Medial calcification in the arterial wall of smooth muscle cell‐specific Smpd1 transgenic mice: A ceramide‐mediated vasculopathy

Abstract Arterial medial calcification (AMC) is associated with crystallization of hydroxyapatite in the extracellular matrix and arterial smooth muscle cells (SMCs) leading to reduced arterial compliance. The study was performed to test whether lysosomal acid sphingomyelinase (murine gene code: Smpd1)‐derived ceramide contributes to the small extracellular vesicle (sEV) secretion from SMCs and consequently leads to AMC. In Smpd1 trg/SMcre mice with SMC‐specific overexpression of Smpd1 gene, a high dose of Vit D (500 000 IU/kg/d) resulted in increased aortic and coronary AMC, associated with augmented expression of RUNX2 and osteopontin in the coronary and aortic media compared with their littermates (Smpd1 trg/SMwt and WT/WT mice), indicating phenotypic switch. However, amitriptyline, an acid sphingomyelinase (ASM) inhibitor, reduced calcification and reversed phenotypic switch. Smpd1 trg/SMcre mice showed increased CD63, AnX2 and ALP levels in the arterial wall, accompanied by reduced co‐localization of lysosome marker (Lamp‐1) with multivesicular body (MVB) marker (VPS16), a parameter for lysosome‐MVB interaction. All these changes related to lysosome fusion and sEV release were substantially attenuated by amitriptyline. Increased arterial stiffness and elastin disorganization were found in Smpd1 trg /SMcre mice as compared to their littermates. In cultured coronary arterial SMCs (CASMCs) from Smpd1 trg/SMcre mice, increased Pi concentrations led to markedly increased calcium deposition, phenotypic change and sEV secretion compared with WT CASMCs, accompanied by reduced lysosome‐MVB interaction. However, amitriptyline prevented these changes in Pi‐treated CASMCs. These data indicate that lysosomal ceramide plays a critical role in phenotype change and sEV release in SMCs, which may contribute to the arterial stiffness during the development of AMC.


