Osr1 orchestrates posterior second heart field cell migration for outflow tract formation

Abstract

Congenital heart defects, particularly those affecting the outflow tract (OFT), represent a significant cause of neonatal morbidity and mortality. The genetic and cellular mechanisms underlying OFT abnormalities, especially the migration of second heart field (SHF) cells, remain poorly understood. In this study, we identify two distinct migratory pathways for Osr1-expressing posterior SHF cells directed toward the OFT and inflow tract (IFT). We show that Osr1 regulates two key ligand-receptor signaling axes, Hedgehog–Ptch1 and Cxcl12–Cxcr7, both of which are essential for the directed migration of Osr1⁺ cells toward the OFT. We establish that Smo and Ackr3 (encoding CXCR7) are direct transcriptional targets of Osr1. Functional studies demonstrate that Osr1-Hh signaling governs SHF cell migration and OFT development, while Osr1-Cxcl12 signaling likely acts in synergy to support this process. These findings underscore the pivotal role of Osr1+ pSHF cells in directing heart pole development and reveal crucial ligand-receptor interactions involved in cell migration, potentially guiding future therapeutic targets for congenital heart defects.

Study identifies two migratory paths for Osr1+posterior SHF cells to OFT and IFT in heart development. Osr1 regulates Hh-Ptch1 and Cxcl12-Cxcr7 signaling, with Smo and Ackr3 as direct targets, essential for outflow tract development and congenital heart defects mechanisms.

Introduction

Congenital heart defects (CHDs) are common birth anomalies^1,2, with outflow tract (OFT) defects comprising one-third of CHDs and leading to significant morbidity and mortality across all ages³. OFT defects, such as double outlet right ventricle (DORV) and overriding aorta (AO), are characterized by the mispositioning of the aorta over the right ventricle, accompanied by a ventricular septal defect. However, the details of OFT septation and alignment remain unclear due to the complexity of OFT development at the cellular, genetic, and molecular levels.

The formation of the cardiac OFT is a complex process involving interactions among various cell types within the heart and extracardiac cells. At embryonic day 8 (E8), cells originating from the second heart field (SHF) pharyngeal mesoderm migrate into the anterior part of the heart^4–8. These SHF cells contribute to the conal portion of the OFT endocardial cushion adjacent to the right ventricle^9–13. Abnormal OFT formation can result from disrupted genes responsible for SHF morphogenesis^5,14–22. Failure to incorporate SHF-derived myocardial and endocardial cells into the arterial pole of the heart leads to a shortened OFT and arterial pole misalignment^20,23,24. Recent studies have highlighted several signaling pathways, including Wnt/PCP^25,26 and Hedgehog (Hh) signaling^25,26, in SHF cell deployment. However, additional information is necessary to comprehend how cells are prompted to migrate.

Odd-skipped-related 1 (Osr1) is a vertebrate zinc finger transcription factor homologous to Drosophila’s Odd-skipped (Odd)²⁷. Despite its evolutionary conservation, OSR1’s role in human heart development remains unclear. In humans, OSR1 is expressed in mesenchymal stem cells and linked to congenital heart defects (CHDs)^28,29. In mice, Osr1 marks the posterior second heart field (pSHF)^30,31, with knockout mice developing atrioventricular septal defects (AVSDs), with dilated atria and hypoplastic venous valves^30,32. At the venous pole, Osr1 functions as a downstream target of Tbx5 and works with Tbx5-Hh signaling to regulate SHF cell proliferation^30–32. While Osr1’s role in the inflow tract (IFT) is well-established, its involvement in OFT development is not yet explored.

In this study, we explore the role of Osr1 in OFT heart development, particularly its regulation of SHF cell migration toward the IFT or OFT. We found that a subset of Osr1+ cardiac precursors in the anterior second heart field (aSHF) and the OFT originates from the pSHF. Additionally, we identified two key transcriptional targets of Osr1 and highlighted the importance of the Osr1-Ackr3 and Osr1-Hh signaling pathways in the development of OFT. These insights enhance our understanding of the molecular and cellular mechanisms in OFT formation and open avenues for potential therapeutic interventions.

Results

Single-cell spatiotemporal mapping identifies two distinct migratory paths and multiple fates in Osr1-expressing SHF cells

Our prior studies indicated that Osr1 is crucial for SHF heart development. To further explore SHF Osr1+ cells, we dissected E9.5 and E10.5 SHF into anterior (A), mid (M), and posterior (P) segments, following a previously established method³³, and performed single-cell RNA-sequencing (scRNA-seq). The precise microdissection of the A, M, and P regions was validated through the region-specific expression of marker genes, such as Tbx5, Tbx1, and Hand2, as visualized using violin plots (Suppl. Fig. 1A, B). We also collected cells from the OFT region (O) of E10.5 embryos. After filtering out low-quality cells, we analyzed 2120 cells at E9.5 and 8000 cells at E10.5. t-SNE clustering at E9.5 revealed distinct transcriptomic patterns among the anterior (A), middle (M), and posterior (P) SHF segments, confirming pre-patterning along the anterior-posterior (A–P) axis (Fig. 1A-b). At E10.5, cells from the O segment also formed a distinct cluster, with partial overlapping with the anterior segment, suggesting that the O segment represents cell similarity of the O and A regions (Fig. 1B-b). We then performed cell-cell similarity analysis to match cells from two embryonic days. The similarity analysis revealed an increasing correlation between E9.5 and E10.5 regions along the progression from P to O (P → M → A → O). When examining the posterior cells from E9.5 (Fig. 1C, yellow bar), we observed that 40% of the cells in E10.5 regions O and A show high correlation with E9.5 region P cells (Fig. 1C). This indicates substantial cell migration from region P to region A between E9.5 and E10.5. Osr1+ cells exhibited greater migratory potential than the average pSHF cells. Similarity analysis showed that over 70% of Osr1+ cells in E10.5 region O and more than 85% in region A were most closely correlated with E9.5 region P cells (Fig. 1D), underscoring Osr1 as a key driver of pSHF cell migration. While the dynamic movement of cells during cardiac morphogenesis is well-established³⁴, our findings reveal a novel subpopulation of pSHF cells that contributes to the OFT by migrating through the aSHF.

Fig. 1. Clustering analysis revealed SHF progression at single-cell spatiotemporal resolution.

A Multi-characteristic clustering analysis of SHF cells from E9.5. a Schematic of the microdissection of the embryo SHF from E9.5 (three regions, A, M, P). b All SHF cells are embedded on the t-SNE dimension reduction map to show distributions in E9.5 A, M, and P SHF regions. c Spectral t-SNE plots of all single-cell transcriptomes from E9.5 embryo SHF cells. d Osr1+ cells embedded on t-SNE dimension reduction map. B Multi-characteristic clustering analysis of SHF cells from E10.5. a Schematic of the microdissection of the embryo SHF from E10.5 (four regions, O, A, M, P). b All SHF cells are embedded on the t-SNE dimension reduction map to show distributions in E10.5 O, A, M, and P SHF regions. c Spectral t-SNE plots of all single-cell transcriptomes from E10.5 embryo SHF cells. d Osr1+ cells embedded on t-SNE dimension reduction map. C Single-cell-based correlation between all SHF cells from E9.5 and E10.5. D Single-cell-based correlation between Osr1+ SHF cells from E9.5 and E10.5. E Cluster identities, cell type annotation, and matching between SHF cells from both embryonic days.

