Transcriptome Profiling Reveals That The SM22α-regulated Molecular Pathways Contribute To Vascular Pathology

Objective: SM22α, also named transgelin, is a22kDa protein abundantin contractile SMCs, and physically associates with actin filament bundles.The expression of SM22α is down-regulated in vascular diseases includingatherosclerosis, abdominal aortic aneurysms and hypotension. Loss of SM22αin apolipoprotein E knockout (ApoE-/-) mice led to enlarged atheroscleroticlesions. Disruption of SM22α promoted arterial inflammation andchondrogenic conversion of VSMCs through activation of ROS-mediatedNF-κB pathways. Our recent studies revealed that overexpression of SM22αinhibited VSMCs proliferation and neointimal formation induced by ballooninjury via blockade of the Ras-ERK1/2signaling pathway. More recently, wedemonstrated that SM22α, as a PKCδ-regulating and PKCδ-regulated adaptorprotein, modulated vascular oxidative stress in vitro and in vivo throughPKCδ-p47phox axis via itself phosphorylation at Ser181site. All of theseliteratures suggest a strong correlation between decreased expression ofSM22α and vascular diseases. However, the question still exists whether thisdecreased expression of SM22α actively contributes to the pathogenesis ofvascular diseases.Methods: In this study, we generated the aortic transcriptomes ofSM22α-/-and SM22α+/+mice by RNA-Seq (RNA-Sequencing) to reveal thefunction and regulatory mechanism of SM22α. Next, we validated the analysisfrom RNA-Seq and found some non-coding RNAs by RNA-Seq. Weidentified one of the long non coding RNA TUG1, and studied the mechanismof TUG1in VSMCs.Results:1Transcriptome profiling analysis of the aortas of SM22α-/-and SM22α+/+mice 1.1Differentially expressed genes generated by RNA-SeqPutative differentially expressed genes generated by RNA-Seq wereidentified (2fold change cut-off). Using these criteria, there were a total of1398genes differentially expressed between SM22α-/-and wild-type mice. Ofthe1398differentially expressed genes,959(56.0%) genes were up-regulatedwhile439(25.6%) genes were down-regulated. Although the majority of thesesignificantly altered genes were protein-coding genes, approximately315(18.4%) were unknown/hypothetical function genes. We randomly selected12genes from those with both different expression patterns and interestingfunction, for qRT-PCR validation. Fold changes from qRT-PCR werecompared with RNA-Seq expression analysis results.Among the differentially expressed genes identified by SM22α-knockout,we found hemoglobin (Hb), apolipoprotein C-I (ApoCI) and G0/G1switchgene2(G0s2) were highly expressed. Specially, ApoCI is one of the mosthighly expressed (based on FPKM value) and significantly changed genescaused by SM22α-knockout.1.2GO analysis of differentially expressed genesGO annotated differentially expressed genes mainly belonged to thethree functional clusters (biological process, cellular component, andmolecular function). The differentially expressed genes in the cluster ofbiological process were found to be mainly related to signal transduction,developmental processes, transport and cell communication. Cellularcomponent GO terms of differentially expressed genes were related to theplasma membrane, synapse, extracellular region and cell junction. Themolecular function of GO terms were mainly involved of binding, catalyticactivity, receptor activity and structure molecular activity.1.3Protein classification of differentially expressed genes using PANTHERProtein classification of differentially expressed genes using PANTHER,ranked receptors to the top, which is very similar to that obtained frommolecular function GO database. Subdivision of the receptors class showed“G-protein coupled receptor” and “cytokine receptor” were the top two class receptors. In addition, transporter, hydrolase, signaling molecular and nucleicacid binding were also found as the major protein classes. Furthermore,protein classes more than50counts were also classified to enzymemodulators, transcription factors, cell adhesion molecules, cytoskeletalproteins, defense/immunity