Cezanne is a critical regulator of pathological arterial remodelling by targeting β-catenin signalling

Abstract Aims Pathological arterial remodelling including neointimal hyperplasia and atherosclerosis is the main underlying cause for occluding arterial diseases. Cezanne is a novel deubiquitinating enzyme, functioning as a NF-кB negative regulator, and plays a key role in renal inflammatory response and kidney injury induced by ischaemia. Here we attempted to examine its pathological role in vascular smooth muscle cell (VSMC) pathology and arterial remodelling. Methods and results Cezanne expression levels were consistently induced by various atherogenic stimuli in VSMCs, and in remodelled arteries upon injury. Functionally, VSMCs over-expressing wild-type Cezanne, but not the mutated catalytically-inactive Cezanne (C209S), had an increased proliferative ability and mobility, while the opposite was observed in VSMCs with Cezanne knockdown. Surprisingly, we observed no significant effects of Cezanne on VSMC apoptosis, NF-κB signalling, or inflammation. RNA-sequencing and biochemical studies showed that Cezanne drives VSMC proliferation by regulating CCN family member 1 (CCN1) by targeting β-catenin for deubiquitination. Importantly, local correction of Cezanne expression in the injured arteries greatly decreased VSMC proliferation, and prevented arterial inward remodelling. Interestingly, global Cezanne gene deletion in mice led to smaller atherosclerotic plaques, but with a lower level of plaque stability. Translating, we observed a similar role for Cezanne in human VSMCs, and higher expression levels of Cezanne in human atherosclerotic lesions. Conclusion Cezanne is a key regulator of VSMC proliferation and migration in pathological arterial remodelling. Our findings have important implications for therapeutic targeting Cezanne signalling and VSMC pathology in vascular diseases.

For Cezanne plasmid transfection, control (pHM6), wild-type (pHM6-Cez), or 38 mutated Cezanne (pHM6-Cez-C209S) plasmids generated in our previous 39 study 7 were transfected into VSMCs (1.0µg per 10 6 VSMCs) using TurboFect 40 Transfection Reagent (Thermo Fisher Scientific Inc) according to the 41 manufacturer's instructions. 42 For plasmid and siRNA co-transfection, respective control or gene over-43 expression plasmids (1.0µg per 10 6 VSMCs), and/or non-target (si-NT) or 44 gene-specific siRNAs (si-CCN1 or si-Ctnnb1) (50 nM, final concentration) 45 were co-transfected into VSMCs using TransIT-X2 Transfection Reagent 46 (Geneflow Limited, UK) according to the manufacturer's instructions, and 47 described in our previous study 3 . Briefly, VSMCs (1.5~2.0 x 10 5 per well) were 48 seeded into six-well plate 24 hours prior to transfection. Ten µl of siRNAs 49 (10µM in stock) and appropriate amount of respective plasmid (200ng of 50 plasmid) were mixed with 250µl of serum free DMEM in a sterile Eppendorf 51 tube, followed by adding 7.5µl of TransIT-X2 reagent. After incubated at RT 52 for 30 mins to allow the complexes to form. The TransIT-X2/plasmids/siRNAs 53 complexes were added dropwise in circular motions to ensure all the cells 54 being covered by the mixture. The transfected cells were cultured overnight 55 prior to medium change for serum starvation. MISSION esiRNA are a 56 heterogeneous mixture of siRNAs that all target the same mRNA sequence, 57 resulting in a highly specific and effective gene silencing. All siRNAs 58 (EHUEGFP for si-NT, EMU026621 for si-CCN1, and EMU047621 for si-59 Ctnnb1) were purchased from Sigma. With an optimum condition (clean and 60 healthy VSMCs at exponential growth phase with 50~70% confluent), a 61 satisfactory transfection efficiency (>60%) is normally achieved with primary 62 VSMCs in our Laboratory using the transfection protocols.

