B119
The use of cell line models to predict patient response to chemotherapy in advanced colorectal cancer
Wendy Allen, Vicky Coyle, Puthen Jithesh, Irina Proutski, Leanne Stevenson, Cathy Fenning, Gail Stewart, Daniel Longley, Patrick Johnston
Queen's University, Belfast, UK
Background
The development of drug resistance remains a major obstacle in the effective treatment of advanced colorectal cancer.
Method
To address this problem, we have previously generated a panel of chemotherapy-resistant HCT116 CRC cell lines by repeated exposure to increasing concentrations of drug over a period of months. These model cell lines have been used to identify the key genes mediating resistance to these chemotherapies. We have utilised the Colorectal DSA™ (Almac Diagnostics) to carry out transcriptional profiling of our HCT116 drug-resistant cell line panel. In addition, expression analyses of pre-treated (5FU/Irinotecan) metastatic liver biopsies have also been carried out and correlated with patient response.
Results
Data analysis identified panels of in vitro genes whose expression is acutely altered in the parental setting following drug treatment and also basally deregulated in the resistant cells. The correlation between in vitro samples and clinical samples in terms of gene expression and pathway analysis has been examined. In addition, an assessment of the feasibility of using in vitro expression data to separate patient samples based on their response to treatment has been assessed using principal components analysis (PCA). Finally, we developed a predictive classifier using the in vitro-derived gene profiles and demonstrated that a 12 gene in vitro-derived model predicted the response of colorectal cancer patients to 5-FU/CPT-11 with 85% accuracy.
Conclusion
This study demonstrates the utility of cell line models as surrogates in the development of gene signatures for the prediction of response to chemotherapy in cancer patients
