Single Cell Sequencing Reveals Mechanisms of Persistent Truncus Arteriosus Formation after PDGFRα and PDGFRβ Double Knockout in Cardiac Neural Crest Cells

Abstract

Persistent truncus arteriosus (PTA) is an uncommon and complex congenital cardiac malformation accounting for about 1.2% of all congenital heart diseases (CHDs), which is caused by a deficiency in the embryonic heart outflow tract’s (OFT) septation and remodeling. PDGFRα and PDGFRβ double knockout (DKO) in cardiac neural crest cells (CNCCs) has been reported to cause PTA, but the underlying mechanisms remain unclear. Here, we constructed a PTA mouse model with PDGFRα and PDGFRβ double knockout in Pax3⁺ CNCCs and described the condensation failure into OFT septum of CNCC-derived cells due to disturbance of cell polarity in the DKO group. In addition, we further explored the mechanism with single-cell RNA sequencing. We found that two main cell differentiation trajectories into vascular smooth muscle cells (VSMCs) from cardiomyocytes (CMs) and mesenchymal cells (MSs), respectively, were interrupted in the DKO group. The process of CM differentiation into VSMC stagnated in a transitional CM I-like state, which contributed to the failure of OFT remodeling and muscular septum formation. On the other hand, a Penk⁺ transitional MS II cluster closely related to cell condensation into the OFT septum disappeared, which led to the OFT’s septation absence directly. In conclusion, the disturbance of CNCC-derived cells caused by PDGFRα and PDGFRβ knockout can lead to the OFT septation disorder and the occurrence of PTA.

1. Introduction

Congenital heart diseases (CHDs) are the most common congenital disabilities, mainly resulting from the disruption of discrete subsets of cardiac progenitor cells [1]. Persistent truncus arteriosus (PTA) is a rare but complex congenital cardiac malformation accounting for about 1.2% of all CHDs [2]. About 80% or more of children born with PTA die within the first year of life without surgical intervention [2,3,4,5]. PTA is marked by a single great artery arising from the base of the heart, with a ventricular septal defect (VSD), leading to the mixing of systemic, coronary, and pulmonary blood flow. These traits are believed to result from the failure of proper separation of the cardiac outflow tract (OFT).

Neural crest cells (NCCs) are a group of multipotent, migratory cells that delaminate from the neural tube’s dorsal part and contribute to the development of various tissues and organs [6,7,8]. The subpopulation of NCCs contributing to the heart is called cardiac neural crest cells (CNCCs) [9,10]. These cells were firstly reported to play important roles in cardiovascular development, including OFT septation, which ensures the connection between the embryonic heart chambers and vascular network, by Le Lievre et al. [11]. Initially, a solitary tube called the truncus arteriosus was formed, then gradually remodeled into two tubes, giving rise to the aorta and pulmonary arteries [12] (Figure 1A). This remodeling is one of the most sensitive processes in heart morphogenesis. For instance, faulty septation of the OFT accounts for 30% of all CHDs. This prompts us to better understand the cellular and molecular cues that CNCCs participate in the OFT separation during development [13].

CNCCs require multiple environmental signals, including wingless/integrated, bone morphogenetic proteins, and fibroblast growth factors to be specified, migrate, proliferate, differentiate, and survive when they leave the neural tube and migrate into the pharynx and the OFT [14,15,16]. Combined with the above signaling cues, members of the ephrin family of ligands and receptors, semaphorins (including receptors plexin-A2, plexin-D1, and Nrp1 and its ligands Sema6a, Sema6b, and Sema3c) and connexin 43, define specific routes for the migration of CNCCs into the OFT [17,18]. These signal pathways and genes are closely related to the regulation of CNCCs in the normal development of the OFT.

Platelet-derived growth factor receptor (PDGFR) is a surface tyrosine kinase receptor of the PDGF family. It is involved in regulating cell chemotaxis, proliferation, and differentiation. In addition, it plays an essential role in regulating fetal growth and development, wound repair, and tumorigenesis [19]. PDGFR has two subforms: PDGFRα and PDGFRβ [20]. As reported by Richarte et al. [21], embryos with PDGFRα conditional knockout in Wnt1⁺ NCCs developed PTA (penetrance 25%) and VSD (penetrance 75%), while those with PDGFRβ conditional knockout showed no obvious abnormity. Simultaneous knockout of PDGFRα and PDGFRβ in Wnt1⁺ CNCCs could induce PTA with 100% penetrance, suggesting that PDGFRα and PDGFRβ may modulate the involvement of Wnt1⁺ CNCCs in the OFT separation process. However, despite exploring the phenotype and cellular mechanisms of PDGFRα and PDGFRβ double knockout (DKO), the study did not include further mechanistic investigation at the molecular level after PDGFRα and PDGFRβ DKO [21]. This research explored the changes after PDGFRα and PDGFRβ DKO in CNCCs. Given that gene expression might provide information essential to understanding PTA, we first constructed a PTA embryo mouse model with PDGFRα and PDGFRβ simultaneous knockout in Pax3⁺ CNCCs and elucidated the pathological consequences. Importantly, we performed single-cell RNA-sequencing of OFT cells to systematically elucidate the key cell behaviors and molecular mechanisms in this process. Our results provided an OFT cell profile at the septation stage, as well as cytological and molecular changes after PDGFRα and PDGFRβ knockout in CNCCs, highlighting that the disturbance of cell differentiation led to the failure of septal-bridge construction as a possible mechanism of PTA malformation.

