MLN4924 inhibits Pi-induced calcium deposition in VSMCs.
To investigate the potential involvement of NEDD8 in vascular calcification (VC), we initially assessed the effect of MLN4924 on Pi-induced calcium deposition in VSMCs. Treatment with MLN4924 resulted in a significant dose-dependent reduction in Pi-induced calcium deposition (Fig. 1a). Alizarin red S staining revealed that mineralization, characteristic of Pi-induced VC, was effectively inhibited by MLN4924 treatment (Fig. 1b, c). Pi treatment led to decrease in the expression of SM22a and SMA, genes associated with the smooth muscle differentiation, whereas upregulating the expression of osteogenic genes RUNX2 and ALP. However, these changes were counteracted by MLN4924 treatment (Fig. 1d). In addition, we assessed cell viability using the MTT assay and observed that MLN4924 did not affect cell viability in both Pi-treated VSMCs (Supplementary Fig. 1a).
We further assessed whether Pi can induce the NEDDylation by western blot analysis. Formation of poly-NEDD8 was increased in high phosphate-treated VSMCs (Fig. 1e). Overexpression of NEDD8 potentiated Pi-induced calcium deposition in VSMCs (Fig. 1f and Supplementary Fig. 1b), while transfection of NEDD8-ΔGG, a mutant incapable of conjugation, inhibited Pi-induced calcium deposition (Fig. 1g and Supplementary Fig. 1c). These findings collectively suggest that the NEDD8 pathway plays a significant role in VC processes, while MLN4924 effectively prevents Pi-induced VC.
NEDD8 is conjugated PARP-1 in VC
To undercover which proteins undergo NEDD8-mediated modifications, we conducted immunoprecipitation-based proteomic analysis using an anti-NEDD8 antibody, followed by the affinity purification and liquid chromatography-mass spectrometry. In Pi-treated RVSMCs, NEDD8-conjugated proteins were isolated and the NEDDylation was confirmed through Coomassie blue staining. The gel bands that increased in the Pi lane and decreased in the lane with Pi and MLN4924 were cut out and subjected to LC-MS/MS to identify the NEDDylated proteins (Fig. 2a). We have identified NEDD8-conjugated candidate proteins, and used the PANTHER tool (PANTHERdb.org) to examine gene ontology (GO) analysis, which provided a comprehensive overview of the biological processes and molecular functions regulated by NEDDylation under the VC condition. Supplementary Fig. 2a illustrates the major cellular components, biological process, and molecular functions. According to the enriched GO annotation, the cellular components can be divided into two categories: cellular anatomical entity and protein-containing complex. Additionally, the identified NEDD8-conjugated candidate proteins were categorized into several major biological processes including cellular process, biological regulation and response to stimulus as well as diverse molecular function encompassing binding, catalytic activity, and transcription regulator activity. Further assessment of these identified proteins through bioinformatics analysis led to the selection of several proteins including HSP90, PARP-1, EEF2, NBR1 and DDB1 as NEDD8-conjugated candidate (Supplementary Fig. 2b). Among these NEDDylated candidate proteins, we decided to focuse on Poly (ADP-ribose) polymerase-1 (PARP-1) as a candidate protein in the context of vascular calcification. Although the roles of PARP-1 in vascular calcification were known, the significance of its posttranslational neddylation had not been extensively investigated. Therefore, we aimed to elucidate the role of neddylation of PARP-1 in the context of vascular calcification.
To validate the presence of NEDD8-conjugated PARP-1, we confirmed the interaction between PARP-1 and endogenous NEDD8. This interaction resulted in smear bands on the Western blots of Pi-treated VSMCs (Fig. 2b). Overexpression of NEDD8 led to the appearance of a higher molecular weight band for PARP-1, indicative of NEDDylation. This smear pattern was abolished when using NEDD8ΔGG, a conjugation-defective mutant due to a Gly-75/76 deletion (Fig. 2c). These findings established that poly-neddylation of PARP-1 with NEDD8 is intricately associated with vascular calcification.
Visualization of NEDD8-PARP-1 interaction
Proximity ligation assay (PLA) is useful for the visualization of interacting proteins32. The PLA assay revealed that NEDD8 binds to PARP-1 in the nucleus. However, this binding was disrupted when MLN4924 was treated (Fig. 2d). Furthermore, MLN4924 treatment attenuated Pi-induced PARP-1 neddylation and Runx2 expression (Fig. 2e).