Unsupervised clustering of gene expression profiles identified ten clusters at E9.5 (Fig. 1A-c and Suppl. Fig. 1C) and 12 at E10.5 (Fig. 1B-c and Suppl. Fig. 1D), with some clusters showing high similarity of global gene expression between the two stages (Fig. 1E). Excluding cluster 5, which consisted primarily of necrosis cells, Osr1+ cells at E9.5 were predominantly found in clusters 0 (13.9%), 2 (53.1%), and 4 (16.1%) (Fig. 1A-d), and at E10.5 in clusters 2 (51.4%) and 3 (27.3%) (Fig. 1B-d). At E9.5, cluster 2, designated as “migratory progenitors,” expressed migration-associated genes (Dbn1, Pdgfra, Sfrp1, Wnt2) alongside cardiac progenitor markers (Tbx5, Gata4, Gata6, Hoxb4, and Osr1). Cluster 0, enriched for epithelial markers (Tjp2, Cd9, Cldn7, Cldn9, and Cdh1), was identified as epithelial cells, while cluster 4, characterized by extracellular matrix genes (Cdh2, Tgfbi, Col3a1, Col1a1, Col26a1, and Col1a2), was classified as “fibroblast-like” (Suppl. Data 1). At E10.5, although the specific identities of clusters 2 and 3 were not fully defined, both showed enriched expression of cardiac progenitor genes, including Hoxb4, Ptch1, Smo, Isl1, Gata6, and Osr1 (Suppl. Data 1).

Trajectory analysis of Osr1⁺ SHF cells at E9.5 and E10.5 revealed two distinct migratory paths. These trajectories, forming an ~90-degree angle in the principal component plots, suggest that the two Osr1⁺ cell populations migrate independently (Fig. 2A, D). The vertical trajectory was primarily composed of Osr1⁺ cells from the posterior (P) region, consistent with proximal migration toward the inflow tract (IFT). In contrast, the horizontal trajectory included cells from all four regions (P, M, A, and O), indicating a population of Osr1⁺ cells migrating toward the OFT (Fig. 2A, D). To further assess the cell fate of the migratory cells, we projected the Osr1+ cells from E9.5 and E10.5 to a spatial map of the E11.5 heart as constructed from the spatial transcriptomic data of the MOSTA study³⁵. The heart regions of the MOSTA map were delineated by cardiac-specific markers like Tbx5, Myh6, and Myh7 (Suppl. Fig. 2).

Fig. 2. scRNA-seq analysis identified two distinct migratory trajectories of Osr1+ cells.

A Trajectory analysis identified two distinct migratory paths for Osr1+ cells at E9.5. Cells from the regions “A, M, and P” are plotted on the trajectory map. B Osr1+ cells at E9.5 along the horizontal path were mapped to E11.5 MOSTA spatial BINs. Results were shown for clusters 0, 2, and 4. C Osr1+ cells at E9.5 along the vertical path were mapped to E11.5 MOSTA spatial BINs. Results were shown for clusters 0, 2, and 4. D Trajectory analysis identified two distinct migratory paths for Osr1+ cells at E10.5. Cells from the regions “O, A, M and P” are plotted on the trajectory map. E Osr1+ cells at E10.5 along the horizontal path were mapped to E11.5 MOSTA spatial BINs. Results were shown for clusters 2 and 3. F Osr1+ cells at E10.5 along the vertical path were mapped to E11.5 MOSTA spatial BINs. Results were shown for clusters 2 and 3. In (B–F) the cardiac region was outlined with a red dashed curve, inferred from the spatial distribution of biomarker Myh6 as in Suppl. Fig. 1D.

Analysis of the predominant cell clusters along the OFT (horizontal) trajectory at E9.5 and E10.5 (Fig. 2A, B, D, E, marked by a red frame) revealed their main distribution in the dorsal mesocardium and OFT at E11.5. While clusters 0 and 4 at E9.5 and cluster 3 at E10.5 were the most prominent, all five clusters contributed to OFT development. At E9.5, Osr1+ cells from section A remained in the dorsal mesocardium and distal OFT, while those from sections M and P migrated to the proximal OFT and ventricular wall. This indicates the target migration of Osr1+ SHF cells to specific cardiac regions. Furthermore, since about 90% of Osr1+ cells from section M at E10.5 likely originated from section P at E9.5 (Fig. 1D), this supports the idea that Osr1+ cells from the pSHF are crucial for OFT formation.

To clarify the cellular identity of E11.5 OFT-region descendants derived from E9.5 Osr1⁺ SHF cells, we analyzed E11.5 cell bins from the MOSTA embryos that exhibited high transcriptomic similarity to E9.5 Osr1⁺ cells. This analysis identified a prominent cell subpopulation, representing 52% of the Osr1⁺ cell bins, characterized by expression of both smooth muscle and ventricular cardiomyocyte genes, including Myh6, Myh7, Tnnt2, Tnni3, Myl3, Mybpc3, Tnni1, Acta2, and Myl9 (Suppl. Table 1).

Osr1 is required in the SHF for OFT alignment

We previously found Osr1 essential for atrial septum development in the SHF³⁰. Our bioinformatics analysis led us to re-evaluate Osr1 knockout (KO) embryos surviving past E14.5. Besides the reported AVSDs (18/21, 86%, Fig. 3A and Table 1), Osr1-KO (Osr1^LacZ/GCE) embryos also showed OFT misalignments, including DORV (14/21, 66.7%, Fig. 3A-middle column and Table 1) and OA (3/21, 14%, Fig. 3A-right column and Table 1). Examination at E11.5 revealed that while wild-type and Osr1-het (Osr1^LacZ/+) embryos showed a 90-degree ventral twist of the proximal OFT (Fig. 3B-a, b). This twist was absent in Osr1-KO embryos, which instead had a leftward twist (Fig. 3B-c). Additionally, WT and Osr1-het embryos exhibited normal looping and fusion of OFT cushions to form a spiraling septum³⁶ (Fig. 3B-g, h), whereas Osr1-KO embryos developed a parallel aortopulmonary septum (Fig. 3B-i).

Fig. 3. Osr1 is required for the OFT development.

A Histological examination of the hearts of Osr1 and Osr1 embryos at E14.5. B Grand view of the heart of WT, LacZ/+ GCE/LacZ fl/fl fl/+ Osr1 or Osr1 embryos at E11.5 (a–f) and E12.5 (g–i). C Histological examination of the hearts of Osr1, Osr1; fl/fl Mef2cAHF::Cre and Osr1; Mef2cAHF::Cre embryos at E14.5. RV right ventricle, ao aorta artery. Magnification: A, C = 50X; B = 100X.

Table 1.

Incidence of OFT defect in Osr1 mutant embryos^a

Genotype	OFT defect	Total	Type	vs. control	p value
Osr1LacZ/LacZ	17	21	DORV, OA	Osr1LacZ/+(2/16)	0.00004
	18	21	ASD (1o, AVSD)	Osr1LacZ/+(1/16)	<0.00001
Osr1fl/fl; Mef2cAHF::Cre	0	12	ASD (1o, AVSD)	Osr1fl/fl(0/10)	1
	7	12	DORV, OA		0.0017
Osr1fl/fl; Tnt-Cre+/-	1	9	DORV, OA	Osr1fl/fl(0/12)	0.1184
bOsr1fl/fl; Wnt1-CreERT/+	0	6	DORV, OA, ASD	Osr1fl/fl(0/10)	1
Osr1fl/fl;Nfatc1-Cre+/-	0	6	DORV, OA, ASD	Osr1fl/fl(0/10)	1
bOsr1fl/fl;Gli1-CreERT2/+	2	7	ASD (1o, AVSD)	Osr1fl/fl(0/12)	0.0251
	4	7	DORV, OA	Osr1fl/fl(1/12)	0.01977
cOsr1fl/fl;Gli1-CreERT2/+	2	7	ASD (1o, AVSD)	Osr1fl/fl(0/7)	0.063
	0	7	DORV, OA		1
bSmofl/+;Osr1GCE/+	9	19	DORV	Smofl/+ (0/15)	0.0009
	4	19	ASD (1o, AVSD)		0.0293
bSmofl/+;Ptenfl/+Osr1GCE/+	5	17	DORV, OA	Smofl/+;Osr1GCE/+(4/19)	0.2691
	0	17	ASD		0.0448
bSmoM2fl/+;Osr1GCE/+	0	9	DORV, OA	Osr1GCE/+(0/10)	1
	0	9	ASD (1o, AVSD)		1
bSmoM2fl/+;Osr1LacZ/GCE	3	10	DORV	Osr1GCE/LacZ(17/21)	0.0056
	0	10	ASD (1o, AVSD)	Osr1GCE/LacZ(18/21)	< 0.00001

^aWhen a given type of heart defect was not observed in the control embryos, the Chi-square test with one tail was applied; otherwise, the Chi-square test with two tails was used to test if the incidence was either higher or lower in the treatment group than the control group.