2. Materials and Methods

2.1. Animal Models

Embryonic 11.5-, 12.0-, 12.5-, 14.5-, and 17.5-day-old, and neonatal 1-day-old mice of both sexes were used in our study. The C57BL/6J WT mice (No. 213) were obtained from Vital River Laboratory Animal Technology Co. Ltd. (Beijing, China). The Pax3^cre/+ (B6;129-Pax3^tm1(cre)Joe/J, JAX 005549), PDGFRα^fl/fl (B6.Cg-PDGFRa^tm8Sor/EiJ, JAX 006492), PDGFRβ^fl/fl (129S/SvJae-PDGFRb^tm11Sor/J, JAX 010977) and Rosa26^tdTomato (B6;129S6-Gt(ROSA)26Sor^{tm14(CAG-tdTomato)Hze}/J, JAX 007908) mice were acquired from the Jackson Laboratory (Bar Harbor, ME, USA). We first obtained Pax3^cre/+; PDGFRα^fl/+; PDGFRβ^fl/fl and PDGFRα^fl/fl; PDGFRβ^fl/fl as the parental generation through hybridization and genotype identification. Then, we constructed PDGFRα and PDGFRβ-conditional-double-knockout mice from the hybridization of Pax3^cre/+; PDGFRα^fl/+; PDGFRβ^fl/fl and PDGFRα^fl/fl; PDGFRβ^fl/fl parental generation mice. The littermates were used as the control group. Embryos with PDGFRα conditional knockout developed various cardiac malformations, including PTA (penetrance 25%) and VSD (penetrance 75%), while simultaneous DKO of PDGFRα and PDGFRβ (DKO) could induce PTA with 100% penetrance. Our research aims to explore the mechanism of PTA based on stably constructing the mouse model. Thus, transcriptomes of single-cell RNA sequencing were captured from control and DKO group embryos, as described below. Lineage tracing of CNCC-derived cells was performed using Pax3^cre/+; PDGFRα^fl/fl; PDGFRβ^fl/fl; Rosa26^tdTomato genotype mice, obtained from the hybridization of Rosa26^tdTomato; Pax3^cre/+; PDGFRα^fl/+; PDGFRβ^fl/fl and Rosa26^tdTomato; PDGFRα^fl/fl; PDGFRβ^fl/fl mice. All experiments involving animals were conducted in accordance with the Guide for the Use and Care of Laboratory Animals [22]. All animal protocols were approved by the Institutional Animal Care and Use Committee (IACUC), Fuwai Hospital, Chinese Academy of Medical Sciences (FW-2021-0044).

In determining the stage of embryonic development, noon on the day of plug detection after sexually mature male and female (6–8 weeks of age) mice were caged together at a ratio of 3:1 was considered E0.5. Pregnant females were euthanized to harvest embryos at E12.5 for single-cell RNA sequencing and section-based immunofluorescence experiments, at E11.5, 12.0, 12.5, 14.5, and 17.5 for confirmation of cardiac phenotype, and at P0 for statistics of birth phenotypes. Phenotyping criteria at E17.5: (1) PTA + VSD-aorticopulmonary septum is absent and replaced by the truncus arteriosus. All PTA were combined with VSD in our models; (2) DORV-aorticopulmonary septum exists, but both aorta and pulmonary artery originate from the right ventricle, with or without VSD, respectively defined as DORV + VSD and isolated DORV; (3) isolate VSD-ventricular septum defect exists, without any aorta or pulmonary artery defect; (4) Normal-no obvious cardiac structural abnormality exists. The sample size of embryos used for single-cell transcriptome analysis at E12.5 was selected to obtain cell numbers equivalent to the estimated cell numbers in the OFT region.

2.2. Embryo Dissection and Single-Cell Library Generation

The entire OFT region was micro-dissected at E12.5. Embryos were dissected in cold PBS and placed in PBS with 1% FBS (Gibco, 10437010, Life Technologies, Carlsbad, CA, USA) solution on ice until dissociation. Embryo DNA was extracted and used for genotyping to distinguish the control and DKO embryos before further microdissection of OFT regions.

Dissected OFT tissue was dissociated into single-cell suspension with Pierce Primary Cardiomyocyte Isolation Kit (ThermoFisher Scientific, Waltham, MA, USA, 88281). According to protocol, OFT tissue was first rinsed with HBSS (Gibco, 14025076, Life Technologies, Carlsbad, CA, USA) for 5 min. Then each OFT was added to the 1.5 mL centrifuge tube with preset enzyme solution (20 μL recombinant isolation enzyme 1 with papain and 1 μL isolation enzyme 2 with thermolysin) and digested for 30 min in the 37 °C water bath. 150 μL DMEM (Gibco, 10313021, Life Technologies, Carlsbad, CA, USA) containing 10% HFBS (Gibco, 10437085, Life Technologies, Carlsbad, CA, USA) was added to each tube, and the cells were fully dissociated by blowing 25–30 times with sterile straw manually instead of using the gentleMACS Dissociator. The cell suspension was centrifuged at 300× g for 5 min. After discarding the supernatant, cells were resuspended in 100 μL HBSS with 0.04% BSA (Gibco, GP22900, Life Technologies, Carlsbad, CA, USA). Cell viability, aggregation rate, and concentration were measured by Countstar (ABER, Aberystwyth, UK) to ensure sample quality.

Single-cell libraries from this suspension were generated according to the manufacturer’s instructions in the Chromium Single Cell 3′ Reagent Kit (10X Genomics, Pleasanton, CA, USA) v.3 User Guide in the 10X Genomics Chromium controller. The cell capture efficiency of the Chromium controller is ~50%. Thus, we loaded all cells dissected from embryos. Additional components used for library preparation include the Chromium Single Cell 3′ Library and Gel Bead Kit (10X Genomics, Pleasanton, CA, USA) v.3 and the Chromium Single Cell 3′ Chip kit (10X Genomics, Pleasanton, CA, USA) v.3. Our sequencing data have been submitted to Sequence Read Archive (SRA). The bioproject number is PRJNA803335, and the submission number is SUB11005893.

2.3. Single-Cell Transcriptome Library Preparation and Sequencing

Libraries were prepared according to the manufacturer’s instructions using the Chromium Single Cell 3′ Library and Gel Bead Kit (10X Genomics, Pleasanton, CA, USA) v.3 and Chromium i7 Multiplex Kit (10X Genomics, Pleasanton, CA, USA). Final libraries were sequenced on an Illumina NovaSeq 6000 System (Illumina, San Diego, CA, USA). Developmental-timing-matched control and PDGFRα and PDGFRβ-conditional-knockout libraries from each litter were pooled and sequenced in the same lane. Sequencing parameters were selected according to the Chromium Single Cell (10X Genomics, Pleasanton, CA, USA) v.3 specifications. Four libraries were sequenced to a mean read depth of at least 50,000 aligned reads per cell.

2.4. Processing of Raw Sequencing Reads

Raw sequencing reads were processed using the Cell Ranger (10X Genomics, Pleasanton, CA, USA) v.4.4.0 pipeline from 10X Genomics (https://support.10xgenomics.com, accessed on 25 October 2020). In brief, raw base call files generated by Illumina sequencers were demultiplexed into reads in FASTQ format using the “cellranger mkfastq” pipeline. Then, the reads of each library were processed separately using the “cellranger count” pipeline to generate a gene-barcode matrix for each library. During this step, the reads were aligned to the mouse (Mus musculus) reference genome (version: mm10). Cell barcodes and UMIs associated with the aligned reads were subjected to correction and filtering. The resulting gene-cell UMI count matrices for each sample (control and DKO, E12.5) were then concatenated into 1 matrix using the “cellranger aggr” pipeline, which also normalized the libraries to the same sequencing depth.

