If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
State Key Laboratory of Cardiovascular Disease, Fuwai Hospital, National Center for Cardiovascular Diseases, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, ChinaThe Cardiomyopathy Research Group at Fuwai Hospital, Beijing, China
Corresponding author at: State Key Laboratory of Cardiovascular Disease, Fuwai Hospital, National Center for Cardiovascular Diseases, Chinese Academy of Medical Sciences and Peking Union Medical College, 167A Beilishi Road, Xi Cheng District, Beijing 100037, China.
State Key Laboratory of Cardiovascular Disease, Fuwai Hospital, National Center for Cardiovascular Diseases, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, ChinaThe Cardiomyopathy Research Group at Fuwai Hospital, Beijing, ChinaDepartment of Cardiovascular Surgery, Fuwai Hospital, National Center for Cardiovascular Diseases, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China
Overview the application of single-cell RNA-sequencing technologies for studying heart valve development and disease.
•
Presentation of limitations and strengths of single-cell RNA-sequencing application in valvular science.
•
Review of novel cell subtypes implicated in heart valve development and disease using single-cell RNA-sequencing methods.
Abstract
From highly aligned extracellular fibrils to the cells, a multilevel ordered hierarchy in valve leaflets is crucial for their biological function. Cardiac valve pathology most frequently involves a disruption in normal structure-function correlations through abnormal and complex interaction of cells, extracellular matrix, and their environment. At present, effective treatment for valve disease is limited and frequently ends with surgical repair or replacement with a mechanical or artificial biological cardiac valve, which comes with insuperable complications for many high-risk patients including aged and pediatric populations. Therefore, there is a critical need to fully appreciate the pathobiology of valve disease in order to develop better, alternative therapies. To date, the majority of studies have focused on delineating valve disease mechanisms at the cellular level. However, the cellular heterogeneity and function is still unclear. In this review, we summarize the body of work on valve cells, with a particular focus on the discoveries about valve cells heterogeneity and functions using single-cell RNA sequencing. We conclude by discussing state-of-the-art strategies for deciphering heterogeneity of these complex cell types, and argue this knowledge could translate into the improved personalized treatment of cardiac valve disease.
Cardiac valve development is initiated by an epithelial–mesenchymal cell transformation (EMT) of endothelial cells in the atrioventricular (AV) canal. EMT begins in response to transforming growth factor-β (TGF-β). Firstly, the endothelial cells lose luminal-abluminal polarity, then extend filopodia and migrate into extravascular space as valvular interstitial cells (VICs) [
In the late embryonic stage, the cardiac valves comprise the atrioventricular valves (mitral valve and tricuspid valve) and the semilunar valves (aortic valve and pulmonary valve), each of which possesses a three-layer architecture: the collagenous fibrosa layer (Fbs), the intermediate glycosaminoglycan-rich spongiosa layer (Spg), and the elastin-enriched atrialis/ventricularis [
]. The three stratified layers of ECM within the valve leaflets are essential for stretch, compression, and strength in response to changes in the hemodynamic load. This ECM architecture is established during embryonic development and postnatal growth, and maintained by VICs that reside within the core of the valve leaflet. Overlying the ECM and VICs is a layer of valvular endothelial cells (VECs) that provide a physical barrier against the hemodynamic environment. Cells interact with ECM [
Calcified aortic valve disease (CAVD) is the most common aortic valve disease, affects 12.6 million people worldwide, and brings a huge burden to individuals and society [
]. CAVD pathogenesis is unclear and is characterized by inflammatory infiltration, fibrotic extracellular matrix synthesis, leaf thickening and increased firmness, calcified mineral deposition, and neovascularization [
]. DMVD is defined as myxomatous degeneration, which is characterized by leaflet thickening, increased leaflet cellularity, proteoglycan accumulation, and collagen fragmentation.
CAVD and DMVD are important causes of disease burden among older adults. At present, the only choice for the treatment of these valve diseases is valve prosthesis implantation. The search for therapeutics and early diagnostics is challenging as valve disease presents in multiple pathological stages. In addition, the development of valve disease involves several highly plastic multipotent resident cell populations [
]. Therefore, defining the diseased valve specific cell subclusters may provide insight into disease cause and therapeutic options.
