SIX of one and half-a-dozen USP2 regulating inflammation

The sine oculis (SIX) transcription factors prevent excessive inflammatory responses. Yi et al. now show that their levels are regulated by a deubiquitinase, USP2, that is upregulated by inflammatory stimuli and prevents proteasomal turnover of SIX proteins.

Inflammation is essential to fight infection, but the need for a highly sensitive, rapidly amplifying response carries with it the risk of an over reaction.¹ If the exquisite regulation of this process fails, the cure can be worse than the disease. Evolution has hedged the process with multiple levels of control and Yi et al. now show how the activity of USP2, a deubiquitinase (DUB), results in the stabilization of SIX transcription factors, allowing them to repress inflammatory gene transcription.²

Earlier work from the same group showed that the non-canonical NF-κB inducing kinase NIK/MAP3K14 upregulates SIX1/2 proteins via induction of SIX transcription and inhibition of SIX protein degradation.³ However, the mechanism of how NIK activation results in SIX stabilization was unclear. It turns out that, rather than directly inhibiting an E3-ubiquitin ligase, NIK/MAP3K14 works indirectly by activating the DUB USP2 thereby reversing the constitutive proteasomal degradation of SIX proteins (Fig. 1).

Yi et al. screened 75 known DUBs for the regulation of SIX protein levels. The fact that both SIX1 and SIX2 were clearly reduced in USP2^–/– cells and cells treated with the USP2 inhibitor ML364,⁴ provided Yi and colleagues with sufficient justification to generate knock-in mice mUsp2^FlagKI. As predicted for an inflammatory regulator, USP2^iso1 was nicely upregulated from almost undetectable levels in primary BMDMs by inflammatory stimuli (e.g., LPS/TNF) and in fibroblasts following Listeria infection and TWEAK stimulation, although USP2^iso2–4 were not. All four human USP2 protein-encoding variants⁵ contain the DUB domain but with different localization and interactions, therefore it is not surprising that they might be differently regulated.

Interestingly, it seems that SIX deubiquitylation occurs in the nucleus. Chromatin immunoprecipitation experiments indicate that USP2 is bound to inflammatory gene promoters. This suggests that SIX proteins are targeted for degradation at the inflammatory gene promoters where they inhibit transcription and that this is antagonized by USP2 in the same location. Indeed, inflammatory genes are upregulated in USP2 knockout cells, which is further exacerbated by cytokine stimulation, and genes upregulated in SIX1/2 double knockout cells were similarly upregulated in USP2 knockout cells.

To further explore the role of SIX1 in vivo, Yi et al. generated Six1^flox/flox mice. Despite whole body deletion of Six1 causing perinatal lethality, tamoxifen-induced, Cre-ER^T2-mediated deletion did not result in noticeable weight loss over 20 days in adult mice. Given the fact that SIX1 seems to be a potent suppressor of inflammatory cytokine production, it will be interesting to perform a more detailed and long-term analysis to see whether low-level systemic inflammation is present and whether it increases over time. Six1 knockout mice challenged with influenza A virus showed a rapid and non-recoverable weight loss while about half of the wild-type mice recovered, although also undergoing weight loss. Usp2 deletion in Usp2^flox/flox mice via CMV^cre did not affect viability, but they were similarly sensitive to influenza infection. Yi et al. were then able to demonstrate the importance of SIX1-USP2 regulatory pathway specifically in myeloid cells, where a similar sensitivity to influenza infection was observed.

Like all good research, the work by Yi et al. raises new questions, more indeed than can be detailed here. Firstly, as discussed, SIX1/2 are key players in several developmental pathways and after the completion of embryonic development, SIX1 is downregulated. However, SIX1 is frequently upregulated in multiple human cancers, and together with its transcriptional activator EYA1, SIX1 can increase cancer stem cell numbers and cell motility, and induce epithelial-mesenchymal transition, all factors that account for the observed correlation with increased aggressiveness and poor prognosis.⁶ Since an inflammatory environment is known to be an enabling characteristic of tumorigenesis,⁷ one wonders whether the potential ability of SIX1-overexpressing tumors to reduce their own inflammatory response provides them with any advantage in this environment.

Secondly, while loss of Six1 in mice is perinatally lethal, mutations in SIX1 and its co-factor EYA1 have been identified in ~50% of individuals with Branchio-otic and Branchio-oto-renal syndromes, which are autosomal dominant disorders with birth defects including malformations of the ear, kidney and branchial arches.⁸ These syndromes are rare but systemic inflammation is not remarked on as a clinical characteristic, nor was it reported in the adult mice with conditional Six1 deletion used in this study, yet it might be worth further exploration.

Lastly, USP2 has been indicated to regulate a diverse range of pathways. With reference to cytokine production, two points are worth considering. Firstly, TNF signalling, which upregulates USP2, is activated by E3 ligases, such as cIAP1 and TRAF2, and restricted by well-known DUBs, such as A20. USP2 has also been reported to be recruited to and reduce K63-linked ubiquitylation on RIPK1, and thus USP2 knockdown enhanced TNF-induced canonical NF-κB activation. Thus, it will be interesting to explore more deeply how these functions of USP2 and the TNF-USP2 negative feedback loop play out in real time to regulate canonical and non-canonical NF-κB signallings.

Besides TNF signalling, the role of USP2 in circadian rhythm is perhaps the most intriguing. Usp2 mRNA is expressed in a circadian manner in many tissues and its levels are damped in Clock mutant mice, suggesting that it plays a key role in the core mechanism of the circadian oscillator, possibly by directly regulating PERIOD1.⁹ Cytokine responses also show circadian regulation,¹⁰ and therefore the study by Yi et al. raises the possibility that USP2 plays a role in this feature of cytokine regulation too; it might be time to check.