Nuclear Control of Vascular Smooth Muscle Cell Plasticity during Vascular Remodeling

Abstract

Control of vascular smooth muscle cell (SMC) gene expression is an essential process for establishing and maintaining lineage identity, contractility, and plasticity. Most mechanisms (epigenetic, transcriptional, and post-transcriptional) implicated in gene regulation occur in the nucleus. Still, intranuclear pathways are directly impacted by modifications in the extracellular environment in conditions of adaptive or maladaptive remodeling. Integration of extracellular, cellular, and genomic information into the nucleus through epigenetic and transcriptional control of genome organization plays a major role in regulating SMC functions and phenotypic transitions during vascular remodeling and diseases. This review aims to provide a comprehensive update on nuclear mechanisms, their interactions, and their integration in controlling SMC homeostasis and dysfunction. It summarizes and discusses the main nuclear mechanisms preponderant in SMCs in the context of vascular disease, such as atherosclerosis, with an emphasis on studies employing in vivo cell-specific loss-of-function and single-cell omics approaches.

Graphical abstract

Basic, translational, and clinical research has led to breakthrough discoveries, enhancing knowledge and understanding of vascular diseases and providing a therapeutic arsenal for treating these conditions. International health organizations highlighted the advances and improvements in vascular disease diagnosis, therapeutic interventions, and management.^1,2 Yet, vascular diseases and their complications still represent the leading cause of mortality worldwide. In the United States, the prevalence of vascular disease (including hypertension) is at 49.2% in adults aged ≥20 years and increases with age in both male and female patients.¹ This high prevalence is associated with the combined incidence of modifiable (eg, hypertension, diabetes, dyslipidemia, obesity, smoking) and nonmodifiable (eg, age, sex, and genetics) risk factors.3, 4, 5, 6 These environmental and genetic factors have been implicated in altering vascular smooth muscle cell (SMC) phenotypes, functions, and contribution to vascular disease, at least partially through nuclear transcriptional and epigenetic processes. SMCs are contractile cells that display physiological and pathologic plasticity and retain the ability to modify their phenotypes in response to acute or chronic changes in environmental cues. The development of technologies, including cell fate mapping animal models and single-cell omics profiling, has profoundly changed our understanding of SMC behavior. Unambiguous in vivo tracking revealed that SMCs undergo a spectrum of phenotypic transitions characterized by the alteration in expression of the SMC gene repertoire and the gain in expression of genes related to the function and identity of other cell types.^7,8 The combination of in vivo fate mapping and single-cell transcriptomics further refined the characterization of SMC phenotypic modulation and transcriptional changes during vascular diseases.9, 10, 11, 12, 13, 14 Modifiable and nonmodifiable factors associated with cardiovascular disease (CVD) occurrence have been shown to influence SMC phenotype directly. Genomic variations determined by genome-wide association studies have been associated with vascular traits, CVD occurrence, and complications. Although the coronary artery disease–associated 9p21 locus was identified and studied since 2007,15, 16, 17 multi-ancestry, multi-ethnicity, and multitrait studies have identified additional associations between common or rare genomic variations and vascular diseases.¹⁸ Functional genomic studies demonstrated that many of these genomic variants directly influence SMC phenotype and function.¹⁹ Application of genome editing in induced-pluripotent stem cell-derived SMCs on the 9p21 disease locus provided unambiguous evidence of the direct causal relationship between the presence of the risk allele and aberrant SMC phenotype, consistent with atherosclerosis progression.²⁰ Integration of genome-wide association studies with single-cell omics assays further deciphered the functional role of significant genomic variants or single-nucleotide polymorphisms. Several groups have established a link between genomic variants, changes in chromatin accessibility determined by assay for transposase-accessible chromatin with sequencing, and transcription factor binding site accessibility.^21,22 On the other hand, acute or chronic environmental cue changes can lead to profound changes in epigenetic and transcriptomic programming and subsequent phenotypic changes. For example, inflammation, a hallmark of many vascular diseases, induces a phenotypic transition in SMCs distinct from ones caused by growth factors.²³ A bidirectional regulatory interaction between inflammation and epigenetics controls this SMC proinflammatory phenotype. Although inflammation modulates cell epigenetic landscapes,²⁴ differential histone modifications and DNA methylation are associated with increased expression of proinflammatory cytokines, amplifying further inflammation.²⁵ These examples illustrate the complex link between environmental cues, epigenomic regulators, and transcriptional regulators in controlling SMC gene expression and influencing their phenotypes and functions. In this review, we will summarize the evidence of the role of transcriptional and epigenetic mechanisms responsible for changes in SMC behavior and functions in vascular remodeling. As an inherently plastic cell type, we will also discuss the concept of epigenetic control of SMC lineage identity and memory and underline challenges and future directions in studying the complex epigenetic and transcriptional interplay in SMCs and vascular disease.

Transcriptional Regulators of Vascular Smooth Muscle Cell Phenotypic Modulation

Decades of research draw a complex yet incomplete picture of the regulatory network controlling the SMC contractile function and disease- and context-dependent phenotypic modulations. This section will focus on in vivo studies demonstrating the role of transcriptional mechanisms in regulating SMC function and biasing their fate and behavior in CVD (Figure 1).

Figure 1.

Transcriptional control of smooth muscle cell (SMC) phenotypic transitions in atherosclerosis. Schematic representation of the main SMC phenotypic transitions reported in atherosclerotic lesions with transcription factors promoting or inhibiting these transitions. There is still a lack of knowledge regarding the precise transcriptional control of several phenotypic states and their switch from one state to another. Note the figure focuses on the transcription factors discussed in the review's corresponding section. This schematic was generated with Biorender.com (Toronto, ON, Canada). KLF4, Kruppel-like factor 4; OCT4, octamer-binding transcription factor 4; PTEN, phosphatase and tensin homolog; RUNX2, Runt-related transcription factor 2; SEM, stem cell, endothelial cell, monocyte; SRF, serum response factor; TCF21, transcription factor 21; ZEB2, zinc finger E-box binding homeobox 2.

