repairing or replacing damaged glial and vascular components after radiation injury seems feasible.Highly potent neural progenitor populations, such as glial-restricted progenitors and human neural progenitor cells transduced with GDNF, have been shown to replace endogenous dysfunctional glia in myelin disease models or amyotrophic lateral sclerosis. 35,36Moreover, the feasibility of an oligodendrocyte replacement strategy after low-dose radiotherapy has been demonstrated but without tumor eradication. 17In our study, we take this approach a step further by using a higher radiation dose (60 Gy delivered as three daily fractions of 20 Gy each) to completely eliminate the tumor and investigate the possibility of preventing late-stage white matter and vascular damage through a regenerative strategy.
An additional strength of our study is the use of immunocompetent mice, which better mimic the tumorhost interactions observed in patients.Most human tumors are established in immunode cient recipient mice (nude/rag2/scid) to prevent xenograft rejection, but this approach lacks an adaptive immune system, limiting its relevance for studying anti-tumor treatment.In contrast, our use of C57Bl/6J immunocompetent mice under a tolerance induction regimen allowed us to investigate the effects of aggressive radiotherapy in a more clinically relevant brain tumor model.To accomplish this, we have developed a successful strategy to induce tolerance for human brain tumor xenografts using antibodies blocking co-stimulatory signaling on T-cells. 11This approach not only allowed us to gain insights into the biological characteristics and molecular mechanisms of GBM but also provided a suitable platform for investigating more effective therapeutic strategies, such as allogeneic stem cell transplantation, for this devastating disease.
Overall, we propose a completely new approach for treating brain tumors with the perspective of using regenerative medicine methods.We have demonstrated that supportive cellular components in the brain (glia and vasculature) are radio-sensitive but also have the potential to be replaced, which could open entirely new treatment paradigms for neoplastic conditions in the brain.This approach could be applied to chemotherapy, where targeted intraarterial delivery techniques are now available but cannot be fully exploited due to excessive toxicity.Finally, our ndings shed light on the consequences and therapeutic targets of extremely high doses of radiation, which could be relevant in the context of nuclear accidents or catastrophes.Furthermore, we utilized immunocompetent mice, which provide a more clinically relevant model for investigating brain tumor-host interactions and testing therapeutic strategies.Overall, our study offers new insights into the use of regenerative medicine in treating brain tumors and provides a foundation for future research in this area.

Declarations
Ethics approval and consent to participate All procedures involving live animals were approved and performed in accordance with the Guide for the Care and Use of Laboratory Animals at the local Animal Care and Use Committee.

Figures
Figures