High-resolution atlas of the developing human heart and the great vessels

High-resolution atlas of the developing human heart and the great vessels | bioRxiv /* */ /* */ <!-- <!-- /*! * yepnope1.5.4 * (c) WTFPL, GPLv2 */ (function(a,b,c){function d(a){return"[object Function]"==o.call(a)}function e(a){return"string"==typeof a}function f(){}function g(a){return!a||"loaded"==a||"complete"==a||"uninitialized"==a}function h(){var a=p.shift();q=1,a?a.t?m(function(){("c"==a.t?B.injectCss:B.injectJs)(a.s,0,a.a,a.x,a.e,1)},0):(a(),h()):q=0}function i(a,c,d,e,f,i,j){function k(b){if(!o&&g(l.readyState)&&(u.r=o=1,!q&&h(),l.onload=l.onreadystatechange=null,b)){"img"!=a&&m(function(){t.removeChild(l)},50);for(var d in y[c])y[c].hasOwnProperty(d)&&y[c][d].onload()}}var j=j||B.errorTimeout,l=b.createElement(a),o=0,r=0,u={t:d,s:c,e:f,a:i,x:j};1===y[c]&&(r=1,y[c]=[]),"object"==a?l.data=c:(l.src=c,l.type=a),l.width=l.height="0",l.onerror=l.onload=l.onreadystatechange=function(){k.call(this,r)},p.splice(e,0,u),"img"!=a&&(r||2===y[c]?(t.insertBefore(l,s?null:n),m(k,j)):y[c].push(l))}function j(a,b,c,d,f){return q=0,b=b||"j",e(a)?i("c"==b?v:u,a,b,this.i++,c,d,f):(p.splice(this.i++,0,a),1==p.length&&h()),this}function k(){var a=B;return a.loader={load:j,i:0},a}var l=b.documentElement,m=a.setTimeout,n=b.getElementsByTagName("script")[0],o={}.toString,p=[],q=0,r="MozAppearance"in l.style,s=r&&!!b.createRange().compareNode,t=s?l:n.parentNode,l=a.opera&&"[object Opera]"==o.call(a.opera),l=!!b.attachEvent&&!l,u=r?"object":l?"script":"img",v=l?"script":u,w=Array.isArray||function(a){return"[object Array]"==o.call(a)},x=[],y={},z={timeout:function(a,b){return b.length&&(a.timeout=b[0]),a}},A,B;B=function(a){function b(a){var a=a.split("!"),b=x.length,c=a.pop(),d=a.length,c={url:c,origUrl:c,prefixes:a},e,f,g;for(f=0;f<d;f++)g=a[f].split("="),(e=z[g.shift()])&&(c=e(c,g));for(f=0;f<b;f++)c=x[f](c);return c}function g(a,e,f,g,h){var i=b(a),j=i.autoCallback;i.url.split(".").pop().split("?").shift(),i.bypass||(e&&(e=d(e)?e:e[a]||e[g]||e[a.split("/").pop().split("?")[0]]),i.instead?i.instead(a,e,f,g,h):(y[i.url]?i.noexec=!0:y[i.url]=1,f.load(i.url,i.forceCSS||!i.forceJS&&"css"==i.url.split(".").pop().split("?").shift()?"c":c,i.noexec,i.attrs,i.timeout),(d(e)||d(j))&&f.load(function(){k(),e&&e(i.origUrl,h,g),j&&j(i.origUrl,h,g),y[i.url]=2})))}function h(a,b){function c(a,c){if(a){if(e(a))c||(j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}),g(a,j,b,0,h);else if(Object(a)===a)for(n in m=function(){var b=0,c;for(c in a)a.hasOwnProperty(c)&&b++;return b}(),a)a.hasOwnProperty(n)&&(!c&&!--m&&(d(j)?j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}:j[n]=function(a){return function(){var b=[].slice.call(arguments);a&&a.apply(this,b),l()}}(k[n])),g(a[n],j,b,n,h))}else!c&&l()}var h=!!a.test,i=a.load||a.both,j=a.callback||f,k=j,l=a.complete||f,m,n;c(h?a.yep:a.nope,!!i),i&&c(i)}var i,j,l=this.yepnope.loader;if(e(a))g(a,0,l,0);else if(w(a))for(i=0;i (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0];var j=d.createElement(s);var dl=l!='dataLayer'?'&l='+l:'';j.src='//www.googletagmanager.com/gtm.js?id='+i+dl;j.type='text/javascript';j.async=true;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-M677548'); Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search New Results High-resolution atlas of the developing human heart and the great vessels View ORCID Profile Semih Bayraktar , View ORCID Profile James Cranley , View ORCID Profile Kazumasa Kanemaru , View ORCID Profile Vincent Knight-Schrijver , Maria Colzani , Hongorzul Davaapil , Jonathan Chuo Min Lee , Krzysztof Polanski , Laura Richardson , Claudia I. Semprich , Rakeshlal Kapuge , Monika Dabrowska , Ilaria Mulas , Shani Perera , Minal Patel , Siew Yen Ho , Xiaoling He , Richard Tyser , Laure Gambardella , Sarah A. Teichmann , Sanjay Sinha doi: https://doi.org/10.1101/2024.04.27.591127 Semih Bayraktar 1 Wellcome-MRC Cambridge Stem Cell Institute, Jeffrey Cheah Biomedical Centre, Cambridge Biomedical Campus, University of Cambridge , Cambridge, UK 2 Department of Medicine, University of Cambridge , Cambridge, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Semih Bayraktar James Cranley 3 Cellular Genetics Programme, Wellcome Sanger Institute, Wellcome Genome Campus , Hinxton, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for James Cranley Kazumasa Kanemaru 3 Cellular Genetics Programme, Wellcome Sanger Institute, Wellcome Genome Campus , Hinxton, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Kazumasa Kanemaru Vincent Knight-Schrijver 1 