Endogenous sulfur dioxide alleviates collagen remodeling via inhibiting TGF-β/Smad pathway in vascular smooth muscle cells

The study was designed to investigate the role of endogenous sulfur dioxide (SO2) in collagen remodeling and its mechanisms in vascular smooth muscle cells (VSMCs). Overexpression of endogenous SO2 synthase aspartate aminotransferase (AAT) 1 or 2 increased SO2 levels and inhibited collagen I and III expressions induced by transforming growth factor (TGF)-β1 in VSMCs. In contrast, AAT1 or AAT2 knockdown induced a severe collagen deposition in TGF-β1-treated VSMCs. Furthermore, AAT1 or AAT2 overexpression suppressed procollagen I and III mRNA, upregulated matrix metalloproteinase (MMP)-13 expression, downregulated tissue inhibitors of MMP-1 level, and vice versa. Mechanistically, AAT1 or AAT2 overexpression inhibited phosphorylation of type I TGF-β receptor (TβRI) and Smad2/3 in TGF-β1-stimulated VSMCs. Whereas SB431542, an inhibitor of TGF-β1/Smad signaling pathway, attenuated excessive collagen deposition induced by AAT knockdown. Most importantly, ectopically expressing AAT or exogenous addition of 100 μM SO2 blocked AAT deficiency-aggravated collagen accumulation in TGF-β1-stimulatd VSMCs, while no inhibition was observed at 100 μM ethyl pyruvate. These findings indicated that endogenous SO2 alleviated collagen remodeling by controlling TGF-β1/TβRI/Smad2/3-mediated modulation of collagen synthesis and degradation.


Results
Endogenous sulfur dioxide was associated with the inhibition of TGF-β1-induced collagen remodeling in VSMCs. Sulfur dioxide could be endogenously produced from L-cysteine in mammals through transamination by AAT. To investigate the effect of endogenous SO 2 on collagen remodeling in VSMCs, we overexpressed two isozymes of AAT, AAT1 and AAT2 in VSMCs respectively. Transfection of VSMCs with AAT1 or AAT2 plasmid could significantly increase AAT1 or AAT2 protein expression as compared with vehicle ( Fig. 1a). In accordance, the endogenous SO 2 level was obviously elevated in VSMCs transfected with AAT1 or AAT2 plasmid (Fig. 1b). HPLC-FD assay also showed that much higher SO 2 content in the supernatant from VSMCs transfected with AAT1 or AAT2 plasmid than vehicle (Fig. 1c). And the concentration of SO 2 corrected by the number of cells was also increased significantly by AAT overexpression (Fig. 1d). Both immunofluorescence method and Western blot analysis showed that TGF-β 1 could upregulated protein expression of collagen I and III in VSMCs, while AAT1 or AAT2 overexpression could significantly inhibit the TGF-β 1-induced collagen I and III expression (Fig. 1e,f). Therefore, these results indicated that endogenous SO 2 might suppress TGF-β 1-induced excessive collagen deposition in VSMCs.

Endogenous sulfur dioxide deficiency aggravated the TGF-β1-induced collagen remodeling in
VSMCs. To further investigate the potential causal role of endogenous SO 2 in vascular collagen remodeling, we used shRNA to knock down AAT1 or AAT2. Specific knockdown of AAT1 or AAT2 was verified at protein level by Western blot analysis (Fig. 2a). In agreement, transfection with AAT1 shRNA or AAT2 shRNA significantly decreased endogenous SO 2 in VSMCs, as well as the SO 2 level in the supernatant (Fig. 2b,c). As compared with control shRNA, endogenous SO 2 silencing greatly exacerbated TGF-β 1-induced collagen I and III expression in VSMCs (Fig. 2d,e). These data probably supported a significant role of endogenous SO 2 in the regulation of collagen remodeling in VSMCs.
Endogenous sulfur dioxide likely inhibited the TGF-β1-induced collagen remodeling in VSMCs via suppressing collagen synthesis. Collagen synthesis is one major link in collagen expression regulation, mainly reflected in the transcription and translation of procollagen gene. Real-time quantitative PCR (RT-PCR) analysis showed that TGF-β 1 stimulation increased the mRNA levels of procollagen I and III, while AAT1 or AAT2 overexpression could significantly inhibit the TGF-β 1-induced mRNA expression of procollagen I and III (Fig. 3a,b). In contrast, AAT1 or AAT2 knockdown further exacerbated it (Fig. 3c,d). These results indicated that endogenous SO 2 likely inhibited the collagen remodeling in VSMCs via inhibiting collagen synthesis.