^bTMX at E7.5 and E8.5.

^cTMX at E8.5.

To determine the specific cell lineage in which Osr1 plays a critical role in OFT development, mouse embryo hearts with lineage-specific inactivation of Osr1 in the SHF, myocardium, cardiac neural crest (CNC), or endocardium were examined for the occurrence of heart defects (Fig. 3, Suppl. Fig. 3, and Table 1, p = 1). The study found that OFT misalignment with DORV was present in 50% of Osr1^fl/fl; Mef2cAHF::Cre embryos, but not in their littermate controls (6 out of 12, Fig. 3C and Table 1). We replicated our previous report showing that 75% (9 out of 12) of Osr1^fl/fl; Mef2cAHF::Cre embryos exhibited AVSD³⁰ (Fig. 3C). These findings suggest that Osr1 is necessary for the SHF for both IFT and OFT development.

To investigate whether the deletion of Osr1 affects the proliferation or survival of SHF-derived cells, potentially leading to abnormalities in OFT extension and twisting, we analyzed the percentage of proliferating and apoptotic cells in the OFT region of Osr1-KO embryos compared to Osr1^GCE/+ embryos at E11.5. Our findings revealed no significant difference in the proportion of proliferating cells between the two groups (Suppl. Fig. 4A–C). Additionally, no apoptotic cells were detected in the OFT at E11.5 (Suppl. Fig. 4D–F). These results ruled out the possibility that disruptions in cell proliferation and survival within the OFT caused the OFT defects.

Migratory Osr1+ cells toward the pulmonary trunk are committed as early as E8

To determine when Osr1+ cells are committed to contributing to the OFT, we employed genetic inducible fate mapping (GIFM)³². In embryos treated with TMX at E7.5 or E8.5, marked β-galactosidase expression was detected in the pulmonary artery but not the aortic artery, as well as the primary atrial septum (PAS) and the dorsal mesenchymal protrusion (DMP) as previously reported³⁰ (Fig. 4A). GIFM of R26R^fl/+;Osr1^GCE/+ embryos (with TMX administration at E7.5 and E8.5) was further conducted daily, from E9.5 to E13.5, to track migration of Osr1+ SHF cells that contribute to the OFT (Fig. 4B). We observed that Osr1+ descendant cells were prominently present in the pSHF but absent from the OFT at E9.5 (Fig. 4B-a and f). By E10.5, these cells were located along the dorsal splanchnic mesoderm (SpM), with some extending into the rostral SpM (Fig. 4B-b and g). At E11.5, they traversed the medial common cardiac OFT (Fig. 4B-h) and reached the pulmonary artery wall by E12.5 (Fig. 4B-i). By E13.5, after OFT separation was complete, Osr1+ cells were consistently present in the atrium and the pulmonary artery wall but were not found in the aortic artery wall (Fig. 4B-j). Additionally, the Osr1-marked cells continually contribute to the atrial wall and the atrial septum during the stages (Fig. 4B).

Fig. 4. Osr1 deletion caused a migration problem of the SHF-derived Osr1+ cells.

A The R26R^fl/+;Osr1^GCE/+ embryos were given TM (75 mg/kg) at E7.5, E8.5, or E9.5, and β-galactosidase expression was evaluated in R26R^fl/+;Osr1^GCE/+ embryos at E14.5. Magnification = 50X. B Osr1-expressing cells migrated from the SHF into the atrial septum and pulmonary trunk between E9.5 to E13.5. TMX was given to the R26R^fl/+;Osr1^GCE/+ embryos at E8.5. The location of the SHF-derived Osr1+ cells, indicated by β-galactosidase expression, was evaluated at daily intervals from E9.5 to E13.5. C The tdTomato^fl/+; Osr1^LacZ/GCE embryos and the tdTomato^fl/+;Osr1^GCE/+ embryos were given TMX at E7.5 and E8.5, and the distribution of tdTomato+ cells was evaluated at E10.5. a, c: Grand view of the embryonic heart; b, d: Sections of OCT-embedded tissues. D a–d Osr1 deletion resulted in ectopic migration of the SHF-derived cells. The tdTomato^fl/+; Osr1^LacZ/GCE embryos and the tdTomato^fl/+;Osr1^GCE/+ embryos were given TMX at E7.5 and E8.5, and the location of tdTomato+ cells was evaluated at E12.5. a–d: Grand view of the embryonic heart. (e, f) LacZ staining was performed on Osr1^LacZ/GCE and Osr1^LacZ/+ embryos at E12.5. EOsr1 deletion inhibits cell migration, leading to reduced wound healing area. FOsr1 deletion inhibits the invasion of cells into the lower chamber. Data were expressed as mean ± SD. P values were defined as ***P < 0.001, *P < 0.05. LV left ventricle, RV right ventricle, LA left atrium, RA right atrium, ao aorta artery, pt pulmonary trunk.

The GIFM was further used to investigate whether Osr1 deletion caused any migration defects of the SHF-derived Osr1+ cells in tdTomato^fl/+; Osr1^LacZ/GCE (Osr1-KO) embryos compared to tdTomato^fl/+;Osr1^GCE/+ (Osr1-het) at E10.5 and E12.5 (Fig. 4C, D). At E10.5, Osr1+ cells were present in the atrium wall and dorsal SpM of both Osr1-het and Osr1-KO embryos. However, Osr1-KO embryos lacked marked cells in the superior OFT (Fig. 4C, green arrow), with more tdTomato-labeled cells in the rostral SpA instead (Fig. 4C, orange arrow). By E12.5, a left-right positional switch between the PA and AO was observed (Fig. 4D-e vs. f), occurring specifically at this stage. At this point, Osr1-het cells were primarily located in the pulmonary trunk, while Osr1-KO cells were predominantly found in the AO. This ectopic relocation of Osr1-expressing cells was also observed in Osr1^GCE/+ and Osr1^GCE/LacZ mouse embryos, which carry the Osr1^eGFPCreERt2 “knock-in” allele driven by Osr1 promoter/enhancer elements³⁷ (Suppl. Fig. 5A).

These findings collectively indicate a migratory defect in Osr1+ SHF cells. To further investigate, we isolated Osr1+ cells based on GFP expression from dissociated SHF tissue dissected from Osr1^GCE/+ and Osr1^GCE/LacZ embryos. Subsequent analysis revealed altered expressions of key genes associated with tight junctions, cell adhesions, tissue connectivity, and cell movement at both E9.5 and E10.5 stages (Suppl. Fig. 5B), providing molecular evidence for the impaired migratory behavior of these cells during cardiac development.

To validate the role of Osr1 in cell migration, we performed a wound healing assay using NIH3T3 cells. A confluent monolayer of cells was scratched with a pipette tip, and the unclosed wound area was quantified after 24 hours. Knockout of Osr1 using CRISPR/Cas9 technology significantly impaired wound closure, suggesting a reduced migratory capacity (Fig. 4E). We further investigated this effect using a transwell migration assay in vitro. WT and Osr1-KO NIH3T3 cells were seeded in the upper chamber. After 24 h of incubation, we quantified the cells that had migrated to the lower side of the porous membrane. Compared to control cells, a significantly lower number of Osr1 KO cells invaded the lower chamber, indicating that Osr1 knockout perturbs cell migration (Fig. 4F).