In this review, we first introduce the development of cardiac valve and the two most common valve diseases. Next, we address the question of whether valve cells may comprise several subsets based on current preclinical and clinical evidence. Then, we demonstrate the function of different valve cell subsets in normal or diseased condition. Finally, we conclude by discussing tentative state-of-the-art strategies for deciphering heterogeneity in these difficult cell types, and how this knowledge could promote diagnosis, treatment, and future precision valve medicine.
Contribution of valve cells to aortic valve disease
Valvular endothelial cells
VECs line the surfaces of the leaflets. Although VECs express CD31, von Willebrand factor (vWF), and VE-cadherin, in pigs, they display different transcriptional and proliferative profiles compared with aortic endothelial cells (ECs) [
]. Furthermore, the VECs on both sides of the aortic valve have completely different phenotypes. Differential gene expression analysis determined that the fibrous layer has lower expression of vascular and valve calcification inhibitors (osteoprotegerin, OPG; C-type natriuretic peptide, CNP; parathyroid hormone, PTH) [
Human pulmonary valve progenitor cells exhibit endothelial/mesenchymal plasticity in response to vascular endothelial growth factor-a and transforming growth factor-beta2.
]. Although VECs and EMT are well recognized, the contribution of VEC subpopulations to calcific valve disease and their focal genesis still needs to be clarified.
Valvular interstitial cells
VICs, the most abundant cell type which locate in each layer of the valve, originate from VECs and highly express ECM proteins (COL1A1 and COL3A1). VICs display regional, adhesional, and synthetic heterogeneity [
]. In human normal aortic valve, a special subpopulation of the VICs express both hematopoietic (CD34+) and mesenchymal stem cell markers (PDGFRa+) akin to interstitial Cajal-like cells/telocytes (TCs), which might be involved in the maintenance of local microenvironment resisting to pathologic remodeling [
Identification of CD34+/PGDFRα+ valve interstitial cells (VICs) in human aortic valves: association of their abundance, morphology and spatial organization with early calcific remodeling.
]. A pathological VIC (VICp) mesenchymal-like phenotype is confirmed by CD90+/CD73+/CD44+ expression and forms microvessels by differentiating into perivascular cells [
], pointing to a novel function of VICp in valve vascularization during CAVD. Additional cell lineage studies from pathological and control valves and appropriate in vivo models are needed to fully demonstrate that VICs transdifferentiate into perivascular cells and that this occurs in vivo in calcific valves.
Recent studies have reported dipeptidyl peptidase-4 (DPP-4) [
] participate in tissue repair and osteogenic induction in porcine models.
Immune cells
Almost all valve leukocytes (CD45+) are myeloid cells, consisting of at least two differentially located macrophage subsets and dendritic cells. After birth, the number of CCR2+ macrophages increase, which is consistent with the infiltration population of monocytes [
Macrophages might contribute to the calcification. In human calcified aortic valves, CD206+ M2 and HO-1+ Mox, which express BMP2, are osteoblast-like cells candidates [
]. Besides, inflammatory M1 macrophages may drive a myofibroblast-to-osteogenic intermediate porcine VIC phenotype, which may mediate the switch from fibrosis to calcification during aortic valve stenosis progression [
]. A summary of aortic valve cells in normal or pathological conditions is presented in Fig. 1. Detailed information is summarized in Table 1.
Fig. 1A summary of aortic valve cells in normal or pathological condition. The figure shows heterogeneity (indicated by same cell types in distinct colors) and transition (indicated by single line, blue arrows, and different cell types) of subpopulations and their capability to adjust their phenotype to the diseased environment.
Human pulmonary valve progenitor cells exhibit endothelial/mesenchymal plasticity in response to vascular endothelial growth factor-a and transforming growth factor-beta2.
Modulation of transforming growth factor-β signaling and extracellular matrix production in myxomatous mitral valves by angiotensin II receptor blockers.
