Review

Despite the importance of metastasis to breast cancer survival and morbidity, our understanding of the cellular and molecular mechanisms underpinning breast cancer metastasis is formative. While it is still debated whether the metastatic spread of a tumour is controlled by ‘metastasis genes’ distinct from those driving primary tumourigenesis, recent studies have identified genes that regulate organ-specific metastasis of experimental breast cancer models (see [Reference Minn, Teruya-Feldstein and Weinberg1–Reference Yang, Mani and Donaher3] as examples). In their recent paper in Nature, Ma et al. [Reference Ma, Teruya-Feldstein and Weinberg4] add a twist to this list of metastatic regulators by describing a role for the miR-10b microRNA in the control of breast cancer invasion and metastasis.

MicroRNAs are small regulatory RNAs that control gene expression by repressing the translation and/or enhancing the degradation of target mRNAs through a process known as RNA interference (RNAi) [Reference Bartel5]. Discovered in 1993 in nematodes, microRNAs have revolutionised our view of cell and molecular biology as they control numerous processes including many aspects of embryonic development and disease states including cancer [Reference Kent and Mendell6]. Expressed as precursor RNAs either from their own promoters or within introns of other genes, they are enzymatically processed to the mature (~20–30 nucleotide) active microRNA.

Due to the promiscuous nature of their binding to target mRNAs, microRNAs are thought to control the expression of many target genes, thus acting to integrate large-scale gene expression programs. The practical implication of this promiscuity is that identifying the key targets of microRNAs in a given process is far from facile. In this paper, Ma et al. make the extraordinary achievement of identifying not only a new role for microRNAs in metastasis but also the downstream targets of this microRNA and the upstream transcription factor that directs the expression of this microRNA.

The story begins with Ma et al. investigating a panel of candidate microRNAs previously identified as altered between normal breast tissue and breast cancer samples [Reference Iorio, Ferracin and Liu7]. After examining the expression of eight microRNAs across a panel of immortalised or transformed breast epithelial cell lines, they found that three (miR-155, miR-9 and miR-10b) were upregulated in transformed cells. Interestingly, miR-10b was highly expressed only in metastatic cancer lines.

To investigate further, Ma et al. then used in vitro assays with cell lines in which they manipulated the expression of miR-10b. They showed that while miR-10b had no impact on cellular proliferation or death, it was both sufficient and necessary for the invasive behaviour of breast cancer cell lines in vitro. They then tested this result in vivo by transplanting poorly invasive breast cancer cell lines into the cleared mammary fat pad of immune-deficient mice. While control cells formed well-encapsulated tumours, cells overexpressing miR-10b invaded the fat pad, entered blood vessels and within 6 weeks formed distant metastases in the lung and peritoneum of their hosts. Interestingly, both groups of tumours grew at approximately the same rate. Thus, in vivo, as in vitro, miR-10b plays a role in metastatic progression but not primary tumour growth and may indeed be a ‘metastasis gene’.

To understand how miR-10b might be regulated during tumour progression, Ma et al. turned to a transcription factor previously identified as a gene driving epithelial to mesenchymal transdifferentiation (EMT) and metastasis: Twist [Reference Yang, Mani and Donaher3]. In a series of elegant experiments in vitro, they showed that when overexpressed, Twist directly binds upstream of the MIRN10B gene and promotes its expression. Intriguingly, blocking this increase in miR-10b impinged only upon the ability of Twist to drive invasion, whereas cells still underwent EMT. Thus miR-10b is induced by Twist and mediates its effects on invasion, not EMT.

Ma et al. then turned their attention to identifying the mRNAs regulated by miR-10b. From the ~100 mRNAs computationally predicted to be targets of miR-10b, they focussed their attention to the homeobox D10 gene (HOXD10), which has been implicated previously in the suppression of invasion and migration. They showed that overexpression of miR-10b repressed the translation of HOXD10 mRNA. Not happy to rest at this result, the authors set out to identify the effector/s downstream of Hoxd10, since others have previously shown the regulation by Hoxd10 of several factors controlling cellular migration. They observed that in cells overexpressing miR-10b, Hoxd10 levels decline and the expression of one of these factors, RhoC, increases. Further, overexpression of Hoxd10 or depletion of RhoC is sufficient to block the pro-invasive effects of miR-10b overexpression.

Taken together, these studies define a pathway from elevated levels of the transcription factor Twist, which upregulates the microRNA miR-10b. miR-10b then binds to the HOXD10 mRNA, repressing its translation. Finally, reduced Hoxd10 protein leads to increased RhoC and cellular migration/invasion.

Fascinatingly, the MIRN10B gene is located within a cluster of HOX genes, including HOXD10, suggesting that this pathway may have evolved to function in a coordinated manner. However, MIRN10B is in a different location within the mouse and human HOX clusters. It is not clear whether Twist binds to the same regulatory elements upstream of human and murine MIRN10B, even though mouse and human cells are used interchangeably for some of the cell biology and gene expression analyses reported.

Finally, the authors test the relevance of these findings to human disease. Across a panel of breast cancers, 5/5 women with non-metastatic disease had lower miR-10b expression in their tumours than matched normal controls, while ~50% of women with metastatic disease had elevated levels of miR-10b in their primary tumour. As with this group’s previous study of Twist [Reference Yang, Mani and Donaher3], perhaps the weakest aspect of this report is its clinical data. The current study evaluates miR-10b expression in only 23 human breast cancers of unspecified histological subtype and gene expression category. It is not reported whether any correlations between Twist, miR-10b, HoxD10 and RhoC expression are found. Furthermore, the expression of miR-10b between the non-metastatic and highly metastatic patients barely reaches statistical significance, presumably due to the small sample sizes and/or technical issues. Given that a previous study conflictingly identified miR-10b as downregulated in breast cancer compared to normal [Reference Iorio, Ferracin and Liu7], further work is required to test the validity of these findings in clinical specimens.

The primary importance of this study is that it is the first demonstration of a role for microRNAs in metastatic progression and also provides further evidence for the existence of genes whose role in cancer is confined to regulation of metastatic success.