NEDDylation of PARP-1 Affects PARP-1 Activity and VC
We found that PARP-1 NEDDylation was increased in vascular calcification. Given that PARP-1 is associated with vascular calcification and its enzymatic activation involves poly (ADP)-ribosylation (PARylation)28 and then PARP-1 catalyze the addition of poly (ADP-ribose) (PAR) to substrate proteins via cleavage of NAD+ 33, it is curious whether PARP-1 NEDDylation can affect the enzymatic activation of PARP-1 in association with vascular calcification. Thus, we explored whether PARP-1 NEDDylation affects its enzymatic activity in VC. PAR polymer expression was detected by its ability to bind to boronate shown in Fig. 2f. Pi treatment induced PAR polymer expression, evidenced by the smearing of boronate-conjugated precipitates (second lane). This smearing was diminished by MLN4924 treatment, which confirms its generation through NEDDylation. Cobb et al., reported that following DNA damage-induced VC, Runx2 undergoes PARylation dependent on PARP-1 activity34. As shown in Fig. 2f (bottom band), Runx2 was pulled down as a PARylated protein, which was blocked by MLN4924 treatment. Direct measurement of PARP-1 activity using a colorimetric assay (PARP universal colorimetric assay methods) revealed that Pi-induced PARP-1 activity was blunted by MLN4924 treatment (Fig. 2g). These findings underscore the dependency of PARP-1 activation on PARP-1 neddylation during the progression of VC.
MLN4924 mitigates VD3-induced VC
In pursuit of evaluating the therapeutic prospects of NEDDylation reversal in vascular calcification in vivo, we introduced a mouse model of vitamin D3-induced calcification and administered intraperitoneal injections of MLN4924 (10mg/kg) every other day for six days. The experimental timeline is outlined in Fig. 3a. For calcification assessment, we employed alizarin red S staining of the entire aorta. The results presented in Fig. 3b demonstrate that MLN4924 treatment effectively counteracted vitamin D3-induced VC when compared with the control group. Calcium levels were quantified in both arteries and serum of mice. Remarkably, administration of MLN4924 resulted in a substantial reduction in calcium deposition within the arteries of vitamin D3-treated mice, as shown in Fig. 3c. However, no significant alteration was observed in serum calcium levels in vitamin D3-treated mice upon MLN4924 treatment (Fig. 3d), Importantly, NEDD8-conjugated PARP-1 induced by vitamin D3, along with its enzymatic activity indicated by poly(ADP-ribose) expression were effectively suppressed by MLN4924 treatment (Fig. 3e, f). Additionally, MLN4924 administration attenuated PARP-1 activity in mice subjected to vitamin D3 injection (Fig. 3g). As visualized by Alizarin Red S staining, vitamin D3-induced calcium deposition, appeared as reddish regions in the aortic media wall, was conspicuously absent by MLN4924 administration (Fig. 3h). Moreover, the interaction of PARP-1 with NEDD8 induced by vitamin D3 was blunted by MLN4924 (Fig. 3i). Collectively, these findings firmly establish the efficacy of MLN4924 in ameliorating vascular calcification through the inhibition of PARP-1 NEDDylation. Furthermore, these findings underscore MLN4924 as a promising candidate for potential therapeutic interventions against VC.
Cbl-b mediates NEDD8 binding to PARP-1 in VC
Given that many posttranslational modifications involve the final conjugation of small molecular moieties, facilitated by E1, E2, and E3 ligases35, with target specificity often guided by E3 ligases, we directed our attention toward deciphering the specific E3 ligase responsible for PARP-1 neddylation. Mammalian cells house several hundred E3 ligases, with CBL-3, RBX1, and Fbxo11 serving as representative E3 ligases that oversee NEDDylation12. Consequently, we next embarked on identification of the specific E3 ligase involved in PARP-1 neddylation in the context of VC. For the identification of dysregulated E3 ligase, we employed our previous microarray analysis (GSE74755) results of rat VSMCs treated with Pi30. (Supplementary Fig. 3a). Baculoviral IAP Repeat Containing 3 (Birc3), Casitas B–lineage lymphoma protein b (Cbl-b), Ring-Box 1 (Rbx1), Mouse double minute 2 homolog (MDM2), Mouse double minute 4 homolog (MDM4), Ring Finger protein 7 (Rnf7), and Ring Finger Protein 111 (Rnf111) were dysregulated. Among candidates, we previously reported that MDM2 mediates ubiquitination of HDAC1 during the VC process30, Thus we have excluded MDM2 as a E3 ligase for the neddylation process of PARP-1. Next, we quantified mRNA level changes of the dysregulated E3 ligase genes using quantitative real-time PCR in Pi-induced VSMCs. Notably, Birc3, Cbl-b, Rbx1, and Rnf111 were significantly upregulated (Supplementary Fig. 3b). Subsequently, using siRNA to knockdown these E3 ligases, we assessed the effect on PARP1- neddylation. Remarkably, only the deletion of Cbl-b effectively attenuated PARP-1 neddylation, whereas the loss of other E3 ligases like Birc3, Rbx1, and Rnf111, had negligible effects on PARP-1 neddylation (Fig. 4a and Supplementary Fig. 4a). PLA further showed the dissociation of NEDD8 from PARP-1 upon Cbl-b knockdown (Fig. 4b). Immunoprecipitation analysis confirmed the interaction between Cbl-b and PARP-1 (Fig. 4c and Supplementary Fig. 4b). Knocking-down of Cbl-b had no effect on the viability of A10 cells in Pi. (Supplementary Fig. 4c). Given the observation of PARP-1 neddylation influencing its activity in vascular calcification, we proceeded to explore whether Cbl-b played a role in regulating PARP-1 activity. Indeed, Cbl-b knockdown significantly dampened Pi-induced PARP-1 activity in VSMCs (Fig. 4d).
