The endocardium possesses remarkable plasticity to differentiate into other cardiovascular cell types including fibroblasts and coronary endothelium. However, their ability to differentiate into hematopoietic populations remains debated with multiple contradicting reports. Supporting the view of the endocardium contributing to hematopoietic development, our work reveals for the first time the function of the endocardium in zebrafish as a niche for hematopoietic stem and progenitor cells (HSPCs). Using single-cell RNA sequencing, we identified a subpopulation (10%) of endocardial cells that expressed hematopoiesis-promoting genes. High-resolution confocal and spinning disk imaging uncovered hematopoietic cells, mainly consisting of HSPCs, attached to the luminal surface of cardiac ventricular and atrial endothelial cells. These cells remain stably integrated on a long term in the developing heart from 60 hours post fertilization until at least 10 days post fertilization. Cell tracing of different vascular beds using the photoconvertible Kaede protein as well as 3D longitudinal light-sheet imaging of the beating heart reveal that cardiac-residing hematopoietic cells derive from the endocardium de novo and the dorsal aorta, the first vascular tissue source of HSPCs. Furthermore, we found emergence of HSPCs in hearts that were extracted and cultured ex vivo, even without external hematopoietic sources. Maintenance of HSPCs in the heart depends on the adhesion factors Integrin α4 and Vcam1, which also mediate HSPC homing into other hematopoietic niches, but is independent of cardiac contraction-induced mechanical forces or cardiac trabecular architecture. Finally, blocking primitive erythropoiesis led to an augmented population of cardiac-residing HSPCs, suggesting that the endocardium functions as a hematopoietic cell reservoir. Together, our work uncovers the endocardial tissue as a de novo source of hematopoiesis and a newly discovered niche for the HSPCs.

Blood flow-derived wall shear stress (WSS) is critical for heart valve formation. However, the mechanism by which blood flow translates into spatiotemporal regulation of valve development remains poorly understood. Here, we uncover a WSS-specific mechanosensor, the primary cilium, on the endocardial cells lining the heart lumen in vivo. We discover a role for primary cilia in the spatial regulation of cushion formation, the first stage of valve development, by regionally controlling endothelial to mesenchymal transition (EndoMT) via modulation of Kruppel-like Factor 4 (Klf4). KLF4 is a mechanosensitive transcription factor that we find negatively correlates with EndoMT progression. In regions of high WSS, we find selective loss of endocardial cilia, correlating with KLF4 downregulation and permissive EndoMT. Mouse embryos constitutively lacking cilia (cilia KOs) exhibit a blood-flow dependent increase in KLF4 expression, independent of upstream laterality defects. These results suggest that ciliary presence and subsequent loss in response to cardiac function are needed for downregulation of KLF4 and EndoMT progression, providing a method to spatially restrict cushion growth. Further, KLF4 dysregulation in cilia KOs correlates with significantly impaired EndoMT progression and cushion cellularization. SingleNuc RNAseq on isolated e9.5 wild-type and cilia KO hearts revealed that hearts lacking cilia fail to progress from mid- to late-EndoMT, corresponding to the timepoint where Klf4 is downregulated in wild-type hearts. Cilia KO hearts retain endothelial markers during EndoMT and fail to express mesenchymal and EndoMT genes that are regulated by KLF4. Gene Ontology terms for mesenchyme and valve development are downregulated in cilia KO EndoMT, while terms for vascular identity/integrity are upregulated. Finally, signaling cascades necessary for EndoMT and upstream of KLF4 expression, such as TGFβ, are significantly altered in cilia KO hearts. Taken together, these data identify a novel mechanosensory role for endocardial primary cilia in driving EndoMT localization and progression via regulation of KLF4 expression during cushion development.

In mammals, endothelial cells (ECs) are required for normal vasculogenesis and angiogenesis. The ubiquitously expressed cytoplasmic adaptor genes, Crk/Crkl, are required for normal cardiovascular development. Their molecular roles in mesoderm/ECs for early embryonic and postnatal vascular development, however, are unclear. To determine functions of Crk/Crkl, we first used Mesp1Cre for gene inactivation in the mouse. We found that mesodermal inactivation of both Crk/Crkl caused embryonic lethality by mid-gestation with severe vascular defects. Although vasculogenesis appeared normal, early angiogenesis was disrupted, leading to disorganized vascular networks. Next, we specifically inactivated Crk/Crkl in ECs in vivo using Tie2-Cre and found there was defective integrity of blood vessels in mid-gestational embryos. In addition, we inactivated Crk/Crkl in postnatal ECs using Cdh5-CreERT2 and examined angiogenesis in the retina. We found that the mutant retinas exhibited defective sprouting together with reduced vascular branch point densities in the superficial vascular plexus. To examine whether Crk/Crkl are required for in vitro angiogenesis, we isolated primary ECs from embryos and performed vascular tube formation assays. We found that ECs lacking Crk/Crkl failed to form vascular tube structures while control ECs were able to form a tube network, suggesting that Crk/Crkl are required for proper angiogenesis. To further investigate the altered transcriptome when Crk/Crkl are inactivated, we also performed scRNA-seq of the Mesp1Cre lineage and found that there was selective upregulation of angiogenesis and cell migration related genes in ECs in the Crk/Crkl mutants, including NOTCH signaling genes such as Dll4 and Hey1. Bioinformatic analysis identified a relative increase in the number of differentiated angiogenic ECs (Pecam1). Consistent with this, we identified an expansion of Dll4 expressing cells within abnormal arteries in vivo. Also, our data indicates that there is dysregulated expression of lineage genes that promote EC differentiation, causing accelerated cell fate progression during EC differentiation. Overall, we suggest that Crk/Crkl are crucial for regulating embryonic and postnatal angiogenesis.

