Deletion of IκB‐Kinase β in Smooth Muscle Cells Induces Vascular Calcification Through β‐Catenin–Runt‐Related Transcription Factor 2 Signaling

Background Vascular calcification was previously considered as an advanced phase of atherosclerosis; however, recent studies have indicated that such calcification can appear in different situations. Nevertheless, there has been a lack of mechanistic insight to explain the difference. For example, the roles of nuclear factor‐κB, a major regulator of inflammation, in vascular calcification are poorly explored, although its roles in atherosclerosis were well documented. Herein, we investigated the roles of nuclear factor‐κB signaling in vascular calcification. Methods and Results We produced mice with deletion of IKKβ, an essential kinase for nuclear factor‐κB activation, in vascular smooth muscle cells (VSMCs; KO mice) and subjected them to the CaCl2‐induced aorta injury model. Unexpectedly, KO mice showed more calcification of the aorta than their wild‐type littermates, despite the former's suppressed nuclear factor‐κB activity. Cultured VSMCs from the aorta of KO mice also showed significant calcification in vitro. In the molecular analysis, we found that Runt‐related transcription factor 2, a transcriptional factor accelerating bone formation, was upregulated in cultured VSMCs from KO mice, and its regulator β‐catenin was more activated with suppressed ubiquitination in KO VSMCs. Furthermore, we examined VSMCs from mice in which kinase‐active or kinase‐dead IKKβ was overexpressed in VSMCs. We found that kinase‐independent function of IKKβ is involved in suppression of calcification via inactivation of β‐catenin, which leads to suppression of Runt‐related transcription factor 2 and osteoblast marker genes. Conclusions IKKβ negatively regulates VSMC calcification through β‐catenin–Runt‐related transcription factor 2 signaling, which revealed a novel function of IKKβ on vascular calcification.

V ascular calcification was previously considered as a degenerative process seen in advanced atherosclerosis lesions, but recent studies have indicated that such calcification can appear in a different manner. Calcification not only accompanies with atherosclerosis but also accompanies diabetes mellitus or renal failure in the absence of atherosclerosis. 1,2 Moreover, recent clinical studies have revealed that statins, a medicine widely used for atherosclerosis, promote coronary calcification, [3][4][5][6][7][8] which clearly indicates that calcification stands on a different basis from atherosclerosis. 9 However, the difference has not been fully explained because the mechanism of vascular calcification is poorly understood. For example, the roles of nuclear factor-jB (NF-jB) in the formation and progression of calcification have not been adequately assessed, although its roles in atherosclerosis have been well studied. 10,11 NF-jB is a family of transcription factors that plays critical roles in inflammation and atherosclerosis, and reports have indicated that it is a target of statins. 12,13 NF-jB generally exists as a homodimer or heterodimer in the cytosol that is bound to the inhibitor of jB (IjB). 14 In response to a wide variety of stimuli, including inflammatory cytokines, IjB is phosphorylated and degraded via the ubiquitin pathway, which is followed by NF-jB translocation to the nucleus and activation of transcription. The serine phosphorylation of IjB is mediated by a large multiunit complex containing 2 catalytic subunits, IKKa and IKKb, and the regulatory subunit IKKc or NF-jB essential modulator (NEMO). 15 Of these subunits, IKKb is the essential kinase that mediates IjB phosphorylation. 15 In this study, we produced mice with a deletion of IKKb in vascular smooth muscle cells (VSMCs), a cell type contributing to calcification, 16 and subjected them to CaCl 2 -mediated aortic injury, which is an established animal model for vascular calcification. 17,18 Unexpectedly, we observed more calcification in the aorta of IKKb-deleted mice and found the activation of b-catenin and Runt-related transcription factor 2 (Runx2) as a possible mechanism. Furthermore, we present, for the first time, that the kinase-independent function of IKKb has an effect on the inactivation of b-catenin, which leads to suppressing the calcification of VSMCs. These results indicate the protective effect of IKKb on vascular calcification, which at least partly explains the difference between the smooth muscle calcification and atherosclerosis observed in clinical settings.

Methods
The data, analytic methods, and study materials will not be made available to other researchers for purposes of reproducing the results or replicating the procedure because some materials are used for other unpublished projects.