| INTRODUC TI ON
Vascular calcification is the build-up or accumulation of apatite calcium salts in the media and/or intima of arteries that has been associated with ageing, chronic kidney disease, diabetes mellitus and atherosclerosis. 1 It has been reported that the arterial calcification pathology mimics the bone formation process and that the earliest phase involves the osteogenic differentiation of vascular SMCs. [1][2][3][4] Various human 5 and animal 5,6 studies have reported that osteogenic conversion of SMCs in the medial region appears prior to mineralization in arterial medial calcification (AMC), suggesting a critical role for smooth muscle cell (SMCs) phenotypic transition in this vascular pathologic change. 7 It is known that both intimal and medial vascular calcification may be mediated by a common mechanism, namely the large increases in extracellular vesicles (EVs) in the vascular interstitial space, in particular, the small extracellular vesicles (sEVs) (with size of 40-100 or to 140 nm). These sEVs are mainly produced and secreted from arterial SMCs. 8 However, role of sphingolipids (SLs) such as ceramide in particular lysosomal ceramide in SMCs and associated pathogenic role in AMC is still poorly understood.
In response to various physiological stimuli or a mineral imbalance, vascular smooth muscle cells (VSMCs) secrete sEVs or exosomes, which act to nucleate calcium phosphate (Ca/P) crystals in the form of hydroxyapatite. 2,4,9 More recently, sphingolipid-mediated signalling took a central stage in understanding the regulation of matrix vesicles (MVs) or exosome release and vascular calcification. It has been reported that activation of sphingomyelin phosphodiesterase 3 (SMPD3, neutral sphingomyelinase) and cytoskeletal rearrangements in synthetic VSMCs led to multivesicular body (MVB) trafficking and elevated exosome secretion, which is a hallmark of vascular calcification or related vascular diseases. 10 Sphingolipids belong to class of lipids located in the plasma membrane and at intracellular organelle membranes, which not only have structural roles but also carry signalling function. Ceramide (CER) and sphingosine-1-phosphate (S1P) are the major SLs that act as signalling molecules, control various cellular processes, such as cell growth, adhesion, migration, senescence, cell death and inflammatory response. [11][12][13] Notably, several SL metabolites have been associated with the development of several pathologies, including diabetes, cancer, microbial infections, neurological syndromes and cardiovascular disease. [14][15][16] Our laboratory previously reported that acid sphingomyelinase (ASM) plays an important role in glomerular injury 17,18 and inflammasome activation. 19,20 Acid sphingomyelinase (ASM), a lysosomal enzyme, is present in lysosomes and secretory lysosomes, and its fusion with plasma membrane results in the release of ceramide on the outer leaflet of the plasma membrane that serves to re-organize and cluster receptors and signalling molecules. 21,22 This re-organization of receptors and associated signalling molecules mediates various effects of ceramides at the cellular level. 23 Bianco et al 24 in 2009 revealed that acid sphingomyelinase (ASM) is a key enzyme involved in P2X 7 -dependent microvesicle biogenesis at the surface of glial cells (microglia and astrocytes) via activation of P38 MAP kinase. This causes translocation of ASM from lysosomes to the plasma membrane outer leaflet, where it catalyses CER formation from sphingomyelin (SM). 24 Further, they observed that SM to CER conversion perturbs membrane curvature and fluidity, favouring budding of multivesicles. 25 In the human macrophage cell line, U937, it was found that activation of ASM in response to oxidized LDL-containing immune complexes contributes to the release of IL-1β in association with exosomes, 26 and inhibiting ASM pharmacologically (desipramine) or genetically (ASM siRNA) reduced exosome and IL-1β secretion. Recently, a study reported that activation of ASM by cigarette smoke (CS) in lung endothelial cells leads to ceramide production in these cells which then results in the release of EVs. However, treatment with imipramine, a functional inhibitor of ASM or ASM deletion, markedly decreased CS-induced EV production. 27 These findings were validated by ASM (Smpd1 −/− ) knockout mice which showed reduced levels of EVs in plasma following CS exposure, whereas mice with overexpressed ASM in endothelial cells display increased levels of circulating EVs. Li et al, in human macrophages, found that CS may promote microvesicle shedding through an ASM-dependent pathway via activation of p38 MAPK. 24,28 It was shown that the increase in lysosomal ceramides causes lysosomal dysfunction through the activation of cathepsins. 29 On contrary to the well-characterized function of surface acid sphingomyelinase and ceramides, the role of lysosomal acid sphingomyelinase and ceramide is poorly understood. PDMP, a ceramide analogue, is a promising target for preventing cardiovascular diseases such as atherosclerosis and cardiac hypertrophy 30,31 and for suppressing osteoclastogenesis by inhibiting glycosphingolipid synthesis. 31 MVBs. We report here that SMC-specific overexpression of ASM leads to extensive AMC, phenotype change and arterial stiffness upon receiving high dose of vitamin D injections. Administration of amitriptyline, a pharmacological inhibitor of ASM, to these animals ameliorates AMC. Kossa staining kit (ab150687, Abcam) and Alizarin Red S Solution (TMS-008-C, EMD Millipore) were used for detection AMC.

| Primary culture of mouse CASMCs and Alizarin Red S staining
Mouse CASMCs were isolated as described previously. 34 CASMCs were cultured in Dulbecco's modified Eagle's medium (DMEM, Gibco), supplemented with 10% FBS (Gibco) and 1% penicillin-streptomycin (Gibco) in humidified 100% air and 5% CO 2 mixture at 37°C. Cells were prepared in 6-well plates overnight, and 70%-80% confluent cells were treated with or without high phosphate (P i ) (3 mmol/L) 35 and then incubated for 2 weeks for in vitro calcification model. 36 CASMCs were rinsed with PBS, fixed with 4% paraformaldehyde for 15 minutes and washed three times with diH 2 O. Then, cells were incubated with 1 mL of 1% Alizarin Red S for 5 minutes, washed with diH 2 O and visualized using a phase microscope. CASMCs were also treated with ASM inhibitor amitriptyline (20 μmol/L) 37 and incubated for 24 hours.

| Isolation of sEVs
Small extracellular vesicles were isolated by differential ultracentrifugation from CASMC culture medium as described previously. 10 As mentioned above, after 70%-80% confluence, CASMCs were incubated with or without P i (3 mmol/L) 35 and also with ASM inhibitor amitriptyline (20 μmol/L) 37

| Nanoparticle tracking analysis (NTA)
Nanoparticle tracking analysis (NTA) was used to analyse the CASMC-derived sEV using the light scattering mode of the NanoSight LM10 (Nano Sight Ltd.