Contribution of valve cells to mitral valve disease
Increased inflammatory cells composed of macrophages and T cells have been described in human myxomatous mitral valves. Endocardially derived macrophages play a phagocytic and antigen presenting role, as well as valve remodeling role [
]. Diseased mitral valves of Marfan syndrome (MFS) mice exhibited a marked increase of infiltrating (MHCII+, CCR2+) and resident macrophages (CD206+, CCR2−), along with increased chemokine activity and inflammatory extracellular matrix modification [
], suggesting sterile inflammation as a novel paradigm to disease progression, and monocytes as a viable candidate for targeted therapy in DMVD. Interestingly, immune cells have been detected in the valve leaflets of both primary and secondary DMVD [
], challenging the traditional classification of DMVD as a “non-inflammatory” degenerative valve disease; additional studies are needed to determine what initiates leaflet thickening and matrix accumulation prior to monocyte infiltration as a result of Fbn1 mutations in the context of MFS.
Human VICs up-regulate TGF-β signaling and differentiate into myofibroblast as the DMVD progresses [
Modulation of transforming growth factor-β signaling and extracellular matrix production in myxomatous mitral valves by angiotensin II receptor blockers.
]. Therefore, attenuation of TGF-β signaling by angiotensin II receptor blockers may represent a mechanistically based strategy to modulate the pathological progression of DMVD in patients. Much is known about the cellular changes and the contribution of activated myofibroblasts to the valve ECM remodeling. However, little is known on how activated myofibroblasts survival might contribute to DMVD. Recently it was found that activated myofibroblasts survival in DMVD canine valves might be a consequence of heightened resistance of VICs to apoptosis [
Ischemic mitral regurgitation is a complication after myocardial infarction (MI), which can induce adaptive mitral valve (MV) response. This response may be beneficial initially, but eventually lead to lobular fibrosis and MV dysfunction. A small population of VECs (~2.5%) express CD45, which is significantly increased after myocardial infarction (MI) in sheep [
], thus the contribution of CD45+ VECs to mitral valve adaptation and fibrosis post-MI warrants investigation. More comprehensive information on understanding cell behavior is summarized in Table 1.
Single-cell RNA sequencing
The above makes us have a preliminary understanding of the heterogeneity of valve cells and the contribution of various subtypes to valve diseases by using conventional experimental methods. The in-depth study of various valve cells using single-cell RNA sequencing (scRNA-seq) may further promote our understanding in cardiac valve homeostasis and disease.
scRNA-seq analysis mainly includes unsupervised cell population clustering, pseudotime analysis, and intercellular communication analysis. It enables researchers to zoom in on cell subclusters and conduct more in-depth studies on them, revealing subtle changes unique to each cell, greatly advancing the field of transcriptomics. The unbiased analysis of cellular differences can identify cellular heterogeneity and even rare populations. Pseudotime analysis introduces a new concept of pseudotime measurement, which can use unsupervised time ordering of cell transcriptome at specific time points of development, cell differentiation, and disease progression. The intercellular communication analysis can also reveal cell-cell interactions, which could suggest unexpected cellular involvement in the development mechanism of valve or the pathogenesis of valvular disease. Those data are hardly revealed by conventional experimental methods [
]. Over the past decades, a sudden increase in the development and application of scRNA-seq techniques has allowed the identification of cellular phenotypic alterations during cardiovascular development and disease progression [
scRNA-seq provides a compilation of gene expression characteristics to further detail the complexity of valve development. Valve development is a result of a balance between cell proliferation and apoptosis. During human embryonic valve development, atrioventricular VICs and semilunar VICs have distinct expression signatures, the valve cell apoptosis begins at embryonic 20 W (E20W) and decreases at E25W [
], providing a deeper understanding of in vivo human fetal cardiac valve development and may help improve our ability to differentiate and mature functional valve cells from pluripotent stem cells in vitro. During murine postnatal valve development, the VECs population includes three spatially distinct subpopulations that are similar at postnatal 7 (P7) and P30. One highly expresses Hapln1 and displays a gene signature related to cell migration and actin filament polymerization [
]. A collagen-producing (Col) VICs population and a glycosaminoglycan-producing (GAG) VICs population are identified at P7, while Tcf21+ VICs and antigen-presenting (Cd74+) VICs express genes involved in antigen processing or defense response respectively, are identified at P30. These suggest that VICs have an unappreciated heterogeneity in valve ECM production and homeostasis after birth [
Integrating lineage tracing and scRNA-seq provides a robust strategy for establishing and testing models of how individual stem cells differentiate and self-renew over time. Combining scRNA-seq with lacZ reporter mice highlight Adamts19 as a novel marker for VICs, indicating that Adamts19 is required for proper valve maturation and maintenance [
]. A Wif1+ cardiac fibroblast subtype locates at the valve interstitial in PdgfraGFP mice, showing transcriptional similarity to skeletal muscle perimysial cells [
]. The similarities observed between perimysial and VICs may suggest common molecular regulators of cardiac valve calcification and heterotopic ossification of the skeletal muscle. Combining scRNA-seq with gene ablation technology in mice, Pitx2 is revealed to be required for the proper differentiation or specification of endocardial cells [
]. However, the mechanisms leading to this heterogeneity are unclear. Further study of VECs in mice may help understand why different VECs exhibit varying susceptibility to developing pathogenic characteristics.