Wellcome-MRC Cambridge Stem Cell Institute, Jeffrey Cheah Biomedical Centre, Cambridge Biomedical Campus, University of Cambridge , Cambridge, UK 2 Department of Medicine, University of Cambridge , Cambridge, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Vincent Knight-Schrijver Maria Colzani 1 Wellcome-MRC Cambridge Stem Cell Institute, Jeffrey Cheah Biomedical Centre, Cambridge Biomedical Campus, University of Cambridge , Cambridge, UK 2 Department of Medicine, University of Cambridge , Cambridge, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site Hongorzul Davaapil 1 Wellcome-MRC Cambridge Stem Cell Institute, Jeffrey Cheah Biomedical Centre, Cambridge Biomedical Campus, University of Cambridge , Cambridge, UK 2 Department of Medicine, University of Cambridge , Cambridge, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site Jonathan Chuo Min Lee 1 Wellcome-MRC Cambridge Stem Cell Institute, Jeffrey Cheah Biomedical Centre, Cambridge Biomedical Campus, University of Cambridge , Cambridge, UK 2 Department of Medicine, University of Cambridge , Cambridge, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site Krzysztof Polanski 3 Cellular Genetics Programme, Wellcome Sanger Institute, Wellcome Genome Campus , Hinxton, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site Laura Richardson 3 Cellular Genetics Programme, Wellcome Sanger Institute, Wellcome Genome Campus , Hinxton, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site Claudia I. Semprich 3 Cellular Genetics Programme, Wellcome Sanger Institute, Wellcome Genome Campus , Hinxton, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site Rakeshlal Kapuge 3 Cellular Genetics Programme, Wellcome Sanger Institute, Wellcome Genome Campus , Hinxton, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site Monika Dabrowska 3 Cellular Genetics Programme, Wellcome Sanger Institute, Wellcome Genome Campus , Hinxton, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site Ilaria Mulas 3 Cellular Genetics Programme, Wellcome Sanger Institute, Wellcome Genome Campus , Hinxton, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site Shani Perera 3 Cellular Genetics Programme, Wellcome Sanger Institute, Wellcome Genome Campus , Hinxton, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site Minal Patel 3 Cellular Genetics Programme, Wellcome Sanger Institute, Wellcome Genome Campus , Hinxton, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site Siew Yen Ho 4 Cardiac Morphology Unit, Royal Brompton Hospital and Imperial College London , UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site Xiaoling He 5 John van Geest Centre for Brain Repair, Cambridge University , Cambridge, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site Richard Tyser 1 Wellcome-MRC Cambridge Stem Cell Institute, Jeffrey Cheah Biomedical Centre, Cambridge Biomedical Campus, University of Cambridge , Cambridge, UK 6 Department of Physiology, Development and Neuroscience, University of Cambridge , Cambridge, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site Laure Gambardella 1 Wellcome-MRC Cambridge Stem Cell Institute, Jeffrey Cheah Biomedical Centre, Cambridge Biomedical Campus, University of Cambridge , Cambridge, UK 2 Department of Medicine, University of Cambridge , Cambridge, UK 3 Cellular Genetics Programme, Wellcome Sanger Institute, Wellcome Genome Campus , Hinxton, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site Sarah A. Teichmann 3 Cellular Genetics Programme, Wellcome Sanger Institute, Wellcome Genome Campus , Hinxton, UK 7 Department of Physics, Cavendish Laboratory, University of Cambridge , Cambridge, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site For correspondence: st9{at}sanger.ac.uk ss661{at}cam.ac.uk Sanjay Sinha 1 Wellcome-MRC Cambridge Stem Cell Institute, Jeffrey Cheah Biomedical Centre, Cambridge Biomedical Campus, University of Cambridge , Cambridge, UK 2 Department of Medicine, University of Cambridge , Cambridge, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site For correspondence: st9{at}sanger.ac.uk ss661{at}cam.ac.uk Abstract Full Text Info/History Metrics Supplementary material Preview PDF Abstract The human heart and adjoining great vessels consist of multiple cell types vital for life, yet many remain uncharacterised molecularly during development. Here, we performed a high-resolution profiling of the heart and great vessels during the first and second trimesters, defining 63 cell types with distinct identity and location-specific signatures. We reveal previously