Endogenous sulfur dioxide possibly inhibited the TGF-β1-induced collagen remodeling in
VSMCs via promoting collagen degradation. Collagen degradation is the other major link in collagen expression regulation. The key molecules regulating the collagen degradation in VSMCs are matrix metalloproteinase-13 (MMP-13) promoting collagen degradation, and tissue inhibitor of metalloproteinase-1 (TIMP-1) suppressing collagen degradation. RT-PCR and Western blot analysis showed that TGF-β 1 downregulated the MMP-13 mRNA and protein levels in VSMCs, whereas increased the TIMP-1 mRNA and protein levels ( Fig. 4a-c). AAT1 or AAT2 overexpression in VSMCs could obviously elevate the TGF-β 1-downregulated MMP-13 mRNA and protein levels, and suppress the TGF-β 1-upregulated TIMP-1 mRNA and protein levels ( Fig. 4a-c). In contrast, AAT1 or AAT2 knockdown could further inhibit the TGF-β 1-downregulated mRNA and protein expressions of MMP-13, and further promote the TGF-β 1-upregulated mRNA and protein expressions of TIMP-1 (Fig. 4d-f). These results suggested that endogenous SO 2 possibly inhibited the collagen remodeling in VSMCs via promoting collagen degradation.
Endogenous sulfur dioxide might inhibit the Smad2/3 signaling pathway during TGF-β1-induced collagen remodeling in VSMCs. Considering TGF-β /Smad2/3 signal is the key pathway in regulating collagen remodeling, next we observed whether this signaling pathway was involved in the regulation of endogenous SO 2 on collagen remodeling in VSMCs. Western blot analysis showed that TGF-β 1 promoted the phosphorylation of Smad2 and Smad3 as compared with control, while AAT1 or AAT2 overexpression  (a) Protein expression of AAT1 and AAT2 in VSMCs transfected with 2 μ g of control shRNA (sh-Con), AAT1 shRNA (sh-AAT1) or AAT2 shRNA (sh-AAT2) for 48 h. (b) Representative fluorescent staining (blue) of endogenous SO 2 in VSMCs transfected with Con, AAT1 or AAT2 shRNA for 48 h. Scale bar, 20 μ m. (c) SO 2 content in supernatant from VSMCs transfected with Con, AAT1 or AAT2 shRNA for 48 h. (d) Collagen I and III expression in VSMCs by confocal images. VSMCs in coverslips were transfected with Con, AAT1 or AAT2 shRNA before TGF-β 1 (10 ng/ml) stimulation for 24 h. Scale bar, 20 μ m. (e) Representative Western blot and quantification of collagen I and III in VSMCs transfected with Con, AAT1 or AAT2 shRNA before TGF-β 1 treatment. ***P < 0.001, **P < 0.01 or *P < 0.05 compared with Con shRNA, ## P < 0.01 or # P < 0.05 compared with Con shRNA + TGF-β 1 group (ANOVA). Data are represented as mean ± SD (n = 5).
Endogenous sulfur dioxide deficiency promoted collagen remodeling mediated by TGF-β1/ Smad signaling pathway in VSMCs. To further explore whether TGF-β 1/Smad pathway mediated the regulatory effect of endogenous SO 2 on collagen remodeling in VSMCs, we applied SB431542, an inhibitor of TGF-β 1/Smad signaling pathway, to VSMCs transfected with AAT1 or AAT2 shRNA. RT-PCR showed SB431542 almost completely abolished the AAT1-or AAT2-silencing induction in mRNA expression of procollagen I and III (Fig. 7a,b). Moreover, enhanced collagen I and III protein levels evoked by AAT1 or AAT2 knockdown were also greatly repressed by SB431542 (Fig. 7c,d), suggesting endogenous SO 2 likely inhibited collagen remodeling at least in part by blocking the activation of TGF-β 1/Smad signaling pathway.