Two ligand-receptor pairs, Cxcl12-Ackr3 and Hh-Ptch1, are identified along the migratory route leading to the OFT

To gain insight into the diversity among migratory Osr1+ cells, we conducted a comparison of differentially expressed genes (DEGs) between Osr1+ cells found along the IFT (vertical) path and those along the OFT (horizontal) path (Suppl. Data 2). At E9.5, our findings showed that the IFT path was enriched in well-reported genes for atrium development, such as Tbx20, Tbx5 (Fig. 5A-a), Nkx2-5, Gata4, and Wnt2 (Fig. 5A-b), as well as genes for cardiomyocytes, such as Tnnc2, Myl1, Tnni3, Acta1, Acta2, Tnnt2, Myl4, and Myl7. The trajectory map at E10.5 exhibited a comparable pattern to that of E9.5 for genes like Tbx5 (Fig. 5B-a) and Wnt2 (Fig. 5B-b).

Fig. 5. Multiple genes show various patterns along the migratory trajectory paths.

A (a–e) Cells positive for Tbx5 (a), Wnt2 (b), Cxcl12 (c), Ihh/Shh (d), and Ptch1 (e) are embedded on the trajectory map from E9.5. B (a–e). Cells positive for Tbx5 (a), Wnt2 (b), Cxcl12 (c), Ihh/Shh (d), and Ptch1 (e) are embedded on the trajectory map from E10.5.

Along the OFT path, DEGs that were highly enriched included genes for cell migration and cell-cell adhesion, such as Cxcl12 (Fig. 5A-c, B-c), Cxcl15, Ihh (Fig. 5A-d, B-d), Shh (Fig. 5A-d, B-d), Ptch1 (Fig. 5A-e, B-e), Wnt7b, Vtn, Krt7, Krt20, Klf5, and Foxa2, as well as DEGs for asymmetry regulation, such as Ihh, Shh, and Ptch1. Like the IFT path, the enriched expression of Cxcl12, Ihh, Shh, and Ptch1 along the OFT path was more pronounced at E9.5 compared to E10.5.

Upon closer examination, we observed a change in the density of Shh and Ihh, which are ligands of Ptch1, along the OFT path (Fig. 5A-d, B-d). This suggests that Hh signaling may play a potential role in guiding the migration of Osr1+ cells. These Hh ligand-expressing cells, which were mainly from the regions P and M, occupied the distal half of the OFT path. Interestingly, we also found that Osr1+ cells, which expressed the chemokine ligand Cxcl12 and originated from regions A at E9.5 (Fig. 5A-c), occupied the proximal half of the OFT path. The combined presence of Hh+ cells and Cxcl12+ cells ensured comprehensive coverage of the OFT route at E9.5. These findings suggest that the migration of Osr1+ cells along the OFT is directed by a combination of signaling cues.

Ackr3 is identified as a direct transcriptional regulatory target of Osr1

We performed a bioinformatic analysis to explore the role of the Cxcl12 chemokine in SHF cell migration by examining its receptor, Ackr3, which encodes CXCR7/GPR159, in the SHF. We analyzed Ackr3 + SHF cells to trace their migratory trajectory, revealing their presence along the OFT path at E9.5 (Fig. 6A). Projecting Ackr3+ and Cxcl12+ SHF cells from Osr1+ clusters at E9.5 onto the E11.5 MOSTA revealed the role of Cxcl12+ cells for developing OFT (Fig. 6B). However, only Ackr3+ cells from the largest cluster (cluster 2) contributed to the OFT development (Fig. 6C), indicating a unique role for this cluster.

Fig. 6. Ackr3 is identified as the direct transcription target of Osr1 in OFT development.

A Cells positive for Ackr3 are embedded on the trajectory map from E9.5. BCxcl12+ cells at E9.5 along the horizontal path were mapped to E11.5 MOSTA spatial BINs. Results were shown for clusters 0, 2, and 4. The cardiac region was outlined with a red dashed curve, inferred from the spatial distribution of biomarker Myh6, as in Suppl. Fig. 1D. CAckr3+ cells at E9.5 along the horizontal path were mapped to E11.5 MOSTA spatial BINs. Results were shown for clusters 0, 2, and 4. The cardiac region was outlined with a red dashed curve, inferred from the spatial distribution of biomarker Myh6, as in Suppl. Fig. 1D. DAckr3 expression in the SHF tissue of the Osr1^GCE/+ and Osr1^GCE/LacZ heart by real-time PCR. E IHC staining of the Cxcr7 in WT and Osr1 knockout embryos at E10.5. F Schematic of the mouse Ackr3 genomic locus including Osr1-binding regions, and the cloned genomic fragments used for Osr1 regulation assays (luciferase reporter assay and ChIP-qPCR). Enrichment of Osr1-responsive Ackr3 genomic fragments in the SHF by Osr1 ChIP-qPCR. Results are presented as mean ± SEM; n = 3; *p < 0.05. G Osr1-stimulated firefly luciferase activity in Ackr3 fragments. Results are presented as mean ± SEM; n = 4; ***p < 0.001, compared with pGL4. H The transgene Dorothy-Gal4 (Dot-Gal4) marks the embryo’s hematopoietic system and pericardial cells. However, disorganized pericardial cells were shown in the Drosophila embryo with Odd or AstC-r2 knockdown by RNAi, respectively. Created with BioRender.com.

qPCR analysis revealed a ~30% reduction in Ackr3 mRNA levels in SHF cells isolated from Osr1-KO embryos compared to Osr1-heterozygous (Osr1-het) controls at both E9.5 and E10.5 (Fig. 6D), suggesting that Ackr3 expression in Osr1-KO SHF is ~30–35% of WT levels. However, since mRNA levels do not always correlate with protein expression, we performed IHC staining on E10.5 embryos to assess Ackr3 protein levels. We revealed robust Cxcr7 expression on the cell membrane of SHF cells in WT embryos, whereas Osr1-KO embryos exhibited significantly reduced Cxcr7 expression (Fig. 6E). To account for the inherent heterogeneity within the SHF population, we quantified the percentage of Cxcr7-expressing cells by analyzing SHF cells using ImageJ. The analysis showed that only 28.7 ± 5.2% of SHF cells in Osr1-KO embryos expressed Cxcr7, compared to 93.6 ± 2.7% in WT embryos at E10.5 (p value < 0.001). These findings highlight a significant downregulation of Cxcr7/Ackr3 in the absence of Osr1, suggesting Ackr3 as the direct transcriptional target of Osr1.

The Ackr3 loci were analyzed for potential Osr1-responsive elements using the overlap of evolutionary conservation and the Osr1 consensus binding motif. ChIP analysis of wild-type SHF tissues identified the Ackr3-R1 region as enriched with Osr1 binding (Fig. 6F and Suppl. Table 2). A luciferase reporter assay showed higher luciferase activity with Osr1 binding to Ackr3-R1 (Fig. 6G and Suppl. Table 3). These results suggest that Osr1 transactivates Ackr3, crucial for regulating Cxcl12 chemokine signaling in SHF cells.

Disrupting AstC-r2, the Drosophila homolog of Ackr3, results in abnormal migration of pericardial cells

The genes guiding embryonic heart development are conserved from Drosophila to humans³⁸. In Drosophila, Odd, the homolog of Osr1, is expressed in pericardial cells from stage 12/2, which migrates toward the dorsal midline³⁹. To test Ackr3’s role in Drosophila, we used a Gal4-UAS-based RNAi system^40,41. Odd expression overlaps with Dorothy-Gal4 (Fig. 6H)⁴². Depleting Odd or AstC-R2 (the Drosophila homolog of Ackr3) in pericardial cells caused misalignment and disorganization (Fig. 6H), indicating migration defects. These results highlight conserved regulatory relationships between Osr1 and Ackr3 homologs across species, suggesting a possible role for Osr1 in modulating CXCL12-CXCR7 signaling during heart development.

