SATB1 reprogrammes gene expression to promote breast tumour growth and metastasis

Hye-Jung Han, Jose Russo, Yoshinori Kohwi & Terumi Kohwi-Shigematsu

Nature 452, 187-193 (13 March 2008) | doi:10.1038/nature06781; Received 11 September 2007; Accepted 22 January 2008

Significance

With the advent of gene expression microarrays, researchers have found gene expression patterns in the early stages of tumors that correspond to poor prognosis (metastasis and recurrence) and good prognosis (remission) in patients treated under the same quality of care. This observation suggests that the aggressiveness of a particular tumor may have been genetically determined well before it is even detected by physicians. It is of obvious clinical significance that such “metastasis genes” are discovered. In fact, many new FDA approved gene expression based assays such as MammaPrint and Oncotype DX have already identified set of genes that correspond with patient prognosis. Such knowledge can help an oncologist determine the appropriate course of treatment. This information can be especially pertinent when one considers the fact that most chemotherapy agents are associated with cytotoxic side effects.

Summary of Experiments

In this paper the authors identified one such metastasis promoting gene, special AT-rich sequence binding protein 1 (SATB1), whose mRNA and protein expression levels correspond to poor prognosis in breast cancer primary tumors, cell lines and animal models (Figure 1D). In their paper, the authors show that SATB1 is both necessary and sufficient for breast cancer metastasis making it a gene of great potential interest in breast cancer diagnostics.

Upon mRNA expression analysis of various breast cancer cell lines, the investigators found that SATB1 is only expression ion metastatic cancer cell lines. This observation is followed up with evidence from immunostaining and tissue microarrays of primary tumors for SATB1 in which the presence of SATB1 corresponded with poorly differentiated infiltrating tumors and poor patient prognosis. The authors also note that SATB1 expression levels are independent of any established prognostic markers such as tumor stage, histological grade or nodal stage. Next, in an in vitro SATB1 knockdown experiment of MDA-MB-231 cells (an aggressive breast cancer cell line), reduced expression of SATB1 corresponded with decreased proliferation, decreased tumor colony formation, reduced invasiveness and greater tissue organization -- all characteristics of less aggressive, “well-behaved” tumors (Figure 2B,C,D,E).

In addition, SATB1 depleted cells were implanted into athymic mice and the degree of metastasis was measured. In comparison with the control group, mice injected SATB1 down regulated MDA-MB-231 cells shown virtually no tumor growth in the lung nodules (a measure of metastatic tumor growth). This illustrated that SATB1 is necessary for metastasis. The converse experiment was also performed in which SATB1 was upregulated in a less aggressive breast cancer cell line (SKBR3). In this experiment mice injected with the SATB1 overexpressing SKBR3 cells showed significantly increased formation of metastatic nodules as compared with the control group Figure 4 A,B,C). This illustrated that SATB1 is sufficient for metastasis.

Since SATB1 is though to be a chromatin remodeling protein, it is thought to control the transcription of a large number of genes. To determine which genes were affected, microarray experiments were performed on MDA-MB-231 cells and MDA-MB-231 SATB1 shRNA cells. As expected a number of metastasis/invasion promoting genes such as metastatin, VEGFB, TGF-beta and ERBB were found to be upregulated. In addition the expression of many ECM components were found to be misregulated which explains the abnormal cell morphology. To discriminate between the primary targets of SATB1 and downstream effects, a chromatin immunoprecipitation (ChIP) experiment was performed followed by real-time PCR of various loci of interest. This experiment in combination with assessment of histone acetylation status suggested that SATB1 associated with chromatin remodeling proteins such as p300 and HDAC1 to regulate gene expression through epigenetic modification.

Together, all of these studies suggest that SATB1 is a genome modifying protein involved in tumorigenesis that promotes metastasis.

Critique

The paper does an excellent job of presenting a series of logical experiments which each supports the notion that SATB1 causes metastasis. However, the authors do not mention what screen was used to identify SATB1 as such an important metastatic promoting protein. After all, many microarray experiments have identified hundreds genes that were simply up or down-regulated in metastasis, the majority of which are “passenger” genes which are not metastasis-promoting. Lastly, it would be interesting to investigate what causes differences in the expression levels of SATB1 in primary tumors.