Exploiting tissue on-chip technology to predict treatment responses in HNSCC

Culture models have advanced significantly in recent years, especially for studying disease in 3-dimensions. However, these models have rarely been used to correlate in vitro response with the clinical response in diseases like head and neck cancer (HNC). The work in this thesis aimed to use a microfluidic model that could maintain biopsies of tissue in a viable state to investigate response of tissue to clinically relevant drug and radiation treatments; changes in the secretome; and whether this type of tissue-on-chip technology could predict patient outcome in response to treatment. In total biopsies of HNC tumour were collected from 18 HNC patients, following informed, written, consent. The biopsy was sub-divided and either fixed immediately upon excision or maintained in a tissue-on-chip device for 72 h. During incubation the tissue was treated with chemo- (Cisplatin, Cetuximab, Nivolumab and a triple combination of the three), radio- (5 x 2Gy) and chemo-radiotherapeutic (Cisplatin + 5 x 2Gy) treatment regimens. The tissue response was measured by immunohistochemistry with a proliferation marker (ki67), an apoptotic marker (cleaved-PARP), and nuclear area was quantified to measure tissue integrity. The secretome of the tissue was analysed using a multiplex assay to analyse 108 different cytokines simultaneously, and ELISA (IL-8 and VEGF). Clinicopathological data and patient status (alive or deceased since last appointment) were recorded.The tissue response and cytokine expression between patients showed a large amount of variability, however generally there was less proliferation, more cell death and similarity in the cytokine profiles of the biopsies treated with an anti-cancer treatment than without. HNC control tissue-maintained ex vivo showed a marginal decrease in proliferation and marginal increase in cell death; thus, demonstrating that the tissue remains functional and largely unaffected by the perfusion model itself. IL-8 and VEGF expression was found to have three distinct expression profiles (sporadic, like the overall average, higher expression in the 24 h and 48 h, and a decrease after 72 h). However, due to the large inter-patient variation no statistically significant differences were detected. Nivolumab and the triple combination seemed to give the largest impact on proliferation, cell death and tissue integrity and, although not significant in comparison to the control, they were statistically significant when compared to the ‘pre’ tissue. The study has shown the feasibility of using this perfusion platform in a tissue and secretome analysis for patient outcome in HNC, which, with a larger patient cohort and more standardised treatment regimen criteria, has potential for investigating correlation(s) between an in vitro model and clinical response to the same treatment regimen.