Smo is identified as a direct transcriptional regulatory target of Osr1

Our trajectory analysis has revealed enriched expression of Ptch1, receptors of hedgehog (Hh) ligands, within the Osr1+ cells along the OFT path (Fig. 5), suggesting a critical role of this signaling pathway. t-SNE clustering of SHF cells at E9.5 shows that 75.2% of Osr1+ cells were Osr1⁺Ptch1⁺ cells, and over 80% located in the P region, with minimal presence in the A or M regions. By E10.5, although the posterior region still harbors the majority, the proportion of Osr1⁺Ptch1⁺ cells in the P region decreased to 71.4% (Suppl. Fig. 6). This progressive anterior shift between E9.5 and E10.5 highlights the pivotal role of Hh signaling in guiding the migration of Osr1⁺ pSHF cells during cardiac morphogenesis.

We focused on the Hh-Ptch1 ligand-receptor pair along the OFT path (Fig. 5A-d, e, B-d, e). At E9.5, Shh/Ihh+ cells from clusters 0, 2, and 4 were sparsely distributed along the OFT, whereas Ptch1+ recipient cells were abundant (Fig. 7A). By E10.5, both Shh/Ihh+ and Ptch1+ cells were exclusively present along the OFT (Fig. 7B).

Fig. 7. Osr1 plays upstream of Hh-signaling for atrial septation and OFT development.

A Shh/Ihh+ or Ptch1+ cells at E9.5 along the OFT path were mapped to E11.5 MOSTA spatial BINs. Results were shown for clusters 0, 2, and 4. The cardiac region was outlined with a red dashed curve, inferred from the spatial distribution of biomarker Myh6, as in Suppl. Fig. 1D. BShh/Ihh+ or Ptch1+ cells at E10.5 along the OFT path were mapped to E11.5 MOSTA spatial BINs. Results were shown for clusters 2 and 3. The cardiac region was outlined with a red dashed curve, inferred from the spatial distribution of biomarker Myh6, as in Suppl. Fig. 1D. C Relative expression of genes involved in Hh-signaling at SHF of the WT vs. Osr1 knockout embryos evaluated by RT-PCR. * p < 0.05, #p < 0.1. D Schematic of the mouse Smo genomic locus, including Osr1-binding regions and the cloned genomic fragments used for Osr1 regulation assays (luciferase reporter assay and ChIP-qPCR). Enrichment of Osr1-responsive Smo genomic fragments in the SHF by Osr1 ChIP-qPCR. Results are presented as mean ± SEM; n = 3; *p < 0.05. E Schematic of the mouse Gli1 genomic locus including Osr1-binding regions and the cloned genomic fragments used for Osr1 regulation assays (luciferase reporter assay and ChIP-qPCR). Enrichment of Osr1-responsive Gli1 genomic fragments in the SHF by Osr1 ChIP-qPCR. Results are presented as mean ± SEM; n = 3; #p < 0.1. F Osr1 binding to Smo genomic fragment (Smo-R5) stimulated firefly luciferase activity. Results are presented as mean ± SEM; n = 4; ***p < 0.001, compared with pGL4. G Histological examination of the embryonic hearts of Osr1^fl/fl and Osr1^fl/fl; Gli1-Cre^ERT2/+ at E14.5. H Histological examination of the embryonic hearts of Osr1^GCE/+, Smo^fl/+;Osr1^GCE/+ and PTEN^fl/+ Smo^fl/+;Osr1^GCE/+ at E14.5. I. Histological examination of the embryonic hearts of Osr1^GCE/LacZ, SmoM2^fl/+;Osr1^GCE/+, and SmoM2^fl/+;Osr1^GCE/lacZ at E14.5. LV left ventricle, RV right ventricle,ao aorta artery, pt pulmonary trunk.

We hypothesized that Hh signaling disruption in Osr1-KO embryos affects cell migration. Osr1-KO embryos showed significantly reduced Smo and increased Ihh expression (p < 0.05), with elevated Shh and Gli1 levels of marginal significance (p < 0.1) (Fig. 7C). ChIP-qPCR revealed Osr1 binding more intensely to the Smo-R5 promoter (Fig. 7D and Suppl. Table 2, p < 0.05) than to Gli1-R3 (Fig. 7E and Suppl. Table 2, p < 0.1), and a luciferase assay confirmed Osr1’s role as a transactivator of Smo (Fig. 7F and Suppl. Table 3).

Previous studies have shown that Hh-signaled SHF progenitors contribute to the pulmonary artery^43–46. In Osr1^fl/fl; Gli1^Cre-ERT/+ embryos treated with TMX at E7.5 and E8.5, 53.8% developed DORV, compared to 1 out of 12 controls (Fig. 7G and Table 1). However, TMX treatment at E8.5 alone did not cause DORV. Additionally, primum ASD was present in two of seven treated embryos (p = 0.0251, Fig. 7G and Table 1). These results highlight the critical role of Osr1 in Hh-receiving cells for OFT and IFT development.

To explore the genetic interaction between Osr1 and Hh signaling, we examined Smo^fl/+;Osr1^GCE/+ embryos. TMX-induced Smo deficiency in Osr1-het cells led to DORV in nine out of 19 embryos by E14.5, whereas Smo^fl/+ or Osr1^GCE/+ embryos had normal AO and PT (p = 0.0009, Fig. 7H and Table 1). Additionally, four Smo^fl/+;Osr1^GCE/+ embryos showed primum ASD or AVSD, which were absent in Smo^fl/+ or Osr1^GCE/+ embryos (p = 0.0293, Fig. 7H and Table 1). These findings confirm the functional interaction between Hh signaling and Osr1 in OFT and IFT development.

Overactivation of Hh signaling rescues OFT development in Osr1-KO embryos

Previous reports have demonstrated that Osr1 null SHF has cell proliferation defects³⁰. We also observed reduced expression of Ccnd2, Cdk2, Cdk4, and Cdk6 in the SHF of Osr1-KO embryos at E10.5 (Suppl. Fig. 5A). In SHF of Smo^fl/+;Osr1^GCE/+ embryos, Cdk2 and Cdk6 were also reduced, with increased expression of proliferation inhibitors Pten and Cdkn1a (Suppl. Fig. 5B), indicating similar proliferation defects.

We tested if reducing Pten expression could reverse ASD or DORV in Smo^fl/+; Osr1^GCE/+ embryos, based on previous findings that lower Pten can correct SHF proliferation defects^47–49. In Pten^fl/+;Smo^fl/+;Osr1^GCE/+ hearts, ASD was fully restored in all 17 embryos (Fig. 7H and Table 1, 0/17). However, five of these embryos still showed DORV or OA (Table 1 and Fig. 7H, p = 0.2691 vs. Smo^fl/+;Pten^fl/+). This indicates that while correcting proliferation defects can rescue ASD, it does not restore OFT development, underscoring the critical role of SHF cell migration issues in OFT defects.

We tested whether continuous Hh signaling activation could rescue ASD or DORV in Osr1-KO embryos. SmoM2^fl/+;Osr1^GCE/+ embryos showed enlarged DMP and hyperplastic ventricles but no DORV/OA or ASDs (Table 1 and Fig. 7I, 8/10). In SmoM2^fl/+;Osr1^GCE/LacZ embryos with activated Hh signaling, seven of ten displayed normal OFT alignment, compared to 4 of 21 in Osr1-KO (Osr1^GCE/LacZ) embryos (Table 1 and Fig. 7I, p = 0.0056). Overactivation also fully rescued ASDs (Table 1 and Fig. 7I). These findings suggest that Osr1 regulates Smo, which in turn modulates Hh signaling in SHF cells for both OFT and IFT development.