64
Generation of mouse CCN1 gene promoter reporter and the WRE mutant 65 The DNA sequence of murine CCN1 gene promoter shown below was 66 synthesized by Genscript Biotech (Nanjing, China), and sub-cloned into the 67 Kpn I and Mlu I sites of the pGL3-basic vector (Promega), designated as 68 pGL3-CCN1. Wnt response element (WRE) mutations were introduced into 69 pGL3-CCN1 by using QuikChange™ site-directed mutagenesis kit (Agilent 70 Technologies) and their respective mutant primers ( pTK-Renilla (20ng/well) was included in all transfection assays as internal 117 control. Dual-luciferase activity assays were conducted 48 hours after 118 transfection using a standard protocol. Relative luciferase unit (RLU) was 119 defined as the ratio of Luciferase versus Renilla activity with that of the control 120 (set as 1.0).

122
Cezanne shRNA lentivirus generation and infection 123 Cezanne shRNA lentiviral particles were generated as described previously 4, 124 11 . Cezanne shRNA lentiviral particles were produced using MISSION shRNA 125 Otud7b plasmids DNA (SHCLNG-NM_001025614, MISSION® shRNA 126 Bacterial Glycerol Stock, Sigma) according to protocol provided. The shRNA 127 Non-Target control vector (SHC002) was used as a negative control (sh-NT). 128 Briefly, 293T cells were transfected with the lentiviral vector and the 129 packaging plasmids, pCMV-dR8.2 and pCMV-VSV-G (both obtained from 130 Addgene), using TurboFect Transfection Reagent (Thermo Fisher Scientific 131 Inc) according to the manufacturer's instructions. The supernatant containing 132 the lentivirus was harvested 48 hours later, filtered, aliquoted, and stored at -133 80°C. shRNA lentiviral infection and Cezanne stable knockdown VSMC 134 generation were performed as described in our previous studies with some 135 modifications 4, 11-13 . Briefly, VSMCs were plated 24 hours prior to infection in 6 136 well-plates at 37°C. One transducing Unit per cell (or 2-3x10 5 /well) of sh-NT 137 or sh-Cezanne lentivirus were added with 10μg/ml hexadimethrine bromide 138 (H9268; Sigma). After incubated for 40~48 hours, the media was replaced 139 with complete media containing 4μg/ml puromycin (P9620, Sigma). For 140 selection of transductants, fresh media containing puromycin was added at 2-141 3 day intervals for 10 days. Stably infected cells with the highest knockdown 142 efficiency were used for functional analysis. 143 144 145 VSMC proliferation assays: manually cell counting & BrdU incorporation 146 assay 147 As described previously 1-6 , VSMCs (5x10 4 per well) cultured in 12 well plates 148 were transfected or infected with respective plasmids/siRNAs or shRNA 149 lentivirus as indicated in the each figures. After cultured overnight, the cells 150 were starved by culturing them in the serum-free DMEM supplemented for 151 further 48 hours, followed by 20%FBS or PDGF-BB (10ng/ml) stimulation for 152 additional 48 hours before trypsinizing and manually counting the cells under 153 hematocytometer. For BrdU incorporation assay, VSMCs were transfected 154 with plasmids/siRNAs or infected with lentiviral particles as indicated in each 155 figure, and were re-cultured (0.75 x10 4 per well) in 96 well plates overnight, 156 followed by serum starvation for 48 hours. Starved VSMCs were re-stimulated 157 with 20% FBS or 10ng/ml PDGF-BB, respectively, for 48 hours. Cell 158 proliferations were evaluated using 5-Bromo-2'-deoxy-uridine (BrdU) Labeling 159 and Detection Kit II (Roche) according to the manufacturer's instructions. The 160 absorbance of the samples was measured by a microplate reader at 405nm 161 (OD405) with reference measurement at 490nm (OD490). Relative 162 Absorbance (A405nm-A490nm) values representing cell proliferation ability were 163 compared between treatments with control sample set as 1.0. 164 165 VSMC Trans-well migration assay 166 Similar to our previous studies 1-6 , VSMCs transfected or infected with 167 respective plasmids/siRNAs or shRNA lentivirus as indicated in the each 168 figures were cultured in serum-free DMEM for 48 hours, and harvested for 169 counting. An aliquot (250,000 cells/200µl) of the cells in serum-free DMEM 170 was dispensed into the trans-well inserts (8µm pore size, Greiner Bio-One Ltd, 171 UK. Item number: 662638) pre-coated with 0.5% gelatin (Sigma, G1393), and 172 DMEM with 20% FBS or 30ng/ml PDGF-BB was placed in the lower chamber. 173 The trans-well plates were incubated at 37°C in a 5% CO2 incubator for 174 12~18 hours. Non-migrated cells in the top insert were carefully removed by 175 cotton swab, and the migrated cells in the bottom side were stained with 176 Crystal Violet dye. Images were captured at five fixed locations (right, bottom, 177 left, up and centre), and migrated cells were counted by two experienced 178 investigators blinded to the treatments. 179 180 VSMC apoptosis (TUNEL) analyses 181 As reported previously 6 , Terminal deoxynucleotidyl transferase dUTP nick end 182 labeling (TUNEL) Assay Kit (11684795910, Sigma) was used to assess 183 VSMC apoptosis by close following the manufacturer's instructions. After 184 staining, images were randomly taken with GFP (Green) and DAPI (Blue) 185 channel, respectively, and pseudo images were created using EVOS FL Auto 186 Imaging System (Thermo Fisher Scientific, UK). TUNEL-positive cells over 187 total cells were counted by two experienced investigators blinded to the 188 treatments. 189

