Glucose starvation impairs DNA repair in tumour cells selectively by blocking histone acetylation

Background and purpose Tumour cells are characterized by aerobic glycolysis and thus have high glucose consumption. Because repairing radiation-induced DNA damage is an energy-demanding process, we hypothesized that glucose starvation combined with radiotherapy could be an effective strategy to selectively target tumour cells. Material and methods We glucose-starved tumour cells (A549, FaDu) in vitro and analysed their radiation-induced cell responses compared to normal fibroblasts (HSF7). Results Irradiation depleted intracellular ATP levels preferentially in cancer cells. Consequently, glucose starvation impaired DNA double-strand break (DSB) repair and radiosensitized confluent tumour cells but not normal fibroblasts. In proliferating tumour cells glucose starvation resulted in a reduction of proliferation, but failed to radiosensitize cells. Glucose supply was indispensable during the late DSB repair in confluent tumour cells starting approximately 13 h after irradiation, and glucose starvation inhibited radiation-induced histone acetylation, which is essential for chromatin relaxation. Sirtinol – an inhibitor of histone deacetylases – reverted the effects of glucose depletion on histone acetylation and DNA DSB repair in tumour cells. Furthermore, a glucose concentration of 2.8 mmol/L was sufficient to impair DSB repair in tumour cells and reduced their clonogenic survival under a fractionated irradiation regimen. Conclusions In resting tumour cells, glucose starvation combined with irradiation resulted in the impairment of late DSB repair and the reduction of clonogenic survival, which was associated with disrupted radiation-induced histone acetylation. However, in normal cells, DNA repair and radiosensitivity were not affected by glucose depletion. Findings: Between October 2012 and June 2015, 257 patients were randomly assigned to weekly offline portal imaging (n = 129) or daily online CBCT IGRT (n = 128). Out of 250 evaluable patients, 96% completed PROs at baseline and 97% at end of RT. Baseline analyses demonstrated balance between groups for baseline characteristics as well as for PROs. In general, patients reported a small degree of side effects at end of RT, and there was no difference between groups for primary outcome (rectal bother scale of QUFW94 1.871 vs 1.884, p = 0.804). In addition, there were no significant differences between groups for any other gastrointestinal or urinary symptom as reported by QUFW94. Health related quality of life analyses (EORTC QLQ 30) demonstrated no differences between groups. Interpretation: In radical RT for prostate cancer, daily CBCT IGRT with reduced PTV margins demonstrated no advantage with respect to patient reported side effects at end of RT as compared to weekly orthogonal offline portal imaging with standard PTV margins.