The SRF-MYOCD Complex

The seminal discovery of myocardin (MYOCD) paved the way for the exponentially expanding field of SMC transcriptional control of gene expression. MYOCD is a cotranscriptional factor selectively expressed in SMCs and cardiomyocytes.26, 27, 28 It functions as a potent coactivator of serum response factor (SRF) that binds to the regulatory DNA element CArG box [CC(A/T)₆GG], present in the promoters and introns of most SMC contractile genes.^26,27 In vivo studies consisting of the global,²⁹ SMC precursor-specific (Wnt1-Cre, Myocd^f/f; Pax3-Cre, Myocd^f/f),³⁰ or adult SMC-specific (Myh11-Cre^ERT2, Myocd^f/f)³¹ knockout of Myocd provide evidence that deficiency in the expression of the cofactor is incompatible with SMC differentiation, SMC contractile protein expression, and maintenance of SMC contractile function. Global knockout of Myocd results in embryonic lethality at embryonic day 10.5.²⁹ Histologic analysis of Myocd^−/− embryos reveal the absence of vascular SMC differentiation. Wnt1-Cre– or Pax3-Cre–mediated ablation of Myocd in neural crest-derived SMCs result in perinatal lethality because of patent ductus arteriosus.³⁰ Although patent ductus arteriosus is the primary vascular abnormality reported, a generalized defect in SMC contractile protein expression is observed in neural crest-derived vascular territories. Myocd expression in adult organisms is also critical for maintaining SMC homeostasis and preventing pathologic SMC transitions.³² Inducible deletion of Myocd in adult SMCs leads to the development of aortic aneurysms, visceral abnormalities, loss of SMC contractile gene expression, and premature death.³¹ SMC-specific ablation of Srf (Myh11-Cre^ERT2 or Tagln-Cre^ERT2, Srf^f/f) cause lethal intestinal pathology because of the impaired contractility of visceral SMCs.33, 34, 35 Dr. Miano's group recently developed an inducible Cre system driven by integrin subunit α8 (Itga8) promoter preferentially activated in vascular SMCs versus visceral SMCs.³³ Vascular SMC-specific deletion of Srf (Itga8-Cre^ERT2, Srf^f/f) induces vascular contractile incompetence, validating the relevance of SRF in controlling SMC contractility in vivo.³³ Together, these studies demonstrate that the SRF-myocardin complex governs the coordinate and cell-selective expression of the SMC contractile gene repertoire. Additional cofactors, like phosphatase and tensin homolog, interact and modulate MYOCD/SRF binding and transcriptional activity.³⁶

Transcriptional Regulators of SMC Dedifferentiation and Plasticity

SMC dedifferentiation is also actively regulated by several key master regulators, including the stem cell pluripotency factor Kruppel-like factor 4 (KLF4),37, 38, 39, 40 octamer-binding transcription factor 4 (Oct4),^10,41 transcription factor 21 (TCF21),^12,42 ETS transcription factor ELK1,^27,37,43 and Yes-associated protein 1,⁴⁴ to name a few. These pioneering studies have demonstrated that transcription networks actively and precisely orchestrate SMC phenotypic modulation in response to external signals, allowing SMCs to react to pathophysiological stimuli. Common features of these contractility repressors include their up-regulation in SMCs responding to dedifferentiation external signals, such as platelet-derived growth factor-BB (PDGF-BB), oxidative lipids, hypoxia, and mechanical stress; and the disruption of the MYOCD/SRF/CArG complex through decrease in MYOCD protein level or impairment of the physical interaction between MYOCD and SRF, and their binding to CArG box. SMC fate mapping combined with SMC-specific knockout of these transcriptional repressors provide evidence of the dynamic nature of SMC dedifferentiation in vivo and the fact that SMCs display a spectrum of phenotypes beyond the classically defined two-stage phenotypic switching (ie, differentiated/dedifferentiated). There is a consensus regarding the extensive plasticity of SMCs in atherosclerosis and their transitions to transitional/pioneer, fibroblast-like, macrophage-like, foam cell–like, and osteoblast/chondrocyte-like states.^13,14,45,46

SMC-specific deletion of Klf4 (Myh11-Cre^ERT2, Klf4^f/f) indicates that loss of expression of Klf4 does not prevent SMC dedifferentiation and plaque investment but biases SMC phenotype within the lesion. Klf4 knockout reduces SMC-derived transitional and macrophage-like populations within the plaque.⁸ The result of Klf4 deletion in SMCs is an overall decrease in atherosclerotic plaque size and an increase in indexes of plaque stability, including fibrous cap thickness and increased smooth muscle α-actin (Acta2)⁺ SMCs within the fibrous cap. Alencar et al¹⁰ further identified several key processes dependent on KLF4 expression during the atherosclerosis progression, including the transition of SMCs to a galectin 3 positive (Lgals3⁺) osteochondrogenic and procalcifying state. Importantly, the initial activation of Lgals3 identifies pioneer cell SMC responsible for early plaque investment and subsequent transition into several distinct phenotypic states. Multipotent SMC-derived mesenchymal stem cell (MSC)-like populations were also reported in other independent studies.^9,47,48 For example, Chen et al⁴⁷ discovered that loss of transforming growth factor-β signaling reprogrammed the SMCs to an MSC-like state that generated osteoblasts, chondrocytes, adipocytes, and macrophages during the progression of aorta aneurysm via single-cell RNA sequencing. They also identified that the increased level of KLF4 was responsible for the SMC reprogramming and aneurysm formation downstream of the SMC loss of transforming growth factor-β signaling.⁴⁷ Interestingly, SMC-specific knockout of the pluripotency gene Oct4 (Myh11-Cre^ERT2, Oct4^f/f) does not prevent or protect against atherosclerotic plaque development but is instead associated with an increase in plaque size and a decrease in plaque stability indexes. SMC fate mapping studies showed that Oct4 is required for the efficient investment of SMCs into the lesion and the formation of a protective fibrous cap.⁴¹ These studies suggest that transcription regulators do not simply function as the transcription switches governing the expression of contractile gene repertoire but regulate a wide range of cellular behaviors of SMCs that contribute to the etiology of cardiovascular diseases differently.