Knock-In Mice
To generate conditional IKKb (K44A) expressing mice, a CAG promoter, a loxP flanked STOP cassette, and IKKb (K44A) cDNA were inserted into the mouse Gt(Rosa) 26Sor locus by homologous recombination. A 3.4-kb upstream DNA fragment and a 4.3-kb downstream DNA fragment from the unique XbaI site in intron 1 of Gt(Rosa) 26Sor locus were amplified by polymerase chain reaction (PCR) from RENKA ES cell genomic DNA and used as the homology arms for the targeting vector. To construct the expression unit, a chemically synthesized DNA sequence containing a DYKDDDDK-tagged mouse IKKb (K44A), a rabbit b-globin polyadenylation signal, and a VloxP sequence were combined with an frt-flanked PGK_neo cassette (phosphoglycerate kinase 1 promoter-driven neomycin-resistant gene). At the 5 0 region of this expression unit, a CAG promoter and a loxP flanked STOP cassette containing chloramphenicol acetyltransferase cDNA and Vlox43L were inserted. Then, the promoter-expression unit was subcloned between 2 homology arms. As a negative selection marker, an MC1 promoter-driven diphtheria toxin A cassette was placed at the 5 0 region of the homology arm. The resulting knock-in targeting vector contains MC1 promoter-driven diphtheria toxin A, a 3.4-kb 5 0 homologous arm, a CAG promoter, the first loxP site, chloramphenicol acetyltransferase cDNA, polyA, Vlox43L, an frt-flanked PGK_neo cassette, the second loxP site, a DYKDDDDK tag, IKK2 (K44A) cDNA, bGH polyA, VloxP, and a 4.3-kb 3 0 homology arm. This vector was linearized and introduced into RENKA ES cells (C57BL/6 genetic background) by electroporation. After selection using geneticin, the resistant clones were isolated and their DNA

Clinical Perspective
What Is New?
• Inhibition of IKKb, an essential kinase for nuclear factor-jB activation, in vascular smooth muscle cells, augments vascular calcification. • IKKb has a kinase-independent function, and it is involved in the suppression of vascular calcification.
What Are the Clinical Implications?

CaCl 2 -Mediated Abdominal Aorta Injury
An abdominal aorta injury was induced in 8-to 14-week-old KO and WT male mice by periaortic application of CaCl 2 , as described previously, 17,18 with minor modifications. We started with a conventional protocol of 20 minutes CaCl 2 treatment, but too much dilatation and calcification seen in WT aorta interfere with the evaluation of the difference between WT and KO aorta. Therefore, we reduced the CaCl 2 incubation period to 10 minutes. Even in that protocol, 2 weeks after CaCl 2 application, the aortas in both types of mice were significantly dilated when compared with those in the sham-operated on mice. Mice were anesthetized by IP injection with pentobarbital (100 mg/kg) and then underwent a laparotomy. An incision of %2 cm was made, and the small and large intestines were moved away to expose the infrarenal aorta. Small pieces of gauzes soaked with 0.6 mol/L CaCl 2 were placed circumferentially on the infrarenal aorta for 10 minutes. Then, the gauze was removed and the aorta was rinsed with 0.95% NaCl and wiped with a cotton swab. NaCl (0.95%) was substituted for CaCl 2 in sham-operated on mice. The incision was closed, and mice were placed on a warm plate until they had recovered from anesthesia. After the surgical procedure, the animals were allowed free access to food and water until the day of euthanasia.

Aorta Collection and Size Measurement
Two weeks after the operation, mice were euthanized with an overdose of pentobarbital. The vascular system was flushed with cold PBS from the left ventricle. The abdominal aorta was carefully dissected under a microscope, cleaned of extraneous tissue, and then photographed for diameter measurement.