scRNA-seq in valve disease
Genetically modified animals have advanced our understanding of valve development. Yet, pathophysiology of human valve remains poorly understood. By combining scRNA-seq and in vivo approaches, a population of human pre-valvular endocardial cells (HPVCs) can be derived from pluripotent stem cells and faithfully recapitulates valve disease in a dish [
]. Single-cell RNA profiling of human induced pluripotent stem cell (hiPSC)-derived endocardium with an underdeveloped left ventricle identifies a developmentally impaired endocardial population, providing a rationale for considering endocardial function in regenerative strategies [
Pseudotime analysis reveals 6 novel valve-derived stromal cells (VDSCs) in the leaflets of CAVD patients particularly and provides evidence of endothelial to mesenchymal transition (Endo-MT) involved in lesion thickening of the aortic valve leaflet. A pseudotime subset of 2 VEC clusters and 3 VICs clusters (FOShigh, HSPA6high or SPARChigh) showed 11 differentiation states in the cell populations. Pseudotime differentiation trajectory analyses revealed that the differentiation of VICs and VECs was earlier than that of the VDSCs. Pathway enrichment revealed that Endo-MT was highly correlated to AGE-RAGE signaling pathway in diabetic complications and PI3K-Akt signaling pathway. These results suggested that the Endo-MT process was active during aortic valve calcification and thickening [
Cell-type transcriptome atlas of human aortic valves reveal cell heterogeneity and endothelial to Mesenchymal transition involved in calcific aortic valve disease.
]. At present, there is no direct evidence that Endo-MT is involved in the progress of human CAVD. Based on the results of this pseudotime differential trajectory, the authors clearly observed VECs distributed at the earliest differentiation state during the process of CAVD production. The transition cell population was identified by trajectory analysis combined with cell localization experiment, which further proved that Endo-MT was involved in the progress of CAVD. However, it is necessary to establish an appropriate preclinical CAVD model using transgenic mice. Cre/loxP lineage tracing technology can use markers identified by scRNA-seq to confirm the fate of cells associated with CAVD in future studies. Further, marker genes and their regulators of transit cell population should be functionally examined, for example, using Crisp-Cas9 technology. The cellular heterogeneity in cardiac valve explored by scRNA-seq is presented in Fig. 2. More comprehensive information on understanding cell behavior by scRNA-seq is summarized in Table 2.
Fig. 2The cellular heterogeneity in cardiac valve explored by single-cell RNA sequencing. The different conditions (development, regeneration, healthy, and calcification) with involved cell types. The figure shows heterogeneity (indicated by same cell types in distinct colors) and transition (indicated by single line, blue arrows, and different cell types) of all these subpopulations and their capability to adjust their phenotype to the environment.
Until today, there is no drug therapy to cure CAVD or MDVD. The only effective treatment is heart valve repair or replacement. Up to 300,000 valve replacements are performed worldwide every year [
]. Strikingly, the prevalence of heart valve disease increases with population growth and aging. It is expected that the number of patients requiring valve replacement surgery will triple by 2050 [
]. Since the 1960s, cardiac valve substitutes have experienced a gradual evolution.