Role of Cbl-b in VC and its mechanistic Insights
Because the role of Cbl-b in VC had not been previously reported, we proceeded to evaluate its effects by overexpression of HA-Cbl-b in VSMCs. Intriguingly, overexpression of Cbl-b significantly potentiated Pi-induced calcium deposition in a dose-dependent manner (Supplementary Fig. 4d). Furthermore, protein expression of Cbl-b was increased in a time-dependent manner bothe in VSMCs treated with Pi (Supplementary Fig. 4e) and vitamin D3-induced VC mouse models (Supplementary Fig. 4f). Notably, knockdown of Cbl-b significantly reduced the intensity of alizarin red S staining in Pi-induced VC (Supplementary Fig. 4g, h). Consistently, the augmentation of calcium deposition induced by Pi was blocked by si-Cbl-b transfection (Fig. 4e). Next, we evaluated the effects of Cbl-b knockdown by siRNA on VC in mouse models. We intravenously injected 50µg of si-Cbl-b, combined with a transfection reagent, twice during the VC induction periods following three consecutive days of vitamin D3 injection into mice (Fig. 4f). The knockdown of Cbl-b was successfully achieved in various tissues, including the aorta, brain, kidney, liver and lung (Supplementary Fig. 5a). This led to a reduction in VD3-induced calcium accumulation in the aorta, but not in serum levels (Fig. 4g and Supplementary Fig. 5b). Additionally, the loss of Cbl-b in VC resulted in blunted PARP-1 neddylation and Runx2 expression (Fig. 4h). As anticipated, the mineralization of vascular smooth muscle from the aorta in the Vitamin D3 group was inhibited following the injection with si-Cbl-b (Fig. 4i). Collectively, these results underscore the pivotal role of the E3 ligase Cbl-b in mediating PARP-1 neddylation and its regulatory role in VC.
Cbl-b E3 ligase activity governs PARP-1 Neddylation in VC
Given that Cbl-b E3 ligase belongs to the RING finger family, which engages E2 enzymes to facilitate substrate ubiquitination36, we examined whether E3 ligase activity of Cbl-b is essential for PARP-1 neddylation. It was previously reported that C373 and W400 of Cbl-b are critical for its E3 ligase activity37. We generated these functionally inert forms of Cbl-b that lacks E3 ligase activity and next subjected them to immunoprecipitation with anti-PARP-1 antibody. As anticipated, high Pi exposure potentiated PARP-1 neddylation (Fig. 5a) and consequently, PARP-1 activity (Fig. 5b). These effects were pronounced upon transfection with Cbl-b WT, but not Cbl-b C373A or Cbl-b W400A. Interestingly, contrary to Cbl-b WT, both Cbl-b C373A and Cbl-b W400A failed to increase in mineralization and calcium deposition (Fig. 5c, d). Besides, Cbl-b WT and Cbl-b E3 ligase activity dead mutants (C373A and W400A) did not alter cell viability in Pi with A10 cells (Supplementary Fig. 6a)
Next, we interrogated whether blocking of E3 ligase activity via a peptide spanning crucial residue C373 and W400 could impede PARP-1 neddylation and, consequently, vascular calcification. Sequence analysis of different species revealed that the C373 and W400 site of Cbl-b are highly conserved (Supplementary Fig. 6b). To visualize the localization of these peptides, fluorescein isothiocyanate-conjugated nuclear localization signal (NLS) sequence 38 were added. The synthetic peptide C373, but not W400, effectively entered the nucleus in VSMCs (Supplementary Fig. 6c). Cbl-b C373 peptide did not affect cell viability in Pi with A10 cells (Supplementary Fig. 6d). In the subsequent experiments, we utilized C373 peptide. NEDD8 conjugated with PARP-1 is cleaved by treatment of Cbl-b C373 peptide in VSMCs under Pi conditions (Fig. 5e). Moreover, both PARP-1 neddylation and its activity were also blunted by C373 in Pi-induced VC (Fig. 5f, h). Further corroborating these findings, the Cbl-b C373 peptide significantly attenuated the propensity for calcium deposition and alizarin red S staining in Pi-treated VSMCs (Fig. 5i, k). Taken together, these observations underscore the pivotal role of Cbl-b E3 ligase activity, particularly at residue C373, in promoting PARP-1 neddylation within the context of vascular calcification. Additionally, the C373-spanning peptide emerges as a potential candidate for thwarting vascular calcification, offering novel therapeutic avenues for its prevention or treatment.