2.5. Cell Filtering and Cell-Type Clustering Analysis

We sequenced the transcriptomes of 7506 cells captured from control and 8529 cells captured from DKO embryos in total. Further filtering and clustering analyses of these cells were performed with the Seurat v.3.6 R package, as described in the tutorials (http://satijalab.org/seurat/) [23]. For each aggregated dataset (control and DKO), cells were normalized for genes expressed per cell and total expression, then multiplied by a scale factor of 10,000 and log-transformed. Cells that were of low quality or represented doublets were excluded from our analyses—this was achieved by filtering out cells with (1) greater than 25,000 and fewer than 600 genes; (2) greater than 4000 and fewer than 350 UMI counts; (3) greater than 20% mitochondrial gene proportion; and (4) greater than or equal to 40% ribosomal gene proportion using Seurat (Figure S9 and Table S3). We then performed a generalized linear regression on all genes to eliminate technical variability due to the number of genes detected, the number of UMIs, and the cell cycle stage (ScaleData function). Two thousand highly variable genes in the dataset were computed and used as input for principal component analysis. Significant principal components were used for downstream graph-based clustering, and uniform manifold approximation and projection (UMAP) [24] dimensionality reduction was used to project these populations in 2 dimensions. For clustering, the resolution parameter (resolution = 0.5), which indirectly controls the number of clusters, was approximated based on the number of cells according to Seurat guidelines; a vector of resolution parameters was passed to the FindClusters function, and the optimal resolution that established discernible clusters with distinct marker gene expression was selected. To identify marker genes, the clusters were compared pairwise for differential gene expression using the Wilcoxon rank-sum test for single-cell gene expression (FindAllMarkers function, min.pct = 0.25, min.diff.pct = 0.25, return.thresh (adjusted p-value cut-off) = 0.05). To detect differentially expressed genes between DKO and control groups, we conducted Wilcoxon rank-sum test implemented in “FindMarkers” function (min.diff.pct = 0.1, return.thresh (adjusted p-value cut-off) = 0.05). To conduct gene function enrichment analysis, we chose Metascape biological process database as the resource on gene functions. Enrichment analyses were carried out with an adjusted p-value threshold of 0.05.

2.6. RNA Velocity Analysis

We performed RNA velocity analysis using scVelo (version 0.2.2) [25] for inferring cell trajectories. The combined loom file was normalized and log transformed with function scvelo.pp.filter_and_normalize. The first and second order moments for each cell across its nearest neighbors were computed with function scvelo.pp.moments (n_neighbors = 30, n_pcs = 30). Then, the velocities were estimated by running the likelihood-based dynamical model with function scvelo.tl.velocity (mode = “deterministic”), and the velocity graph was constructed with scvelo.tl.velocity_graph. RNA velocities were then recomputed with functions scvelo.pp.neighbor, scvelo.tl.velocity and scvelo.tl.velocity_graph. To visualize the RNA velocities, the UMAP coordinate information and cluster assignments were extracted from the Seurat analysis output. Velocities were projected into 2D on top of the previously computed UMAP coordinates.

2.7. CellChat Analysis

To study the communicating interactions between cells and identify the mechanism of the communicating molecules at a single-cell resolution, R package “CellChat” (version 1.0.0) was applied to cells involved in 6 cell clusters, including MSI, CM I, VSMC I, MS II, CM II and VSMC II [26].

2.8. Immunofluorescence Experiments

Each immunofluorescence experiment was replicated at least twice for identifying spatial expression of genes and 3 times for quantification of in situ signal for differentially expressed genes in the PDGFRα and PDGFRβ-conditional-knockout analysis. For immunofluorescence experiments performed on E12.5 heart sections: Embryos were washed 5 times in PBS after overnight fixation in 4% formaldehyde and stored in 75% ethanol indefinitely until embedding. Embryos were embedded in paraffin processed using standard protocols and embedded for transverse sectioning. Tissue slices were serially sectioned at 5 μm intervals, mounted on slides, and stored at room temperature until the immunofluorescence experiment for paraffin-embedded sections. Sections were imaged with a Zeiss Axio Observer Z1 inverted epifluorescence microscope (Carl Zeiss, Jena, Germany) with Zeiss Axiocam MRm and PCO.edge sCMOS (PCO Imaging) monochrome cameras run by Zeiss Zen imaging software (Carl Zeiss, Jena, Germany). Catalog numbers (Abcam, Cambridge, UK) for Immunofluorescence antibodies used in this study: Anti-PDGFR α: ab234965; Anti-PDGFR beita: ab69506; Anti-α smooth muscle Actin: ab8211; Anti-GM130: ab52649.

2.9. In Situ Hybridization Experiments

Each immunofluorescence experiment was replicated at least twice for identifying spatial expression of genes and 3 times for quantification of in situ signal for differentially expressed genes in the PDGFRα and PDGFRβ-conditional-knockout analysis. For in situ hybridization experiments performed on E12.5 heart sections: hearts were washed 5X in PBS after overnight fixation in 4% formaldehyde and stored in 70% ethanol (VWR, Arlington Heights, IL, USA, 89125–186) indefinitely until embedding. Embryos were embedded in Histogel (Thermo Scientific, Waltham, MA, USA, R904012) and paraffin processed using standard protocols and embedded for transverse sectioning. Tissue slices were serially sectioned at 5 μm intervals, mounted on slides, and stored at room temperature until initiation of the RNA Scope (ACD, Santa Ana, CA, USA) protocol for paraffin-embedded sections (User manual catalog number 322452-USM). Sections were imaged with a Zeiss Axio Observer (Carl Zeiss, Jena, Germany). Z1 inverted epifluorescence microscope (Carl Zeiss, Jena, Germany) with Zeiss Axiocam MRm and PCO edges CMOS (PCO. Imaging, Kelheim, Germany) monochrome cameras run by Zeiss Zen imaging software (Carl Zeiss, Jena, Germany). The catalog numbers for RNA Scope probes used in this study were: Penk, 318761-C2; Sema3c, 441441.

2.10. Statistics and Reproducibility

Standard statistical analyses were performed using GraphPad Prism 8. The number of replicates, statistical test used, and test results were described in the figure legends. For fluorescence quantification, corrected total fluorescence values were log-transformed before t-tests were conducted to satisfy the prerequisite assumptions of normality. For all quantifications, p values were calculated using 2-tailed tests. The mean ± s.e.m. was reported. The level of significance in all graphs is represented as follow: * p < 0.05 and ** p < 0.01. When representative results were presented to indicate expression patterns of genes in wild-type embryos, at least 2 independent embryos were analyzed. Differential proportion analysis (DPA, developed by Farbehi et al.) was used to analyze whether changes in the proportion of populations were greater than expected by chance [27].

