Head and neck cancer is a common malignancy with approximately 600 000 new cases yearly. Disappointingly, the overall survival has not increased over the last decades. The concept of personalized medicine, i.e. to treat every patient with an individually planned treatment regime has gathered increased interest, but requires the establishment of novel biomarkers that can predict treatment response.

The aim of this thesis is to propose novel predictive single markers or combinations of markers of response to radiation, cisplatin and cetuximab. The general methodology is to evaluate common differences of cell lines resistant to radiation, cisplatin or cetuximab compared to sensitive counterparts.

In paper I, we analysed the expression of 14 proteins involved in growth control and/or apoptosis by western blot and related them to intrinsic radiosensitivity (IR) in nine cell lines. No factor had a significant correlation to IR on its own. A combination of EGFR, survivin, Bak, Smad4, and Hsp70 had the best correlation to IR (R=0.886, p=0.001). Additionally, we analysed the presence of p53 mutations in the cell lines. All cell lines had at least one missense, splice site or loss of transcript mutation. To be able to combine protein expression and presence of p53 mutations we created a system designated the number of negative points (NNP). With this system we could extract that expression of EGFR, survivin, and p53 missense or splice site mutations had the best correlation to IR (R=0.990, p<0.001).

In paper II we conducted a gene expression microarray analysis of three cell lines, from which common deregulations in two cisplatin resistant cell lines was compared to a cisplatin sensitive cell line. From a bioinformatic approach of gene ontology and molecular network analysis, we defined a transcriptional profile of 20 genes. Finally, key findings were analysed in a larger panel of cell lines, where high MMP-7 expression correlated with higher cisplatin resistance.

Paper III compared 4 cell lines with high IR to a radiosensitive equivalent. Using a similar bioinformatic approach as paper II, we established a transcriptional profile of 14 genes. Analysis in a larger panel of cell lines revealed that FN1 expression predicts higher IR.

Paper IV establishes the cetuximab sensitivity of 35 cell lines of which 12 were resistant and five were sensitive to cetuximab. After whole genome gene copy number analysis of five cetuximab resistant and five cetuximab sensitive cell lines, and verification of key findings in a larger cell line panel, the results show that the amplification of the YAP1 gene is coupled to cetuximab resistance.

In summary, this thesis proposes a number of novel markers of resistance to radiation, cisplatin, and cetuximab which could influence treatment choice in the future, following verifications in primary tumor material.