SM22α deficiency: promoting vascular fibrosis via SRF-SMAD3-mediated activation of Col1a2 transcription following arterial injury

Vascular fibrosis, characterized by increased Type I collagen expression, significantly contributes to vascular remodeling. Our previous studies show that disrupting the expression of SM22α (aka SM22, Tagln) induces extensive vascular remodeling following arterial injury, involving oxidative stress, inflammation, and chondrogenesis within the vessel wall. This study aims to investigate the molecular mechanisms underlying the transcription of Col1a2, a key fibrotic extracellular matrix marker. We observed upregulation of COL1A2 in the arterial wall of Sm22−/− mice following carotid injury. Bioinformatics and molecular analyses reveal that Col1a2 transcription depends on a CArG box in the promoter, activated synergistically by SRF and SMAD3. Notably, we detected enhanced nuclear translocation of both SRF and SMAD3 in the smooth muscle cells of the injured carotid artery in Sm22−/− mice. These findings demonstrate that SM22 deficiency regulates vascular fibrosis through the interaction of SRF and the SMAD3-mediated canonical TGF-β1 signal pathway, suggesting SM22α as a potential therapeutic target for preventing vascular fibrosis.


Introduction
Vascular brosis, characterized by a thickened vessel wall due to excessive collagen deposition, is a signi cant factor in cardiovascular diseases such as hypertension, atherosclerosis, and aneurysm [1].
Since vascular brosis can be induced by in ammation and chondrogenesis [5,20,21], we examined the expression of a type I collagen Col1a2 (a brotic marker) in injury-induced vessel wall of Sm22 -/-mice.
Here we report the high expression of Col1a2 in arteries of Sm22 -/-mice in response to vascular injury.Mechanistically, we nd that serum response factor (SRF) [22,23] and its interacting partner SMAD3 [24] bind to the Col1a2 locus to synergistically activates the transcription of Col1a2 in the vessel wall.This study sheds light on the molecular mechanisms underlying vascular brosis induction.

Materials and Methods
Artery injury by carotid denudation.The mouse carotid denudation protocol was approved by the Institutional Animal Care and Use Committee (IACUC) at Wayne State University.The animal procedures conform with the NIH guidelines (Guide for the care and use of laboratory animals).Generation and characterization of Sm22 −/− mice at mixed C57BL/6x SV129 genetic background as described previously [25]: this mouse strain (named as Tagln tm1LiW /Mmnc or SM22 knockout) was deposited in the Mutant Mouse Resource and Research Centers (MMRRC, ID#67375, SM22 Knockout) supported by NIH (Stock Number: 067375-UNC).The mouse carotid denudation injury model was described before [15,16].Brie y, carotid denudation was carried out on male Sm22 −/− mice and their wild type littermates of 18-20 weeks of age, as described [26].The mice were sacri ced two weeks after surgery.The 3mm carotid segments covering the part from 2 mm to 5 mm proximal to the carotid bifurcation were harvested and embedded in OCT medium (Tissue-Tek).Around 100 frozen slides were made for each mouse with triplicate sections on each slide at 8 µm thickness.The carotids of ve Sm22 −/− mice and their Sm22 +/+ littermates were stored separately in RNAlater reagent (Ambion) at 4°C for no more than 1 week before RNA extraction for qRT-PCR.Immunohistochemistry (IHC) analysis.Six slides, in the order of one every 15 consecutive slides, from each mouse were subject to H&E staining to screen sections with most prominent in ammatory responses.Immunohistochemistry was performed on the properly selected consecutive frozen slides using VECTASTAIN Elite ABC Kit (Vectorlabs).Brie y, air-dried slides were xed in methanol containing 0.3% H 2 O 2 for 10 minutes and serum blocked for 20 minutes.The following incubation steps of primary antibody, secondary antibody, ABC reagent and DAB substrate were performed according to the manufacturer's protocol.The slides were counterstained with hematoxylin.The primary antibodies (1:50 dilution) against procollagen COL1A2 (Santa Cruz, sc-8787), SRF (Santa Cruz, sc-335) and SMAD3 (Abcam, ab28379) were used.The secondary antibodies without the primary antibodies were used as the negative control.Morphometric analysis.For each antibody, both 100X and 400X images were taken using a Leica DM4000B microscope.Images of adventitia and media were separated using Photoshope 7.0 software.Semi-quantitative analyses of positive signals in adventitia and media were performed on 100X images using color segmentation and integrative optical density function in the Image-Pro software.