Discussion

The SHF cells play a vital role in the development of OFT. Mutations in genes associated with the SHF can lead to significant defects in the arterial pole, accounting for approximately one-third of congenital heart defects in newborns. A human study has identified a strong association between the rs12329305 T polymorphism in the OSR1 gene and congenital heart malformations, highlighting the critical role of Osr1 in heart development²⁹. In this study, we provide further insights into the functional role of Osr1 in regulating the migration of pSHF cells toward the OFT, uncovering the underlying molecular mechanisms. Using trajectory analysis of scRNA-seq data from SHF microdissections at E9.5 and E10.5, we identified two distinct migratory pathways of Osr1+ SHF cells. Integrating these findings with a spatial transcriptomic atlas revealed that the migratory Osr1+ cells contribute to the development of either the OFT or the IFT. Notably, Osr1 deletion resulted in conditions such as ventricular septal defects (VSD), DORV, or OA, linked to irregular cell migration. We identified key ligand-receptor pairs, including Cxcl12-Cxcr7 and Hh-Ptch1, as essential for the migration of Osr1+ pSHF cells. Additionally, we determined that Smo and Ackr3 are downstream targets of Osr1 transactivation, establishing the functional regulation of Osr1-Hh signaling and Osr1-Cxcl12 signaling in cell migration and OFT development. These findings highlight the critical role of Osr1+ pSHF cells in guiding heart pole development and elucidate important ligand-receptor interactions involved in cell migration, which may inform potential therapeutic targets for congenital heart defects.

Previous research has underscored the importance of aSHF cells in the formation of the right ventricle and OFT myocardium, leading to substantial investigation into aSHF progenitors^{26,50–52,19}. In our study, we identified that over 30% of OFT and aSHF cells present at E10.5 likely originate from pSHF cells at E9.5, with a prominent representation among Osr1+ cells. Our findings indicate that over 70% of Osr1+ aSHF cells at E10.5 are derived from pSHF cells traced to E9.5, underscoring the essential role of directed migratory behavior originating from the pSHF in heart development. Our lab was not the first to show that marker genes traditionally associated with aSHF or pSHF do not exclusively contribute to the OFT or IFT, respectively. For example, Tbx1, usually seen as a marker for aSHF, plays a crucial role in regulating the segregation and deployment of progenitor cells within the pSHF, contributing to both inflow and OFT morphogenesis^53–55. Our discovery that a subset of pSHF cells contributes to OFT development aligns with previous reports indicating that Hoxb1+ pSHF cells contribute to the arterial pole^51,56,57. This emphasizes the complex patterning of SHF and the significant role of cell migration, as Osr1+ pSHF cells travel substantial distances to contribute to OFT development.

Our study elucidates the critical role of Osr1 in regulating Osr1+ pSHF cell migration during OFT development. We found that Osr1+ pSHF cells, which contribute to the pulmonary trunk, are specified between E8 and E10. Osr1 deletion in the SHF disrupted this process, leading to OFT malrotation and structural defects such as DORV and OA, driven by aberrant migration of the Osr1+ lineage. While migration persisted in the absence of Osr1, abnormal migration patterns emerged, likely causing cells to miss essential spatiotemporal milestones for proper OFT integration, resulting in misalignment. This is evidenced by the swapped positions of the AO and PA at E12.5 and ectopic Osr1+ cell localization on the AO in Osr1-deleted embryos. Future studies should explore how Osr1 levels influence migration speed, cell adhesion, and migratory force to further clarify its regulatory role in SHF cell migration.

Analogous to limb development^58–60, where Hh signaling directs cell migration, Hh signaling likely guides the migration of Osr1+ cells toward the OFT. The Moskowitz lab has previously demonstrated that Hh signaling from distinct pulmonary and pharyngeal endoderm is required for inflow and outflow septation, respectively⁴⁶. While showing Shh/Ihh+ cells presenting exclusively along the SHF and the OFT, we also showed that Osr1+ cells on the OFT migratory path enriched Ptch1 expression, and these Hh-receiving cells contribute to the OFT, validated by spatial transcriptomics and supported by the occurrence of DORV/OA in Osr1-deleted embryos. We identified Smo as a downstream target of Osr1, and combined heterozygosity for Osr1 and Smo resulted in OFT misalignment, underscoring Osr1’s role as an upstream regulator of Hh signaling. Notably, hyperactivation of Hh signaling in Osr1KO embryos partially rescued OFT defects, with only three out of ten embryos exhibiting DORV compared to 17 out of 21 without Hh overactivation (P < 0.0056, Table 1). These findings highlight the crucial role of Hh signaling in regulating SHF cell migration during OFT development.

Our study identifies Ackr3 (encoding CXCR7) as a direct transcriptional target of Osr1, as evidenced by reduced Cxcr7 expression in Osr1-deficient SHF cells and confirmed through ChIP and reporter assays. Although the precise contribution of Cxcl12-Cxcr7 signaling to the OFT defects in Osr1 mutants remains to be fully elucidated, our findings suggest that Osr1 modulates chemokine responsiveness in SHF cells. Trajectory analyses reveal that the expression domains of Shh/Ihh and Cxcl12 are spatially complementary, while their corresponding receptors, Ptch1 and Ackr3, are enriched along the SHF migratory path toward the OFT. Partial rescue of OFT defects in Osr1 knockout embryos by Hh-pathway overactivation suggests that additional ligand-receptor interactions, such as Cxcl12-Cxcr7, may coordinate SHF cell migration. Further, evolutionary conservation is reflected by defective pericardial cell migration in Drosophila lacking Astc-r2, the homolog of Ackr3, and by heart defects in Ackr3 knockout mice, which exhibit cardiac hypertrophy and outflow valve malformations, as reported in two independent studies^61,62. Building on previous studies implicating chemokine-receptor axes such as MIF-CXCR2 in SHF migration⁶³, our results imply a complex signaling network, potentially involving Cxcl12–Cxcr7 acting in parallel with the Osr1-Hh/Ptch1 pathway. Further studies using Ackr3 and Ptch1 double-knockdown mouse models will be critical to determine whether and how these pathways synergistically function in guiding SHF cell migration, contributing to OFT development.

In conclusion, these findings shed light on the intricate mechanisms governing SHF cell migration, particularly emphasizing a subset of migratory Osr1+ pSHF lineages in OFT development. By exploring the roles of Osr1 in Hh signaling and its regulation of Ackr3, we deepen our understanding of the processes essential for proper OFT alignment. This research not only enhances our comprehension of congenital heart defects but also identifies potential pathways for developing improved treatment strategies.

Methods

Mouse lines

All mouse experiments were performed in a mixed B6/129/SvEv background and around the age of 2–3 months. The Osr1^fl/+ mouse line has been described before (Lan et al., 2011). The SmoM2^fl/+, Osr1^Cre-ERT2/+, Osr1^lacZ/+; Gli1^CreERT2/+, Mef2cAHF::Cre, Tie2^Cre/+, Smo^fl/+ mouse lines were obtained from Jackson Laboratory (USA). TnT^Cre/+ mouse line was from Dr. Yiping Chen’s lab (Tulane University, New Orleans). Nfat1c^Cre/+ mouse line was from Dr. Bin Zhou’s lab (Albert Einstein College of Medicine, Bronx, NY). Mouse experiments were completed according to a protocol reviewed and approved by the Institutional Animal Care and Use Committee of Texas A&M University and the University of North Dakota, in compliance with the USA Public Health Service Policy on Humane Care and Use of Laboratory Animals.

Mouse experiments were completed according to a protocol (IACUC 2020-0161) reviewed and approved by the Institutional Animal Care and Use Committee of Texas A&M University, in compliance with the USA Public Health Service Policy on Humane Care and Use of Laboratory Animals. We have complied with all relevant ethical regulations for animal use

Animals were housed in a temperature-controlled (22 ± 2 °C) and humidity-controlled (50 ± 10%) facility with a 12-h light/dark cycle (lights on at 7:00 a.m.). Mice were housed in individually ventilated cages with ad libitum access to a standard chow diet and water. Bedding was changed regularly, and cages were enriched with nesting materials and shelter (e.g., cotton nestlets and plastic huts) to support natural behaviors and improve animal welfare. All husbandry procedures were conducted following institutional animal care guidelines.

Tamoxifen administration

Tamoxifen (TMX) -induced activation of CreERT2 was accomplished by oral gavage with two doses of 75 mg/kg TM at E7.5 and E8.5.

