The regulation of gene activation or repression is a balancing act involving the integration of signals from a complex network, which results in the recruitment of specific chromatin-remodelling complexes and transcription factors. Jung Hwa Kim and colleagues report in Nature that the transcription regulation of the metastasis-suppressor gene KAI1 is regulated in a tissue-specific manner by such a balance, which pivots on the transcription factor nuclear factor-κB (NF-κB).

KAI1 is a known NF-κB-responsive gene, so the team wondered whether NF-κB could regulate the metastasis-suppressive function of KAI1. PCR after reverse transcription of RNA (RT-PCR) showed that KAI1 mRNA is increased in normal as well as tumorigenic but non-metastatic prostate (RWPE) cells following treatment with the NF-κB activator interleukin-1β (IL-1β). However, KAI1 shows no increase in response to IL-1β in a highly metastatic prostate cancer cell line (LNCaP). Restoring expression of KAI1 to the LNCaP cells markedly decreased the number of metastases that formed in the lungs of mice orthotopically injected with these cells.

To tease out the mechanisms of KAI1 transcriptional regulation in prostate cancer, Kim et al. used a chromatin immunoprecipitation (ChIP) assay to identify factors bound to the KAI1 promoter. In RWPE cells, IL-1β treatment led to the release of an NCo-R transcriptional co-repressor complex from the promoter, and the recruitment of TIP60, an androgen-receptor co-activator implicated in the development of prostate cancer. In LNCaP cells, however, once the NCo-R was released, the TIP60 co-activator complex was not recruited.

TIP60 forms a complex with the chromatin-remodelling proteins pontin and reptin, and as the RWPE ChIP experiments showed that histone H3 and H4 were acetylated following IL-1β treatment, the team next focused on the involvement of these proteins in the regulation of KAI1. In the non-metastatic cells, TIP60 and pontin bound to the KAI1 promoter after exposure to IL-1β; in the metastatic LNCaP cells, however, they did not. Instead, reptin was recruited to the promoter along with β-catenin, another transcriptional regulator that interacts with both pontin and reptin.

At the protein level, the metastatic cells showed less TIP60 and more β-catenin than the non-metastatic ones. Together with the ChIP experiment, this indicated a transcriptional balance between repression, mediated by a β-catenin–reptin complex, and activation through a TIP60 complex. This theory was borne out by expression of a constitutively active mutant of β-catenin in the non-metastatic cells that was sufficient to downregulate KAI1. Increasing TIP60 expression restored KAI1 expression. Moreover, a two-step ChIP assay showed that β-catenin–reptin was recruited only to promoters that were TIP60-negative.

The effects of β-catenin, on KAI1 at least, seem to be mediated through NF-κB. Using a minimal promoter containing binding sites for the NF-κB DNA-binding protein p50, the authors found that recruitment of both β-catenin and TIP60 was dependent on p50.

But does this transcriptional balance between TIP60 and β-catenin actually modulate metastasis in prostate cancer cells? To test this, the researchers altered the ratio of TIP60 to β-catenin and measured the ability of the cells to traverse a Matrigel-coated membrane. Overexpression of TIP60 or knocking down β-catenin in IL-1β-treated LNCaP cells does indeed decrease invasion of the Matrigel compared with controls. So it seems that this pathway could be central to the metastatic potential of these cells.