The expression and activity of transcriptional repressors have also been directly associated with single-nucleotide polymorphisms and genomic variants in genome-wide association studies and post–genome-wide association studies. TCF21, encoding the transcription factor TCF21, is located in the 6q23.2 locus, where multiple single-nucleotide polymorphisms have been associated with coronary artery disease. Wirka et al¹² reported that TCF21 is up-regulated in phenotypically modulated SMCs within atherosclerotic lesions in humans and mice. SMC-specific deletion of Tcf21 inhibits SMC phenotypic transition to the fibroblast-like cells, which is associated with a lack of formation and SMC investment of protective fibrous cap in atherosclerosis-prone mice. These modifications in SMC phenotypic states do not lead to changes in the overall lesion size. With a similar approach, Cheng et al⁴⁹ found that Smad3, another gene located within the coronary artery disease–associated locus 15q22.33, is implicated in SMC phenotypic bias in the context of atherosclerosis. SMC-specific loss of SMAD3 induces an increase in lesion size and outward remodeling, associated with SMC phenotypic transitioning toward the expression of genes associated with vascular remodeling, immune cell recruitment, and chondrocyte-like signature.⁴⁹ In another study combining single-cell RNA sequencing and single-cell assay for transposase-accessible chromatin with sequencing, the transcription factor zinc finger E-box binding homeobox 2 (ZEB2), located in the coronary artery disease–associated genomic region at 2q22.3, was also identified as a central regulator of SMC phenotypic transitions in atherosclerosis.⁵⁰ ZEB2 controls SMC phenotype by regulating the chromatin accessibility of genes involved in key signaling pathways, including the transforming growth factor-β and Notch pathways.⁵⁰ SMC-specific loss of Zeb2 promotes SMC phenotypic transition to a chondrocyte-like state and impairs the acquisition of fibroblast-like features without significantly increasing plaque burden.⁵⁰

In summary, a large body of research has depicted a complex regulatory network of transcription factors responsible for controlling SMC phenotype and phenotypic evolution during CVD (Figure 1). The precise transcriptional networks promoting or inhibiting specific phenotypic states and transitions between states still need to be defined and functionally evaluated. Further implementation of in vivo cell fate tracing, single-cell sequencing, and genome/epigenome editing studies will undoubtedly enable researchers to unbiasedly, comprehensively, and more precisely characterize transcription signaling pathways and functional outcomes on perturbation at the single-cell level.

Epigenetic Mechanisms Control Vascular Smooth Muscle Cell Gene Expression

Epigenetics is an ensemble of molecules and biochemical processes regulating gene expression without alteration of the DNA sequence,⁵¹ including the following: chemical modifications of DNA basis (eg, cystine methylation),⁵² post-translational modifications of histone tail amino acids,⁵³ noncoding RNAs [eg, miRNA, long noncoding RNA (lncRNA), and circular RNA],^54,55 and histone subunit variant deposition (eg, H2A.Z.).⁵⁶ These nuclear or cytoplasmic processes influence gene expression by promoting or inhibiting chromatin conformation and accessibility, transcription factor binding, gene activation, transcription, RNA splicing, stability, degradation, and translation rate. All these epigenetic programs have been implicated in regulating SMC differentiation, phenotypic modulation, and behavior under various physiological and pathologic conditions. This section will focus on the nuclear epigenetic mechanisms in controlling SMC phenotype.

Histone Modifications

Histone octamers (consisting of two copies of histone H2A, H2B, H3, and H4) and wrapped DNA (147 bp) constitute the chromatin core particle by forming nucleosomes. Each histone subunit's N-terminal tail can undergo many post-transcriptional modifications, including but not limited to acetylation, methylation, phosphorylation, ubiquitination, SUMOylation, and lactylation. These modifications regulate secondary and tertiary chromatin structure by modulating the level of chromatin compaction, DNA accessibility, and protein binding and recruitment. Different histone modifications often coreside at a given genomic locus, a process referred to as the histone code, conceptualized by Dr. Allis.⁵³ The integration of the histone code with other epigenetic mechanisms, such as DNA methylation, dictates the downstream effects on gene activation and transcription regulation by the specific recruitment of transcriptional factors, cotranscriptional factors, and chromatin remodeling complexes. The histone modification landscape is dynamically regulated by the catalytic activities of histone-modifying enzymes (HMEs), which either add or remove post-translation modifications on the histone tails. Among these post-translational modifications, lysine methylation and acetylation are the most studied modifications that display different functions.⁵⁷ HMEs can perform post-translation modifications of histones and nonhistone proteins⁵⁸ or regulate gene expression independently of their catalytic activities.⁵⁹ The multifaceted functions of HME complexify the interpretation of studies involving their manipulation. Histone acetylation is associated with transcriptional activation and is regulated by histone acetyltransferases (HATs) and histone deacetylases (HDACs). HATs use acetyl-CoA to add an acetyl group to the lysine residues. This weakens the interaction between nucleosomes and the DNA by neutralizing a positive charge on the lysine residues, thereby increasing chromatin decompaction and accessibility. Histone methylation, on the other hand, displays distinct regulatory functions determined by the following: i) the histone subunit on which the methylated lysine residue is located, ii) the position of the methylated lysine residue on the histone tail, and iii) the degree of methylation (monomethylation, dimethylation, or trimethylation). All these parameters dictate the activating or repressing role and the genomic distribution of the histone methylation (Table 1). Histone methylation is dynamically regulated by families of methyltransferases and demethylases. The complexity in studying epigenetic mechanisms resides in the nonselectivity of many HMEs, which can perform modification of several histone tail residues with opposite effects on gene regulation. For example, the histone demethylase lysine-specific demethylase 1 performs demethylation of the H3K4me1/2 (associated with gene activation) and H3K9me1/2 (associated with gene repression).⁶⁰ Pharmacologic or genetic inhibition of such enzymes will have complex effects on the histone modification landscape and gene regulation. Besides these limitations, the roles of histone modifications in regulating vascular SMC gene expression have been extensively studied.

