Methodological Approaches towards Personalised Cancer Medicine

Abstract: Despite advances in diagnostics and treatments, many cancer patients have poor survival rates. Tumours develop drug resistance followed by metastasis, and survivors suffer from treatment side-effects. Omics techniques, targeted treatments and immunotherapy offer the prospect of individually adapting treatments for optimal efficacy and minimal side-effects. This requires integration of biomolecular, clinical and drug data to successfully predict optimal treatments for every patient. The aim of this thesis was to evaluate and apply different methodologies important for personalised treatment in a variety of cancer settings. The first paper showed that E6/E7 mRNA detection through RT-NASBA is more accurate and sensitive than DNA genotyping for classifying HPV infection in cervical adenocarcinoma, showing RNA analysis to be preferable for identifying high-risk patients. In paper II, HPV16 E2 and E5 mRNA expression in oropharyngeal cancer was analysed in relation to clinical outcomes and tumour immunology. Neither down-regulation of HLA class I nor CD8+ T-cell infiltration, both indicators of good prognosis, were dependent on E2 or E5. However, absence of E2 was related to poor progression-free survival. This allows E2 expression to be combined with HLA class I and CD8+ T-cells when stratifying patients with good prognosis for milder treatment. The third paper screened combinations of growth factors and drugs for impact on proliferation of breast cancer cells, creating a two-dimensional space to simulate tumours in different signalling states interacting with drugs. In MDA-MB-231 cells, TGF-β in combination with EGF and oestrogen inhibited growth, with the effect strengthened by Tamoxifen. In MCF7 cells, Tamoxifen inhibited growth when added to both EGF and oestrogen. In paper IV, the immunoproteome in urinary bladder cancer was analysed. Proteomics and network analysis of regulatory (Treg) and effector T-cells (Teff) of lymph nodes showed that Tregs in sentinel nodes (SN) up-regulate growth and immune signalling networks. IL-16, previously not shown to be expressed by Tregs, was predicted as central to SN-Treg signalling. IL-16 expression in Tregs was validated, shown to be higher in lymph nodes than peripheral blood and inhibited by tumour cell supernatant. In conclusion, this thesis has shown methods to improve patient stratification in cervical adenocarcinoma and HPV-positive oropharyngeal cancer. The utility of proliferation screenings replicating tumour heterogeneity for optimising drug combinations was demonstrated, and finally, lymph node proteomics revealed individual differences in T-cell signalling, important for optimising immunotherapy. Integration of these and other methods will be key to arrive at personalised cancer medicine – application of the optimal treatment combination for every patient.