Microdissection of the embryo SHF

To obtain the SHF splanchnic mesoderm for use in single-cell RNA sequencing or quantitative real-time PCR, E9.5 or E10.5 embryos were dissected to posterior-, anterior-, and middle- sections^31,33,64. Microdissected SHF fragments were dissociated by collagenase digestion for single-cell sequencing library preparation (10X Genomics). For fluorescence-activated cell sorting, SHF was identified and dissected under a fluorescence dissecting microscope. The dissociated single cells were sorted using a FACSAria IIu cell sorter (BD Biosciences).

Real-time PCR

Total RNA was extracted from the PSHF regions of the mouse embryo’s hearts using the RNeasy Mini Kit (QIAGEN), according to the manufacturer’s instructions. Two hundred nanograms of total RNA were reverse transcribed using a SuperScript^TM III Reverse Transcriptase kit from Invitrogen. qPCR was performed using a POWER SYBR Green PCR master mix from Applied Biosystems. Results were analyzed using the delta-delta Ct method with GAPDH as a normalization control.

Fluorescence-activated cell sorting

The GFP signal was identified using a fluorescence dissecting microscope. GFP+ and control SHF tissue from each E9.5 or E10.5 embryo was dissected and individually placed in 1.5 ml microcentrifuge tubes containing 50 μl sort buffer (PBS with 2% FBS and 25 mM HEPES) on ice. The sort buffer was removed carefully, and 50 μl 0.25% trypsin-EDTA was added to each tube. Tissues were digested into single cells by incubating at 37 °C on Thermomixer R (Eppendorf) for 10 min, along with pipetting up and down every 2 min to facilitate tissue dissociation. Cells were then pelleted at 1000×g for 3 min at 4 °C and washed with 100 μL of sorting buffer. Centrifugation was repeated, and the cells were resuspended in 500 μL of sort buffer.

Cells were sorted using FACSAria IIu (BD Biosciences). The parameters used depended on each run, but generally, the voltage for FSC, SSC, and FITC was set to 126–132, 226, and 445–451, respectively. The population of interest was first selected using an SSC-A vs. FSC-A gate (P1), and singlets were selected using an SSC-A vs. FSC-W gate (P2). The cutoff for calling a cell GFP+ or GFP- was defined by selecting the maximum FITC-A value generated by cells in the GFP- sample. Cells that fell in the intersection of P1 and P2 and with FITC-A greater than the cutoff were sorted into 200 μl lysis buffer from the RNA extraction kit. Approximately 1000 to 4000 cells were collected from each E9.5 embryo, and 7000 to 28,000 cells were collected from each E10.5 embryo.

Genetic inducible mate mapping

To perform fate-mapping of Osr1-expressing progenitors in the SHF, we utilized the Osr1^CreERT2/+; tdTomato^fl/+ mouse line, which allows precise temporal control of Cre recombinase activity through tamoxifen administration. This inducible system ensures that Cre-mediated recombination, and thus tdTomato reporter expression, is activated only in Osr1-expressing cells during the desired developmental window. Specifically, pregnant females carrying Osr1^CreERT2/+; tdTomato^fl/+ embryos were administered tamoxifen via intraperitoneal injection at embryonic days E7.5 and E8.5. This timing was chosen to selectively label Osr1-expressing progenitors contributing to the SHF, minimizing labeling of progenitors from other regions, such as the first heart field or non-cardiac tissues. Embryos were subsequently harvested at later stages (e.g., E9.5 to E12.5) for analysis, allowing us to trace the migration and differentiation of labeled Osr1+ SHF cells during cardiac development.

Single-cell RNA sequencing

SHF were harvested from WT mice at E9.5 and E10.5. After dissection, hearts were minced and subjected to cell dissociation using trypsin. After digestion, the samples were centrifuged at 500 × g for 10 min. After removing the supernatant, the cells were resuspended and washed in PBS containing RNase inhibitor, and the cell number was measured. Subsequently, cells were processed using the 10x Genomics Chromium Controller to encapsulate them with gel beads in emulsion. Library preparation followed the protocol provided by 10x Genomics. Sequencing was performed using NextSeq (Illumina). All sequence reads were aligned to the mm10 mouse reference genome.

Chromatin immunoprecipitation

SHF tissues from E9.5 and E10.5 wild-type embryos were collected and pooled in PBS containing Complete Mini EDTA-free Protease Inhibitor Cocktail (Sigma-Aldrich) on ice. Tissues were crosslinked in 1% formaldehyde in PBS on the rotator for 10 min. Crosslink was quenched by adding glycine to a final concentration of 0.125 M and incubating for 5 min. Tissues were pelleted, washed once with PBS, and stored at −80 °C until sufficient material was acquired for one chromatin immunoprecipitation (ChIP). Approximately 100 or 60 SHF tissues were used for one E9.5 or E10.5 ChIP, respectively. Tissues were pooled into Sonication Buffer (0.5% SDS, 20 mM Tris, pH 8.0, 2 mM EDTA, 0.5 mM EGTA, with freshly added 0.5 mM PMSF and Protease Inhibitor), homogenized using Tissue Grinder (Axygen), and incubated for 30 min on ice for cell lysis. Chromatin was sheared for 12 min into 200–1000 bp fragments using the S220 sonicator (Covaris) and the High Cell program, then diluted 5-fold in IP Buffer (0.5% Triton X-100, 2 mM EDTA, 20 mM Tris, pH 8.0, 150 mM NaCl, 10% glycerol, with freshly added 0.5 mM PMSF, and protease inhibitor). After 1 h of pre-clear at 4 °C using Dynabeads Protein G (Life Technologies), chromatin was incubated with OSR1 antibody (sc-376529X, Santa Cruz) with rotation at 4 °C overnight. Chromatin-antibody complexes were captured using Dynabeads Protein G and washed with Low Salt Buffer (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris, pH 8.0, 150 mM NaCl), High Salt Buffer (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris, pH 8.0, 500 mM NaCl), LiCl Buffer (1 mM EDTA, 10 mM Tris, pH 8.0, 0.25 M LiCl, 1% NP-40, 1% sodium deoxycholate), and TE Buffer (10 mM Tris, pH 8.0, 1 mM EDTA). Chromatin-antibody complexes were then eluted with Elution Buffer (1% SDS, 0.1 M NaHCO₃), and reverse-crosslinked overnight at 65 °C using NaCl. RNA and proteins were digested by incubating with RNAse A (Thermo Scientific) at 37 °C and Proteinase K (Invitrogen) at 65 °C, respectively. DNA was then purified using phenol-chloroform followed by ethanol precipitation. QPCR was performed to quantify the IP pulldown genomic DNA. The signal from the ChIP sample is normalized to the input sample (the fraction of starting chromatin used in the ChIP). Genomic regions with potential Osr1-binding sites and negative control sites are listed in Supplementary Table 2.

Luciferase reporter assay

Regulatory regions were cloned upstream of a firefly luc2 gene in the pGL4.23 reporter vector (Promega). pcDNA3 Osr1 was a gift from Paul Danielian (Addgene plasmid # 26485; http://www.addgene.org/26485/; RRID: Addgene_26485). About 2 × 10⁴ HEK293 cells were plated per well in a 96-well plate containing 100 µl culture media (DMEM with 5% FBS and 1× Antibiotic-Antimycotic). After 24 h, reporter vectors were transfected into the cells, with or without the Osr1 vector, using FuGENE HD Transfection Reagent (Promega). Cells were lysed and assayed 24 h after transfection using the Dual-Luciferase Reporter Assay System (Promega). Student’s t-test was performed to determine statistical significance, and p < 0.05 was considered significant. Genomic regions with potential sites tested are listed in Supplementary Data 1. Primers used for site-specific mutation and subcloning are listed in Supplementary Table 3.