Table 1.

List of Epigenetic Modifications, Their Main Effects on Gene Expression, and the Epigenetic Modifiers Adding or Removing These Modifications

Modification	Main effect on transcription	Modified residues	Enzymes adding histone PTM	Enzymes removing histone PTM
Histone methylation	Activation	H3K4me2/3	MLL1 (KMT2A)	LSD1/KDM1A (H3K4me2 only)
			MLL2 (KMT2B)	LSD2/KDM1B (H3K4me2 only)
			MLL3 (KMT2C)	JARID1A (KDM5A)
			MLL4 (KMT2D)	JARID1B (KDM5B)
			SET1A (KMT2E)	JARID1C (KDM5C)
			SET1B (KMT2F)	JARID1D (KDM5D)
			SMYD1	NO66
			SMYD2
			SET7/9
			PRDM9
		H3K79me2/3	DOT1L
	Repression	H3K9me1/2/3	SUV39H	LSD1/KDM1A
			G9a	JHDM2
			GLP	JHDM3
			SETDB1
			PRDMs
		H3K27me1/2/3	EZH1	JMJD3
			EZH2
	Transcriptional priming and poising (promoters and enhancers)	H3K4me1/2	KMT2 family	KDM1 and KMD5
		H3K79me1	DOT1L
	Regulation of alternative transcription and splicing	H3K36me1/2/3	SETD2/3	JHDM2
			NSDs	JHDM3
			SMYD2
			ASH1L
Histone acetylation	Activation	H2Aac	HATs	HDACs
		H2Bac
		H3ac
		H4ac
DNA methylation	Repression	5mC	DNMT1	TET1
			DNMT3A	TET2
			DNMT3B	TET3
	Activation	5hmC	TET1
			TET2
			TET3

ASH1L, ASH1 like histone lysine methyltransferase; DNMT, DNA methyltransferase; DOT1L, disruptor of telomeric silencing 1-like; EZH2, enhancer of zeste 2 polycomb repressive complex 2 subunit; GLP, G9a-like protein; HAT, histone acetyltransferase; HDAC, histone deacetylase; 5hmC, 5-hydroxymethylcytosine; 5mC, methylated cytosine; JARID, Jumonji and AT-rich interaction domain; JHDM, JmjC domain-containing histone demethylation; JMJ, Jumonji; KDM, lysine demethylase; KMT, lysine methyltransferase; LSD1, lysine-specific demethylase 1; MLL, mixed-lineage leukemia; NO66, nucleolar protein 66; NSD, nuclear receptor binding SET domain protein; PRDM, PR/SET domain; PTM, post-translational modification; SET, Su(var)3-9, enhancer-of-zeste and trithorax; SMYD, SET and MYND domain; SUV, suppressor of variegation; TET, 10-11 translocation.

Dr. Owens's group published one of the first reports investigating the role of histone modifications in the control of the SMC contractile gene expression in the early 2000s.⁶¹ In a model of retinoic acid–induced SMC differentiation using SMC precursor cells (P19-derived A404 cells),⁶² they discovered that SRF is not recruited on the CArG box regions of SMC contractile genes in A404 cells within intact chromatin despite high expression of SRF in undifferentiated precursors and the strong binding to CArG box by gel shift assay. Indeed, retinoic acid–induced SRF recruitment on these loci is associated with the enrichment in H3 and H4 acetylation in differentiated SMCs. These studies suggest that SRF’s regulatory role depends not only on its expression level but on the histone acetylation-mediated control of SRF DNA binding site accessibility. The role of HATs and histone acetylation during SMC differentiation was further confirmed. Trichostatin A–mediated HDAC inhibition induced a marked acceleration of retinoic acid–induced SMC differentiation.⁶³ Conversely, the inhibition of p300 catalytic activity, rather than p300 expression level, blunted SMC contractile gene expression. Further in vitro studies have shown an association between HAT recruitment and SMC differentiation, whereas HDAC-mediated histone deacetylation was observed in PDGF-BB–mediated SMC phenotypic modulation.^43,64 Mechanistically, HATs and HDACs interact with central differentiation or dedifferentiation transcription factors and cofactors, including paired like homeodomain 2 (Pitx2),⁶⁵ MYOCD,⁶⁶ and KLF4.³⁷ A recent study by Dr. Martin's group demonstrated that the HATs p300 and CREB-binding protein (CBP), rather than being redundant, have distinct functions in regulating SMC transcriptional programs.⁶⁷ Remarkably, SMC-specific deletion of these enzymes, which similarly perform histone H3 and H4 acetylation, had opposite effects on the regulation of SMC contractility and neointima formation in vivo. Although p300 mediates SMC contractility and SMC contractile gene acetylation, CBP promotes KLF4 expression and SMC dedifferentiation.⁶⁷ HDAC9 is implicated in various vascular diseases. Single-nucleotide polymorphisms in HDAC9 causing elevated expression of the deacetylase are significantly associated with increased CVD and atherosclerosis risk.⁶⁸ Mechanistic studies, including global and SMC-specific knockout, have demonstrated that HDAC9 participates in SMC phenotypic modulation in atherosclerosis, aortic dilation, and arterial stenosis by promoting the loss of SMC contractile gene expression.68, 69, 70 However, unselective targeting of HDAC9 with pan-HDAC inhibitors induces an unexpected detrimental effect on aortic diameter and structure in a mouse model of aortic aneurysms.⁷¹ These studies provide further evidence of the implication of histone acetylation in regulating SMC function but also highlight the difficulties of therapeutically targeting histone-modifying enzymes with pleiotropic effects or without selective inhibitors.⁷¹