Analysis of the Col1a2 promoter.The transcriptional factor binding sites of the Col1a2 promoter of Homo sapiens, Mus musculus, Norway rat, Canis lupus and Equus caballus were analyzed using bioinformatic tools from the Genomatix (http://www.genomatix.de/) and UCSC genome browser (https://genome.ucsc.edu/,2019 year).The sequence of CArG box (the SRF binding site) in the promotor of Col1a2 in Homo sapiens, Mus musculus, Norway rat, Canis lupus and Equus caballus were compared using VISTA (VISTA tools (lbl.gov)) and UCSC genome browser (https://genome.ucsc.edu/,2019 year).
The ChIP-seq data analyses from GTRD-V20.06database (GTRD (biouml.org))and UCSC genome browser (https://genome.ucsc.edu/,2019 year) were performed to con rm the binding of SRF and SMAD3 in the promotor region of Col1a2 in Homo sapiens, Mus musculus, Norway rat, Canis lupus and Equus caballus.The sequences of these promoters were marked with the number started from the transcription starting site.
Real-time RT-PCR (rtRT-PCR).Total RNA from carotids was extracted and puri ed using RNeasy Fibrous Tissue Kit (Qiagen), and total RNA from tissue culture cells was extracted and puri ed using RNeasy Kit (Qiagen).The cDNA was synthesized using the Superscript II reverse transcriptase (Invitrogen).Real-time PCR was performed using SYBR Green on a StepOnePlus system (Applied Biosystems).GAPDH and snRNA U6 were used as internal controls in ΔΔCt method.All PCR primers were designed to cover at least 2 exons.
Plasmid Constructions and Mutagenesis The Col2a1 reporter containing the 1342 bp of the mouse Col2a1 gene was PCR cloned into the NheI/HindIII sites of pGL3-basic luciferase vector (Promega).The CC in the putative CArG box (CCAAACTTGG) was mutated to AT by PCR mutagenesis using the QuikChange™ site-directed mutagenesis kit (Stratagene, La Jolla, CA).Luciferase reporter plasmids were prepared for transfection using the plasmid maxi kit (Qiagen).
Promoter reporter luciferase assay.The 10T1/2 cells were seeded in a 96-well plate at a density of 5x10 4 /ml, 150µl per well and incubated for 24 hours at 37ºC with 5% CO2 to get 80-90% con uence.
Transfection was performed using Lipofectamin with Plus Reagent (Invitrogen) based on the following plasmid ratio: the Col1a2 promoter reporter (Col1a2_luc) or the Col1a2 promoter reporter with CArG box mutation (Col1a2_CArG_mut-luc) 50 ng, internal control renilla luciferase reporter (pRL_CMV, Promega) 1ng, pCGN_SRF or pCGN mock 10 ng, XFIF-Smad3 or XFIF mock 10 ng [24].The luciferase activity assay was conducted 24 hours after transfection using the Dual-Luciferase Reporter Assay System (Promega, E1910) on a Veritas™ Microplate Luminometer (Promega).Each transfection was carried out in triplicates on the same plate and the analyses were based on experiments on three independent experiments.

Chromatin immunoprecipitation assay (CHIP). 10T1/2 cells after pCGN-SRF transfection for 48 hours
were prepared for CHIP assay using the EZ-ChIP kit (Millipore).The sonicated reaction mixtures were subject to pull-down by 1 µg of either a rabbit anti-SRF antibody (Santa Cruz, sc-335) or a control rabbit IgG antibody (Invitrogen) followed by the reversal of cross-link.The primer pair for both conventional PCR and real-time PCR was 5'-AGTGAAGCGGGACTGGACA and 5'-GGCTTTCGAGGGGGAACTC; the PCR product containing the CArG box was 201 bp.
Electrophoresis mobility shift assay (EMSA).Nuclear fraction was harvested from COS-7 cells using the NE-PER reagent (Pierce) 48 hours after transfection with pCGN-SRF expression plasmid.The 20-bp CArG box probe from the Col1a2 promoter was 5'-TGCTTCCAAACTTGGCAAGG with 5' IRDye labeling (originally synthesized by LI_COR, now by IDT); the CArG box mutant oligo sequence is 5'-TGCTTCCCAACTTGGCAAGG.The antibody used for supershift was rabbit anti-SRF (Santa Cruz, sc-335).The reaction mixtures of EMSA were prepared using the Odyssey Infrared EMSA Kit (LI_COR) according to the manufacturer's protocol.The reaction mixtures were resolved on a 6% DNA retarding gel (Invitrogen) followed by visualization on an Odyssey Dual-Mode Imaging system (LO_COR).
