The Human-Specific and Smooth Muscle Cell-Enriched LncRNA SMILR Promotes Proliferation by Regulating Mitotic CENPF mRNA and Drives Cell-Cycle Progression Which Can Be Targeted to Limit Vascular Remodeling

Supplemental Digital Content is available in the text.

analyse protein-coding gene differential expression. For the atherosclerosis RNAseq, we 82 obtained between 57 and 72M paired end read, allowing a deeper coverage. Gene 83 quantification (read count and FPKM) was obtained using RSEM (options: -bowtie2 -paired-84 end) 2 , based on GENCODE annotation (Release 25). For both RNA-sequencing experiments 85 the differential expression was performed utilizing edgeR 4 . We considered a threshold of 86 absolute Fold Change >=1.5 and adjusted pvalue<0.05 to identify significant changes 87 between two conditions. 88 Sample clustering was evaluated using the Principal component analysis (PCA) tool available 89 in DESeq2 on the regularized log transformed data. The 3D plot was obtained using the rgl R 90 package. 91 The gene ontology analysis was done using topGO 6 on enriched genes over a 92 background of expressed genes (FPKM>2 in at least one condition). Fisher's exact test was 93 used to calculate the p-values. 94 Both data set are deposited in the Gene Expression Omnibus (GEO) repository, study 95 number GSE120521 for the atherosclerosis RNAseq and GSE117608 for SMILR RNAseq. 96 97 Fluorescent Imaging 98 HSVSMC were transfected with siControl, siSMILR, or siAURKB, and plated at 1 x 99 104 cells per well in a 6-well plate containing coverslips, before being quiesced and then 100 stimulated with IL-PDGF for 96h. They were then fixed in 4% paraformaldehyde (PFA) for 101 10min and then permeabilised with 0.1% Triton X for 2-3min before adding Phalloidin-102 iFluor 488 reagent CytoPainter (Abcam) for 20min and imaged using Andor Revolution XDi 103 spinning disk confocal microscope. HSVSMCs infected with FUCCI virus as described 104 above were also imaged using the Andor Revolution XDi spinning disk confocal microscope. 105

Flow cytometry 107
Single-cell suspensions of FUCCI-infected HSVSMCs were fixed for 5 min in 70% 108 Ethanol, followed by an incubation in PBS/1% BSA containing for 1h at 4˚C

RNA-antisense Probe Pulldown 140
Antisense DNA probes were designed targeting the full length sequence of SMILR 1 141 or GFP (used as negative control) using the online designer at 142 https://www.biosearchtech.com/stellaris-designer. Probes mapping to tandem repeat 143 sequences or to homologous genes were excluded. 10 probes of 40bp were generated for 144 SMILR and 5 for GFP (Sequences in Supplementary Table 5). The SMILR probes were split 145 into two sets based on their relative positions along the gene sequence and were accordingly 146 referred to as even or odd capture oligo sets. All probes were biotinylated at the 3'end with an 147 extended 15 atom triethylene glycol spacer arm (IDT). 148 SVSMCs stimulated with IL-PDGF for 48h were rinsed once with PBS before 149 crosslinking with 2% formaldehyde at 37C for 10min. Crosslinking was quenched with 150 500mM glycine at 37˚C for 5min and then the cells were washed with ice cold PBS. 151 Subsequently, cells were scraped in ice cold PBS + 0.5% BSA Fraction V and moved into 152 Falcon tubes and centrifuged and pelleted at 1,000g. Thereafter, the supernatant was removed 153 and the pellet was resuspended in scraping buffer. At this point, cells were counted and the 154 resuspended pellet was divided into falcon tubes of 20 million cells each. These were then 155 pelleted again at 2,000g and snap frozen and stored in -80C for subsequent steps.
To prepare lysates for pulldown, cell pellets were quickly thawed in a 37 °C water 157 bath and resuspended in cell lysis buffer (50 mM Tris, pH 7.0, 10 mM EDTA, 1% SDS, and 158 added just before use: dithithreitol (DTT), phenyl-methylsulphonyl fluoride (PMSF), and 159 RNAse inhibitor) on ice for 10 min. Cell lysates were then diluted in double the volume of 160 hybridization buffer (500 mM NaCl, 1% SDS, 100 mM Tris, pH 7.0, 10 mM EDTA, 15% 161 formamide, and added just before use: DTT, PMSF, protease inhibitor, and RNAse inhibitor. 162 Probes (50 pmol each) were added to lysate, which was mixed by end-to-end rotation at 4°C 163 overnight. Streptavidin-magnetic C1 beads were washed three times in cell lysis buffer, 500μl 164 of beads was added to hybridization reaction and the whole reaction was mixed for 30 min at 165 37°C. Beads-biotin-probes-RNA adducts were captured by magnets (Invitrogen) and washed 166 five times with a wash buffer (2×SSC, 0.5% SDS, fresh PMSF added). After the last wash, 167 buffer was removed and beads were resuspended in 200μl RNA proteinase K buffer (100 mM 168 NaCl, 10 mM Tris, pH 7.0, 1 mM EDTA, 0.5% SDS) with fresh 1 mg ml−1 proteinase K 169 added (Ambion). After incubation at 50°C for 45 min, followed by boiling for 10 min, RNA 170 was spiked with 75ng of total C. elegans RNA to use AMA1 as a reference gene as 171 previously described 1 was isolated using the chloroform isoamyl approach. Eluted RNA was 172 subject to DNase treatment (TURBO DNase-free kit, Ambion) followed by qRT-PCR for the 173 detection of enriched transcripts. 174 175

