ZBTB18 restricts chromatin accessibility and prevents transcriptional adaptations that drive metastasis.

Metastases arise from rare cancer cells that successfully adapt to the diverse microenvironments encountered during dissemination through the bloodstream and colonization of distant tissues. How cancer cells acquire the ability to appropriately respond to microenvironmental stimuli remains largely unexplored. Here, we report an epigenetic pliancy mechanism that allows cancer cells to successfully metastasize. We find that a decline in the activity of the transcriptional repressor ZBTB18 defines metastasis-competent cancer cells in mouse models. Restoration of ZBTB18 activity reduces chromatin accessibility at the promoters of genes that drive metastasis, such as Tgfbr2, and this prevents TGFß1 pathway activation and consequently reduces cell migration and invasion. Besides repressing the expression of metastatic genes, ZBTB18 also induces widespread chromatin closing, a global epigenetic adaptation previously linked to reduced phenotypic flexibility. Thus, ZBTB18 is a potent chromatin regulator, and the loss of its activity enhances chromatin accessibility and transcriptional adaptations that promote the phenotypic changes required for metastasis.

The tumor cell-secreted matricellular protein WISP1 drives pro-metastatic collagen linearization.

Collagen linearization is a hallmark of aggressive tumors and a key pathogenic event that promotes cancer cell invasion and metastasis. Cell-generated mechanical tension has been proposed to contribute to collagen linearization in tumors, but it is unknown whether other mechanisms play prominent roles in this process. Here, we show that the secretome of cancer cells is by itself able to induce collagen linearization independently of cell-generated mechanical forces. Among the tumor cell-secreted factors, we find a key role in this process for the matricellular protein WISP1 (CCN4). Specifically, WISP1 directly binds to type I collagen to promote its linearization in vitro (in the absence of cells) and in vivo in tumors. Consequently, WISP1-induced type I collagen linearization facilitates tumor cell invasion and promotes spontaneous breast cancer metastasis, without significantly affecting gene expression. Furthermore, higher WISP1 expression in tumors from cancer patients correlates with faster progression to metastatic disease and poor prognosis. Altogether, these findings reveal a conceptually novel mechanism whereby pro-metastatic collagen linearization critically depends on a cancer cell-secreted factor.