SMC contractile gene promoters are also enriched in histone methylation, reflecting their activation status. In contractile SMCs, SMC marker gene promoters are enriched in H3K4me3 and H3K79me2, in addition to H3K9ac and H4ac. During phenotypic modulation, this profile of histone modification switches with the disappearance of histone acetylation and H3K4me3 and the enrichment in H3K9me3 and H3K27me3 combined with histone deacetylation.⁷² In contrast, H3K4me2 is present on the SMC contractile gene promoters, irrespective of their activation status.^72,73 The implications of these observations are discussed in Histone Modifications Govern SMC Lineage Identity and Memory. Besides SMC contractile gene activation, differential histone methylation drives smooth muscle transition to a proinflammatory phenotype in vitro and in vivo. H3K9me3 enrichment on proinflammatory genes is decreased in SMCs derived for db/db mouse aortas, which is directly associated with the activation of inflammatory pathways and production of proinflammatory cytokines.²⁵ Similarly, Harman et al⁷⁴ reported a reduction of H3K9me2 on proinflammatory genes (eg, Il6, Mmp3, and Mmp12) in injured and atherosclerotic vessels associated with the activation and increased expression of these genes. The H3K9 methyltransferase G9A/GLP (G9a-like protein) inhibition further potentiated SMC proinflammatory phenotype by preventing H3K9me2 enrichment. Overall, there is an increasing need for more comprehensive, cell- and context-specific investigations of individual histone modifications and HMEs to elaborate efficient epigenetic reprogramming strategies to bias SMC phenotypes and functions during vascular disease and remodeling.

DNA Methylation

DNA methylation is commonly associated with gene repression and chromatin compaction. DNA methyltransferases DNMT3A and DNMT3B perform de novo methylation of unmodified cytosines, whereas DNMT1 maintains the methylation pattern between cells during replication.^75,76 Although DNA methylation was believed to be a stable and irreversible process, the discovery of the 10-11 translocation (TET) enzymes mediating active DNA methylation reshaped the understanding of DNA methylation control of gene expression.⁷⁷ On binding to methylated DNA, TETs oxidize methylated cytosine into 5-hydroxymethylcytosine, 5-formylcytosine, and 5-carboxylcytosine, which is further converted into unmodified cytosine via DNA repair pathways.^78,79 Both DNA methylation and hydroxymethylation have been implicated in SMC phenotype regulation.

Alteration of DNA methylation profile has been investigated in the context of vascular remodeling and disease. Yet, whether DNA methylation plays a protective or detrimental role is cell and context dependent and remains to be further examined. Global DNA hypomethylation has been reported in human and mouse atherosclerotic lesions and neointimal SMCs after arterial injury, concomitant with decreased DNMT activity. Conversely, DNA hypermethylation of atheroprotective, cell-cycle inhibition, and prodifferentiation genes is found in SMCs in atherosclerosis and vascular injury, suggesting that DNA methylation could promote detrimental vascular remodeling. Similar discordance is found in studies employing DNA methylation pharmacologic inhibition with 5-aza-2′-deoxycytidine (5-Aza). Strand et al⁸⁰ showed that 5-Aza treatment promotes SMC differentiation and attenuates SMC proliferation and migration via up-regulation of phosphatase and tensin homolog. In vivo, administration of 5-Aza markedly inhibits vascular injury-mediated neointima formation. Meanwhile, 5-Aza–mediated DNMT1 inhibition is also associated with aortic stiffness and SMC phenotypic transition to a procalcifying cell.⁸¹ Administration of 5-Aza combined with CaCl₂ promotes aortic stiffness, including increased arterial rigidity, elastic modulus, and medial calcification (visualized by Alizarin red staining). Although these studies provide evidence that systemic treatment with 5-Aza impacts vascular remodeling and SMC behavior, they do not directly assess the primary role of DNA methylation and DNA methylation inhibition in SMCs in vivo, as the drug likely impacts other cell types that could secondarily affect SMCs.