Immunoblotting 190
Equal amount of protein was separated by SDS-PAGE with 4%~20% Tris-191 Glycine gel (Invitrogen, Carlsbad, CA, USA) and subjected to standard 192 Western blot analysis. The blots were subjected to densitometric analysis with 193 Image J software. Relative protein expression level was defined as the ratio of 194 target protein expression level to α-tubulin or GAPDH expression level with 195 that of the control sample set as 1.0. 196 197 Real time quantitative PCR (RT-qPCR) analysis 198 RT-qPCR was performed as previously described [8][9][10] . Briefly, total RNAs were 199 isolated from cells using TRI reagent (Sigma) according to the manufacturer's 200 instructions, and subjected to DNase I (Sigma) digestion to remove potential 201 DNA contamination. Reverse transcription was performed using an Improm-202 II TM RT kit (Promega, Madison, WI, USA) with RNase inhibitor (Promega), and 203 Random primers (Promega). The resultant cDNA was diluted to a working 204 concentration of 5ng/μl and stored at -20ºC for future using. Relative mRNA 205 expression level was defined as the ratio of target gene expression level to 206 18S expression level, respectively, with that of the control sample set as 1.0. 207 Primers were designed using Primer3-BLAST (National Center for 208 Biotechnology Information, USA) and the sequence for each primer was listed 209 in supplementary Table S1.

211
Immunoprecipitation (IP) assays. VSMCs transfected with control or 212 respective Cezanne over-expression plasmids were washed and harvested in 213 ice-cold lysis buffer (10 7 Cells/ml). After lysed and centrifuged in a 214 microcentrifuge at 4°C for 20 minutes, the supernatant was carefully collected 215 and placed in a fresh tube kept on ice. Equal amount of samples (40-50 μg) 216 were diluted into 1 ml immunoprecipitation buffer, and incubated with 2.5µg 217 anti-Ubiquitin (Rabbit IgG, ab7780) antibody, or equal amount of rabbit IgG at 218 4°C overnight under gentle rotation. After then, 70~100μL of the protein A-219 coupled Sepharose beads were added into each sample, and incubated at 220 4°C for 4 hours under gentle rotation. After washed the beads with washing 221 buffer three times, the immunoprecipitates were eluted from the beads using 2 222 x SDS loading buffer, and subjected to standard Western blot analysis.