Predicting SMILR-RNA interactions 176
Sum energy based ranking method was used to predicate SMILR and RNA interactions 8 . We 177 first filtered the top100 list based on their expression levels in IL1-PDGF stimulated vSMCs 178 in our control RNAseq samples (i.e. siControl and Null LNT). Genes falling below the 179 expression threshold of an FPKM < 1 were discarded. We then considered all the remaining 180 expressed genes and filtered through those that were found to be modulated following SMILR knockdown or overexpression. However, as we know that SMILR's upregulation and 182 function is dependent on IL1-PDGF stimulation, we further filtered the genes that were 183 differentially regulated with IL1-PDGF from 0.2% based on RNAseq data from Ballantyne el 184 al. 2016, 1 .

STAU1 Pulldowns 207
SVSMCs stimulated with IL-PDGF for 48h were rinsed once with PBS before 208 crosslinking with 2% formaldehyde at 37˚C for 10min. Crosslinking was quenched with 209 500mM glycine at 37˚C for 5min and then the cells were washed with ice cold PBS. 210 Subsequently, cells were scraped in ice cold PBS + 0.5% BSA Fraction V and moved into 211 Falcon tubes and centrifuged and pelleted at 1,000g. Thereafter, the supernatant was removed 212 and the pellet was resuspended in scraping buffer. At this point, cells were counted and the 213 resuspended pellet was divided into falcon tubes of 20 million cells each. These were then 214 pelleted again at 2,000g and snap frozen and stored in -80C for subsequent steps. 215 Cell pellets were thawed at 37˚C and resuspended in cell lysis buffer (50 mM Tris, pH 216 7.0, 10 mM EDTA, 1% SDS, and added just before use: dithithreitol (DTT), phenyl-217 methylsulphonyl fluoride (PMSF), and RNAse inhibitor) on ice for 10 min. 13. Samples 218 were then precleared with Dynabeads G (ThermoFisher) at 4˚C for 1 h using an inversion 219 shaker that rotates the tubes end-to-end. 220 Then using a DynaMag-2 (ThermoFisher), beads were collected against the magnet 221 and the supernatant was split evenly into two fresh Eppendorf. In each of the two tubes, fresh 222 Dynabeads G, cOmplete Mini, EDTA-free Protease Inhibitor Cocktail tablets (Roche) and 223 RNAse inhibitors were added, and then either 4µg of mouse IgG (Abcam) or Stau1 224 antibodies (Santa Cruz Biotechnology) were added. The samples were then rotating at 4˚C for 225 2-3 h before being washed using the DynaMag-2 5 times using NET-2 buffer (50 mM Trsi-226 HCl pH 7.4, 200 mM NaCl, 0.05% NP40, and cOmplete Mini, EDTA-free Protease Inhibitor 227 Cocktail tablets (Roche), and RNAse inhibitors). After the last wash, buffer was removed and 228 beads were resuspended in 200μl RNA proteinase K as described previously. After 229 incubation at 50°C for 45 min, followed by boiling for 10 min, RNA was isolated using the chloroform isoamyl approach. Eluted RNA was subject to DNase treatment (TURBO DNase-231 free kit, Ambion) followed by qRT-PCR for the detection of enriched transcripts. Systems, Erlangen, Germany). Both of these tracers have been used by our group and others 248 for plaque imaging and highlight high-risk actively calcifying 9 and inflamed or hypoxic 249 atherosclerotic plaques and were administered as previously described in 1,9 . PET tracer 250 uptake was quantified using an OsiriX workstation (OsiriX version 3.5.1 64-bit; OsiriX 251 Imaging Software, Geneva, Switzerland). PET/CT image data were reviewed for evidence of 252 tracer uptake, image quality and registration. The CT angiogram was examined to establish 253 plaque presence, location and characteristics. Regions of interest were then drawn on three 254 adjacent 3-mm PET slices to incorporate the internal carotid artery plaque. Three ROI were then drawn around adjacent healthy portions of carotid artery and the lumen of the SVC to 256 derive control values for "normal" arterial uptake and the blood pool respectively. Arterial 257 standardized uptake values (SUV) were recorded and also indexed to blood pool activity thus 258 giving a target-to-background-ratio (TBR). 259 At the time of surgery, plaques were collected immediately following excision and 260 photographed. Two-millimeter diameter core biopsy specimens for RNA analysis were taken 261 from regions of focally high uptake on PET and from normal tissue at the periphery of the 262 endarterectomy specimen. These, along with the main specimen, were immediately frozen 263 and placed in an -80°C fridge for subsequent batch analysis.