| Alizarin Red S staining, Von Kossa staining and immunohistochemical analyses
Alizarin Red S staining was used to detect AMC. Paraffin-embedded artery samples were deparaffinised with alcohol and xylene.
After three washes with distilled water, the arteries were stained with 1% Alizarin Red S for 5 minutes and washed with acetone for 20 seconds followed by acetone-xylene wash for further 20 seconds. The positively stained area showed a reddish colour. Von Kossa staining was performed to detect mineralization. 42
Frozen aortic tissue slides were fixed in acetone and then incu-

| Calcium and phosphate assay
Blood concentrations of calcium and phosphate were measured using commercially available kits ab102505, Abcam, USA, and ab65622, Abcam, USA, as described by manufacture's protocol.

| Elastin staining
Elastin staining of the aortic tissues was performed using commercially available kits ab150667, Abcam, USA, as described by manufacture's protocol.

| Non-invasive, in vivo measurement of aortic pulse wave velocity (PWV)
Pulse wave velocity was measured 3 weeks post-Vit D injection or vehicle using a high-resolution Doppler ultrasound instrument (Vevo2100, Visual Sonics), as described. 45 Briefly, 2% isoflurane was used to anaesthetize the mice prior to mounting on a heated (37°C) platform to monitor electrocardiogram (ECG), heart rate (HR), respiratory rate, and to eliminate movement artefacts. Abdominal hair was removed using hair removal cream (Veet). Isoflurane was reduced to 1.5% with slight adjustments in order to maintain a heart rate (

| Statistical analysis
All of the values are expressed as mean ± SEM. Significant differences among multiple groups were examined using two-way ANOVA followed by Duncan's test. P < .05 was considered statistically significant.

| Characterization of Smpd1 trg /SM cre transgenic mice
To confirm the role of sphingolipid, ceramide in AMC, we employed an animal model, namely SMC-specific Smpd1 transgenic mice (Smpd1 trg /SM cre ) which caused overexpression of Smpd1 gene in SMCs that encodes acid sphingomyelinase (ASM) to produce ceramide via hydrolysis of sphingomyelin. Smpd1 trg /SM cre were generated by crossing the Smpd1 gene knock-in mice (with floxed blocker for its transcription) 46  ing SMC-specific overexpression of Smpd1 gene ( Figure 1B). We observed co-localization of GFP with SM22-α in the coronary arterial wall of Smpd1 trg /SM cre /ROSA mice ( Figure 1C). We observed markedly increased immunostaining of CER, ASM and Cre in the aortic wall of Smpd1 trg /SM cre mice as compared to Smpd1 trg /SM wt and WT/ WT mice ( Figure 1D). As shown in Figure 1E, in SMC-specific Smpd1 transgenic mice, largely increased co-localization (yellow spots) of α-SMA (green) vs ASM (red) or ceramide (green) vs SM22-α (red) was observed in Smpd1 trg /SM cre mice compared with Smpd1 trg /SM wt and WT/WT mice.  Figure S1A and B, indicating that both genetic and pharmacological inhibition of ASM reduced arterial calcification.

| Aortic calcification and SMC phenotype transition in Vit D-treated
The present study also showed that phenotypic switch in SMCs during AMC was characterized by decreased expression of the VSMC lineage marker, smooth muscle 22α (SM22-α) ( Figure 3A) and up-regulation of both OSP ( Figure 3C) and RUNX2 ( Figure 3E) while as amitriptyline treatment prevented this phenotype change.
As shown in the bar graph ( Figure 3B), it is clear that SM22-α ex-    Figure 5B,D,F, the SMC phenotype was changed in coronary arterial media towards more dedifferentiated or osteogenic status when Smpd1 gene was specifically overexpressed in these SMCs, which was prevented due to ASM inhibition by amitriptyline.