unreported cell types, including for the pericardium and the ductus arteriosus. At the ventricles, we identified signatures involved in establishing the trabeculation-compaction and right-left axes of ventricular cardiomyocytes. At the vessels, we distinguished the constituents belonging to either coronary or great vessels. We confirmed our findings and revealed nuanced signatures with specific zonation patterns. Collectively, we provide a comprehensive human cardiac developmental atlas for enhanced understanding of function in health and disease. Introduction The heart is the first organ to function during embryonic development. Successful cardiogenesis requires the differentiation and coordination of multiple different cell types. A deeper understanding of the transcriptomic and epigenomic landscape of the unique cellular constituents of the developing human heart will help us understand how these cells coordinate during development and how this can go awry to cause disease, congenital or adult-onset cardiac complications. To fully understand the cellular and molecular mechanisms at play during this cardiogenesis, we must first define the repertoire of cell types in the developing heart. With the advent of single-cell transcriptomic and epigenomic technologies, alongside spatial transcriptomics, many researchers have leveraged the mapping of cellular signatures to better understand organs, including the developing human heart 1 – 8 . While the studies offered valuable insights into cardiac development, a clear resolution of cellular identities remains unaddressed due to factors including a limited number of samples, inadequate coverage of developmental stages (focusing on either the 1st or 2nd trimester), or a lack of spatial localization. In this paper, and our accompanying paper 9 , we aimed to address these limitations by providing a high-resolution spatio-temporal and multi-omic atlas of the developing human heart and great vessels. To achieve this, we performed single-cell RNA-sequencing using 21 hearts across the first and second trimesters 9 . In this paper, we establish the diverse array of cell types that make up the first and second trimester human heart and shed light into the previously unreported cell types, including the pericardium or the cells of the ductus arteriosus. We identify distinct signatures that constitute the left-right, and compact-trabeculated axes of cardiomyocytes. Within the conduction system, we resolved the sinoatrial node and atrioventricular node pacemaker cells, alongside proximal and distal cardiac conduction system cells. Through targeted dissection and sequencing of samples, we differentiated the signatures between similar cellular constituents across coronary and great vessels and identified enhanced Notch signalling directed towards coronary smooth muscle cells compared to their great vessel counterparts. Our analysis indicates that cells of the great vessels collectively are enhanced in synaptic assembly. We infer increased prostaglandin signalling, alongside reduced endothelin sensing, at ductus arteriosus smooth muscle cells to maintain the patency of ductus arteriosus during development. Finally, we resolve the distinct signatures that surround the myocardium, including the cells of the pericardium. Results Atlas of the developing human heart and great vessels To investigate cellular composition, heterogeneity and interactions in the developing human heart, we collected a total of 21 human fetal hearts between the ages of 4 and 20 post-conception weeks (PCW) (12 female, 9 male, Supplementary Fig. 1A, B ) that underwent single-cell RNA or single-nucleus RNA sequencing ( Fig. 1A ). At the time points studied, the hearts were small enough to sequence the whole organ, which enabled us to capture the full breadth of cellular heterogeneity, ensure that we included rare populations, and explore cell types at different stages of maturity. To gain regional insight, some of these hearts were dissected into several distinct anatomical regions, such as node, great vessels, ductus arteriosus, or pericardium, prior to sequencing ( Fig. 1A ). Download figure Open in new tab Supplementary Fig. 1: Samples overview a,b - Sex inference of the samples. XIST counts and the percentage of Y chromosome counts for single-cell (a) and single-nuclei (b) processed samples. c- UMAP embedding of the trimester, sex, week, region, mid- and coarse-grain information of the cells. d- Matrix plot depicting a global view of the fine-grain level cell type similarity based on the