The methylcytosine dioxygenase TET2 has been recently implicated in atherosclerosis and CVD. Somatic mutations impacting TET2 and causing a decrease in its expression are associated with clonal hematopoiesis and acceleration of atherosclerosis progression.⁸² Besides regulating hematopoietic cell homeostasis, TET2 plays a central role in controlling SMC differentiation and phenotypic state. Liu et al⁸³ were the first to report that TET2 directly controls SMC contractile gene expression by regulating the 5-hydroxymethylcytosine/methylated cytosine ratio on these loci. TET2 expression decreases during dedifferentiation and phenotypic modulation, including treatment with PDGF-BB. In vivo, global TET2 overexpression and knockdown causes abrogation or exacerbation of neointima formation, respectively. Interestingly, TET2-dependent hydroxymethylated cytosines (5-hydroxymethylcytosine) persist in medial SMCs, implying that 5-hydroxymethylcytosine could display intrinsic regulatory properties by the selective recruitment of transcriptional and epigenetic complexes. Overexpression or activation of TET2 by vitamin C prevents SMC apoptosis and synthetic phenotypes involved in transplant vasculopathy, suggesting enhancing TET2 activity could be a potential therapeutic strategy in pathologic SMC-mediated vascular remodeling.⁸⁴

Long Noncoding RNA

With advancement of sequencing techniques for lncRNA gene annotation, transcript quantification, and isoform identification (total RNA sequencing, cap analysis of gene expression sequencing, and global run-on sequencing) as well as functional characterization (cross-linking and immunoprecipitation sequencing and RNA immunoprecipitation sequencing), several functionally relevant noncoding RNAs have been identified in regulating SMC phenotype and behavior during vascular remodeling. lncRNAs are a class of noncoding RNAs with lengths >200 nucleotides. lncRNAs act through different mechanisms, depending on their subcellular localization. Although cytosolic lncRNA can regulate miRNA availability, mRNA stability, and translation rate, nuclear lncRNAs modulate the recruitment of transcriptional complex and epigenetic modifiers and directly impact chromatin remodeling, higher chromatin conformation, as well as transcription initiation and elongation.⁸⁵ lncRNAs can express tissue-, cell-, and context-dependent expression. In this section, we summarize evidence of the contribution of nuclear lncRNA in regulating SMC gene expression, although cytosolic noncoding RNAs (lncRNAs, miRNAs, and circular RNAs) also participate in SMC homeostasis and plasticity regulation.86, 87, 88, 89, 90

The gene CARMN, located upstream of the miR143/miR145 cluster, encodes the lncRNA CARMEN (cardiac mesoderm enhancer-associated noncoding RNA) selectively enriched in SMCs.⁹¹ CARMEN expression and nuclear localization are required to maintain a contractile phenotype.⁹² Mechanistically, CARMEN binds to MYOCD, but not SRF, and potentiates myocardin activation of CArG/SRF-dependent genes. A lack of CARMEN expression results in a blunted activation in MYOCD-overexpressing SMCs. Loss of CARMEN expression in SMCs exacerbates atherosclerotic plaque formation and injury-induced neointima formation because of uncontrolled SMC phenotypic modulation.92, 93 Increased lncRNA metastasis-associated lung adenocarcinoma transcript 1 (MALAT1) expression is associated with SMC contractile gene down-regulation through interaction in the nucleus with a chromatin remodeling complex in thoracic aortic aneurysms.⁶⁹ MALAT1 binds the histone deacetylase HDAC9 and is essential for HDAC9 nuclear translocation and the formation of a ternary complex consisting of MALAT1, HDAC9, and the chromatin remodeling enzyme Brahma-related gene 1 (Brg1). Recruitment of the MALAT1/HDAC9/Brg1 complex on the SMC contractile gene promoters is associated with an enrichment in the repressive histone modification H3K27me3. Global genetic deletion of Malat1 in a mouse model of thoracic aortic aneurysm (Fbn1^C1041G/+ mice) improves the disease's key features, including aortic dilation and SMC contractile gene expression. These examples highlight the essential roles of nuclear lncRNA in regulating gene expression in SMCs via interaction and partnering with other epigenetic mechanisms.

Nuclear Encoding of Smooth Muscle Cell Lineage Identity and Memory

Evidence of Lineage Memory in Vascular SMCs

SMC phenotypic modulation is an evolutionarily conserved process essential for adaptive and maladaptive remodeling. SMC plastic nature has been recognized as a hallmark of vascular disease. Nevertheless, the process of dynamic activation and repression of the SMC contractile gene repertoire has evolved to allow SMCs to respond to changes in environmental cues for optimal tissue perfusion, oxygenation, and nutrient distribution. Adequate participation of SMCs in adaptive vascular remodeling requires the following: i) the down-regulation of the SMC contractile apparatus genes and contractile function, ii) the activation of genes involved in cell proliferation, migration, or extracellular matrix production, and iii) the reactivation of the SMC contractile genes and restoration of contractile function. This process of dedifferentiation-redifferentiation has been characterized in vitro in response to starvation or on PDGF-BB withdrawal.^63,72 In vivo fate mapping studies in atherosclerosis and pulmonary hypertension have identified SMC-expressing ACTA2 and smooth muscle myosin heavy chain (MYH11) in remodeled vessels in location, inferring their prior proliferation and migration from the existing medial layer.^8,94,95 This unique property among muscle cells questions the mechanisms allowing SMCs to retain memory of their original lineage while dedifferentiated and not expressing their canonical repertoire of contractile genes.