Histology and Immunohistochemistry
Approximately 2 to 3 mm of infrarenal aorta was fixed with 4% paraformaldehyde overnight, embedded in paraffin, cut into 4-lm cross sections. Then, cross sections were deparaffinized 3 times in xylene and dehydrated in 80%, 90%, and 100% ethanol for 5 minutes each. Elastin was visualized using the fluorescence properties of hematoxylin and eosin. 20 Alizarin Red staining was performed to detect calcification for both the in vivo and in vitro samples. For immunohistochemical staining, cross sections were boiled in 10 mmol/L citrate buffer for antigen retrieval, followed by avidin/biotin. Sections were then incubated with primary antibodies that recognize SM22a, IKKb, Runx2, and active bcatenin overnight at 4°C. The secondary antibody IgG/biotin goat anti-rabbit or goat anti-mouse was incubated for 1 hour at room temperature, followed by diaminobenzidine. Sections were visualized using a BIOREVO BZ-9000 Keyence microscope.

Cell Culture
VSMCs from the aorta of littermate mice were isolated on the basis of a protocol described by Villa-Bellosta and Hamczyk. 21 Briefly, aortas were perfused with cold PBS. The aorta from the aortic arch to the iliac bifurcation was isolated and incubated in collagenase type 2 solution for 15 minutes at 37°C in 5% CO 2 . Then, adventitia was rolled off the medial layer. Medial layers were minced and further incubated for 2 hours at 37°C with agitation. Resuspended cells were centrifuged for 5 minutes at 300g. The supernatant was aspirated, and the cell pellet was resuspended in an appropriate volume of DMEM supplemented with 10% fetal bovine serum (FBS). Finally, cell suspensions were cultured in collagenase type 1 coated culture dishes and maintained in DMEM supplemented with 10% FBS, 100 U/mL of penicillin, and 100 lg/mL of streptomycin in a 5% CO 2 /water-saturated incubator at 37°C. Each kind of VSMCs was isolated from 1 mouse of KO, KA, KD, and littermate WT, and individual samples that used in vitro experiments were harvested from separately cultured dishes of cells.

Western Blotting
VSMCs were harvested from culture dishes using cell scraper tools. Protein extraction was performed on ice and using a cell lysis buffer (9803S; Cell Signaling) with a protease inhibitor cocktail. After the total protein concentration was measured, the same amount of extracted protein was loaded for SDS-PAGE immunoblot analysis and transferred to nitrocellulose membranes. Next, membranes were blocked with 5% nonfat milk or 5% BSA in Tris buffer solution containing 0.05% Tween-20 for 1 hour at room temperature, with gentle agitation. The membrane was immunoblotted with primary antibody overnight at 4°C. The membrane was washed 3 times with Tris buffer solution containing 0.05% Tween-20 for 10 minutes each. This was followed by a 1-hour incubation with a horseradish peroxidase-conjugated secondary antibody. Blots were visualized using the ECL Western Blotting Detection kit (GE Healthcare). Each band was normalized by the corresponding value of b-actin or GAPDH as an internal control. Densitometric analysis was performed by ImageJ Software. All experiments were performed at least in triplicate.

Quantitative Real-Time PCR
Total RNA was extracted from cultured VSMCs using an RNAqueous total RNA isolation kit (AM1912; Thermo Fisher Scientific K.K., Yokoyama, Japan) according to the manufacturer's instructions. cDNA was generated from total RNA using The High Capacity RNA-to-cDNA Kit (4387406). Quantitative PCR was performed using the QuantiTect SYBR Green PCR Kit and real-time PCR system (StepOnePlus; Applied Biosystems Japan, Tokyo, Japan). Relative mRNA level of ostrix, alkaline phosphatase, and osteocalcin was normalized to the Rn18s mRNA level in the same sample.

Statistical Analysis
Experimental results are expressed as the meanAESD. Groups were compared using the 2-tailed Student t test or 1-and 2way ANOVA, followed by the Tukey test. P<0.05 was considered statistically significant.

Generation of Mice With Deletion of IKKb in VSMCs
To inactivate NF-jB signaling in VSMCs, we produce mice with a deletion of IKKb selectively in smooth muscle cells (KO mice), as previously described. Mice were born according to mendelian frequency, and there was no significant difference in growth between WT and KO mice ( Figure 1A). Immunohistochemistry analyses indicated that IKKb was successfully knocked out in SM22a-positive VSMCs in IKKb knockout mice ( Figure 1B). Furthermore, we performed cultured VSMCs from the aorta of KO and littermate control mice and confirmed that IKKb is deleted in VSMCs of KO mice ( Figure 1C).

