EGFR/ERBB/HER family of receptor tyrosine kinases are frequently mutated, and hence, aberrantly active in cancer. Currently, numerous ERBB-targeted therapies are used to treat various types of cancer, using activating mutations or amplification of EGFR and ERBB2 as predictive biomarkers. Yet, it remains unknown whether cancer patients with activating ERBB4 mutations could benefit from ERBB4-targeted therapy. This is despite previous reports on activating ERBB4 mutations that may drive cancer growth and the availability of clinically used second-generation pan-ERBB inhibitors that potently block also ERBB4. The unclear role of ERBB4 in cancer is largely due to the highly context-dependent and partly opposing functions of the different ERBB4 isoforms, as well as due to the high diversity of cancer-associated ERBB4 mutations lacking obvious mutational hotspots.

The aim of this thesis was to clarify the role of ERBB4 in cancer and to evaluate the potential of using ERBB4 mutations as predictive biomarkers. To elucidate ERBB4 signaling mechanisms in cancer, the interactome of the two cancer-predominant isoforms of ERBB4 was analyzed in breast cancer cells. VAV3 was identified as a novel effector of ERBB4-mediated signaling promoting cancer cell migration. To address the clinical significance of the hundreds of different ERBB4 mutations found in cancer patients, two complementary approaches were employed: an unbiased high-throughput screen and a focused functional characterization of the newly emerging hotspot ERBB4 mutations, of which many are paralogous to known oncogenic mutations in other ERBB genes. These two approaches identified 14 novel gain-of-function ERBB4 mutations of which the five most potent were mechanistically characterized. These most potential activating driver mutations were also targetable with clinically used pan-ERBB inhibitors, and thus, could potentially serve as predictive biomarkers.

Taken together, the results of this thesis have clarified the rationale for targeting ERBB4 in cancer. These findings can facilitate clinical interpretation of ERBB4 mutations and warrant further studies of using activating ERBB4 mutations as predictive biomarkers for pan-ERBB inhibitor therapy.