3. Results

3.1. Simultaneous Knockout of PDGFRα and PDGFRβ in Pax3⁺ NCCs Induced PTA

To ablate PDGFRα and PDGFRβ in CNCCs, we first constructed a Pax3^cre/+; PDGFRα^fl/fl; PDGFRβ^fl/fl transgenic mouse strain (Figure 1C) [10]. Cell lineage tracing of CNCC-derived cells was achieved using a ubiquitous fluorescent Cre reporter allele, Rosa26^tdTomato, in which Cre-mediated recombination labels cells with a membrane-targeted red fluorescent protein (RFP) [28]. The pattern of cell recombination in the OFT cushions of E12.5 control embryos carrying either Cre driver matched with the pattern of colonizing CNCC described by previous lineage analyses [29]. In situ hybridization on histological sections was used to monitor PDGFRα and PDGFRβ expression and validate their deletion efficiency on CNCCs. As shown in Figure 1D, PDGFRα and PDGFRβ were ubiquitously expressed in all layers of control embryos. In Pax3^cre/+; PDGFRα^fl/fl; PDGFRβ^fl/fl; Rosa26^tdTomato recombined embryos, the CNCC-derived cells displayed a substantial reduction in PDGFRα and PDGFRβ levels, while the surrounding tissues remained PDGFRα and PDGFRβ positive (Figure 1D). PDGFRα and PDGFRβ play an important role in the spatial distribution and function of CNCC-derived cells.

Genotype identification of surviving embryos at E17.5 showed that genes were inherited according to the Mendelian law (Supplementary Figure S1). Embryonic lethality was observed in mice with simultaneous knockout of PDGFRα and PDGFRβ in Pax3⁺ CNCCs at P0, but not in single knockouts (Supplementary Figure S1). Coronal sections of E17.5 embryo mouse hearts were harvested to identify the cause of embryonic lethality. The phenotypic confirmation at E17.5 with multiple methods (stereoscope, microscope, and 3D OFT reconstruction) all showed that loss of PDGFRα and PDGFRβ in Pax3⁺ CNCCs resulted in PTA with 100% penetrance (Figure 1E), consistent with a previous report [21]. We also performed phenotype analysis in mice with single-gene knockout of PDGFRα (Pax3^Cre/, PDGFRα^fl/fl) or PDGFRβ (Pax3^Cre/+; PDGFRβ^fl/fl). PDGFRα knockout mice carried a series of conotruncal malformations, including PTAs (31%), the double outlet of the right ventricle (DORV, 53%), and isolated VSDs 8% (Figure 1F and Supplementary Figure S2). In comparison, most PDGFRβ mice remained normal except for a few isolated VSDs (5%) (Figure 1F and Supplementary Figure S3). The phenotypes of single and double knockout groups indicated that conditional PDGFRα or PDGFRβ DKO in CNCCs could lead to conotruncal malformations and have a functional compensation between PDGFRα and PDGFRβ (Supplementary Table S1). In addition, we observed that the conotruncal phenotypes of Pax3^Cre/+; PDGFRα^fl/+; PDGFRβ^+/+ embryos were normal (Supplementary Figure S4). This indicated that the specific knockdown of PDGFRα in CNCCs did not affect the development of the OFT.

Since the phenotype of PTA is highly correlated with abnormality in OFT separation, we checked OFT separation of embryos at different stages (E11.5, E12.0, E12.5, and E14.5) after simultaneous knockout of PDGFRα and PDGFRβ in CNCCs. To describe the morphological progression of OFT separation, we analyzed 65 hearts by hematoxylin-eosin (HE) staining (E11.5: 16; E12.0: 18; E12.5: 21; and E14.5: 10). Horizontal OFT sections showed that the PDGFRα and PDGFRβ DKO and control group embryos were normal, and no gross morphological defects were detected at E11.5 and E12.0. At E12.5, DKO embryos showed penetrant morphological OFT defects, and the cushion was poorly matched and failed to form the septal bridge structure. At E14.5, OFT separation between the aorta and pulmonary artery was completely absent in DKO mice (Figure 1G). Phenotypic examination of transverse sections at three distinct distal-proximal levels of OFT (distal, medial, and proximal) also showed that the lack of septal bridge structure in the medial level led to the separation failure in the distal level directly (Figure 1H). This indicated that PDGFRα and PDGFRβ combined knockout in CNCCs could lead to disorder of septal bridge formation and absence of OFT separation, which may trigger the induction of the PTA phenotype. In addition, time course phenotypic changes of OFT were observed for the first time from E12.5 and gradually developed into PTA from E12.5 (Figure 1G). Thus, it is a crucial time point for studying this process’s cellular and molecular mechanisms, which are suggested by the results of early phenotypic confirmation.

3.2. Condensation Failure of CNCC-Derived Cells Due to Disturbance of Cell Polarity after Pdgfrα and Pdgfrβ Knockout

Next, we next investigated the cellular mechanisms, including cell migration, proliferation, apoptosis, and polarity in OFT separation after PDGFRα and PDGFRβ DKO. First, whole-mount immunostaining in CNCC lineage-tracing mice revealed that RFP⁺ CNCCs reached similar OFT levels in E12.5, indicating that cell migration was not affected by PDGFRα and PDGFRβ DKO (Supplementary Figure S5). Next, we analyzed the proliferation of cells within the conotruncal region of embryos at E12.5 to assess the proliferation rate of CNCC-derived cells in PDGFRα and PDGFRβ DKO and control embryos. Quantifying cell proliferation in tissue sections, using antibodies against cell proliferation-related markers Ki67 and phospho-histone H3 (pH3) demonstrated no significant change in the number of RFP⁺ CNCC-derived proliferating cells compared to littermate controls (Supplementary Figures S6 and S7). Terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labeling (TUNEL) assay revealed similar apoptosis rates in both groups at E12.5 (Supplementary Figure S8). These results indicated that PDGFRα and PDGFRβ do not control the proliferation or survival of CNCCs. Finally, we wondered whether OFT morphogenetic defects in the DKO group resulted from differences in cell polarity because cell polarity may reflect the cell condensation in OFT cushions, an important segment in septum formation [13]. We performed Golgi marker GM130 and nuclear DAPI immunostaining, determined the polarity direction of OFT cells through the connecting line between Golgi and nuclear center (taking the craniocaudal axis of mice as 0°), and statistically analyzed the position and orientation of all cells at the same level (Figure 2A). In the control group, the polarity of OFT cushion cells on the left side was more concentrated between 120° to 150°, which was consistent with the fusion angle of OFT cushions. However, the cellular polarity was not apparent in the DKO group, suggesting that OFT cell polarity was potentially disrupted in DKO embryos (Figure 2B and Supplementary Table S2).