Scientific RepoRts | 6:19503 | DOI: 10.1038/srep19503 AAT1-scilencing VSMCs, and transfecting with AAT2 plasmid normalized AAT2 expression in AAT2-knocked down cells (Fig. 8a). Moreover, ectopic AAT1 or AAT2 expression significantly upregulated the inhibited SO 2 level in AAT deficient VSMCs (Fig. 8b). Of note, AAT1 or AAT2 overexpression substantially abolished the excessive deposition of collagen I and III induced by AAT1 or AAT2 knocked-down in the TGF-β 1-treated VSMCs (Fig. 8a). And the enhanced mRNA levels of procollagen I and III by AAT1 or AAT2 deficiency were also retarded by ectopic AAT1 or AAT2 expression (Fig. 8c,d). These data verified that the effects produced by AAT shRNA constructs on collagen remodeling resulted from AAT deficiency.

Discussion
With the increasing understanding about the function of hydrogen sulfide as gaseous signaling molecule in cardiovascular system, more and more attention has been paid to the physiological and pathophysiological functions of SO 2 , another important gasotransmitter in cardiovascular regulation. Recent studies have found that SO 2 can be endogenously generated from sulfur-containing amino acid metabolism in human body 7 . L-cysteine can be oxidized into L-cysteine sulfinate by cysteine dioxygenase, the latter can be transformed into β -sulfinylpyruvate by AAT, then spontaneously decomposes to pyruvate and SO 2 8 . A part of SO 2 in vivo can quickly combine with water to form sulphite, while other exists in gaseous form. Previous Studies found that exogenous SO 2 donor inhibited the hypoxic pulmonary vascular remodeling 9 . In addition, Liu et al. reported that SO 2 suppressed aortic VSMCs proliferation 15 . However, it's still unclear about whether endogenous SO 2 regulates collagen remodeling in VSMCs, and its underlying mechanisms. Therefore, we overexpressed SO 2 generating enzyme, AAT in VSMCs or knocked down AAT with shRNA, demonstrating endogenous SO 2 inhibited the TGF-β 1-induced excessive collagen deposition in VSMCs and its possible mechanisms.
Maintenance of extracellular matrix (ECM) is one of the major functions of VSMCs. Disturbance of ECM homeostasis leads to excessive ECM deposition, much of which is produced by VSMCs, is a common consequence of cardiovascular diseases such as atherosclerosis, hypertension, and restenosis after coronary angioplasty [16][17][18] . Most of ECM proteins within vascular walls are collagen I and III 18 . Therefore, searching for endogenous molecules regulating collagen remodeling in VSMCs, can help to protect from the occurrence and development of cardiovascular disease. SO 2 , a gaseous signaling molecule, can be endogenously generated in many cells, including VSMCs 19 . To explore the effect of endogenous SO 2 on collagen remodeling in VSMCs, we used immunofluorescence staining and Western blot to detect the protein expression of collagen I and III in VSMCs. The data showed that endogenous SO 2 significantly inhibited the TGF-β 1-induced collagen I and III protein levels.
Collagen remodeling contains two major links, namely excessively increased collagen synthesis and/or reduced collagen degradation. Collagen synthesis reflects in the transcription and translation of procollagen gene. So next we detected the effect of endogenous SO 2 on collagen synthesis. We found that endogenous SO 2 markedly suppressed the TGF-β 1-induced procollagen I and III mRNA levels, suggesting that endogenous SO 2 could inhibit collagen synthesis in VSMCs.
The balance between degradation of ECM is guaranteed by MMPs and TIMPs 20,21 . MMPs consist of a family of Zn-dependent endopeptidases that degrade ECM and basement membrane. They take part in tissue remodeling, cell infiltration and tumor invasion. For example, MMP-2, MMP-9 and MMP-13 are reported to have pivotal roles in vascular remodeling in several disease states [20][21][22] . MMP-2 mainly degrades type-IV, V, VII, X and XI collagen, gelatin, fibronectin, laminin and elastin. MMP-9 mainly degrades type-IV, V, VII, X collagen, gelatin, elastin, laminin, proteoglycan, fibronectin and entacin. Of note, MMP-13 degrades type-I, II, III, IV, IX, X and XIV collagen, fibronectin, tenascin-C, proteoglycan and gelatin 22 . MMPs activation can be suppressed by their endogenous inhibitors, the TIMPs. Four types of TIMPs have been found, of which TIMP-1 is the important inhibitor of MMP-1, -3, -9 and -13 23 . Therefore, in this study we detected the effect of endogenous SO 2 on MMP-13 and TIMP-1 mRNA and protein expressions in VSMCs. We found that TGF-β 1 downregulated the mRNA and protein expressions of MMP-13, and upregulated the mRNA and protein levels of TIMP-1. Endogenous SO 2 could dramatically increase the TGF-β 1-downregulated MMP-13 level, and decrease the TGF-β 1-upregulated TIMP-1 level, suggesting endogenous SO 2 could promote collagen degradation.