SMC Heterogeneity and Transcriptional Identity

Besides their inherent capacity to modulate their phenotype, vascular SMCs are characterized by their notable heterogeneity.⁹ One of the main factors driving SMC heterogeneity is their embryonic origin (Figure 2). SMCs arise from distinct embryonic sources: the neural crest (ascending aorta, aortic arch, and carotid arteries), the second heart field (aortic root and ascending aorta), the proepicardium (coronary arteries), the somites (descending aorta), and the splanchnic mesoderm (abdominal aorta), to name the principal vascular territories.^96,97 Interestingly, memory of the embryonic origin is retained in adult SMCs as the embryonic origin correlates with the establishment of distinct transcriptional profiles. For example, differential expression of transcription factors in SMCs derived from the ascending and descending aorta has been reported. Ascending aorta-derived SMCs express exclusively the transcription factor heart- and neural crest derivatives-expressed protein 2 (Hand2), whereas SMCs from the descending aorta express homeobox A7 (Hoxa7).⁹ Beyond transcriptional diversity, SMCs from different embryonic origins display distinct behaviors during vascular disease. A compelling example is the differential behavior of neural crest– and second heart field–derived SMCs in the ascending aorta. Lineage-tracing studies have demonstrated that neural crest– and second heart field–derived SMCs colocalize in the ascending aorta while having spatially distinct distribution within the medial layer.⁹⁸ This region with mixed SMC populations is more susceptible to the development of aortic dilation and aneurysms. Several lines of evidence suggest that the colocalization of neural crest– and second heart field–derived SMCs drives detrimental remodeling. First, several transcriptomic studies have established that SMCs from different embryonic origins undergo different phenotypic modulation in the aneurysmal aorta.^99,100 For example, second heart field–derived and neural crest–derived SMCs display distinct transcriptomic profiles in the ascending aorta from Fbn1^C1041G/+ mice.¹⁰⁰ In these animals, second heart field–derived SMCs express extracellular matrix–related genes, whereas neural crest–derived SMCs activate genes related to the osteochondrogenic program. Second, the embryonic origin determines the responsiveness of SMCs to transforming growth factor-β or angiotensin-II, central pathways involved in aortic aneurysm formation and progression.^101,102

Figure 2.

Epigenetic control of smooth muscle cell (SMC) lineage identity and memory. The schematic represents SMC lineage determination and differentiation from different embryonic territories and precursors. In adult organisms, SMCs arise from various precursors, including neural crest, second heart field, and somites (other embryonic origins are not represented for clarity). A working hypothesis is priming of SMC contractile genes (SCGs) by H3K4me2 occurs early during lineage determination and biases undifferentiated cells toward SMC differentiation. H3K4me2 can be retained in dedifferentiated SMC during phenotypic modulation. Maintenance of this lineage memory mechanism could serve for SCG reactivation. In contrast, SCG-specific H3K4me2 demethylation leads to impaired contractility and a profound loss of lineage identity associated with exacerbated plasticity and transdifferentiation capabilities.

Histone Modifications Govern SMC Lineage Identity and Memory

The contribution of epigenetic programming during cell specification and lineage determination has been extensively studied. Early observations using electron microscopy showed that embryonic stem cells (ESCs) contain a homogenously loosely packaged chromatin structure. In contrast, differentiated cells present heterogeneous chromatin condensation by establishing euchromatic and heterochromatic regions.¹⁰³ Chromatin ultrastructure and organization changes are associated with the differential distribution and abundance of DNA and histone modifications during cell differentiation. In ESCs, DNA methylation is highly dynamic and consists of cycles of active methylation and demethylation.¹⁰⁴ More importantly, expression of the TET enzymes is required for both ESC self-renewal and preventing hypermethylation of the pluripotency factor genes,^78,105 and proper cell lineage specification.106, 107, 108 Concomitantly, a differential histone modification enrichment is observed between ESCs and lineage-committed cells with the selective distribution of active or repressive histone modifications for targeted and lineage-dependent gene activation or repression.¹⁰⁹

H3K4 methylation plays a central role during lineage determination and cell differentiation. First, H3K4 methylation is enriched in bivalent chromatin domains in association with the repressive histone modification H3K27 methylation.¹¹⁰ Bivalent domains poise developmental and lineage-specific genes and enhancers for context-dependent activation. Transition of bivalent chromatin domains from H3K4me/H3K27me to H3K4me/H3K27ac coincides with the activation of poised developmental genes and cell lineage commitment.¹¹¹ Conversely, further repression of bivalent chromatin domains on genes involved in alternative differentiation paths is mediated by the recruitment of additional repressive histone modifications and DNA methylation.¹¹² Second, developmental gene poising, selective activation, or repression implies that epigenetic modifiers and HMEs are essential in mediating cell differentiation and acquiring lineage identity. Genetic knockout of H3K4 methyltransferase mixed-lineage leukemia does not affect H3K4 methylation global abundance and maintenance of ESC stemness but affects skewed and delayed ESC differentiation.^113,114 Of interest, cardiac precursor-specific deletion of the H3K4me1/H3K4me2 methyltransferase lysine methyltransferase 2D (Kmt2d) alters cardiac development with defects of H3K4me2 enrichment on genes associated with cardiomyocyte-specific function, supporting a role for H3K4me2 in lineage-specific gene activation during development.¹¹⁵ Third, the H3K4 demethylase, lysine-specific demethylase 1 (LSD1), removes H3K4me2 from bivalent chromatin domains for further repression. Lysine-specific demethylase 1 deletion in ESCs disrupts pluripotency and drives precocious endodermal and mesodermal differentiation.¹¹⁶ Therefore, histone modifications, including H3K4 methylation, dynamic distribution, and chromatin structure reorganization, are required for lineage determination and cell differentiation. Yet, whether these mechanisms play a role in maintaining lineage identity and memory in adult organisms has not been extensively studied, despite their potential relevance with respect to cell plasticity or reversible phenotypic changes, as observed in SMCs.