expression of highly variable genes. Download figure Open in new tab Figure 1: Atlas of the developing human heart and the great vessels a- Overview of study design and data modalities. Single-cell data were generated from a total of 21 human fetal hearts between the ages of 4 and 20 PCW. Targeted dissection was performed on some of the samples indicated on the upper row. b- UMAP embeddings of gene expression of 63 cell types of the developing human heart and the great vessels, spread across 14 mid- and 6 coarse-grains. Fine-grained cell type annotations are provided adjacent to the UMAP embeddings, grouped under their respective mid-grains. c- Changes in the composition of the heart and the great vessels across 6 and 13 PCW, depicted as percentages, with a focus on cardiomyocytes (in colour). To reduce sampling bias, only the samples that were not subjected to targeted dissection were used. For visual clarity, lines were fitted using a smoothing spline (number of knots = 5). After quality control, 297,473 cells and nuclei were retained for further analysis. Cell type annotation was performed in gene expression space, with batch correction using scVI 10 , 11 . A coarse-grained annotation identified 6 major cell types representing cardiomyocytes, endothelium, epicardium, mesenchymal cells, leukocytes and neural cells. Iterative rounds of annotation revealed 14 mid-grained and 63 fine-grain cell types. Wherever possible cells were annotated to match a specific function or anatomical location, such as the coronary and great vessel constituents ( Fig. 1B ). No adipocytes were detected in our samples, in agreement with previous work showing they are first observed in the third trimester of development and later than we sampled 12 . Using the temporal power of the atlas, we also assessed the changing cellular composition within the whole heart nuclei RNA-seq samples. Our main observation was a decrease in the relative abundance of non-myocyte populations coupled with a rise in the fraction of cardiomyocytes, especially those of the compact myocardium ( Fig. 1C ). This suggests that the late first trimester stage of development is important for the expansion of compact myocardium. Cardiomyocyte diversity in development Subclustering of cardiomyocytes yielded a total of 12 fine-grain cell types, grouped into three mid-grained cell types based on the anatomy and the function; atrial and ventricular cardiomyocytes (aCM and vCM), and cardiac conduction system (CCS) cells ( Fig. 2A ). Using data from Cranley et al. 9 these profiles were validated through cell2location mapping to their corresponding positions using spatial transcriptomics, which highlighted the distinct localisation of atrial and vCMs (left, right, compact and trabeculated) ( Fig. 2B ). Concurrently, annotations across CCS were validated by their distinct spatial mapping ( Fig. 2C ), as well as through immunohistochemistry ( Fig. 2D ). TNNT2 appeared as the most sensitive cardiomyocyte marker, regardless of the subtype of cardiomyocytes ( Supplementary Fig. 2 ). Download figure Open in new tab Supplementary Fig. 2: Marker expression across the atlas Dot plot displaying the differentially expressed markers associated with each coarse-grain level cell label, alongside the expression profiles of markers (right) referenced across this article. Download figure Open in new tab Figure 2: Cardiomyocyte diversity a- UMAP embedding of cardiomyocyte gene expression displaying fine-grain cell type labels. b- cell2location spatial mapping of working cardiomyocytes on 16 PCW heart spatial transcriptomics section. c- cell2location spatial mapping of SAN and AVN pacemaker cells, alongside proximal VCS cells, on a 7 PCW heart spatial transcriptomics section. d- Immunofluorescence validation of the CCS. TBX3 expression is observed at SAN pacemaker cells, AVN pacemaker cells and proximal CCS, while SHOX2 expression is observed at SAN pacemaker cells only. e- Dot plot depicting the markers differentially expressed between the right and left atrial cardiomyocytes. f- Dot plot depicting the markers differentially expressed between the cardiomyocytes of the right/left and compact/trabeculated myocardium. g- Dot plot depicting the markers differentially expressed between the fine-grain cell labels of the CCS cells. h- Proportion of cycling cardiomyocytes across mid-grains, indicating a reduction in cycling cardiomyocyte abundance through development. PITX2, a regulator of left-right asymmetry associated with the left atrium 13 , was a specific marker of left