Suppression of NF-jB Activity in IKKb-Deficient VSMCs
Next, we examined the NF-jB activity in IKKb-deficient VSMCs. IKKb activates NF-jB via the phosphorylation and degradation of IjB and eventually phosphorylates p65, a major subunit of NF-jB. 15 VSMCs of the aorta from KO and control mice were stimulated by interleukin-1b to activate NF-jB signaling, and we found that the phosphorylation of p65 was significantly suppressed in KO cells ( Figure 1D). These results confirm that NF-jB is suppressed in VSMCs from KO mice.

Deletion of IKKb in VSMCs Promotes Aortic Calcification in CaCl 2 -Induced Abdominal Aorta Injury
To elucidate the roles of IKKb/NF-jB signaling in vascular calcification, we subjected WT and KO mice to a CaCl 2induced aorta injury model. Unexpectedly, KO mice showed more significant calcification in aorta, and those aortas were more dilated compared with WT mice (Figure 2A through 2C). Histologically, the aorta of KO mice revealed extensive damage in the medial layer and more disruption and elongation of the medial elastic lamellae, possibly because of the significant calcification ( Figure 2D). Consistent with these in vivo results, Alizarin Red staining of 4-week cultured VSMCs showed remarkable calcium deposition in KO VSMCs but not in WT VSMCs ( Figure 2E).

Deletion of IKKb Induces Transdifferentiation of VSMCs to Osteoblast-Like Cells via Activating the b-Catenin-Runx2 Signaling Pathway
To determine the reason why injured aortic wall and cultured VSMCs of KO mice showed more calcium deposition, we investigated the underlying molecular mechanism. Recently, it is widely accepted that pathways controlling vascular calcification and bone remodeling are similar. 22,23 Several studies have shown that the activation of b-catenin, which is the prototypic mediator and marker of canonical Wnt signaling, is necessary for osteoblast formation and is involved in the calcification of VSMCs by promoting the expression of osteocyte genes, which leads to the transdifferentiation of VSMCs to osteoblast-like cells. 24,25 Therefore, we investigated the activity of b-catenin in cultured VSMCs and found that the expression of the nonphosphorylated active form of b-catenin in the whole cell lysate of VSMCs was significantly augmented in KO cells when compared with that in WT cells ( Figure 3A). We also assessed active b-catenin in nuclear extraction to confirm its translocation to the nucleus for osteocyte gene expression and found that more active b-catenin was located in the nucleus of KO VSMCs (Figure 3B). Next, we examined Runx2 expression in cultured VSMCs. Runx2 is a major transcription factor playing a crucial role in VSMC transdifferentiation into osteoblast-like cells 26,27 by regulating osteoblast genes, and it is regulated by b-catenin. 24 Consistent with the b-catenin results, the expression of Runx2 was significantly more increased in KO VSMCs compared with that in WT VSMCs in both nucleus and cytoplasm ( Figure 3B). These results indicate that the ablation of IKKb in VSMCs accelerates calcification through upregulation of b-catenin and Runx2 signaling.

Constitutively Active IKKb Protects VSMCs From Calcification via the Downregulation of b-Catenin and Runx2 Expression
To confirm the mechanism previously described, we examined the effect of kinase-active IKKb on vascular calcification as opposite situation. We produced mice with a knock-in of kinase-active IKKb specifically in VSMCs (KA mice), as described previously, 19 and subjected them to a CaCl 2induced aorta injury model. Histologically, the aorta of KA mice showed an intact medial layer with negligible disruption in the medial elastic lamellae ( Figure 4A), and no calcification was observed in the medial layer of the aorta after CaCl 2 ( Figure 4B). Western blot analysis of cultured VSMCs from KA mice confirmed the upregulation of IKKb and phosphorylated p65 ( Figure 4C and 4D), indicating the constitutive activation of NF-jB in cultured VSMCs from KA mice. Consistent with in vivo results, Alizarin Red staining of 4-week cultured KA VSMCs showed an absence of calcium deposits ( Figure 4E). Next, we examined b-catenin and Runx2 and found a significant decrease in both active and total b-catenin and Runx2 in KA-VSMCs when compared with that in WT ( Figure 5A). Furthermore, nuclear location of Runx2 was significantly decreased in KA VSMCs ( Figure 5B). These results indicated that overexpression of kinase-active IKKb negatively regulates b-catenin-Runx2 signaling in KA VSMCs, which were opposite to the results seen in KO VSMCs.