Prosthetic heart valves
Currently available valve replacement devices include mechanical heart valve and biological heart valve. Within 10 years after the first heart valve replacement, more than half of patients have complications. These complications include embolic events, bleeding, valve obstruction (due to thrombosis or pannus), prosthetic valve endocarditis, structural degradation, perivalvular leakage, hemolytic anemia, and mismatch of prosthetic valves with patients. Each type of substitute valve has its unique complications. For example, thromboembolism- and anticoagulation-related problems are the most common complications after mechanical heart valve replacement.
Although these valve replacement procedures have improved the survival rate and quality of life of many patients, the ideal prosthetic heart valve with abundant availability, immune compatibility, capable of growth, self-repair, functional adaptation, and life-long performance has not been developed.
Tissue-engineered heart valves
Tissue-engineered heart valves (TEHV) can solve the unmet demand for replacement heart valves that can grow, self-repair, and provide life-long performance, in which bioabsorbable biomaterials or allogeneic/xenogeneic matrix are implanted, combined with cells to produce living valve in patients with heart valve disease. TEHV have been explored for almost 25 years [
Transcatheter implantation of homologous “off-the-shelf” tissue-engineered heart valves with self-repair capacity: long-term functionality and rapid in vivo remodeling in sheep.
]. However, they still have major disadvantages, including insufficient stability of leaflet structures and uncontrolled balance between polymer biodegradation and extracellular matrix formation, which eventually leads to the failure of these structures [
]. Raw polymer material can be designed to be closer to the material characteristics of native heart valve, and polymer film can easily form geometric shapes that allow physiological flow. Therefore, polymeric heart valves may be more durable than biological heart valves, while avoiding the long-term anticoagulant treatment required by mechanical heart valves. In addition, typical polymer heart valve manufacturing technologies, such as impregnation molding and compression molding, have been fully studied and widely used in the plastic industry [
], and the commercial application of polymeric heart valves is limited to ventricular assist devices. The remaining mismatch with the complex mechanical properties of natural heart valves, coupled with imperfect valve geometry and surface morphology, seems to be the main potential cause of structural degradation, calcification, thrombosis, and limited overall durability of polymer materials [
In general, surgery is not the optimal solution. New drugs need to be developed to reverse valve disease. However, our lack of medical options stems from our limited understanding of the cellular mechanisms of heart valve disease and how cells interact with ECM to affect valve function. Improving our understanding of the cellular mechanisms that control ECM production and remodeling is the key to understanding valve function and pathological dysfunction. Besides, understanding the mechanobiological principles that control the development, growth, and remodeling of heart valves is essential for the development of effective medical therapies. When we begin to clarify the cellular mechanisms of heart valve disease, we should be able to reasonably design treatments to prevent, reduce, and even reverse progressive heart disease. Based on the structure function correlation, the function of heart valves is determined by the biomechanical properties of ECM, which is determined by the valve cells.
Discussion
By summarizing the body of work on valve cells, with a particular focus on the discoveries about valve cells heterogeneity and functions, it can be seen that the cytological mechanism of cardiac valve disease is complicated. Each cell subtype participates in the occurrence and development of valve diseases. The explosion of single-cell technology provides an efficient solution to identify these heterogeneous subtypes and establish the spatiotemporal dynamic model of valve biology and pathological cell composition. Specific cell subtype markers are expected to become diagnostic criteria, therapeutic targets or prognostic indicators. Current applications and perspectives of scRNA-seq for cardiac valve development and disease are summarized in Fig. 3.
Fig. 3The application of single-cell RNA sequencing (scRNA-seq) for cardiac valve studies. The most popular scRNA-seq is 10xGenomics technology. Valve tissue is dissociated into single cells then undergo microfluidics to be barcoded. Barcoded cells are sequenced and analyzed. Cells are clustered, pseudotime is performed to find the differentiation trajectory. Differential gene expression is conducted to evaluate cell heterogeneity based on cell clusters or specific groups. The use of documented ligand-receptor pathways in available databases can predict intercellular communication between single cells. The application of scRNA-seq in clinical research combined with preclinical models will improve the development and efficacy of personalized treatment for patients with cardiac valve disease.