aCMs ( Fig. 2E ). Interestingly PANCR, a PITX2 -adjacent long non-coding RNA which has been shown to positively regulate PITX2 expression 14 , was highly specific to left aCMs as well, underlining the critical role of PITX2 in left aCM cell identity. Alongside the previously reported BMP10 , the right aCMs expressed NTM and ANGPT1 ( Fig. 2E ). Right atrial-derived angiopoietin has been shown critical for coronary vein formation in mice 15 , which may suggest a similar function in humans. Left and right vCMs were transcriptionally identified; SLC1A3 and CDH13 were enriched in left vCMs, whilst ANKRD1 , MYL2 and GPNMB were enriched in right vCMs ( Fig. 2F ). PRRX1 emerged as the most distinguishing feature of the cardiomyocytes of the right ventricle compared to left. Across the ventricles, we could separate the cardiomyocytes of the compact and trabeculated myocardium, where SGCD , SPOCK1 and SLIT3 were associated with trabeculated cardiomyocytes for both the ventricles ( Fig. 2F ). CCS did not have any marker that could capture all its constituents in a specific manner; CNTN2 , NPTN and NTM, used to mark the entire mouse CCS 16 , did not exhibit the same specificity in the developing human heart ( Supplementary Fig. 2 ). Within the CCS, MYH6 , CACNA1D and NR2F2 marked the sinoatrial node (SAN) and atrioventricular node (AVN) pacemaker cells, whereas MYH7 , NAV1 and LPL marked the ventricular conduction system (VCS) cells. SHOX2 , PDE1A and TENM4 were specific for SAN pacemaker cells, whereas GREB1L , NRXN3 and RSPO were specific to AVN pacemaker cells ( Fig. 2G ). Interestingly, MYH11 , one of the most specific markers of smooth muscle cells, was expressed at the SAN pacemaker cells, in line with previous reports 17 , 18 . Furthermore, NIBAN1 and PHACTR1 were expressed by proximal VCS and, OPCML , NTN1 and IRX2 marked distal VCS cells ( Fig. 2G ). Overall, there was a gradual decrease in the proportion of cycling cells from earliest to latest timepoints ( Fig. 2H ). Coronary vessels and the great vessels Both mural cells and endothelial cells (ECs) within the coronary and great vessels are known to have multiple embryological origins 19 , 20 . These distinct origins along with specific local cues give rise to a wide range of differing vascular cell states that are poorly characterised in human development yet have important roles in vascular development and function. To investigate this diversity in more detail, we examined the endothelium, fibroblasts and smooth muscle cells (SMCs) that were collected at apex and base dissociated samples, which enabled us to identify these cell types based on their distribution. For example, through our targeted dissection strategy ( Fig. 1A ), we could identify coronary SMCs (CSMCs) from their localisation to both apex and base, whereas great vessel SMCs were observed only at the base ( Fig. 3A ). Download figure Open in new tab Figure 3: Coronary and great vessel constituents a- Identification of CSMCs through targeted sampling of fetal hearts. Cells sourced from apex and base dissociated samples only were highlighted in the left panel. b- Validation of distinct expression of PLA2G5 and FOXC1 at CSMC and great vessel SMCs. Scale bars, 50 µM. c, d- Construction of endothelial continuum of coronary circulation based on vascular anatomy (c), with the expression profile of fine-grain in blood vessel endothelial cells along this continuum. e- Split violin plots for the expression of coronary (blue) vs great vessel (orange) enriched genes in vessel constituents. f- The interactions between ECs and SMCs of the coronary and great vessels at the intima-media interface, despite sharing similar cell types, exhibit distinct interactions. g-Differentially expressed markers of DASMCs in comparison to the other mural cells in the atlas. h- Temporal expression pattern of PDE1C across mural cells. i- Spatial mapping of great vessel constituents; mapping absent from myocardium, highlighting the high-resolution nature of annotations and the dataset. j- GO analysis of great vessel constituents have shared terms relating to neural processes. Scaled F-scores for the terms are plotted for all coronary and great vessel constituents. We observed transcriptional differences in these SMC types and validated our finding with immunohistochemistry. For example, coronary and great vessel SMCs expressed PLA2G5 and FOXC1, respectively, in a mutually exclusive manner ( Fig. 3B ). Compared within coronary mural cells, coronary pericytes expressed THBS4 , KCNJ8 and CYGB , whereas CSMCs