Cytotoxicity of ascorbate to cancer cells: mathematical modelling, mechanisms and clinical implications

Abstract

Studies over the last few decades provided supportive evidence for further research on the use of intravenous administration of high-dose ascorbate (0.5 – 1.5 grams per kilogram of body weight) in cancer treatment. Although intravenous ascorbate showed limited anticancer activity in the recent clinical trials, its potential beneficial effects were noted, such as improvement in quality of life, reduction of chemotherapy-associated toxicity, reduction of inflammation and stabilization of cancer progression. These observations were in agreement with the outcomes of three cancer patients who declined chemotherapy and received intravenous ascorbate as an alternative in a preliminary clinical case study (Chapter 4). It was commonly accepted that the cytotoxicity of ascorbate was mediated by hydrogen peroxide (H2O2). However, our analysis and interpretation of the published results were not consistent with this being the sole mechanism of cytotoxicity, and an additional dehydroascorbate-mediated mechanism was proposed in this thesis (Chapter 1). In order to have a holistic understanding of the differential effects of ascorbate treatment in in vitro, animal and human studies, the treatment-induced changes in the intracellular H2O2 concentrations of the targeted cells should be compared. However, precise measurement of intracellular H2O2 concentrations in vivo is difficult. In Chapter 2, a mathematical model was established for estimating the extracellular and intracellular H2O2 concentrations in a biological system that was challenged with H2O2 and/or an agent that could induce H2O2 production (such as ascorbate). Simulation of the mathematical model was carried out with the values of variables estimated from the published scientific literature and the experimental data obtained from this study. Several predictions were made and validated with the experimental results. One simulation predicted a limited increase of intracellular H2O2 concentration in a typical intravenous ascorbate treatment, which should be insufficient to induce cell death in tumours. The mathematical model also provided methods for estimating intrinsic intracellular H2O2 production rate and intracellular H2O2 concentration based on extracellular H2O2 concentration and other cell-dependent parameters, which were easier to be obtained with the current experimental methods. In Chapter 3, it was demonstrated in vitro that the cytotoxicity of ascorbate was mediated by both H2O2 and dehydroascorbate (DHA). The associations between the responses of cells to ascorbate (0 – 30 mM) and some extracellular, cellular and molecular variables, such as cell culture medium, serum, glucose, pyruvate, cell cycle, cell volume and the expression and activities of H2O2 decomposing enzymes catalase and glutathione peroxidase, were investigated. Pyruvate but not glucose showed protective effect on cells in DHA treatment. Increased oxygen reactive species (ROS) levels in the DHA-treated cells was observed. These results supported our hypothesis that DHA might be able to antagonize glucose uptake and modulate the cellular H2O2 decomposing capability, which could allow higher concentrations of intracellular H2O2 to be accumulated. Thus DHA could be potentially used as an adjuvant anticancer molecule in a ROS-mediated treatment of cancer, which was supported with our results that demonstrated synergistic cytotoxicity of DHA and arsenic trioxide on the leukaemia cell line HL-60. Transforming growth factor beta (TGFβ) signalling is a participant in both tumour suppression and tumour progression, and its involvement in regulating H2O2 production and removal has been demonstrated. In Chapter 5, the mRNA expression levels of TGFB1, TGFB2, TGFB3, TGFBR1 and TGFBR2 in primary breast tumours and adjacent normal breast tissues were analysed; and the associations with tumour characteristics and patients’ overall and relapse-free survival time were evaluated, using the public gene expression microarray data and our quantitative real-time PCR validation data. The results indicated a reduced responsiveness of breast tumour cells to TGFβ, a preferential up-regulation of TGFB1 in malignant tumours and a preferential up-regulation of TGFB3 in premalignant tumours. High TGFB2, TGFB3 and TGFBR2 mRNA levels in tumours were generally associated with better prognosis for patients, especially those diagnosed with lymph node-negative diseases. High TGFB1 and TGFBR1 mRNA levels in tumours were associated with poorer prognosis for patients diagnosed with small (diameter ≤ 2 cm) tumours. The results might not only provide prognostic value for patients but also assist in classifying tumours according to their potential responses to TGFβ and selecting patients for TGFβ signalling pathway targeted therapies. With the possibility to modulate H2O2 production and removal by TGFβ signalling pathway targeted therapies in the targeted tumours, it might allow us to develop strategies to improve the efficacy of intravenous ascorbate in selected patients in the future.