Quantification of CENPF protein level in HSV media
Human saphenous veins were harvested and fixed in 4% buffered formalin for 352 paraffin embedding, before being sectioned (5 μm) and mounted onto poly-L-lysine coated 353 slides. The tissue sections were deparaffinised and rehydrated by washing in Xylene for 354 5mins, followed by sequential washes in 100%, 96% and 70% ethanol respectively prior to 355 antigen retrieval in sodium citrate buffer (

Statistical Analysis
Statistical analysis was performed according to figure legends. Data are expressed as 376 mean ± SEM. For samples sizes ≥5 normal distribution was determined by Shapiro-Wilk test 377 followed by student t test or ANOVA. For each biological replicate, control and test 378 conditions were set up and accordingly analysed as pairs. Normal distribution cannot be 379 determined on small samples sizes and samples with n<5 were assumed to be not normally 380 distributed and subjected to Iman and Conover non parametric ranking followed by students t 381 test or ANOVA. Statistical significance is denoted by a P value of less than 0.05 (Graph Pad  LNT versus under quiesced conditions. ns = not significant, * = p < 0.05 by nonparametric 452 Iman and Conover ranked analysis and students t-test, n = 3 biological replicates. First ½ SMILR is denoted as blue and second ½ denoted as green boxes. Full length 553 represents the full sequence of SMILR. B) Biotinylated constructs, denoted in a, or bead only 554 controls were utilised as bait and incubated with protein lysates. Streptavidin beads were 555 utilised to capture bound proteins and washed with wash buffer to remove non-specific 556 interactions. Captured proteins were resolved by SDS-PAGE followed by immunoblotting 557 with Stau1 primary antibody. Arrowhead represents expected size of Stau1 protein (55kDa). Representative images of in situ detection of SMILR in 4 independent biological replicates of 576 plaques in the carotid artery derived from symptomatic patients at carotid endarterectomy. 577 SMILR is present in 4 of 5 plaques with intraplaque haemorrhaging using NBT/BCIP 578 (purple), and present in 3 of 5 plaques with thick fibrous cap ahteroma in varying intensity (i,  579 ii), while signal is absent after in situ hybridization with a scrambled (SCR) probe (iii, iv). 580 Nuclei are  other SMILR downstream targets within the cell cycle network, at day 0, 7 and 14. * = p < 620 0.05, Iman and Conover ranked non parametric analysis followed by students t test (n=3).