expressed ELN , TPM1 and ACTA2 ( Supplementary Fig. 4A ). Our most mature sample (20 PCW) allowed us to confidently dissect the aorta and pulmonary artery during sampling and sequence separately. We observed a pulmonary specification within pulmonary artery SMCs with enriched expression of INPP4B , PXDNL and ALDH1A2 ( Supplementary Fig. 4B ). Download figure Open in new tab Supplementary Fig. 3: Coronary and great vessel constituents a- Volcano plot of the differentially expressed genes between coronary pericytes and CSMCs. b- Pulmonary artery specification of great vessel ECs, SMCs and adventitial fibroblast, through comparing the cells obtained from the targeted dissection of the aorta and pulmonary artery. c, d- Embedding of coronary ECs, presented with fine-grain label, trajectory, and cell cycling information. Lower right panel indicates the expression of cell cycling modules presented in (d) across the embedding, with venous but not arterial cells in proximity to cycling modules. e- Tip and stalk investigation of coronary capillary ECs reveals the tip identity peaks around 6 PCW and gradually decreases. f- The interactions between SMCs and adventitial fibroblasts of the coronary and great vessels at the media-adventitia interface. g- Expression profile of the ECs obtained from the targeted dissection of ductus arteriosus, compared to the ECs obtained from the targeted dissection of the aorta and pulmonary artery of the same donor. h- Mapping of vessel constituents across spatial transcriptomics sections underlines the in-depth annotation of the atlas. First row, great vessel constituents. Second row, coronary vessel constituents. i- Genes enriched in and shared across the great vessel constituents for great vessel and coronary vessel comparison, alongside the expression profile of various extracellular matrix proteins. Download figure Open in new tab Supplementary Fig. 4: Myocardial fibroblasts a, b- Fibroblasts of the heart (a), with distinct transmural localisation patterns (b). c- Dot plot displaying the differentially expressed genes associated with heart fibroblasts. Download figure Open in new tab Supplementary Fig. 5: Neural cells a- Cells of the neural compartment. b- Dot plot displaying the differentially expressed genes associated with neurons. c- cell2location mapping of sympathetic and parasympathetic neurons on Visium slides localises these cells in proximity to nodal pacemaker cardiomyocytes. d- Dot plot displaying the differentially expressed genes associated with glia. Download figure Open in new tab Supplementary Fig. 6: Leukocytes and lymphatic cells a- Fine-grain cell labels of leukocytes. b- Dot plot depicting the differentially expressed genes of all the fine-grain cell types of the myeloid mid-grain. c- Dot plot depicting the differentially expressed genes of all the fine-grain cell types of the lymphoid mid-grain. d- cell2location mapping of lymph node fibroblastic reticular cells and lymphatic ECs on Visium slides localises these cells on a cardiac lymph node, alongside other leukocytes. e- Dot plot depicting the differentially expressed genes of lymph node fibroblastic reticular cells and lymphatic ECs. Using the same framework, we distinguished the endothelial and fibroblast constituents of the coronary and great vessels. We identified great vessel and coronary ECs, with arterial, venous and capillary specifications, which allowed us to construct a continuum of the ECs present within the coronary circulation ( Fig. 3C ). For example, HMCN1 and PLXNA4 were expressed by the ECs of the great vessels, while coronary vessel ECs expressed A2M and ADGRL4 ( Fig. 3D ). It has been suggested that NR2F2 signalling blocks pre-arterial specification and activates cell cycling genes 21 , 22 . We observed that cycling cells of the developing coronary endothelium have a closer transcriptomic identity to the venous and capillary coronary ECs ( Supplementary Fig. 4C,D ). Sprouting angiogenesis is achieved by endothelial tip cells, and within the capillary coronary endothelium, we observed the tip identity peaked around 6 PCW and gradually decreased through our timepoints ( Supplementary Fig. 4E ). In addition to SMCs and ECs, we detected differences in great and coronary vessel adventitial fibroblasts, where great vessel adventitial fibroblasts expressed ROR1 and the mechanosensing ion channel PIEZO2 , where coronary vessel adventitial fibroblasts expressed TBX20 and TCF21 ( Fig. 3E ). Investigating cell to cell communication mechanisms at both the vessel types, we observed