Hyaluronic acid (HA) is a vital component of the extracellular matrix (ECM) providing a fluid filled environment essential for morphogenesis during embryonic development. Levels of HA are highly regulated by its synthesis, via HA synthases, and its degradation by hyaluronidases. The balance of these processes has profound impacts on tissue morphogenesis. To investigate the role of HA degradation in the developing heart, we have generated a knockout mouse model for Cemip2, a cell surface hyaluronidase. LacZ staining shows Cemip2 expression is enriched in endothelial cells of the heart and endocardial loss of Cemip2 (using Tie2:Cre), leads to catastrophic heart defects including enlarged atrioventricular (AV) valves, double outlet right ventricle and ventricular septal defects that result in embryonic loss at midgestation. Cemip2-deficient AV valves fail to undergo remodelling to form leaflets and remain in a more cushion-like state. To investigate the impact of dysregulated ECM on cell-matrix interactions during cardiac valve morphogenesis, single cell sequencing was paired with imaging (including 3D methodologies) on Cemip2-deficient valves and reveals changes in cellular architecture, matrix organisation, and signalling. Cemip2-deficient valves have a specific loss of the condensing mesenchyme population, which normally resides at the edges of the developing valves, and also fail to undergo ECM remodelling of the proteoglycan-rich matrix in this region. Prior to the phenotypic valve expansion in Cemip2-deficient valves, there are profound changes in signalling pathways, known to play a role in valve development, such as changes in phosphorylated Smad1/5/8, a downstream effector of BMP signaling and phosophorylated ERK, a downstream effector of VEGF/FGF signalling. The Cemip2 mutant valve phenotype illustrates that disruption of one matrix component leads to disorganisation of numerous matrix components, aberrant signalling and loss of a subset of mesenchyme cells in the developing valve. Further investigation of this model will provide fundamental knowledge about how cells communicate with the ECM and how the ECM influences cell fate and signalling in the developing heart.

The epicardium, a mesothelial cell layer in all vertebrate hearts, promotes heart regeneration after injury by providing regenerative signals and progenitor cells for rebuilding functional heart tissues. Previous investigations from our lab revealed an injury-activated epicardial progenitor population that gives rise to mesenchymal cells and mural cells to support heart regeneration. Through single-cell transcriptomic analysis, we identified the expression of the transcription factor scxa (scleraxis bHLH transcription factor a) in epicardial progenitors and their progenies, hinting at its potential involvement in epicardial cell fate determination. Reporter assays revealed transient expression of epicardial scxa during coronary angiogenesis, with minimal to negligible expression observed in the adult epicardium. Intriguingly, hypoxia strongly induces scxa expression in the uninjured adult heart. Upon heart amputation injury, scxa expression is induced in epicardial cells at the injury site, forming a scaffold surrounding coronary vessels. Employing lineage tracing tools of tcf21:CreER and scxa:CreER, we found that ventricular scxa+ cells are derived from the epicardial progenitors and give rise to progenies interacting with the developing or regenerating coronary vessels. Further characterization identified one progeny type as a novel perivascular fibroblast population that expresses col18a1a and low-levels of pdgfrb, distinguishing them from pericytes and vascular smooth muscle cells. Notably, double homozygous mutants of scxa and scxb exhibited signs of coronary vessel defects. In summary, our findings suggest that hypoxia-driven scxa contributes to epicardial-to-perivascular fibroblast differentiation, facilitating angiogenesis and regeneration processes.

Neonatal Marfan syndrome (neoMFS) is a rare connective tissue disorder that is genotypically and phenotypically distinct from classical Marfan syndrome (MFS). Importantly, neoMFS is a diagnosis with a 95% mortality within the first two years of life due to rapidly progressing heart failure secondary to severe mitral/tricuspid valve insufficiency. For adults with classical MFS, beta blockers and surgery are gold standard interventions to prevent life-threatening complications from aortic aneurysm; however, medical or surgical interventions are non-existent for infants with neoMFS, who require palliative care and hospice after diagnosis. The high mortality rate of neoMFS infants due to treatment-refractory mitral/tricuspid valve insufficiency therefore highlights the urgent need for novel therapeutics. Based on our ongoing multi-site clinical study in the United States analyzing peripheral blood from neoMFS newborns (n=11) and classical MFS adults (n=56), we found that the severity of mitral insufficiency is positively associated with elevated levels of circulating monocytes. In a mouse model of neoMFS (Fbn1C1041G/C1041G) that recapitulates clinical features including post-natal lethality (by post-natal day 7) and severe mitral/tricuspid valve degeneration, we found that diseased mitral/tricuspid valves exhibited significantly more CD45+CD206- hematopoietic cells, primarily due to the recruitment of CCR2+ monocytes. Diseased valves also exhibit a pro-inflammatory micro-environment comprised of monocytic chemoattractants (e.g., CCL2), activators of innate immunity (e.g., TLR4), metalloproteinases (e.g., MMP12), and immunogenic extracellular matrix components (e.g., Versican). Aortic/pulmonary valves in neoMFS mice remain morphologically normal without signs of degeneration or inflammation. Remarkably, pharmacologic blockade of CCR2 using a small molecule inhibitor (RS504393) injected daily for one week rescues mitral/tricuspid valve phenotype in neoMFS mice, therefore identifying inflammation as a potential therapeutic target in the treatment of life-threatening valve disease in neoMFS.