proteins, oxidoreductases and proteases.1.4Pathway analysis by KEGG, BioCarta and IPAThe most representative terms (higher level of the pathways) in KEGGincluded neuroactive ligand-receptor interaction, hematopoietic cell lineageand cytokine-cytokine receptor interaction. The most enriched pathways inBioCarta were associated with hematopoiesis-related cytokines, nuclearreceptors involved in lipid metabolism and toxicity, and cytokines modulatinginflammatory response. IPA canonical pathway analysis allowed furtherinsights into the molecular processes and pathways involved asSM22α-knockout. According to IPA, the specific enrichment pathways ofdifferentially expressed genes were observed for LXR/RXR activation,atherosclerosis signaling, communication between innate and adaptiveimmune cells, dendritic cell maturation, altered T cell and B cell signaling inrheumatoid arthritis, calcium and amyotrophic lateral sclerosis signaling.Overall, KEGG, BioCarta and IPA pathway analysis were all showed that theregulatory cytokines of hematopoiesis, inflammation and lipid metabolismwere significantly enriched as top ranked pathways.1.5Upstream regulator analysis by IPA and networks analysisUpstream regulator analysis by IPA can identify potential regulatorynodes despite the levels of gene expression. By this analysis, we found6upstream regulators caused by SM22α-knockout. It showed that TNF wasrespectively the target gene of NF-κB, RXRα, JUN and KLF2. CXCR4wasactivated by NF-κB, KLF2and NUPR1. HLA-B was significantlyup-regulated by both NF-κB and JUN. Col1a1and Col3a1were suppressed byNF-κB and NUPR1respectively. These results suggest that the regulators ofinflammation are activated in SM22α-/-mice.We used the IPA system to examine the potential functional networks of differentially expressed genes, and found that these genes were clustered into10significant functional networks. The main functional roles of thesenetworks involved in cell signaling, lipid metabolism and cardiovasculardisease. Diseases and functions analysis by IPA revealed that diseaseprocesses and biological functions caused by SM22α-knockout were mainlyrelated to arteriosclerosis, atherosclerosis, hypotension, disorder of artery,occlusion of artery and vascular disease.2Validation of RNA-Seq analysis: SM22α-/-mice revealed increasedtendency to develop vascular disease2.1SM22α-/-mice showed characteristics of pro-atherosclerosisA total of45genes were enriched in atherosclerosis predicted by IPA,indicating molecular changes by SM22α-knockout may have already initiatedthe early stage of atherosclerosis. Then we selected the enriched genes whichare the markers of early atherosclerosis involved in inflammation (TNF,IL-18) and lipid translocation or metabolism (LPL, ApoAII, Adiponectin,ApoCI) for further analysis. Quantitative real-time PCR showed that thesegene expressions had a strong correlation with SM22α-knockout. TNF wassignificantly up-regulated and NF-κB pathway was activated bySM22α-knockout. To confirm the relationship between loss of SM22α andinflammation, we subsequently examined whether disruption/overexpressionof SM22α affected TNF-α-mediated NF-κB activation in VSMCs. The resultsshowed that knockdown of SM22α using specific siRNA increasedphosphorylation and degradation of IκBα following TNF-α treatment.Disruption of SM22α activated NF-κB complex to translocate into the nucleuswhere it initiated gene expression, while overexpression of SM22α suppressedTNF-α-mediated NF-κB nucleus translocation. These results were inaccordance with the above bioinformation analysis by IPA. Another importantevent in atherosclerosis is dysfunction of lipid metabolism. Network analysisrevealed that lipoprotein ApoCI was a node molecular in cardiovasculardiseases. ApoCI was prominent expressed in the arteries of SM22α-/-micecompared with that in wild-type. Further, we found that suppression of SM22α by siRNA increased ApoCI expression levels in cultured rat VSMCs. Next, weperformed serum lipid assay between SM22α-/-and SM22α+/+mice. Theresults showed the levels of triglycerides