In summary, the above results showed that PDGFRα and PDGFRβ are important modulators of CNCC behavior in the heart and the OFT separation process. After PDGFRα and PDGFRβ DKO, the polarity of OFT cells is affected. Without correct cell polarity concentration in the fusion angle of OFT cushions, loosely arranged cells in the middle of OFT may fail to condense into a compact septal-bridge structure, leading to the septum formation disorder and the emergence of the PTA phenotype.

3.3. Single-Cell Transcriptomic Sequencing and Overview of Developing OFT Cells

Although traditional gene knockout studies have revealed that PDGFRα and PDGFRβ can modulate the involvement of Pax3⁺ CNCCs in the OFT separation process, but the specific mechanisms of gene regulation are still elusive. Given the complexity of the OFT separation process, we used single-cell transcriptomic sequencing to further its exploration.

We isolated OFT cells from embryonic mouse hearts at the OFT separation stage (E12.5). Then, we conduct single-cell transcriptomic sequencing. After stringent quality filtering (details see Methods), we obtained a high-quality transcriptomic dataset of 6337 cells in the DKO group and 5661 cells in the control group (Figure 3A, Supplementary Figures S9 and S10, Tables S3 and S4). Unsupervised clustering identified 10 cell clusters representing distinct cell types or subpopulations (Figure 3B and Supplementary Figure S11).

We performed differential expression analysis between each cluster and all others to identify each cell cluster and assigned a specific cell type to each cluster based on established lineage-specific marker genes (Figure 3C and Supplementary Table S5). Epicardial cells (EP) specifically expressed Upk3b and Upk1b19 [30]. Endocardial cells (EC) I and II clusters highly expressed endocardial markers Ramp2 and Fabp5 [31]. The cardiomyocyte (CM) lineage constituted CM cluster I, II, and III, as they highly expressed CM marker genes, such as Myh7 and Myl4 [32,33]. The vascular smooth muscle cell (VSMC) lineage comprised two closely aligned clusters, namely, VSMC I and II, which highly expressed Cxcl12, a chemokine encoding gene that is highly expressed in the walls of the aorta and pulmonary trunk of the embryonic heart [34] and Rgs5, a gene that is abundantly expressed in arterial smooth muscle cells [35,36]. The mesenchymal lineage included mesenchymal cells (MS) I and II clusters, which specifically expressed mesenchymal marker genes such as Postn and Cthrc1 (Figure 3D) [32].

Significantly, we found significant changes in the cell type proportions following DKO, mainly in CMs and MSs (p < 0.05). In CMs, the proportion of CM I cluster increased, and CM II decreased. Meanwhile, the MS II cluster was absent in the DKO group (Figure 3E).

3.4. The Disturbance of CMs Trans-Differentiation after Pdgfrα and Pdgfrβ Knockout

We bioinformatically isolated three CM clusters from the rest of the cells in the OFT for further analysis (Figure 4A). After PDGFRα and PDGFRβ knockout in CNCCs, the proportion of CM I cluster increased, and CM II decreased significantly (p < 0.05), while CM III remained similar (Figure 3E and Figure 4B). Differentially expressed genes (DEGs) analysis showed that the top 5 marker genes in CM I were Malat1, Itm2a, Actb, Bmp4, and Clu; Mest, Gja1, mt-Atp6, Atp2a2, and Nppb in CM II; and Cited1, Myl2, Smpx, Hopx, and Gyg1 in CM III (Supplementary Table S6). Pathway enrichment analysis pathway for each CM cluster showed that the genes up-regulated in CM II and III were enriched for oxidative phosphorylation and CM contraction, with high expression of Smpx, Nppb, and Myl2 (Figure 4C). By contrast, in CM I, terms was mainly enriched in vascular development, ECM organization, and cell migration, with high expression of Cxcl12 (Figure 4C).

The highly expressed genes in CM I included Cox4i2, Col1a2, Cxcl12, and Eln, which were specifically expressed in VSMCs [37]. In addition, previous studies have found that Sema3c was expressed in CMs of the OFT wall and played a vital role in attracting CNCCs into OFT [9,17,18,38] and promoting the aggregation of CNCCs in primary cultures [17,18,39,40]. It was also reported that Sema3c⁺ CMs were mainly distributed on the pulmonary artery (Pa) side and represented future subpulmonary CMs derived from the second heart field [17,38]. In our data, Sema3c was highly expressed in the CM I cluster, consistent with previous findings (detail see Result 5, Figure 5G). In addition, the Sema3c-Nrp1 signaling has been proposed as a triggering signal for the epithelial-mesenchymal transition (EMT) of the endocardium [39]. Consistent with this, highly expressed marker genes, including Malat1, Bmp4, Clu, Cxcl12, Col1a2, and Mt1, were closely related to the EMT procedure [40,41,42,43,44,45]. These results indicated that the CM I cluster might represent future subpulmonary CMs and had VSMC-like characteristics (Cxcl12⁺), which also played an important role in EMT regulation. CM II cluster highly expressed mature CM markers like Nppb and electrophysiology genes like Atp2a2 represented a relatively mature working CM cluster [46]. While CM III highly expressed Hopx, which contributes to stem cell quiescence, might represent multipotent cardiomyoblasts [47] (Figure 4C,D).