224
Chromatin immunoprecipitation (ChIP) assay. The ChIP assays were 225 performed as described in our previous studies 4, 6 . VSMCs transfected with 226 pHM6, pHM6-Cez or pHM6-Cez-C209S were treated with 1% (v/v) 227 formaldehyde at room temperature for 10 min and then quenched with glycine 228 at room temperature. The medium was removed, and cells were harvested 229 and sonicated. The sheared samples were diluted into 1 ml 230 immunoprecipitation buffer containing 25 mM Tris-HCl, pH 7.2, 0.1% NP-40, 231 150 mM NaCl, 1 mM EDTA, and immunoprecipitation was conducted with 5µg 232 antibody raised against β-catenin (Rabbit IgG, ab32572), together with single-233 strand salmon sperm DNA saturated with protein-G-Sepharose beads. 234 Normal rabbit IgG was used as a control. The immunoprecipitates were eluted 235 from the beads using 100 μl elution buffer (50 mM NaHCO3, 1% SDS). A total 236 of 200 μl proteinase K solution was added to a total elution volume of 300 μl 237 and incubated at 60°C overnight. Immunoprecipitaed DNA was extracted, 238 purified, and then used to amplify target DNA sequences by  Promoter DNA enrichment with specific antibody was calculated using percent 240 input method with that of the IgG control set as 1.0. The relative level of 241 promoter DNA enrichment was defined as the ratio of promoter DNA 242 enrichments in the samples with treatment(s) (pHM6-Cez or oHM6-Cez-243 C209S) to the control samples (pHM6) with that of the control sample set as 244 1.0. PCR amplification of the murine CCN1 gene intron-1 regions were 245 included as additional control for specific promoter DNA enrichment. 246 247 RNA sequencing and data analysis 248 Total RNA was extracted from samples using TRI Reagent ® solution from 249 Sigma, and mRNA was purified from total RNA using oligo (dT) magnetic 250 beads. RNA quality control and cDNA library preparation was performed at 251 our in house Genome Centre at Queen Mary University of London 252 (http://www.smd.qmul.ac.uk/gc/Services/SeqRNA/index.html). The cDNA library 253 quality was determined on the Agilent Bioanalyzer 2100 system, followed by 254 sequencing on Illumina NextSeq 2000 system. Original image data generated 255 from NextSeq was transferred into sequencing reads through base calling, 256 and defined as raw reads. Partek ® Flow ® pipeline was used for sequencing 257 data analysis. Briefly, Adaptor sequences and low-quality sequences were 258 filtered out, and the remaining reads were mapped to the mouse genome 259 mm10 using STAR -2.6.1d, and no more than 2 mismatches were allowed 260 during the mapping read procedure. followed by nuclei staining with 4,6-diamidino-2-phenylindole (DAPI) (1ug/ml). 290 After mounting, the slides were examined using a laser scanning confocal 291 microscope (Zeiss LSM 510 Mark 4) and Zen 2009 image software. The 292 mean fluorescence intensity (MFI) for red (Cezanne, β-catenin and CNN1) 293 and blue (DAPI) fluorescence signal of the selected regions (media and 294 neointima layers, excluding endothelium of murine aortas) from each aortic 295 section were measured with Image J pro software by two experienced 296 investigators blinded to the treatments, and presented as the relative MFI 297 (target proteins over DAPI). Three sections were analyzed per vessel or aortic 298 roots, and averaged. 299 300 Atherosclerosis and Characterization 301 Cezanne transgenic gene-trapped (GT) (Cez GT/GT or Cez -/-, C57BL/6 302 background) mice used in our previous study 14 were crossbred with 303 LDLR −/− mice (C57BL/6 background, bred in house) to generate 304 Cez +/− /LDLR +/− double heterozygous mice. Cez +/− /LDLR +/− double 305 heterozygous mice were bred to produce Cez −/− /LDLR −/− double knockout 306 mice and their control littermates (Cez +/+ /LDLR −/− ). Eight-week-old male mice 307 were fed a high-fat diet (HFD) containing 21% fat, 1.25% cholesterol, and 0% 308 cholate (AIN-76A/Clinton-Cybulsky Cholesterol Series #3-108, 309 T-58R6-1810021, Test Diet Limited) for 12 weeks to induce atherosclerosis as 310 described in our previous study 15 . At the end of protocol, the heart harbouring 311 the aortic roots was carefully isolated and cut from the level above the 312 coronary artery at the base of the heart. The hearts were fixed in 4% 313 formaldehyde, and proceed for paraffin embedding and microtoming. The 314 extent of atherosclerotic lesions of aortic roots, and collagen content within 315 atherosclerotic lesions were analyzed by hematoxylin/eosin (H&E), and Sirius 316 Red staining, respectively. The atherosclerotic plaque, or Sirius Red-stained 317 area (refer to collagen content) in a given image was highlighted and 318 quantified (pixel 2 for lesion size; percentage over the atherosclerotic lesion 319 area for collagen content) by two experienced investigators blinded to the 320 treatments using Image J pro software. Three to six sections were analyzed 321 per aortic roots (or per mouse) and averaged.