Statistics.Five Sm22 −/− mice and ve Sm22 +/+ littermates were used in histology, immunohistochemistry and RT-PCR analyses.Values are means ± SE.Statistical analyses were performed using SPSS13.0software.Student t-test was applied to evaluate differences in all experiments and differences were considered signi cant at p < 0.05.

Results
Carotid injury triggers the upregulation of COL1A2 expression in the medial layer of Sm22 −/− mice.Previous studies from our lab illustrate that SM22 de ciency leads to arterial in ammation and chondrogenesis, accompanied by extracellular matrix (ECM) remodeling in response to carotid injury [15,16,18].Given that brosis can stimulate both in ammation and chondrogenesis, we examined the expression of COL1A2, a brotic marker indicative of type I collagen, in Sm22 −/− mice in comparison to their Sm22 +/+ littermates.Immunohistochemical (IHC) analysis uncovered a noteworthy increase in the expression of COL1A2 in the smooth muscle cell (SMC) layer of the vessel wall, harvested two weeks post-carotid denudation (Fig. 1A).In line with this increased protein expression, Col1a2 mRNA level was also elevated in the injured carotids of Sm22 −/− mice as opposed to their Sm22 +/+ counterparts (Fig. 1B).These ndings suggest that the transcriptional activation of Col1a2 may contribute to the upregulation of Col1a2 expression in Sm22 −/− carotids.
Col1a2 is a putative SRF target gene and SRF binds and transactivates the CArG box in the Col1a2 promoter.To investigate the molecular mechanism that regulates the transcription of Col1a2, we performed bioinformatic analyses using tools including the MatInspector and the VISTA genome browser.We identi ed a transcription module that contains an evolutionally conserved CArG box binding site (Fig. 2A-2B).SRF is known to bind to the CArG box to regulate the transcription of genes involved in a variety of pathophysiological processes [22,23,27].To con rm that SRF could activate the Col1a2 promoter, we cloned the mouse Col1a2 promoter into the pGL3Basic promoter luciferase reporter plasmid and generated a corresponding CArG box mutant.The luciferase reporter assay shows that SRF upregulated the Col1a2 promoter activity by 3.5 fold over its mock control; and that this activation was reduced signi cantly in the Col1a2 promoter with the CArG box mutant (Fig. 2C, middle 2 columns).Consistent with this result, SRF-VP16 drastically increased activities of the wild type the Col1a2 promoter, not the Col1a2 promoter with the CArG box mutant (Fig. 2C, right 2 columns).To determine whether SRF binds to the CArG box region of the Col1a2 promoter in vivo, we performed ChIP assay.Both real-time PCR (Fig. 2D) and regular PCR (Fig. 2D, the inset panel) reveal that SRF antibody precipitated a signi cant amount of Col1a2 promoter chromatin containing the putative CArG box.To further investigate whether SRF can bind directly to the putative CArG box, we performed the EMSA assay using a 20 bp oligo probe containing the CArG box in the Col1a2 promoter.The binding of SRF to the probe was competitively inhibited by the excess amount of unlabeled oligo of the same sequence (Fig. 2E, lane 2) but not by the CArG mutant oligo (Fig. 2E, lane 3).In addition, the SRF-probe complex was disrupted by the SRF antibody (Fig. 2E, lane 4) but not by the IgG control (Fig. 2E, lane 5).Taken together, these results support that Col1a2 is a direct target of SRF and can be activated by SRF binding to the CArG box in its promoter.
SRF and SMAD3 cooperate to activate the Col1a2 promoter.
Our previous study shows that SRF interacts with Smad3 to regulate gene transcription [24].To determine whether SRF and SMAD3 regulate the transcription of Col1a2 in VSMCs, we performed bioinformatic analyses of the Col1a2 promoter.Motif nding analysis of eukaryotic promoter database (EPD) identi ed evolutionarily conserved SRF and SMAD3 binding sites at − 163 and − 233 upstream to the TSS within the Col1a2 promoter in the promoter region of mouse Col1a2 gene (Fig. 3A).ChIP-seq analyses using the Gene Transcription Regulation Database (GTRD) detected the binding of both SRF and SMAD3 to the promoter regions of human COL1A2 and mouse Col1a2 in mouse aortic SMCs.We also observed their binding to the conserved human COL1A2 gene in several human cancer cell lines and embryonic stem cells.The SRF and SMAD3 bind sites found in ChIP-seq databases overlaps with those SRF and SMAD3 binding sites (red and blue boxes) in the promoter region of mouse Col1a2 gene using the Motif nding analysis tool for the eukaryotic promoter database (EPD).