Compared with the control group, up-regulated genes in the DKO group included Hspa1a, Hspa1b, and Malat1 (Figure 4E and Supplementary Table S7). Pathway enrichment analysis revealed that up-regulated genes participated in L13a-mediated translational silencing of ceruloplasmin expression and regulation of cellular protein catabolic process (Figure 4F). In CM II, down-regulated genes included Cited1, Fos, Tnni1, Smpx, and Egr1, whereas up-regulated genes included Mt1, Malat1, Hk2, Bnip3, and Ttn. In CM III, down-regulated genes included Myl1, Cited1, Atox1, Mgst3, and Tnni1, while up-regulated genes included Bsg, Nme2, Hk2, Rpl13a, and Ankrd1 (Figure 4E). Enrichment pathways in CM II included ATP metabolic process and cardiac muscle tissue development, while CM III included ATP metabolic process and GTP hydrolysis (Figure 4F). RNA velocity analysis is a new analytical method that can infer the cell differentiation trend based on the relative abundance of spliced and unspliced transcripts [48]. Our RNA velocity analysis result of CMs (Figure 4G) and all OFT cells (Supplementary Figure S12) revealed that CM II showed progression toward VSMC via CM I. In addition, genes related to vascular smooth muscle development, including Csrp2 and Acta2, were down-regulated in CM I in the DKO group (Figure 4E) and mainly participated in vascular development, energy metabolism, and smooth muscle contraction pathways, which were similar to CM I marker genes (overlapped gene including Csrp2, Cst3, Sfrp1, Itm2a, Sparc, and Ctgf) and enrichment pathways (Figure 4C,F). This may suggest that vascular smooth muscle development of embryonic OFT was impaired. The above findings suggested that CM I was a group of transitional cells along the trajectory for CM-to-VSMC trans-differentiation. After PDGFRα and PDGFRβ knockout, CM trans-differentiation into VSMCs may be disturbed and stagnated in a transitional cell state (CM I), which may lead to the failure of OFT remodeling and muscular septum formation. We conducted an in vivo immunofluorescence verification experiment using CM marker cTnT, VSMC marker α-SMA, and CM I cluster-specific marker Sema3c. The result indicated the existence of a CM I cluster, which co-expressed cTnT, α-SMA, and Sema3c and was mainly distributed on the pulmonary artery (Pa) side. In addition, compared with the control group, the distribution area of CM I in the DKO group seemed to expand larger, which was consistent with our single-cell analysis result that CM-to-VSMC differentiation stagnated in the transitional CM I-like state (Figure 4H). However, the CM-to-VSMC differentiation process during OFT separation still remains controversial and only supported by a few recent studies [37]. More research evidence is needed to confirm the existence of this transition process.

Although our current evidence is not enough to clearly explain the potential regulation mechanism, we still make reasonable speculation about the possible effect of PDGF knockout on CMs. Through CellChat analysis [26], we noticed that CM I cluster played the main role of the sender in the PDGF signaling pathway network, and the target receiver was MS I. The loss of PDGF signaling in MS I due to the crest deletion of PDGFR might regulate the transformation of more CMs into CM I cells through intercellular feedback communication (Supplementary Figure S13).

3.5. Absence of a Key MS Cluster Associated with Septal Bridge Formation after Pdgfrα and Pdgfrβ Knockout

Next, we concentrated on the two mesenchymal clusters (Figure 5A). Compared with the control group, there was a complete lack of MS II in the DKO group (Figure 5A,B). Analysis of differentially expressed genes between two MS subclusters showed that MS-related genes Hapln1 and Twist1 were highly expressed in MS I, with enriched pathways including regulation of epithelial cell proliferation and translation. Twist1 was previously reported to function in post-migratory CNCC-derived MSs to repress pro-neural factors, thus regulating cell fate determination between ectodermal and mesodermal lineages [49]. The absence of Twist1 resulted in aberrant cardiac neural crest morphogenesis and OFT septation [50]. By contrast, VSMC marker genes Cxcl12 and Acta2 were higher in MS II, and enriched pathways included tissue morphogenesis, angiogenesis, and muscle tissue development (Figure 5C,D and Supplementary Table S8). Therefore, we postulated that MS II was much closer to the VSMC state than MS I (Figure 5C,D). MS II may be a group of transitional cells along the trajectory of MS-to-VSMC differentiation [37].

Due to the absence of the MS II cluster in the DKO group, only gene expression and enriched pathways of MS I could be compared. Compared with the control group, down-regulated genes of the MS I cluster included Ptn, Tsix, and Sfrp2, and up-regulated genes included Dbi, Ramp2, and Nme2 (Figure 5E and Supplementary Table S9). Pathway enrichment analysis revealed that down-regulated genes of MS I primarily participated in muscle development, respiratory chain, and oxidative phosphorylation, and up-regulated genes participated in translation and regulation of cellular amide metabolic process (Figure 5F). This indicated that vascular smooth muscle development was impaired in the DKO group.

Previous studies revealed that Sema3c was highly expressed in the CM wall and septum of OFT [9,17,51], and Penk was specifically expressed in the septum [37]. They were both highly expressed in septal-bridge cells, a group of cells located in the center of the OFT septum and involved in septum formation, according to our data. Of note, our analysis revealed that the expression level of Sema3c and Penk was higher in MS II (Figure 5G), indicating that MS II represented septal-bridge cells. The MSs in OFT have two main origins: EndoMT and CNCC derivation. We hypothesized that MS II derived from CNCCs and conducted an in vivo immunofluorescence verification experiment. The result showed that all the Penk⁺ MS II cells had red fluorescence used for CNCC lineage-tracing, which verified our hypothesis (Supplementary Figure S14). Consistent with our findings, a previous study also reported that the Penk⁺ cells were mainly localized in the center of OFT, where the aortopulmonary septum formed, and all the Penk⁺ cells in the heart were derived from CNCCs [29]. When we projected our single-cell data onto previous single-cell data sets of CNCCs [29], we found that Penk⁺ septal bridge cells from our OFT cell and previous CNCC datasets could mix well into a group and disappeared in DKO mice (Supplementary Figure S15). Consequently, RNA velocity analysis of all OFT cells (Supplementary Figure S12) showed that there were two main differentiation trends into VSMCs from CMs and MSs, respectively. Details in Figure 5H revealed that the MS II cluster was a transitional cell state in MS-to-VMSC differentiation. This may indicate that the deletion of MS II in the DKO group is one of the key reasons leading to OFT separation disorder and PTA phenotype. The deletion of PDGFRα and PDGFRβ caused CNCCs to fail to differentiate into septal-bridge cells.

We performed an RNA Scope experiment in situ hybridization to verify the findings suggested by single-cell transcriptomic sequencing. As shown in Figure 5G, Sema3c and Penk were co-stained in the center of the OFT septum, the location of septal-bridge cells in the control group. However, there were few signals of Sema3c and Penk in the DKO group (Figure 5I). These results are consistent with our findings that the MS II cluster with high expression of Sema3c and Penk was absent after PDGFRα and PDGFRβ knockout. Altogether, our data found intermediate cell subpopulations involved in either MS-to-VSMC transition or CM-to-VSMC trans-differentiation, which were affected or absent after PDGFRα and PDGFRβ conditional knockout in CNCCs. This may be the critical mechanism of OFT separation disorder and PTA phenotype.