IKKb Modulates Ubiquitination of b-Catenin in VSMCs
According to previous reports, b-catenin is inactivated by ubiquitination via the ubiquitin-proteasome pathway after it is phosphorylated. 28 Thus, we examined ubiquitination of b-catenin on WT, KO, and KA-VSMCs. Cells were pretreated with proteasome inhibitor MG132 to block ubiquitin-induced degradation by proteasomes and subjected to ubiquitination assay. The results showed that the deletion of IKKb in VSMCs reduced the ubiquitination of b-catenin and, conversely, it was augmented in VSMCs with constitutively active IKKb ( Figure 5C). These results were consistent with the change of active b-catenin in KO and KA VSMCs. We also histologically examined the change of b-catenin and Runx2 in CaCl 2 -injured aortas of WT, KO, and KA mice. As  shown in Figure 5D, CaCl 2 induces staining of active b-catenin and Runx2 in nuclei of smooth muscle layers in KO aorta compared with that in sham group. In contrast, no apparent staining of active b-catenin and Runx2 was observed in both KA and WT aorta, even after CaCl 2 -induced aortic injury. These results are consistent with observations seen in cultured VSMCs.

There is No Apparent Relationship Between Apoptosis and VSMC Calcification
Several reports have shown that apoptosis of VSMCs can induce vascular calcification. 29,30 To investigate the involvement of apoptosis in VSMC calcification induced by IKKb deletion, we performed terminal deoxynucleotidyl transferase dUTP nick end labeling staining on the aortas. There was no remarkable difference in number of apoptotic cells between WT and KO aortas after CaCl 2 treatment. In contrast, terminal deoxynucleotidyl transferase dUTP nick end labeling staining revealed a remarkable increase of apoptotic cells after CaCl 2induced aortic injury in KA mice, compared with WT and KO mice ( Figure 6A). To understand the molecular mechanisms by which overexpression of kinase-active IKKb induced VSMC apoptosis, we examined caspase-9 and caspase-3, crucial mediators of apoptosis, in cultured VSMCs. Western blot analysis of cultured VSMCs demonstrated that expression of both cleaved caspase-9 and cleaved caspase-3 was significantly increased in KA VSMCs compared with those in WT and KO VSMCs ( Figure 6B). Despite the increase of apoptosis at the medial layer of KA mice after CaCl 2 application, no remarkable calcification was observed, as previously shown.
Collectively, these observations suggest that the apoptosis is not the primary cause of VSMC calcification, at least in these mouse models.

Kinase-Independent Function of IKKb Is Involved in Regulating VSMC Calcification
Kinase-dependent function of IKKb in the activation of NF-jB is well known; however, recent reports showed that IKKb has other regulatory functions, independently from NF-jB activation. 31 Furthermore, we previously reported that IKKb regulates other proteins, even in a kinase-independent manner. 31,32 Results previously shown clearly indicated that IKKb negatively regulates calcification, but it is difficult to conclude whether either or both of kinase-dependent or kinase-independent function of IKKb is involved in it, as far as we compare among WT, KO, and KA-IKKb. Indeed, interleukin-1b stimulation did not induce any change in Runx2 or b-catenin in WT cells ( Figure 7A), which implicates the possible involvement of kinase-independent function of IKKb in regulating these molecules.
To make it clear, we produced mice expressing kinasedead mutant (KD) of IKKb in VSMCs on the background of IKKbf/f sm22Cre +/À (KO) mice, according to the previous report showing that point mutation of an amino acid inactivates kinase function 19 (Figure 7B), and we examined cultured VSMCs from KD mice. As shown in Figure 7C, IKKb is highly more expressed in KD VSMCs compared with KO VSMCs. However, the level of phosphorylated p65 in KD cells is equivalent to that in KO cells, which indicated that the kinase activity is successfully lost in KD cells. Surprisingly, Alizarin Red staining showed no remarkable increase of calcium deposit in KD VSMCs compared with that in WT VSMCs ( Figure 7D), contrary to KO VSMCs. Considering the significant calcification in KO VSMCs compared with that in WT VSMCs, it is clearly indicated that the kinase-independent function of IKKb has a suppressive effect on calcification. To confirm the difference in calcification among WT, KO, KA, and KD VSMCs, we performed real-time PCR to examine the expression of major osteoblast marker genes regulating transdifferentiation of VSMCs to osteoblast-like cells. 33,34 As shown in Figure 7E, osterix, alkaline phosphatase, and osteocalcin were significantly upregulated only in KO VSMCs.
Next, we examined b-catenin and Runx2 signaling in KD VSMCs. Consistent with the previous results, active b-catenin and Runx2 in KD VSMCs were significantly downregulated compared with KO VSMCs, clearly indicating that the kinaseindependent function of IKKb is involved in the downregulation of b-catenin ( Figure 8A). We also noticed that active b-catenin and Runx2 in KD VSMCs were significantly more expressed than those in KA VSMCs, despite more expression of IKKb in KD cells, which indicated that kinase-dependent function of IKKb also has a suppressive effect on calcification. Furthermore, we performed b-catenin ubiquitination assay in KD VSMCs and found that b-catenin ubiquitination was upregulated in KD VSMCs ( Figure 8B), which supported the observation that b-catenin is inactivated in KD VSMCs. To confirm the observations further, KO VSMCs were transfected with IKKb-KD and IKKb-KA plasmids. KO VSMCs transfected with IKKb-KD or IKKb-KA plasmids showed a significant decrease in Runx2 expression ( Figure 8C), which is consistent with results seen in transgenic VSMCs.

