Milea Timbergen

326 Investing in in Vivo and in Vitro Models to Understand DTF Biology The clinical course of DTF tumours varies widely. Although several factors have been mentioned to influence risk of recurrence or to reflect tumour aggressiveness such as tumour site, size and type of CTNNB1 mutation, up until now, for a large part, DTF biology is still unknown. The current thesis focuses mainly on the hypothesized biological differences between the most common mutation types of DTF. Our translational research was mainly challenged by the limited number of available biomaterials and corresponding clinical data. Therefore, inevitably given the rarity of DTF, the used study sample sizes were small, potentially contributing to inconclusive or negative results. Moreover, types of analysis were restricted by the use of FFPE material and lack of fresh frozen tissues and micro-array data. Commercially available desmoid cell lines are scarce and culturing of DTF tumour tissue remains challenging. This is mainly due to overgrowth of fibroblastic cells without a CTNNB1 mutation. Currently, clonal expansion techniques to separate tumour cells (containing an CTNNB1 mutation) from stromal cells are being exploited 1 . Although cell lines are a corner stone of many translational studies, one should also consider the role of the tumour microenvironment in DTF pathogenesis and growth. DTF tumours are known to be heterogeneous, and unpublished work from M. Al-Jazrawe and B.A. Alman showed that there is a large variation in the proportion of stromal cells and mutated cells per tumour 1 . The clinical consequences of these differences are still not clear but could explain the observed differences in tumour behaviour in culture, and possibly in patients. The role of the tumour microenvironment is well recognized in cancer, and the cross talk between stromal and neoplastic cells are thought to play an important role in tumour initiation, progression, and invasion 2, 3 . Future studies should therefore acknowledge the contribution of the tumour microenvironment for example by using representative animal models rather than cell lines. A murine model; APC+/APC1638N 1638N and a Xenopus tropicalis model have been developed. Unfortunately, both models are APC -based and may not be representative for the sporadic DTF 4,5 . Currently, mouse models with a mutation in CTNNB1 (T41A and S45F) are being developed for research purposes 6 . These animal models could be of use to study the DTF pathogenesis, microenvironment, and tumour growth. Furthermore, they could be of value when testing the efficacy of existing treatments in vivo . 12