3.6. More VSMC in an Immature State after PDGFRα and PDGFRβ Knockout

As explained above, the cluster proportion and gene expression of CMs and MSs exhibited noticeable changes after PDGFRα and PDGFRβ knockout. We wondered whether VSMCs demonstrated similar changes in the DKO group (Figure 6A–C). DEGs analysis between two VSMC subclusters showed that the top5 marker genes in VSMC I were Col1a2, Fbn1, Eln, Rgs5, and Myh10, highly correlated with mature VSMC markers; while genes abundantly expressed in VSMC II were Tmsb10, Tmsb4x, Actb, H2afz, and Mif, ones with relatively higher expression in fetal MSs (Figure 6D,E and Supplementary Figure S16 and Table S10). Pathway enrichment analysis showed that the highly expressed genes of VSMC I were closely related to molecules associated with elastic fibers, and those of VSMC II were related to actin cytoskeleton organization (Figure 6F). Combing DEG and pathway enrichment analysis results, we inferred that VSMC I was a relatively mature VSMC cluster. Significantly, the proportion of VSMC II increased in the DKO group, indicating that more VSMCs were relatively immature (Figure 6C).

Compared with the control group, pathway enrichment analysis revealed that down-regulated genes of VSMC I in the DKO group (Pclaf, Tsix, Acta2, Rrm2, and Ranbp1) primarily participated in the deoxyribonucleotide biosynthetic process and regulation of mRNA splicing via spliceosome. Meanwhile, down-regulated genes of VSMC II in the DKO group primarily participated in the metabolism of RNA, energy metabolism, and mitochondrial biogenesis (Figure 6G,H and Supplementary Table S11). In combination with cell proportion, gene expression, and enrichment pathway characteristics, RNA synthesis and splicing were affected after PDGFRα and PDGFRβ knockout, leading to a higher proportion of VSMCs in a relatively immature manner VSMC II-like state and OFT septum remodeling disorder.

3.7. Ligand-Receptor Interaction Changes among MS, CM and VSMC Clusters after PDGFRα and PDGFRβ Knockout

In order to further reveal the changes in the molecular regulation mechanisms related to cell differentiation and function in OFT after PDGFR conditional knockout. We identified a total of 77 ligand-receptor (L-R) interactions through CellChat [26] among MS, CM, and VSMC clusters, the involved signaling pathways including integrin, non-canonical Wnt, Tgfβ, Fgf, Pdgf, Notch, Bmp, Ras/MAPK, PI3K/Akt pathway. The control group had 15 specific L-R interactions, and the DKO group had 14 specific L-R interactions (Supplementary Figure S17).

Of the 77 L-R interactions identified, 20 involved integrins. Integrins can provide crucial adhesive and signaling functions through interactions with the extracellular matrix (ECM) like laminin, fibronectin, and collagen, which are important for paracrine signaling, ECM homeostasis, and the intercellular interactions that impact cardiac cell biology and pathophysiological adaptation in disease [52].

Of non-integrins, we noticed that some L-R interaction changes were mainly related to CMs. Firstly, Bmp signaling L-R interactions [Bmp7-(Bmpr1a + Bmpr2), Bmp7-(Acvr1 + Bmpr2), Bmp4-(Bmpr1a + Bmpr2) and Bmp2-(Bmpr1a + Bmpr2)] to CM I were absent in the DKO group. Bmp is a member of the TGF-β superfamily and plays an important role in embryonic heart development. A previous study reported that Bmp signaling could induce SHF-derived CM differentiation and beating in embryonic mice [53]. It was also reported that Bmp regulated CM proliferation along with Notch signaling [54,55]. Consistent with this, the Notch-relevant L-R interactions (Jag1-Notch2 and Dlk1-Notch2) decreased significantly in the DKO group. Notch signaling was also reported to regulate the differentiation from CM to VSMC [37]. In addition, we observed that Agrn-Dag1 interaction from MS I and VSMC I to CM I and II only existed in the DKO group. Agrin (Agrn coding protein) is a component of neonatal ECM, which can attenuate CM maturation and promote CM proliferation [56]. The cooperation effect of Bmp, Notch, and Agrn signaling to CM might lead to the increased proportion of relatively immature CM I-like cells in OFT. Finally, Hspg2-Dag1 L-R interaction was absent in the DKO group. Perlecan (Hspg2 coding protein) modulates CM activity and is essential for normal tissue development and function [57]. Its loss can lead to a high frequency of OFT malformations, such as the complete transposition of great arteries (TGA) [58]. We noticed that F11r-F11r was only relevant to CM I and II and absent in the DKO group, which might represent self-regulation in CMs.

There are also several L-R interaction changes mainly related to MSs and VSMCs. Ackr3 (originally named RDC1 and CXCR7) related L-R interactions (Mif-Ackr3 and Cxcl12-Ackr3) to MSs were absent in the DKO group, indicating that Ackr3 was mainly expressed in MSs and lost in the DKO group. Cxcl12-Ackr3 heteromeric complexes had been demonstrated to be critical to septum and valve morphogenesis during heart development [59,60]. In contrast, Sdc2-related L-R interactions (Ptn-Sdc2 and Mdk-Sdc2) to MSs were increased in the DKO group. Sdc2 can physically interact with members of the Fgf, Vegf, and Tgfβ superfamilies [61,62]. In cell culture, Sdc2 has been implicated in microvascular angiogenesis through the regulation of cell adhesion and cell migration [63]. In addition, the Mdk-Lrp1 expression pattern showed that this interaction only acted on VSMC I and was absent in the DKO group. Lrp1 function in CNCCs is required for normal OFT development, with other cell lineages along the CNC migratory path playing a supporting role [64].

In summary, the identified L-R interaction changes between the control and DKO group revealed fundamental aspects of molecular regulation mechanism in normal and after PDGFR conditional knockout. These findings also provided clues for further research on the regulatory signal network related to OFT separation.

4. Discussion

It is well established that CHDs are a kind of congenital disease with complex etiologies related to genetic and environmental factors. Among them, PTA is a rare but serious developmental conotruncal malformation. This defect is characterized by a single vessel and common valve that originates from the base of the heart supplying the systemic, coronary, and pulmonary circulation. Previous studies have reported that CNCCs in the caudal pharyngeal arches will migrate into the cardiac OFT to form condensed mesenchyme at the junction of the bilateral OFT cushion. This condensed mesenchyme, termed the aorticopulmonary septation complex, divides the OFT into the aorta and pulmonary trunk [65,66]. Disturbance of CNCCs in heart development due to genetic and environmental factors may lead to PTA, but the specific mechanism remains unclear.