Discussion
Recent clinical studies have indicated that vascular calcification is different from atherosclerosis, but the mechanisms of vascular calcification have not been explored adequately. In this study, we show that the deletion of IKKb, an essential kinase for NF-jB activation, in VSMCs promotes vascular calcification. We also show that it is mediated by the upregulation of b-catenin, Runx2, and transcription of osteoblast genes. Furthermore, we found that kinaseindependent function of IKKb is involved in the regulation of such signaling. Considering that the accelerating roles of NF-jB in atherosclerosis have been well analyzed, this finding would mechanistically explain the difference between atherosclerosis and calcification.
The difference between atherosclerosis and vascular calcification has been only empirically suggested, but recent clinical studies have clearly demonstrated it. [4][5][6][7] For instance, Puri et al reported that statins, the most powerful medicine against atherosclerosis, unexpectedly accelerate vascular calcification. 8,35 However, no mechanistic insights have been reported to explain this finding. Considering statin has the activating effect on b-catenin 36 and the suppressive effect on NF-jB, 12,37 the observations in this study would account for it, at least in part.
It was also an unexpected observation that more aortic dilatation was seen in KO mice. It has been reported that aortic calcification induces mechanical wall stress by decreasing biomechanical stability, which also associated with destruction and elongation of the elastic lamella in the medial layer. 38 In addition, mechanical wall stress itself causes expansion and dilatation of the aortic wall. 39 These mechanisms would account for the dilatation of calcified aorta in KO mice.
The molecular mechanisms of calcification in vessel walls and bone are reported to be similar, as indicated by the observation that the phenotypic transdifferentiation of VSMCs into the osteoblast-like cells is seen in vascular calcification. 22,23 Indeed, b-catenin and Runx2, major regulators of bone-related genes, such as osterix (OSX), alkaline phosphatase (ALP), and osteocalcin (OCN), promote the transdifferentiation of VSMCs into osteoblast-like cells. [24][25][26][27] Therefore, the changes of b-catenin, Runx2, and osteoblast marker genes seen in KO VSMCs fundamentally support the histological observation of vascular calcification in vivo and  (t test, n=3). B, Kinase-dead IKKb transgene construct. C, Representative Western blots and densitometric analysis of IKKb/GAPDH, phosphorylated p65/total p65 expression in VSMCs isolated from kinase dead (KD) and IKKb knockout (KO) mice. Bars represent the meanAESD (t test, n=3). D, Representative microscopy images of Alizarin Red staining of 4-week cultured VSMCs isolated from WT and KD mice and quantification of VSMC calcification. Calcification was quantified by ImageJ software. Graph presented is the percentage of positively stained area in the total area randomly selected. Bars represent the meanAESD (t test, n=6). E, Results of quantitative real-time PCR (qRT-PCR) for the expression of various osteogenic-related genes (osterix, alkaline phosphatase [ALP], and osteocalcin) in WT, KD, KO, and kinase active IKKb (KA) cells that were normalized to the Rn18s mRNA level. WT samples used in RT-PCR were from littermate of KO mouse. Cells used in qRT-PCR were cultured for 2 weeks in normal medium with 10% fetal bovine serum. Bars represent the meanAESD (1way ANOVA, n=4). **P<0.01, ***P<0.001. in vitro from the molecular level. Furthermore, considering the similarity between bone formation and vascular calcification, it is interesting that previous reports on bone research showed that statin can enhance bone formation. [40][41][42] Although the roles of b-catenin in vascular calcification and bone formation have been well analyzed, as previously mentioned, those of IKKb/NF-jB signaling have not been adequately unveiled to date. In the bone research, it was reported that the deletion of IKKb significantly enhanced bone formation by activating b-catenin ubiquitination and degradation. 