Strikingly though, there are currently no published scRNA-seq datasets containing rheumatic valvular disease and infective endocarditis, be it in mice or humans. In our experience, valves from patients with such valvular diseases are seriously calcified and full of bacteria. Such specimens contain a large number of mineralized granules, impure particles, and bacteria that are difficult to filter, resulting in insufficient cell viability and purity. Flow cytometry can be used to improve the cell quality of scRNA-seq, but a large number of cells will be lost after this step, and the supply of valves itself is insufficient. Experimenters need to constantly improve the method of separating single cells to achieve the best sequencing state.
scRNA-seq allows us to better understand how genes affect the whole process of individual traits by affecting the phenotype of cell subsets. However, most sequencing techniques can only obtain one-dimensional information from a batch of cells, and other histological data will be lost. It is necessary to obtain a variety of histological information of a single cell, which means that we can establish links between different histological data to better describe cell function and its internal regulatory process. Combining several of these dimensions of data into an integrated multi-omics analysis of the same single cell will generate unprecedented knowledge in the field of valve medicine.
scRNA-seq requires fresh cell suspension, thus the spatial location information of cells is destroyed. However, the spatial location of cells plays a critical role in the progression of diseases. Therefore, spatial transcriptome technology has been developed, a new technological breakthrough that reveals the genetic activity in tissues, and shows the precise regional localization of these activity signals. Further, valvular interstitial cells are mainly derived from endocardial cells. However, single cell analysis needs to distinguish the valve leaflets because the cell sources of the mesenchymal cells in the leaflets are different, which may lead to valve cell heterogeneity. Specifically, the cells of the posterior aortic valve leaflet and the anterior pulmonary valve leaflet originate from TNNT2+ cells, the second heart field progenitor cells [
]. Therefore, spatial transcriptome is a good choice to explore the difference of leaflets. However, unlike myocardium, the healthy valve leaflet is too thin to extract enough RNA, and the calcified valve leaflet contains calcium crystals, which will affect the sampling purity. All these cannot be overcome for the time being, but with the continuous development of spatial transcriptomes, there will be better solutions in the future.
The combination of spatial transcriptome and scRNA-seq may change our conceptual understanding of valve biology and disease state, and has far-reaching significance in revealing new valve disease mechanisms and cell type specific pathways. Further study on the functional heterogeneity and intercellular interactions of human valve tissues may clarify the pathophysiological processes related to diseases, and provide a basis for determining new ways of valve disease regulation and developing new treatment strategies for early prevention and even reversal of valve diseases.
CRediT authorship contribution statement
MX.F reviewed recent published papers and drafted the manuscript. JP·S contributed to conception and critically revised the manuscript.
Funding
This work was funded by grants from the National Natural Science Foundation of China (82000225, 31601063), and the Beijing Nova Program (grant no. Z20110006820003).
Declaration of competing interest
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Acknowledgments
We thank Xiao Chen, the staff, and the graduate students of Fuwai Hospital who gave us valuable advice for this project.
Human pulmonary valve progenitor cells exhibit endothelial/mesenchymal plasticity in response to vascular endothelial growth factor-a and transforming growth factor-beta2.
Identification of CD34+/PGDFRα+ valve interstitial cells (VICs) in human aortic valves: association of their abundance, morphology and spatial organization with early calcific remodeling.
Modulation of transforming growth factor-β signaling and extracellular matrix production in myxomatous mitral valves by angiotensin II receptor blockers.
Cell-type transcriptome atlas of human aortic valves reveal cell heterogeneity and endothelial to Mesenchymal transition involved in calcific aortic valve disease.
Transcatheter implantation of homologous “off-the-shelf” tissue-engineered heart valves with self-repair capacity: long-term functionality and rapid in vivo remodeling in sheep.
00072-7&id=ga1.jpg){kind=link}
00072-7&id=gr1.jpg){kind=link}
00072-7&id=gr2.jpg){kind=link}
00072-7&id=gr3.jpg){kind=link}