Scientific RepoRts | 6:19503 | DOI: 10.1038/srep19503 through TGF-β /Smad signaling pathway 3,4,31 . TGF-β 1 could significantly increase the collagen protein expression in VSMCs through Smad2/3 signaling pathway 4,16 . Our present study showed that endogenous SO 2 markedly inhibited the TGF-β 1-induced phosphorylation of Smad2 and Smad3 in VSMCs. In order to further explore how endogenous SO 2 blocked the Smad signaling pathway, we detected the phosphorylation level of Tβ RI. The data indicated that endogenous SO 2 suppressed the phosphorylation of Tβ RI. And SB431542, an inhibitor of TGF-β 1/Smad signaling pathway, could significantly abolish the promoting effect of endogenous SO 2 deficiency on TGF-β 1-induced collagen remodeling in VSMCs. These results suggested that endogenous SO 2 alleviated the collagen remodeling in VSMCs at least in part by inhibiting TGF-β /Smad signaling pathway.
The limitation of the present study involved the employment of exclusively one shRNA construct targeting either AAT1 or AAT2, and compared to only one control for all experiments. Considering the collateral effects that were often produced by such constructs, we used ectopically expressing AAT in knocked-down cells. The data showed that ectopic AAT1 or AAT2 expression upregulated the repressed SO 2 level in AAT knocked down-VSMCs. As a consequence, AAT1 or AAT2 overexpression abolished AAT silencing-induced collagen I and III deposition in TGF-β 1-treated VSMCs possibly via reducing collagen synthesis. These results supported the assumption that AAT1 and AAT2 shRNA used in the present study targeted AAT1 and AAT2 genes, respectively. AAT shRNA-induced exacerbation of collagen remodeling might result from AAT deficiency, other than the off-target effects. VSMCs were co-transfected with Con shRNA and vehicle, AAT1 shRNA and vehicle, AAT2 shRNA and vehicle, AAT1 shRNA and AAT1, or AAT2 shRNA and AAT2 for 48 h, and then stimulated with TGF-β 1 for 24 h. *P < 0.05 compared with Con shRNA + vehicle, # P < 0.05 compared with Con shRNA + vehicle + TGF-β 1 group, & P < 0.05 compared with AAT1 shRNA + vehicle + TGF-β 1 group, and + P < 0.05 compared with AAT2 shRNA + vehicle + TGF-β 1 group (ANOVA). Data are represented as mean ± SD (n = 4-5).
Scientific RepoRts | 6:19503 | DOI: 10.1038/srep19503 Endogenous SO 2 was produced from L-cysteine in a two-step reaction, in which the later step was catalyzed by AAT and resulted in equimolar pyruvate as well 7,8,32 . Previous studies showed that pyruvate could modulate TGF-β signaling 33,34 . However, whether the effects of AAT manipulation on collagen remodeling were caused by SO 2 or pyruvate had not been elucidated. Here, we added exogenous SO 2 derivatives or pyruvate in AAT-knocked down VSMCs. Considering the instability of pyruvate, ethyl pyruvate, a simple aliphatic ester derived from pyruvic acid, was widely implicated in scientific research instead of pyruvate 34,35 . Since SO 2 and pyruvate were generated equimolar from L-cysteine catalyzed by AAT 7,8,32 , VSMCs were pretreated with 100 μ M SO 2 derivatives or 100 μ M ethyl pyruvate in the present study. The results showed that SO 2 derivatives at 100 μ M significantly inhibited AAT silencing-exacerbated collagen I and III deposition in TGF-β 1-treated VSMCs, while no inhibition was observed at 100 μ M ethyl pyruvate. These data demonstrated that SO 2 generated by AAT but not pyruvate played a crucially protective role against collagen remodeling in AAT knocked-down VSMCs with TGF-β 1 stimulation.