Alleviation of Vitamin D3-induced VC through Cbl-b C373 peptide
To further validate the therapeutic potential of the Cbl-b C373-blocking peptide, we examined its effects in VD3-induced VC models. Mice were intraperitoneally administered with Cbl-b C373 peptide (1mg/kg/day) every two day, following a single injection of vitamin D3 in mice. After 9-day of experimental period, mice were sacrificed and assessed for VC effects (Fig. 6a). The uptake of the C373 peptide into the mouse aorta was observed through FITC fluorescence, as shown in Supplementary Fig. 7. Treatment with Cbl-b C373 peptide effectively attenuated the formation of calcified nodules in aorta of vitamin D3-induced VC, compared to the scramble group (Fig. 6b). Calcium content analysis demonstrated that increased calcium deposition induced by vitamin D3 was significantly counteracted by Cbl-b C373 peptide treatment in aorta, but not in serum (Fig. 6c, d). Remarkably, the enhancement of PARP-1 neddylation and PARP-1 enzyme activities such as PAR polymer, instigated by vitamin D3, were effectively mitigated by Cbl-b C373-blocking peptide within the aorta, compared to the scramble control in mice (Fig. 6e, f).
Evidently, the distinct calcium deposition characteristic of vitamin D3 presence was entirely abrogated following the administration of the Cbl-b C373 peptide (Fig. 6g, left panel). Additionally, the typical interaction between PARP-1 and NEDD8 elicited by vitamin D3 was efficiently hindered by the Cbl-b C373 peptide, as illustrated by Fig. 6g (right panel). Taken together, these findings underscore the efficacy of the Cbl-b C373-blocking peptide in vitamin D3-induced vascular calcification through the disruption of PARP-1 neddylation.
PARP-1 neddylation is counteracted by NEDP-1 in VC
The dynamic nature of NEDDylation is well-established, encompassing a reversible process facilitated by NEDP-1, an NEDD8-specific protease 139. In light of this, we embarked on investigating whether NEDP-1-mediated de-NEDDylation could effectively reverse PARP-1 neddylation in VC. Thus, we designed a study centered on the NEDP-1. As anticipated, the introduction of ectopically expressed Cbl-b augmented the conjugation of NEDD8 with PARP-1. Nevertheless, this enhancement was nullified by the overexpression of NEDP-1 (Fig. 7a). For deeper insights, we scutinized the effect of NEDP-1 on PARP-1 neddylation and Poly (ADP-ribose)ylation in Pi-induced VSMCs. Remarkably, the overexpression of NEDP-1 resulted in the suppression of Pi-induced PARP-1 neddylation and poly(ADP)-ribosylation (Fig. 7b, c). Furthermore, the increase in the activity of PARP-1 evident in the Pi-induced VC model was effectively down-regulated by the overexpression of NEDP-1 (Fig. 7d).
Until now, the role of NEDP-1 in vascular calcification has not been reported. Therefore, we sought to investigate this. The expression of NEDP-1 expression gradually decreased in vascular calcification both in vitro and in vivo models (Supplementary Fig. 8a, b). NEDP-1 overexpression did not affect cell viability in VSMCs (Supplementary Fig. 8c). Notably, the transient overexpression of NEDP-1 in VSMCs via transfection significantly exerted a dose-dependent migratory effect on Pi-induced calcium deposition (Fig. 7e and Supplementary Fig. 8d) and mineralization in VSMCs (Fig. 7f, g). Conversely, the depletion of NEDP-1 significantly amplified the accumulation of calcium stimulated by Pi in a dose dependent manner (Supplementary Fig. 8e, f).
Given the function of NEDP-1 in dissociating NEDD8 from target proteins, we postulated that NEDP-1 might exert an antagonistic effect against the functional interplay of neddylated PARP-1 and E3 ligase activity of Cbl-b in the milieu of VC. Armed with this supposition, we proceeded to investigate whether functionality of either PARP-1 or Cbl-b could be effectively counteracted by NEDP-1. Evidently, the overexpression of either PARP-1 or Cbl-b led to an elevation of calcium content and mineralization in the presence of Pi in VSMCs. Notably, this exacerbation was significantly dampened by the simultaneous overexpression of NEDP-1 in a dose dependent manner (Fig. 7h-k). Collectively, these findings underscore the role of NEDP-1 as a pivotal conteractive element against PARP-1 neddylation in VC.