Inflammation has been proposed to promote metastasis, but how these two processes are linked is unknown. Michael Karin and colleagues have now shown a role for I
B kinase 
(IKK), possibly activated by tumour-infiltrating inflammatory cells, in promoting prostate cancer metastasis independently of nuclear factor B (NFB).
IKK is one of two catalytic subunits of the IKK complex. Karin and colleagues crossed mice with a mutation in IKK that prevents its activation (Ikk
AA/AA) to the TRAMP mouse model of prostate cancer metastasis (which expresses SV40 T antigen in the prostate epithelium). The IkkAA/AA;TRAMP mice had slowed tumour growth and increased survival, and also had fewer metastases. The authors then used quantitative reverse transcription PCR to examine the expression of about 40 genes known to be involved in metastasis. The only gene that was consistently different in TRAMP mice expressing IkkAA/AA compared with TRAMP mice was the metastasis suppressor maspin (encoded by Serpinb5). Maspin expression declined in prostate cancers of TRAMP mice but remained high in those of IkkAA/AA;TRAMP mice. Furthermore, cells from IkkAA/AA; TRAMP mice had reduced metastatic potential when injected into the spleens of immunocompromised mice, and metastatic potential was restored by maspin knockdown. These data indicate that maspin might prevent metastasis in mice with inactive IKK.
To assess the role of IKK activation in promoting metastasis, the authors expressed wild-type or activated IKK (IKK(EE)) in IkkAA/AA; TRAMP prostate cancer cells. IKK(EE) significantly reduced maspin expression in these cells. Chromatin immunoprecipitation showed that IKK(EE) was recruited to the endogenous Serpinb5 promoter in human mammary epithleial cells, and endogenous IKK was found in the nucleus in its phosphorylated (activated) form in TRAMP prostate cancer cells. To establish further that nuclear IKK is required to control maspin expression, the authors showed that IKK(EE) with a defective nuclear localization sequence was unable to repress the Serpinb5 promoter. Interestingly, the authors also found that an activated IKK
(a more potent activator of NFB-dependent transcription than IKK(EE)) did not repress Serpinb5, which, taken together with the fact that NFB activation depends on IB phosphorylation in the cytoplasm, suggests that maspin repression occurs independently of NFB.
How is IKK activated? Prostate cancers from older TRAMP mice had higher expression of receptor activator of NFB ligand (RANKL), which is a known activator of IKK, and maspin expression was reduced in TRAMP, but not IkkAA/AA; TRAMP, prostate cells treated with RANKL. Tumours from older mice also had increased infiltrating T cells and macrophages compared with tumours from younger mice, suggesting that the infiltration of tumours by inflammatory cells that express RANKL activates IKK, which in turn represses maspin and promotes tumour metastasis.
Levels of phosphorylated nuclear IKK were increased in higher grade human prostate cancers and inversely correlated with maspin levels, indicating that this pathway might also be present in human prostate cancer.