| Lysosome-MVB interactions and sEV release in the arterial wall of Vit D-treated SMC-specific Smpd1 transgenic mice
The literature reports that sphingolipids such as CER participate in exosome or sEV biogenesis, formation of MVBs and their fusion with plasma membrane causing increased exosome secretion. 47 We observed that co-localization of MVBs (VPS16, green) and lysosomes (Lamp-1, red) in SMCs was much lower in the aortic medial wall of SMC-specific Smpd1 transgenic mice than their littermates (Smpd1 trg /SM wt and WT/WT mice) receiving Vit D injection, while as amitriptyline enhanced this interaction between MVBs (VPS16, green) and lysosomes (Lamp-1, red) in SMCs ( Figure 6A). The co-localization coefficient (PCC) of both markers in Smpd1 trg /SM cre mice was clearly reduced as shown in the bar graph, while as amitriptyline significantly increased the co-localization coefficient of these markers ( Figure 6B). Immunohistochemically, we indeed found that CD63 ( Figure 6C), Annexin-II (AnX2) ( Figure 6E) and alkaline phosphatase (ALP) ( Figure 6G) staining as sEV markers were significantly increased in the coronary arterial wall of Vit D-treated Smpd1 trg /SM cre mice than their littermates (Smpd1 trg /SM wt and WT/WT) which were decreased by amitriptyline as shown in the bar graphs ( Figure 6D,F,H).

| Enhanced calcification and phenotype change in the P i -treated Smpd1 trg /SM cre mice CASMCs in vitro
In our in vitro study, using CASMCs, we determined the effect of Cre-mediated overexpression of Smpd1 gene in high phosphate (P i )induced calcification model. As shown in Figure 7A,B, calcium deposition as shown by the presence of Alizarin Red-stained nodules was significantly increased with P i treatment in CASMCs isolated from Smpd1 trg /SM cre mice as compared to WT/WT cells, which were significantly decreased by ASM inhibition by amitriptyline.
Using real-time PCR, we observed that SMC-specific overexpression of Smpd1 gene induced the osteogenic phenotypic conversion in P i -treated CASMCs, as depicted by significantly increased expression of OSP and RUNX2 ( Figure 7C,D). Together, these results indicate that lysosomal ceramide-sphingolipid contributes to the phenotypic transition during the development of calcification.

| Lysosome-MVB interactions and sEV secretion in the P i -treated Smpd1 trg /SM cre mice CASMCs in vitro
Using confocal microscopy, we observed increased co-localization of VPS16 (MVB marker, green) and Lamp-1 (lysosome marker, red), indicating MVB interactions or even fusion with lysosomes in WT/WT CASMCs. In the CASMCs without P i treatment, overexpression of Smpd1 gene had reduced co-localization of VPS16 vs Lamp-1 (fewer yellows dots) as compared to WT/WT cells, while as P i exposure decreased co-localization of both the markers in CASMCs from WT/WT as well as Smpd1 trg /SM cre ( Figure 8A). However, amitriptyline treatment significantly increased the co-localization of VPS16 vs Lamp-1 (larger yellow spots) in P i -treated CASMCs both in Smpd1 trg /SM cre and WT/WT cells. The bar graphs represent the co-localization coefficient (PPC), exhibiting decreased interaction of lysosomes and MVBs more in P i -treated Smpd1 trg /SM cre CASMCs as compared to WT/WT cells, which was significantly increased by amitriptyline ( Figure 8B).
Quantification of sEVs using a nanoparticle tracking analysis system showed that P i treatment in CASMCs significantly increased secretion of sEVs (<200 nm), as shown by representative 3-D histograms in Figure 8C and more particles in <200 nm in P i -treated CASMCs from SMC-specific Smpd1 transgenic mice, while as amitriptyline significantly decreased P i -induced sEV secretion. A bar graph shows vesicle counts of <200 nm size ( Figure 8D). These data confirm that sEV release increased from SMCs with overexpression of Smpd1 gene even without stimulation by P i , which may drive arterial calcification.
F I G U R E 7 Effect of SMC-specific Smpd1 overexpression on P i -induced calcification and phenotype change in SMC-specific Smpd1 transgenic CASMCs in vitro. A, Representative images showed increased calcium deposition in P i -treated Smpd1 trg /SM cre CASMCs. B, Summarized bar graph showed significant increased P i -induced mineralization in Smpd1 trg / SM cre , which was significantly decreased by amitriptyline (Ami). Summarized bar graph showed mRNA expression of C. OSP and D. RUNX2 in P i -treated CASMCs by RT-PCR. n = 3. Ami, amitriptyline; OSP, osteopontin; P i , high phosphate; RUNX2, runt-related transcription factor 2; SMC, smooth muscle cell; Vehl, vehicle. *P ˂ .05 vs WT/WT Vehl; #P ˂ .05 vs WT/WT P i ; $P ˂ .05 vs Smpd1 trg /SM cre P i by two-way ANOVA followed by Duncan's test

| Overexpression of SMC-specific Smpd1 accelerates arterial stiffness in Smpd1 transgenic mice
Pulse wave velocity directly correlates with arterial stiffness and inversely proportional to arterial distensibility. 48 As shown in Figure 9A,B, SMC-specific overexpression of Smpd1 significantly increased PWV as compared to control WT/WT littermates, suggesting that aortic wall stiffening and remodelling in Smpd1 trg /SM cre mice even before frank aortic medial calcification can be observed.