Since SMAD3 binding site is near the CArG box in the promoter region of Col1a2 gene, we were promoted to examine whether SMAD3 participates SRF-mediated transactivation of the mouse Col1a2 promoter.We thus co-transfected SRF and SMAD3 with the luciferase reporter driven by the Col1a2 promoter.Although SMAD3 alone did not signi cantly increase the promoter activities, combination of SRF and SMAD3 boosted Col1a2 promoter activities about 10 folds while SRF only activates the Col1a2 promoter activities to about 4.5 folds (Fig. 3B).These results demonstrate that SRF and SMAD3 can bind to Col1a2 chromatin to exert synergetic effect on the transcriptional activation of the Col1a2 promoter.
The association between the activation of SRF transcriptional activity, TGF-β signal pathway, and the nuclear translocation of SRF and SMAD3 is well-established.We conducted immunohistochemistry assays on Sm22 −/− mice injured arterial wall to examine the expression of SRF and SMAD3.Our ndings reveal that both SRF and SMAD3 are minimally expressed in the vessel walls of both Sm22 −/− and its wildtype control mice without injury (Fig. 3C).However, two weeks post-injury, we observed a signi cant increase in the expression of both SRF and SMAD3 in the carotid media of Sm22 −/− mice (Fig. 3C).The nuclear presence of SRF and SMAD3 is indicated by arrows.Collectively, these ndings suggest that injury activates the SRF and SMAD3-mediated canonical TGF-β signal pathway, promoting the transcription of Col1a2 in the carotids of Sm22 −/− mice.

Discussion
To demonstrate the active roles of downregulation of contractile proteins in arterial pathogenesis, we and a series of independent studies have demonstrated that loss of SM22 led to various interrelated pathogenic processes ranging from proliferation, migration, in ammation, chondrogenesis to senescence [14][15][16][17][18][19]28].Here we show that SM22 de ciency also induces vascular brosis via the activation of the transcription of Col1a2 by SRF and SMAD3 in Sm22 −/− mice in response to vascular injury (Fig. 4).
This study stems from our pursuit to identify a common regulatory mechanism that links both proin ammatory and chondrogenic phenotypic changes in vascular smooth muscle cells (VSMCs) following injury in Sm22 −/− mice [15, 16].We consider COL1A2, an extracellular matrix (ECM) protein, as a potential candidate.This choice is based on the upregulation of type I collagen expression in the pre-chondrogenic stage, as demonstrated in previous studies [20,21], which contributes to chondrogenesis.Additionally, Type I collagen has the capability to induce a pro-in ammatory phenotype in VSMCs [29].As anticipated, the immunohistochemistry (IHC) and reverse transcription-polymerase chain reaction (RT-PCR) results illustrated a substantial expression of Col1a2 in injured Sm22 −/− carotids (Fig. 1).Given that increased Col1a2 is a marker of vascular brosis, its elevated expression strengthens the association of vascular brosis with the observed proin ammatory and chondrogenic phenotypic changes in VSMCs.This nding suggests that Col1a2 may serve as a valid mediator connecting vascular brosis with in ammation and chondrogenesis during vascular remodeling in Sm22 −/− mice.
Since there was an elevation of Col1a2 transcription in Sm22 −/− carotids after injury, we next sought to characterize the molecular mechanism to regulate the transcription of Col1a2 in VSMCs (Fig. 2-3).The identi cation of a SRF binding site in the Col1a2 promoter from routine bioinformatics analysis offered a clue of possible role of SRF in activating Col1a2 transcription since SRF is shown to promote in ammation [23,30].Col1a2 is a true SRF target gene, which is further con rmed by the promoter reporter, ChIP and EMSA assays (Fig. 2).In agreement with activation of Col1a2 by SRF, IHC analysis revealed prominent SRF nuclear localization in injured Sm22 −/− carotid (Fig. 3C).This result is consistent with the binding of SRF to the Col1a2 promoter reported in cardiac myo broblasts in response to myocardium infarction [31].
It is intriguing to observe that disrupting a contractile protein SM22 also demonstrates a novel role in regulating brosis in addition to its known impacts on VSMC in ammation and chondrogenesis [15,16,18].These ndings support the concept of VSMC plasticity in response to vascular stress, highlighting the central role of SRF in intricately regulating VSMC phenotypic changes through interactions with other key transcription factors such as Myocd, Elk1, NF-κB, Sox9, and Smad3 [18, 24,[37][38][39][40].Given Smad3's a nity for the SRF binding domain in Myocd [37,[41][42][43], further research should examine whether Smad3 competes with Myocd for binding to SRF.This potential competition could reveal mechanistic insights into TGF-β signaling responses during vascular brosis.The molecular mechanisms discovered in this study suggests that SM22 could be a potential therapeutic target for preventing brosis.