Impact of ultra-high dose rate (FLASH) versus conventional radiotherapy on tumor control in wild-type and cGAS-knockout mice

Background: Radiotherapy (RT) is central to breast cancer (BC) management but limited by normal tissue toxicity. Conventional RT (CONV; ≤0.03 Gy/s) controls tumors but often causes skin inflammation, compromises treatment intensity and quality of life. Ultra-high dose-rate FLASH RT (>40 Gy/s) achieves comparable tumor control with markedly reduced tissue injury (“FLASH effect”), yet its mechanism of tumor control and sparing of normal tissue remains unclear. Ionizing radiation activates the cGAS-STING pathway through DNA damage, triggering proinflammatory cytokine production and tissue injury, yet has also been implicated with improvement in the antitumor immune response. Emerging data suggest that FLASH may attenuate or abrogate cGAS-STING signaling leading to reduced inflammation and tissue injury. This study investigated the role of cGAS signaling in mediating the differential effects of FLASH and CONV RT on normal tissue toxicity and tumor control using wild-type (WT) C57BL/6 and cGAS double knockout (cGAS -/- KO) mice. Methods: PYMT117 BC cells were orthotopically implanted into the third mammary fat pad of 6-8-week-old female WT and cGAS-/- mice. Once tumors reached ~50 mm³, mice received a single 30 Gy dose of either FLASH or CONV RT targeted to the tumor site. Tumor growth and skin toxicity (graded 0-5) were measured every other day. Euthanasia was performed upon excessive tumor burden or skin injury. Data were analyzed by two-way ANOVA followed by Tukey's post hoc test ( p < 0.05). Results: Both FLASH and CONV RT significantly reduced tumor volume ( p < 0.001), with complete regression by day 14. Tumor recurrence occurred around day 25 in all groups. In WT mice, both modalities produced comparable tumor control, with CONV showing a slight but non-significant trend toward smaller recurrent tumors and few complete responses but caused severe skin toxicity (score 5) requiring euthanasia by day 50-60. In contrast, FLASH treated mice showed equivalent tumor suppression markedly reducing skin toxicity (score ≤ 3 in 2/10 mice). Notably, the cGAS-/- mice exhibited minimal or no visible toxicity with either modality, except one CONV-treated mouse (score 2). Conclusion: FLASH RT significantly reduces normal tissue toxicity compared to CONV RT while maintaining equivalent tumor control: improved tumor control might be achievable through higher FLASH doses because of the improved therapeutic index over CONV RT. The minimal toxicity observed following FLASH RT, together with the absence of toxicity in cGAS-/- mice, also supports the possibility that FLASH may limit early inflammatory responses and tissue injury by suppression or abrogation of cGAS-STING activation. These findings support FLASH RT as a promising, less-toxic radiotherapeutic approach and highlights the potential of suppression of the cGAS-STING pathway to improve treatment outcomes in BC.