Central role for endothelial human deneddylase-1/SENP8 in fine-tuning the vascular inflammatory response

Abstract

A deeper understanding of the mechanisms that control responses to inflammation is critical to the development of effective therapies. We sought to define the most proximal regulators of the Cullin (Cul)-RING ligases, which play a central role in the stabilization of NF-κB and hypoxia-inducible factor (HIF). In these studies, we identify the human deneddylase-1 (SENP8) as a key regulator of Cul neddylation response in vitro and in vivo. Using human microvascular endothelial cells (HMECs), we examined inflammatory responses to LPS or TNF-α by assessing Cul neddylation status, NF-κB and HIF-1α stabilization, and inflammatory cytokine secretion. HMECs with an intact neddylation pathway showed a time-dependent induction of Cul-1 neddylation, nuclear translocation of NF-κB, stabilization of HIF-1α, and increased NF-κB/HIF-α promoter activity in response to LPS. HMECs lacking SENP8 were unable to neddylate Cul-1 and subsequently were unable to activate NF-κB or HIF-1α. Pharmacological targeting of neddylation (MLN4924) significantly abrogated NF-κB responses, induced HIF-1α promoter activity, and reduced secretion of TNF-α-elicited proinflammatory cytokines. MLN4924 stabilized HIF and abrogated proinflammatory responses while maintaining anti-inflammatory IL-10 responses in vivo following LPS administration. These studies identify SENP8 as a proximal regulator of Cul neddylation and provide an important role for SENP8 in fine-tuning the inflammatory response. Moreover, our findings provide feasibility for therapeutic targeting of the Culs during inflammation.

Figure 1. Influence of LPS and adenosine on Cul-1 neddylation status

A) LPS-stimulation (10μg/ml medium) of human microvascular endothelial cells increases neddylation of Cul-1 protein in time dependent manner. B) Similar to previous findings in mice treated with hypoxic preconditioning, adenosine reduced the LPS induced Cul-1 neddylation in a dose dependent manner. Representative blots of 3 experiments per group.

Figure 2. Generation and validation of lentiviral SENP8 knockdown in endothelia

A Efficiency of lentiviral gene silencing on SENP8 transcript and protein levels was assessed by analysis of two different functional hairpin vectors (clones 38 and 42) compared to a scrambled vector (ctrl). Real-time PCR analysis showed different efficiency levels for both functional vectors, mirrored by immunoprecipitation/western blot analysis of SENP8. B Knockdown of SENP8 was further ascertained in the #42 clone by immunofluorescence, indicating loss of positive staining in the knockdown cell line. C Direct detection of Cullin-conjugated Nedd8 in HMEC SENP8-KD cells in response to LPS over the time course of 6 hours is indicated by western blot against NEDD8. Cells lacking functional SENP8 (KD) show a diminished neddylation response to LPS compared to empty vector (EV) transfected controls. D This impaired general neddylation response was mirrored by abrogated neddylation of Cul-1 in SENP8 knockdown cells. Representative blots of ≥3 experiments.

FIGURE 3. Functional influence of SENP8 on LPS-induced activation of NF-κB and HIF

A Following LPS stimulation, control HMEC cells exhibit an increased translocation of the NFκB subunits p50 and p65 to the nucleus. This effect is abrogated in cells lacking SENP8. B Cells transfected with a constitutive SENP8 overexpressing (OE) vector showed baseline NFκB-luciferase levels comparable to LPS treated control cells transfected with empty vector (EV) only. C MLN4924 (structure homologue of adenosine-monophosphate, AMP) significantly quenched LPS induced NFκB-Luciferase response. D LPS induced HIF-1α stabilization in the nucleus of empty vector cells after 4 and 6 hours of stimulation. This effect is not observable in SENP8 knockdown cells. All numerical data are mean±SEM, n≥3 experiments, **p<0.01, ***p<0.001.

FIGURE 4. Influence of neddylation inhibition on HIF stabilization, HIF hydroxylation, Cul-1 neddylation and HIF activity

A Pretreatment with MLN4924 increases HIF-1α protein in nuclear lysates of HMEC cells stimulated with LPS to higher levels than MLN4924 alone, indicative of synergistic effects of both compounds. LPS induced Cullin-1 neddylation was lost when cells were pretreated with MLN4924. B Treatment with MLN4924 stabilized HIF-1α in its hydroxylated form and allowed for its translocation to the nucleus, implicating effects of MLN4924 on Cullin-2 neddylation. C Luciferase-activity of HRE-Luciferase following 24hr of LPS stimulation mirrored protein results. All numerical data are mean±SEM, n≥3 experiments. *p<0.05, **p<0.01, ***p<0.001.

FIGURE 5. Influence of neddylation inhibition on transcription and translation of inflammatory cytokines and endothelial function

6 hours TNFα treatment of endothelial cells results in the significant upregulation of proinflammatory cytokine transcript and protein, which is absent when cells were pretreated with MLN4924 (A–C). MLN4924 also prevents transcriptional upregulation of ICAM-1 (D) and LPS induced increases of endothelial permeability (E). All numerical data are mean±SEM, n≥3 experiments, *p<0.05, **p<0.01.

Figure 6. Impact of MLN4924 on LPS-mediated HIF activity and cytokine production in vivo

ΔODD mice received vehicle, MLN4924 (3mg/kg BW), LPS (100μg/kg BW) or MLN4924+LPS i.p. for 6hrs. Inflammatory cytokines were measured in serum and HIF-luciferase was measured in renal lysates. A LPS+MLN4924 induced a significant increase of renal HIF cytokines. B–E, G–H Mice pretreated with the neddylation inhibitor MLN4924 prior to LPS showed no significant upregulation of all pro-inflammatory serum cytokines, while maintaining increased levels of the anti-inflammatory cytokine IL-10 (F). All values are mean±SEM with ≥4 animals per group, *p<0.05, **p<0.01, ***p<0.001.

Figure 7. Neddylation pathways influencing NF-κB and HIFα

NFκB pathway (left). Proinflammatory stimuli, such as LPS, facilitate the phosphorylation of IκB and result in the recognition of P-IκB by the Cullin-1-Nedd8-Skp-βTRCP complex, culminating in its polyubiquitination and proteasomal degradation. The rate-limiting step for this is the conjugation of NEDD8 to Cullin-1. Neddylation is achieved through a multi-enzyme process, wherein SENP8 enables cleavage of the NEDD8-precursor and promotes NEDD8 conjugation to the Cullins. Loss of SENP8, or pharmacological inhibition of NEDD8 conjugation by MLN4924, prevents activation of Cullin-1 and thus prevents liberation of NFκB from IκB, resulting in quenched pro-inflammatory signaling. HIFα pathway (right). In contrast to NFκB, HIFα in its hydroxylated form is degraded by the proteasome after ubiquitination through the von Hippel Lindau protein (pVHL). pVHL in its activated form contains neddylated Cullin-2, thereby controlling cellular HIFα levels. Use of MLN4924 prevents Cullin-2 neddylation, and as shown in this study leads to higher levels of the hydroxylated HIFα isoform.