H3K4me2 promoter occupancy is independent of the gene expression status and primes transcriptionally poised genes for subsequent activation. In contrast with H3K4me3, which is enriched at transcription start sites during active transcription, H3K4me2 occupancy spans proximal promoters, transcription start sites, intronic regions, and enhancer regions in a cell type– and cell state–dependent manner.¹¹⁷ D'Urso et al¹¹⁸ used a novel yeast genetic system in which cells express lysine 4 substitution to arginine (H3K4R) or alanine (H3K4A) to assess the role of H3K4me2 in mediating transcriptional memory. They demonstrated that H3K4me2 occupancy was stable during transient gene repression and was required to recruit RNA polymerase II to reactivate poised gene expression. H3K4me2 enrichment on the SMC contractile genes exhibits similar features: i) The presence of H3K4me2 on the SMC contractile genes seems specific to the SMC lineage and has not been reported in other differentiated cell types.⁷² ii) H3K4me2 enrichment occurs in SMC precursors treated with retinoic acid and irrespective of the presence of a functional CArG box, suggesting SMC-specific H3K4me2 occupancy is an early and SRF/myocardin-independent event during SMC differentiation.⁷² iii) The status of H3K4me2 enrichment on the SMC contractile genes is not impacted during transient PDGF-BB treatment, unlike other histone modifications such as histone acetylation. These observations align with the functional studies of d'Urso et al¹¹⁸ and the role of H3K4me2 in maintaining cell transcription memory. iv) H3K4me2 stable enrichment on the MYH11 promoter was detected in vivo in dedifferentiated SMCs, including ACTA2^–, ACTA2^– LGALS3⁺, stem cells antigen-1 (Sca1)⁺, and lipid⁺ SMCs.^8,73,119, 120, 121 Together, these descriptive studies suggest that the stable presence of H3K4me2 on the SMC contractile genes does not preclude their repression during phenotypic modulation but maintains the gene permissiveness for potential reactivation, as observed during transient PDGF-BB treatment or starvation. Recent studies conducted by our group elucidated the functional role of H3K4me2 in regulating SMC contractile gene expression, SMC phenotype, and functions. Selective demethylation of H3K4me2 on the SMC contractile gene was achieved by using a fusion protein construct consisting of the SRF binding domain of MYOCD and the H3K4 demethylase lysine-specific demethylase 1.¹⁴ H3K4me2 editing led to the loss of SMC contractile gene expression and vascular contractile properties, but there was also an alteration in the expression of proteins and pathways required for SMC lineage identity. More importantly, selective H3K4me2 erasure on the SMC contractile gene network induced a gain in H3K4me2 on genes associated with stemness and other developmental programs, further indicating a profound loss of lineage identity in H3K4me2-edited SMCs. Moreover, SMC plasticity was highly exacerbated in SMCs with H3K4me2 editing. Together, these results point towards H3K4me2 as part of a central mechanism controlling the maintenance of SMC lineage identity (Figure 2). Yet, investigation of SMC H3K4me2 enrichment and retention in vivo in healthy and diseased vessels has been partially achieved. Although there is clear evidence of retention of H3K4me2 on the MYH11 and ACTA2 genes in subsets of modulated SMCs in mouse and human disease vessels, nonquantitative and low-throughput methods have been employed, precluding the precise and exhaustive characterization of genome-wide variation in H3K4me2 occupancy during development, adaptive remodeling, acute injury, or chronic vascular disease.

Conclusions and Perspectives

The nucleus integrates and centralizes critical transcriptional mechanisms regulating SMC identity and behavior. In this review, we have discussed the following: first, key transcriptional master regulators orchestrate SMC plasticity and phenotypic transitions in response to specific environmental signals that contribute to vascular disease and remodeling; and second, epigenetic factors dictate chromatin organization conducive to SMC contractile gene expression and SMC differentiation. Alteration of these mechanisms (ie, epigenetic reprogramming) participates in transcriptional and phenotypic changes; third, the combined use of in vivo fate mapping, single-cell transcriptomic, and epigenomic methods in SMC plasticity has helped clarify the diverse states of SMC phenotypes; last, the maintenance of SMC lineage memory and identity involves a complex interplay of epigenetic factors. Yet, the implications of these discoveries in terms of the basic understanding of SMC biology and the translational relevance of transcription factor or epigenetic modifier pharmacologic targeting should be discussed. Most in vitro or in vivo functional studies have identified multiple mechanisms responsible for SMC phenotypic control but are designed to modulate a single factor at the time. Thus, challenges remain in studying the interwovenness of transcriptional and epigenetic mechanisms, as well as their possible redundant, cumulative, synergetic, combinatorial, and hierarchic nature. In addition, most transcription factors and epigenetic mechanisms presented in the review are ubiquitous and not selectively expressed or operational in SMCs. Unexpected results could arise from global and systemic inhibition of transcription factors and epigenetic modifiers. KLF4 is a well-characterized example of transcription factors with divergent effects in different cell types. Although KLF4 promotes SMC transition to atheropromoting phenotype and decreases the overall index of plaque stability, it plays a beneficial role in macrophages by virtue of biasing their phenotype toward an anti-inflammatory state.¹²² The field could leverage the capabilities of systems or computational modeling platforms through machine learning and molecular network mapping tools to identify cell- and context-dependent upstream regulators and downstream targets of these mechanisms. Negi et al¹²³ used computational modeling to identify differentially dependent networks of genes impacted by cancer drugs and the extrapolation of their putative effects and relevance in the context of pulmonary hypertension. A similar approach in integrating the cell-specific effects of pharmacologic targeting of transcription factors and epigenetic modifiers would further expand our understanding of the primary and secondary impact of such strategies. Finally, developing and using efficient tissue- and cell-selective drug or nucleotide delivery systems will greatly enhance the translatability of cell-specific loss-of-function and single-cell studies.

Author Contributions

I.A.A., M.L., and D.G. wrote, reviewed, and edited the article; D.G. acquired funding; and D.G. supervised the work.

Disclosure Statement

None declared.