The role of SUMOylation in Liver X Receptor (LXR) transrepression of inflammatory responses

Liver X Receptors (LXRs) are ligand-activated transcription factors that regulate genes involved in lipid metabolism, as well as those concerned with immune and inflammatory responses. SUMOylation of LXRs has been proposed to be central to the anti-inflammatory activity of LXRs in terms of transrepression of inflammatory gene promoters. However, recent work has suggested that SUMOylation-independent transcriptional activation of the lipid transporter ABCA1 can lead to inhibition of inflammatory signalling pathways, thus the mechanism/s by which LXRs inhibit inflammatory responses remains to be fully understood. Studies presented here demonstrate SUMO modification of LXRs in a pattern which is in contrast to reports in the literature. Both LXR isoforms (α, β) could be SUMOylated by three SUMO variants (SUMO-1, 2 and 3) in the absence of exogenous ligand. Synthetic agonists and antagonist binding was found to reduce LXR SUMOylation. In contrast naturally occurring ligands, (22(R)-hydroxycholesterol and 24(S)-hydroxycholesterol) increased LXRs SUMOylation. Treatment with ligands influenced LXR SUMOylation in a time-dependent manner with effects starting from 8 h, in contrast to reports suggesting that SUMOylation was increased after 1-2 h of treatment. Although the synthetic agonists (GW3965 and T0901317) had the opposite effect in terms of SUMOylation to the oxysterols, all LXR ligands activated the ABCA1 promoter and were inhibited by the antagonist GSK2033. LPS-induced iNOS expression was inhibited in response to GW3965 and was increased with GSK2033 treatment after 24 hours of treatment at which time there was no detectable SUMOylation of LXR. There was a noticeable difference in response to 22(R)HC where it activated LXRs in terms of AbcA1 expression, but at the same time increased inflammatory chemokine expression. Both LXRs were deSUMOylated by SENP1 and SENP2 whereas SENP6 showed specificity for LXRβ. Mutation of putative SUMOylation sites (K328 and K343 in LXRα and K31, K410 and K448 in LXRβ) to arginine demonstrated that all LXRα mutants and the K31R LXRβ mutant at retained their transactivation activity at the ABCA1 promoter. Only mutation of K31 reduced LXRβ SUMOylation whereas none of the LXRα mutants had any effect on levels of SUMOylation. The K410R and K448R mutants showed reduced LXRβ transcriptional activation activity whereas K434 increased LXRα activity without affecting LXR SUMOylation. Additionally, PIAS1 acted as the SUMO E3 ligase for both LXRα and LXRβ in contrast to previous reports suggesting that HDAC4 was the SUMO E3 ligase for both LXRα and β. HDAC4 expression strongly enhanced SUMOylation of LXRβ alone and LXRβ SUMOylation could be reduced by treatment with the HDAC4 inhibitors TSA and LMK235. The precise role of HDAC4 in LXRβ SUMOylation remains to be determined. In these experiments, LXR SUMOylation appeared to have little influence upon their transactivation or transrepression activity and SUMOylation state appeared to be dependent upon the type of ligand present. Further work should examine the impact of LXR SUMOylation/de-SUMOylation upon inflammatory signalling