Measurements using the alkaline comet assay predict bladder cancer cell radiosensitivity

In the UK, bladder cancer is a common urological malignancy, affecting more than 12 500 individuals each year and causing nearly 5000 deaths per annum (CRC, 2001). In all, 90% of bladder cancers are transitional cell carcinomas (TCCs) and ∼30% of bladder tumours are muscle invasive at presentation, a feature that is associated with significant risk of metastasis (30–60%). Patients with organ-confined, muscle-invasive tumours, (T2/3/4a) who are deemed fit, are considered for potentially curative treatment in the form of radical surgery (cystectomy) or radiotherapy. Radical cystectomy (removal of the bladder and urinary diversion) is associated with significant morbidity and a mortality rate of ∼2% (Skinner et al, 1988; Studer et al, 1995). Furthermore, the necessary creation of an abdominal wall stoma (ileal conduit) may impact on quality of life, and erectile dysfunction is virtually universal. The outcome following radical cystectomy varies with tumour stage, with an overall 5-year survival of ∼40%. Radical radiotherapy (RT) is the mainstay of bladder-sparing treatment regimens. It avoids the trauma of major surgery and is considered a primary treatment for patients deemed unfit for surgery. However, RT is itself associated with dose-related complications arising from bowel and bladder being included in the radiation field. Furthermore, while for ∼50% of patients radical RT results in effective local tumour control and acceptable bladder function (Quilty et al, 1986; Jenkins et al, 1989), the remaining patients suffer local recurrence. For this latter group, the decision to treat with RT is disadvantageous, as these patients have been unnecessarily exposed to additional risks of ionising radiation. In addition, the time taken to recognise treatment failure (3–6 months) may provide further opportunity for metastatic spread, before secondary treatment (salvage cystectomy) is undertaken. Also, the morbidity in patients undergoing salvage cystectomy has been found to be significantly higher than in patients undergoing primary cystectomy (Rosario et al, 2000). Consequently, if a patient's bladder tumour RT response could be predicted in advance, RT could be promoted in patients with tumours that are predicted to respond. Conversely, patients with nonresponsive tumours could be identified and offered surgery at an earlier stage. In this way, the overall local control rates could be improved. While much work has been undertaken to develop assays capable of predicting tumour response to RT, none has been successfully applied to clinical practice. However, ex vivo measures of the surviving fraction of tumour cells at 2 Gy (SF2) suggest that intrinsic radiosensitivity (IRS) is a significant factor in determining tumour radiocurability (West, 1995; West et al, 1997). However, the SF2 assay fails to provide information on a time scale appropriate for treatment planning and also suffers from limited success rates; in an unpublished study, <10% of bladder tumours gave rise to colonies on soft agar (McKeown, McKelvey-Martin and Ho, personal communication). In addition, the relationship between clonogenic survival and clinical response is far from proven. The limitations of SF2 have stimulated research into methods to provide a more rapid and complete measure of IRS. DNA is the most important cellular target for the lethal effects of ionising radiation, with double-strand breaks (DSBs) proposed to be the principal lesions responsible for radiogenic cell killing (Ward, 1988; Iliakis, 1991). Unfortunately, the relative yield of radiogenic DSBs is low, and high radiation doses tend to be required to produce measurable levels; doses that are far greater than those used clinically. Indeed, while a recent study using the neutral comet assay to measure DSBs reports an association between DSB manifestation and survival (Price et al, 2000), the correlations were not definite, with one cell line yielding a false impression of radiosensitivity at the high doses (30 Gy) required for the assay. This highlights a benefit of conducting predictive tests of radiosensitivity at clinically relevant doses. In contrast with DSBs, the yield of radiation-induced single-strand breaks (SSBs) is far greater (Ward, 1988), and can be readily measured at low clinically relevant doses of radiation. Furthermore, the mechanisms that are proposed to vary the yield of radiation-induced DSBs formation are also expected to vary the yield of radiation-induced SSBs (Ward, 1990). Consequently, the extent of radiation-induced SSB formation can be considered a valid surrogate marker of radiogenic DSB formation. The alkaline comet assay (ACA) is a highly sensitive method for the assessment of SSBs and alkali labile sites (ALSs), and can readily detect levels of damage induced by clinically relevant doses of radiation (Singh et al, 1994; Singh, 1996). In a previous study of just three bladder cancer cell lines, an inverse correlation was obtained between clonogenic survival and mean tail moment (TM) for comet formation, suggesting that ACA could potentially be used to predict the radio response of the single cell lines (McKelvey-Martin et al, 1998). In the present study, we report our evaluation of ACA as a measure of bladder cancer cell radiosensitivity in vitro using a panel of six bladder cancer cell lines, and demonstrate that the extent of comet formation best reflects bladder cancer cell radiosensitivity; these results are supported by two independent, parallel studies using colorectal tumour cells (Dunne et al, 2003) and bladder tumour cells (McKeown et al, 2003). We also report on a preliminary study to determine the differing ACA radio response of epithelial cells isolated from human bladder tumour biopsies.