Final Report Summary - ISIS (Identification and targeting of somatic changes initiating sporadic cancers)
The general goal of ISIS was to molecularly characterize early disseminated cancer cells (DCCs) to identify shared features suited for therapeutic targeting. The rationale was deduced from following findings: (i) DCCs detected in bone marrow or lymph nodes of cancer patients that do not have manifest metastasis yet are harbingers of an increased risk for metastasis and death. (ii) Some of these early DCCs may display fewer genetic alterations than primary tumor cells that are resected at the time point of bone marrow or lymph node sampling. We therefore reasoned that early DCCs compared to primary tumor cells may inform about the sequence of evolutionary events. With this background, we aimed to molecularly characterize early DCCs, to generate DCC-derived-cell models, to identify molecular targets shared among DCCs and functionally test candidate targets in high-throughput screens and other assays. Over the course of the project we found that DCCs leave the primary tumor very early, with the time point being defined by a very small size of the lesion (often less than 1 mm) and molecularly by a lack of important mutations or alterations in oncogenes and tumor suppressor genes (Hosseini, Nature 2016; Harper Nature 2016; Werner-Klein, Nat Comm, 2018). We further identified that large tumors are less efficient in seeding cells, defined the mechanism, and confirmed that metastases (at least in a mouse model) are mostly derived from early DCCs as opposed to late DCCs. In human melanoma we could show that critical changes are acquired within the distant site and that primary tumors and metastasis-forming DCCs converge independently on the acquisition of some important changes. However, we have not discovered genetic alterations that are shared by all DCCs of a patient until today, which may be caused by the initially low resolution used to characterize the cells. During our attempts to expand early DCCs in cell culture, we noted that success is directly linked to some critical genetic alterations. In their absence, cells could not be expanded. We therefore have not yet achieved to generate models of very early DCCs, however will continue to work on this aim with the insight gained. Lacking suitable models for high-throughput functional screens, we changed direction to characterize the transcriptomes of single bone marrow derived DCCs and identified several highly interesting pathways activated. We already noted inter-patient recurrence of some cell states, which we are currently exploring for novel modes of therapeutic intervention in order to prevent metachonous metastasis.