In conclusion, we discovered an inhibitory effect of endogenous SO 2 on TGF-β 1-induced collagen remodeling in VSMCs via suppressing collagen synthesis and promoting collagen degradation. The underlying mechanism might involve inhibiting the Tβ RI phosphorylation by endogenous SO 2 to block the Smad2/3 signaling pathway activation. These findings suggest that endogenous SO 2 may be a promising therapeutic target for vascular collagen remodeling related cardiovascular diseases such as hypertension. VSMCs were transfected with Con shRNA, AAT1 shRNA or AAT2 shRNA for 48 h, pretreated with SO 2 derivatives (NaHSO 3 /Na 2 SO 3 , 100 μ M) or ethyl pyruvate (EP, 100 μ M) for 1 h, and then stimulated with TGF-β 1 for 24 h. *P < 0.05 compared with Con shRNA, # P < 0.05 compared with Con shRNA + TGF-β 1 group, & P < 0.05 compared with AAT1 shRNA + TGF-β 1 group, and + P < 0.05 compared with AAT2 shRNA + TGF-β 1 group (ANOVA). Data are represented as mean ± SD (n = 4-5).
Scientific RepoRts | 6:19503 | DOI: 10.1038/srep19503 Measurement of SO 2 concentration in VSMCs supernatant by high-performance liquid chromatography with fluorescence detection. VSMCs supernatants were collected for SO 2 content determination. SO 2 concentrations were measured using high-performance liquid chromatography with fluorescence detection (HPLC-FD, Agilent 1200 series, Agilent Technologies, Palo Alto, CA, USA) 9 . Briefly, 100 μ L of VSMCs supernatant was mixed with 70 μ L of 0.212 M sodium borohydride in 0.05 M Tris-HCl (pH 8.5) and incubated at 28 °C for 30 min. The sample was then mixed with 10 μ L of 70 mM mBrB in acetonitrile, incubated for 10 min at 42 °C, and then mixed with 40 μ L of 1.5 M perchloric acid. Protein precipitate in the mixture was removed by centrifugation at 12400× g for 10 min at 25 °C. The supernatant was immediately neutralized by adding 10 μ L of 2 M Tris-HCl (pH 3.0), and centrifuged at 12400× g for 10 min. The neutralized supernatant was used for HPLC-FD. Sulfitebimane was measured by excitation at 392 nm and emission at 479 nm. Quantification was carried out by the standardization of sodium sulfite.
Measurement of endogenous SO 2 content in VSMCs by a fluorescent probe. Endogenous SO 2 in VSMCs was measured using a fluorescent probe (kindly provided by Professor Kun Li, College of Chemistry, Sichuan University, Sichuan, China). The culture supernatant of VSMCs was discarded, and washed with PBS for three times. Then, the cells were stained in the working liquid of fluorescent probe (10 μ M) for 1 h at 37 °C, washed with PBS and then fixed with 4% paraformaldehyde for 15 min at room temperature. After washed with PBS, the cells were detected as blue fluorescent by confocal microscopy.
Phosphorylation of TβRI in VSMCs evaluated by immunoprecipitation. VSMCs were seeded in 100-mm dishes for immunoprecipitation. When cells were grown to 60-70% confluences, they were treated as follows. In the first series, cells were transfected with vehicle, AAT1 or AAT2 plasmid. After 24 h, they were starved in DMEM with 0.5% FBS before TGF-β 1 (10 ng/ml) treatment for 1 h. In the second series, cells were transfected with control shRNA, AAT1 shRNA or AAT2 shRNA for 24 h. Then they were starved for 24 h followed by TGF-β 1 treatment for 1 h.
The cells were then harvested with the antibody against Tβ RI (Santa Cruz Biotechnology, catalog sc-398) before immunoprecipitation with protein A/G agarose beads (Thermo Fisher Scientific, Waltham, MA, USA) 20,36 . The precipitated proteins were resolved by 10% SDS-PAGE and then immunoblotted with antibody against phosphoserine (Abcam, Cambridge, MA, USA) 20 .

Statistical analysis.
Results are presented as mean ± SD. Statistical comparisons were performed with PRISM5 software (GraphPad). Comparisons among more than 2 groups involved one-way ANOVA followed by the Student-Newman-Keuls test for post-hoc comparison as appropriate. P < 0.05 was considered significant.