RNA interference in Drosophila

Drosophila lines were obtained from the Bloomington Drosophila Stock Center at Indiana University. Dot-Gal4 and UAS-GFP homozygous lines were balanced and then crossed with Odd or AstC-R2 RNAi lines. Embryos were placed in a drop of halocarbon oil suspended from a coverslip and evaluated by confocal microscopy⁶⁵.

Bioinformatics and data analysis

Quality control was performed on RNA-seq and scRNA-seq reads from all sequencing runs using FastQC (v0.11.5) (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). RNA-seq Reads were mapped to the mouse reference genome GRCm38.91 (mm10) using HISAT2 (v2.1.0)⁶⁶. Reads that were mapped to multiple locations were removed, and only read pairs that were aligned uniquely and concordantly were retained for further analysis. Transcript quantification and differential gene analysis were performed using the Cufflinks suite (v2.2.1)⁶⁶. Genes with a false discovery rate (FDR) less than 0.05 were considered differentially expressed. Gene ontology (GO) and KEGG pathway analyses were performed using the R package clusterProfiler (v3.10.1)⁶⁷.

scRNA-seq reads were demultiplexed by Cell Ranger 3.0. The raw data were then normalized by an SCTransform⁶⁸ and were initially processed with Seurat for individual time stages. For integrative analysis of both time stages, Osr1 expression cells from both stages were selected, and canonical correlation analysis (CCA) was applied by the following steps⁶⁹: First, it brought both time points into a comparable subspace. Second, nearest-neighbor methods were used to identify paired anchors (cells in similar states at each time point). Finally, these anchors are used to improve the transformation of other cells into the joint subspace. Trajectory analysis of both embryonic days (E9.5 and E10.5) was done by Slingshot⁷⁰ using the R package dyno, and the trajectory of Osr1 cells was plotted in 2D with principal component analysis (PCA). Pearson correlation between cells from both E9.5 and E10.5 was calculated to quantify the percentage of cells corresponding to specific segment labels (O, A, M, P), as well as the similarities between major clusters.

Cell-cell similarity analysis

Cell–cell similarity analysis was performed to identify matched SHF cells between E9.5 and E10.5. This was achieved by calculating the Pearson correlation of global gene expression profiles between cells from the two time points. For each E9.5 cell, the E10.5 cell with the highest correlation was selected as its matched pair.

BIN data analysis, SHF mapping, and integrative analysis

Spatial mapping with single-cell resolution was performed based on the mouse organogenesis spatiotemporal transcriptomic atlas (MOSTA)³⁵. The fundamental spatial unit of MOSTA is BIN, which represents a square with a side length of ~25 μm. E9.5, E10.5, and E11.5 from MOSTA were chosen with multiple sections included. The best section was chosen from each embryonic day based on the spatial pattern of SHF-related or developmental genes, including Osr1, Tbx5, Isl1, Myh6, and Myh7. Once the best section was selected, the Osr1-expressing cells from both embryonic days were mapped to MOSTA E9.5, E10.5, and E11.5. First, a list of common genes in both SHF cells and MOSTA was sorted out. Using the overlapped gene list, the top 100 genes with the highest standard deviation in the expression level across all SHF cells were sorted out. Then, Pearson’s correlation coefficient between all single cells and MOSTA BINs was calculated via the new gene list. Second, BINs in the area surrounding the heart (heart included unless specified) were further selected. For integration, Osr1-expressing cells from both embryonic days were mapped to the confined regions in E9.5, E10.5, and E11.5 from MOSTA, respectively. For Osr1+ cells in the trajectories, those that displayed a distinct migratory pattern were selected and mapped to E11.5 MOSTA BINs. All mappings were categorized into various groups corresponding to the SHF clustering with pre-identified cell types based on the Louvain algorithm, as well as their segment labels based on experimental dissection.

CRISPR/Cas9-mediated knockout of Osr1 in NIH3T3 cells

To generate Osr1-deficient NIH3T3 cells (ATCC), we employed a CRISPR/Cas9-mediated genome editing approach using a GenScript all-in-one CRISPR plasmid system. Guide RNA (gRNA) sequences targeting the mouse Osr1 gene were designed using the GenScript CRISPR design tool to minimize predicted off-target effects and maximize on-target activity. The selected gRNA sequence was cloned into the GenScript CRISPR/Cas9 vector, which co-expresses Cas9 nuclease and an antibiotic (puromycin) selection marker. NIH3T3 cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM; Gibco) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin at 37 °C in a humidified 5% CO₂ atmosphere. Cells were seeded at ~60% confluency one day before transfection. Transfection was performed using FuGENE® HD Transfection Reagent (Promega) according to the manufacturer’s instructions. At 48 h post-transfection, cells were subjected to selection with puromycin for 72 h. Single-cell clones were generated by limiting dilution and expanded for screening. Successful genome editing and loss of Osr1 expression was confirmed at the transcript level by quantitative reverse transcription PCR (RT-qPCR) and at the protein level by Western blotting.

Transwell cell migration assay

Cell migration was assessed using a Transwell migration assay with 8.0-μm pore size polycarbonate membrane inserts (Corning Costar, #3422) in 24-well plates. Both control and Osr1 knockout (KO) NIH3T3 cells were serum-starved in DMEM containing 0.5% FBS for 6 h prior to the assay. A total of 5 × 10⁴ cells in 200 μL of serum-free DMEM were seeded into the upper chambers. Cells were incubated at 37 °C in a 5% CO₂ atmosphere for 24 h. After incubation, non-migrated cells remaining on the upper side of the membrane were removed with a cotton swab, and migrated cells on the underside were fixed with 4% paraformaldehyde for 15 min and subject to HE staining for 20 min. Stained cells were imaged under a light microscope, and the number of migrated cells was quantified.

Wound healing invasion assay

To assess cell motility and invasive behavior, a wound healing assay was performed using both control and Osr1 knockout (KO) NIH3T3 cells. Cells were seeded in six-well plates and cultured in complete DMEM until they reached ~90–100% confluence. A linear scratch was made through the cell monolayer using a sterile 200 µL pipette tip, and cellular debris was gently removed by washing twice with PBS. Cells were then incubated in DMEM containing 1% FBS to minimize proliferation-driven closure. Images of the wound area were captured immediately (0 h) and at 24 h post-scratch. The wound width was measured at multiple points per field using ImageJ software, and wound closure was calculated in wound area relative to the initial scratch width.

Blinding

In this study, the individual responsible for group allocation was aware of the assignments during randomization. The investigators conducting the experiment were [blinded/not blinded], depending on the nature of the intervention. During data analysis, the analyst was blinded, with group identities until statistical analysis was completed.

Statistics and reproducibility

Statistical analyses were performed using Student’s t-test, with a p value less than 0.05 considered statistically significant, unless otherwise specified. All mouse experiments included at least three independent biological replicates (N ≥ 3). Replicates were defined as independent mouse samples subjected to the same experimental conditions in separate experimental runs. Data were presented as mean ± standard deviation (SD) unless indicated otherwise.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

L. Xie and K. Zhang designed the study; L. Liu, H. Cheng, M. Xiang, J. Liu, Y. Qin, and H. Peng conducted experiments; K. Zhang, B. Kidd, JY. Li (Jiangyuan Li), and J. Li (Jing Li) performed the statistical analysis, L. Xie, J. Ji, R. Jiang, RG. Kelly, J. Li, and L. Liu wrote the manuscript.

Peer review

Peer review information

Communications Biology thanks Ian Scott and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Primary Handling Editors: Kaliya Georgieva. [A peer review file is available].

Data availability

Both original data and secondary data supporting the findings of this study are available within the paper and its Supplemental Material. The mouse organogenesis spatiotemporal transcriptomic atlas (MOSTA)⁷¹ data for spatial mapping are available at the following URL: https://db.cngb.org/stomics/mosta/download/. The single-cell RNA-seq datasets used in this study are available through the Gene Expression Omnibus (GEO) database with GEO accession number GSE309215.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s42003-025-09118-0.