| D ISCUSS I ON
In the present study, we describe for the first time that lysosomal ceramide plays a critical role in sEV release, phenotype transition and F I G U R E 8 Lysosome-MVB interactions and sEV excretion in the P i -treated SMC-specific Smpd1 transgenic CASMCs in vitro. A, Representative confocal images showed co-localization of VPS16 (green) and Lamp-1(red) in CASMCs B. Bar graph shows significant decrease in co-localization of VPS16/ Lamp-1 in P i -treated Smpd1 trg /SM cre CASMCs, which was significantly increased by Ami. n = 3. C, High P i treatment increased sEV release from Smpd1 trg /SM cre CASMCs. D, Bar graph shows significantly increased sEV release from CASMCs of SMC-specific Smpd1 transgenic mice, which was significantly decreased through ASM inhibition by Ami, n = 3. Ami, amitriptyline; PCC, Pearson correlation coefficient; P i , high phosphate; SMC, smooth muscle cell; Vehl, vehicle. lipid/ceramide pathway, which triggers more secretion of calcifying exosomes, resulting in arterial calcification. 10 Also, increased ceramide production in cell membrane or cytoplasm via SMPD3 pathway may induce sEV biogenesis leading to arterial calcification, 10,47 which represents a different mechanism from lysosomal regulation  [54][55][56] In this context, in human femoral arterial SMCs it was observed that Ox-LDL-induced matrix mineralization may be mediated by ceramide, 57 which was due to increased neutral sphingomyelinase activity and ceramide levels. GW4869, a neutral sphingomyelinase (N-SMase) inhibitor, significantly reduced Ox-LDL-induced calcification in these cultured SMCs. 57 In the present study, we found that the overexpression of Smpd1 Medial calcification increases arterial stiffness that develop along the concentric elastin lamellae during ageing, detected as continuous linear hydroxyapatite deposits in the absence of inflammatory cells. 58,59 Cellular sphingolipids alterations appear to be a hallmark of ageing; in particular, ceramide could contribute to age-related remodelling of the vasculature. Various studies reported that ceramide promotes collagen deposition, lung and liver fibrosis. 60,61 Moreover, ceramide is known to regulate actin cytoskeleton dynamics and its alteration has been reported in VSMCs from large arteries with age. 62,63 Actin cytoskeleton regulation is an important component of VSMC mechanosensing 64 and the myogenic response, 65 suggesting that ceramide could also play a role in impaired mechano-transduction in small arteries with ageing. In the current study, we tried to investigate whether the increased lysosomal ceramide in SMCs contributes to arterial plasticity by measuring PWV as an index of arterial stiffness. 66 We found that SMC-specific overexpression of Smpd1 significantly increased PWV suggesting that increased lysosomal ceramide due to overexpression of Smpd1 in SMCs contributed to the arterial stiffness during the development of AMC. In this context, study in ApoE −/− mice and rabbits fed a Western diet reported that inhibition of glycosphingolipid synthesis can prevent the development of atherosclerosis and lower arterial stiffness independent of blood pressure. 30 Hence, these studies provide evidence that sphingolipid biosynthesis pathway could be a target for the prevention of arterial stiffness during AMC.
In summary, this study demonstrates remodelling of arteries during AMC that is accompanied by lysosomal enzyme ASM controlling ceramide production in arterial SMCs. Lysosomal overexpression of Smpd1 gene specifically in SMCs may be crucially involved in the secretion of sEVs and phenotypic switch in arterial SMCs, initiating AMC. This suggests sphingolipids may be important mediators of vascular calcification. Given that arterial medial calcification is a major risk factor for cardiovascular disease, our study opens a new area for further research into the mechanisms that underlie vascular remodelling in AMC.

ACK N OWLED G EM ENT
This study was supported by grants from the National Institutes of Health (HL122937, HL057244 and HL075316).

CO N FLI C T O F I NTE R E S T
The authors declare no conflict interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data generated or analysed during this study are included in this article.