Our analysis of the cellular and molecular mechanism of the OFT separation disorder after PDGFRα and PDGFRβ DKO in CNCCs enabled us to propose a model whereby PDGFR-dependent cell trans-differentiations of CNCC-derived cells would set the occurrence of septal bridge formation and OFT septation. First, we constructed a PTA mouse model with a simultaneous knockout of PDGFRα and PDGFRβ in Pax3+ NCCs. The analogy between the phenotype of PDGFRα and PDGFRβ DKO mice and patients suffering from PTA condition enabled us to further explore the cellular and molecular etiology of this complex CHD. Next, we described the changes in cell proliferation, apoptosis, migration, and polarity in the DKO group at the cellular level, verifying and supplementing the findings reported in previous studies [21,36]. Finally, there is still a lack of molecular-level research on malformation caused by PDGFRα and PDGFRβ knockout. Our single-cell sequencing study provided primary insights into cell subsets and regulatory molecules in OFT septation.

It has been described that disruption of PDGFRα signaling can cause defects in CNCC populations, and PDGFRα null, and Patch (Ph) mutant allele mice (a naturally occurring PDGFRα deletion) indicated that cell survival or matrix deposition, but not apoptosis and proliferation, was the primary function of PDGFRα signaling [67,68]. Another study reported that, whereas PDGFRβ is essential in certain cell types during embryonic development, its broader role may be masked because of compensation by PDGFRα [69]. When combining PDGFRα and PDGFRβ specifical knockout in Wnt1⁺ CNCCs, Richarte et al. found a PTA phenotype with 100% penetrance and cellular changes after knockout [21]. In our research, we found no significant difference between DKO and normal embryos concerning CNCC migration, cell proliferation, and apoptosis. Inspired by the study of Darrigrand et al. that used cell polarity to reveal the cytological mechanisms in OFT septation [13], polarity detection of OFT cells taking advantage of GM130 showed that the concentration trend in cushion merge direction was disturbed after PDGFRα and PDGFRβ DKO.

In recent years, multiple single-cell RNA-sequencing studies have focused on the field of heart development and CHDs [13,29,37,70]. In OFT development, changes in cell subpopulation and lineage differentiation are key to uncovering the mechanisms underlying normal development and abnormalities [29,37,70]. As a newly-developing technique, single-cell transcriptomics was worthwhile revisiting the model so that it is possible to separate the cell-autonomous from non-cell-autonomous mechanisms. Two single-cell studies recently published by the Zhou group provided profiles of embryonic cardiac OFT development and involvement of NCCs in heart development, respectively [29,37]. Our research further explored the abnormalities of NCCs and aorticopulmonary septum formation during the process of PTA.

We performed single-cell transcriptomic sequencing of 11,998 mouse OFT cells from developmental stages corresponding to OFT septation (E12.5). The large-scale single-cell data empowered us to unbiasedly and systematically dissect the cellular diversity and heterogeneity during OFT development. We identified 10 cell clusters that could be assigned to 6 cell lineages. We provided molecular signatures for the cell lineages and clusters and identified intermediate cell subpopulations (CM I and MS II), which were likely to be involved in either CM to VSMC trans-differentiation or MS to VSMC differentiation after PDGFRα and PDGFRβ DKO. Particularly, the Penk⁺ cluster MS II, a relatively small MS subpopulation undergoing MS to VSMC transition, was explicitly associated with the fusion of the OFT cushions and was absent in DKO OFT tissues.

The CM to arterial phenotypic change in OFT has been described in a recent single-cell study on OFT development [37,71,72], demonstrating that trans-differentiation of CMs to arterial components might occur during OFT development in embryonic hearts, yet this view still remains controversial. Consistent with this, our single-cell dataset supports this view of CM to VSMC trans-differentiation by identifying cell clusters representing a continuum of cell state transitions. The expression profile comparison analysis demonstrated that CM cluster I could express CM and VSMC markers simultaneously and might represent an intermediate cell state along the trajectory of CM-to-VSMC trans-differentiation. In vivo immunofluorescence verification experiment also proved that there was a group of cells distributed on the pulmonary artery (Pa) side co-expressing CM and VSMC markers in the early development stage. The expression of SMA-mCherry in the embryo and adult mouse organs reported by Armstrong et al. indicated that α-SMA, which is used as a marker of VSMCs, could be expressed in the embryonic heart and lost in adults [73]. This appeared to provide an explanation for the expression of some smooth muscle markers in the embryonic myocardium. Although our study results might suggest the existence of a group of cells in the intermediate state between CMs and VSMCs, further research is needed to confirm this finding. Meanwhile, we observed that cell trans-differentiation seemed to have stagnated in an intermediate state of CM I in the DKO group. Combined with pathway enrichment analysis results, this may indicate that vascular smooth muscle development of embryonic OFT was impaired after PDGFRα and PDGFRβ DKO.

In a previous study, a small MS subpopulation involved in the MS differentiation into VSMC was reported by Liu et al. [37]. In our study, we provided further evidence for the existence of the MS-to-VSMC transition and an intermediate MS subpopulation (MS II). Corresponding to previous studies, MS II highly expressed intermediate MS subpopulation marker Penk and Sema3c. Furthermore, this cluster was mainly condensed into a compact septal-bridge structure, as shown in Figure 5I. All these observations demonstrated the role of cluster MS II in OFT separation formation. Importantly, we found that cluster MS II was absent after PDGFRα and PDGFRβ DKO, which indicated the disturbance of MS differentiation into VSMC and failure of cell condensation into a septal bridge, providing mechanistic insights into OFT septation disorder.

Our single-cell sequencing data analysis of CMs, MSs, and VSMCs revealed that VSMCs could be downstream of the differentiation of both CMs and MSs. First, as two crucial differentiation sources of VSMCs, CMs and MSs, especially the two transitional cell subpopulations CM I and MS II identified in this study, had undergone important changes after PDGFRα and PDGFRβ knockout. Among them, CM differentiation into VSMCs stagnated in a Sema3c⁺ CM I state, while the MS II cluster expressing Penk and Sema3c disappeared in the DKO group. Besides, VSMCs also changed gene expression with a more immature transcriptional state.

Cell polarity may reflect the condensation of OFT cushion cells, which is an important segment in aortopulmonary septum formation. Our analysis results indicated that the cell polarity disorder in the DKO group might be caused by the down-regulation of Sema3c (a key molecule regulating the directional migration of cushion cells) expression in specific cell subpopulations. In addition, differentiation disorders of MSs into VSMCs may also make cushion cells acquire a less polarized morphology. These changes together caused the abnormal condensation of OFT cushion cells, which led to the occurrence of PTA.

There are still some limitations to this research. Using single-cell RNA-sequencing, our study provided new insights into the OFT septation disorder in PTA formation. However, the molecular mechanisms and downstream regulatory networks still need to be further explored and elucidated. In addition, more embryo samples at different developmental stages are needed to reconstruct this developmental process more comprehensively.