43 As its mechanism, Lamberti et al reported that IKKb interacts with and is able to phosphorylate b-catenin, 44 and Chang et al reported that IKKb promotes its ubiquitination. 43 In this study, the nonphosphorylated active form of b-catenin is suppressed and its ubiquitination is augmented in KD VSMCs. The detailed mechanism of how IKKb regulates b-catenin in a kinase-independent manner is still unclear and is to be unveiled. However, this finding would be of high importance considering that kinase-independent function of IKKb has never reported, except by us. 32 b-Catenin is the big target of interest, and it has been researched in many fields, including cancer, fibrosis diseases, and vascular calcification.
Considering that kinase-independent function of IKKb regulates calcification, as shown in this study, the regulation of protein expression of IKKb should be more carefully recognized. Interestingly, IKKb is the target of miR148a, miR503, and miR199a, [45][46][47] which are upregulated micro-RNAs in vascular calcification. 48,49 The suppression of such microRNAs targeting IKKb might be a new strategy against vascular calcification.
The findings in this study explain not only the difference between atherosclerosis and calcification but also the disappointing results reported in clinical studies that examined the effect of aspirin on calcification. Aspirin has a suppressive effect on inflammatory responses, including NF-jB. Indeed, Yin et al reported that aspirin is a specific inhibitor of IKKb. 50 Therefore, people expect that aspirin use would have a protective effect against calcification; however, recent clinical studies have found that coronary calcification was associated with aspirin use 51 and that aspirin had no protective effect on the onset of calcification in either coronary arteries or aortic valves. The anticalcification effect of IKKb revealed in this study might explain these results. Furthermore, we have previously reported that long-term and high concentration of aspirin can suppress the protein expression of IKKb, 52 which might indicate possible modification of kinase-independent function of IKKb by aspirin. These new insights can lead to unveiling a new relationship between vascular calcification and aspirin use.
There are strengths and limitations in this study. The strength is that kinase-independent function of IKKb in suppressing calcification is unveiled. The kinase-dependent functions of IKKb have been well analyzed; however, its kinase-independent function has been rarely reported, except in our previous study. 32 Therefore, its roles in diseases are almost unknown. This study indicated its new function in vascular calcification, which would attract interest of researchers in the field of vascular diseases. On the other hand, the limitation of this study is that we have not evaluated the kinase-dependent roles of IKKb in the regulation of calcification. Considering that KD-IKKb works similarly with KA-IKKb in inhibition of calcification, kinase-independent function would have a major contribution to the inhibitory role of IKKb on calcification. However, it does not exclude the involvement of kinase-dependent function of IKKb on calcification. Some articles have reported that kinase activity of IKKb regulates b-catenin. 43,44,53 Indeed, in Figure 8A, the expression of b-catenin and Runx2 in KA-IKKb cells is less than that in KD-IKKb cells, despite less IKKb protein expression, which implicates the involvement of kinase activity of IKKb in regulation of these molecules. To explore it precisely, transient expression of constitutive active p65 in vitro would be not suitable because long-term cell culture is required for observing calcification. Future study of making CaCl 2 model on mice with modification of p65 or IjB in VSMCs would make it clear.