were up-regulated but adown-regulation of serum total cholesterol. Meanwhile, genes enriched infatty acid metabolism were clustered by IPA software analysis.2.2SM22α-/-mice revealed increased tendency to develop vascular diseaseWe next validated some candidate genes which expressed at the initiationstage of atherosclerosis. We also hypothesized that SM22α-/-mice may havean increased tendency to develop vascular disease compared to the wild typebased on the above molecular information analyzed by us.We detected several molecules of differentially expressed genes based onGO analysis, which included monocyte chemoattractant protein-1(MCP-1),intracellular cell adhesion molecule1(ICAM-1), vascular cell adhesionmolecule1(VCAM-1) and matrix metalloproteinases (MMP-2, MMP-9) inSM22α-/-and SM22α+/+mice arteries by immunohistochemical staining. Theresults demonstrated that the expression of ICAM-1, MMP-2, MMP-9andVCAM-1were significantly elevated, while MCP-1showed no obviousexpression change both in SM22α-/-and SM22α+/+arteries.Next, we established the model of neointimal formation using partialligation of the left carotid arteries of both SM22α-/-and SM22α+/+mice. Theintimal thickness in SM22α-/-mice was more than that of wild type mice afterligation for14days. The expressions of MCP-1, VCAM-1, ICAM-1, MMP-2and MMP-9were all exacerbated in the partial ligation model of SM22α-/-compared with SM22α+/+mice by immunohistochemistry staining. ApoCI wasup-regulated during carotid neointimal formation compared with normalarteries and loss of SM22α increased ApoCl expression, especially in theinjured arteries by partial ligation. These results may imply that disruption ofSM22α influenced cellular adhesion and extracellular matrix degradation,enhanced the phenotype change of VSMCs and increased vascular response toinjury. 3Interaction of long non coding RNA TUG1with EZH2mediatedα-acitn methylation promotes the cortical cytoskeleton formation ofVSMCs3.1Sequence analysis of TUG1in rat VSMCsWe identified and cloned rat TUG1in VSMCs by PCR through sequencealignment and homology analysis from human and mouse. The rat TUG1waslocated in chromosome14and adjacent with MORC family CW-type zincfinger2, which was similar to that of mouse TUG1. The whole sequencelength of rat TUG1was4612bp. TUG1gene of rat contained four exons andhad93.0%and75.5%homology compared with mouse and human,respectively. The expression of TUG1was increased in cultured VSMCs.3.2The expression of cytoplasmic EZH2may participated in the formation ofcortical cytoskeleton formation of VSMCsPrevious studies showed that EZH2was mainly expressed in nucleus. Wefound that EZH2expression increased in both nucleus and cytoplasm ofVSMCs cultured with serum. Immunofluorescent analysis found thecolocalizaiton of EZH2with cortial F-actin in VSMCs, suggesting thatcytoplasmic EZH2may participated in the cortical cytoskeleton formation ofVSMCs.3.3Interaction among TUG1, EZH2and α-actin in VSMCs.Both TUG1and EZH2were found in nucleus and cytoplasm. Todetermine the relationship of TUG1with EZH2, we performed RNA pulldown assay. The results showed TUG1interacted with both EZH2and α-actin.Western blot and imunofluorescent analysis showed an increased cytoplasmicEZH2expression in siTUG1-transfected VSMCs stimulated by PDGF (15ng/mL) for30minutes.3.4TUG1/EZH2-mediated α-actin methylation participated in cortical F-actinformation of VSMCsTo identify whether disruption of TUG1influences F-actinpolymerization, we performed F-actin/G-actin ratio analysis in siCon andsiTUG1-transfected VSMCs. The results indicated that disruption of TUG1accelerated depolymerization of F-actin in VSMCs. The methylation sites of α-actin were predicated using PLMLA. The K193site of α-actin was stronglypredicted with methylation. We performed immunoprecipitation analysis forα-actin methylation using lysine antibody. The data revealed that α-actin wasmethylated, which was decreased by SiTUG1. These findings suggested thatEZH2-mediated methylation of α-actin was dependent on TUG1, and therebypromoted cortical F-actin polymerization.