a pronounced Notch signalling towards coronary vessel SMCs, mediated by JAG1 , JAG2 and DLL4 at intima-media interface ( Fig. 3F ), and by DLK1 at media-adventitia interface ( Supplementary Fig. 4F ). Closure of the ductus arteriosus is essential immediately after birth to ensure proper cardio-pulmonary functioning, as a patent ductus arteriosus (PDA) may lead to respiratory distress and heart failure. We identified the SMCs of the ductus arteriosus (DASMCs) with a distinct expression profile of DCLK2 , PDE1C and TAFA1 compared to other mural cells in the dataset ( Fig. 3G ). The expression of PDE1C increases sharply around 7 weeks into development ( Fig. 3H ) and given it encodes for a phosphodiesterase, this might be implicated in the vasoconstriction of the ductus arteriosus. Interestingly, we observed a specific expression of prostaglandin receptor PTGER4 and muscarinic M2 receptor CHRM2 in DASMCs, with a reduced endothelin receptor expression ( EDNRA and EDNRB ; Fig. 3G ) when compared to other mural cells in the dataset. PTGER4 has been shown to play a role in ductus arteriosus closure in mice 23 . Investigating the ECs obtained from the ductus arteriosus sampling, these cells had an enriched expression of PDE4D , alongside genes implicated in prostaglandin synthesis ( Supplementary Fig. 4G ). Collectively, these suggest a mechanism whereby the duct is primed for closure, yet is kept open due to active prostaglandin signalling and reduced endothelin sensing during development. Our analysis regarding distinct vessel constituents was supported by spatial transcriptomics 9 , where we observed a clear separation in the mapping of relevant cell types across spatial slides ( Fig. 3I , Supplementary Fig. 4H ). We were interested in identifying vessel-wide biological functions shared across the constituents of either vessel type. Overall, ECs, mural cells and fibroblasts of the great vessels had shared ontology terms relating to neurogenesis , sensory organ development or axon guidance , and all the components expressed NLGN1 , with a potential role in synaptic assembly and guiding the formation of the cardiac plexus that sits on the great vessels ( Fig. 3J , Supplementary Fig. 4I ). Pericardium and epicardium The pericardial and epicardial layers remain relatively poorly studied in terms of cellular composition and function despite their essential role in mediating cardiac development and function. We included a targeted dissection of the pericardium from one of our donated hearts (Hst41, Fig. 1A ) allowing us to define cell clusters from this donor, as well as others, belonging to both fibrous and serous layers of the pericardial structure, as well as the parietal and visceral layers of the folded serous pericardium ( Fig. 4A ). Download figure Open in new tab Figure 4: Pericardium and endocardium a- UMAP embedding of epicardial and pericardial fine-grain labels based on gene expression. b- Dot plot displaying the differentially expressed genes associated with epicardial and pericardial cell types. c- Immunohistochemical analysis validates expression of the TSHZ2 at the parietal layer of the serous pericardium. d- cell2location spatial mapping of mesothelial epicardial cells and epicardium derived cells on a 16 PCW spatial transcriptomics section. e- Schematic illustrating the cell types comprising pericardium and the marker genes associated with them. f- Diffusion map embedding of the endocardial and valve cell types. g- Dot plot displaying the differentially expressed genes associated with endocardial and valve cell types. h- cell2location spatial mapping of endocardial and valve cells on a 16 PCW spatial transcriptomics section. i- The cell-cell interactions between the epicardium and the cardiomyocytes of the compact myocardium. j- The cell-cell interactions between the endocardium and the cardiomyocytes of the trabeculated myocardium. Firstly, we found the two previously described mesothelial epicardium or migratory epicardium-derived cell (EPDC) populations of epicardial cells that make up the inner visceral layer of the serous pericardium ( Fig. 4A, B ) 7 . Compared with the other layers of the pericardium, both epicardial populations selectively expressed ALDH1A2 and EZR . These cells also expressed the lubricant protein-coding gene PRG4 , and hyaluronic acid synthesis gene HAS1 ( Fig. 4B ), and interestingly, the interaction between lubricin and hyaluronic has been reported to synergistically enhance anti-adhesive properties 24 . Thus, epicardium-provided lubrication is implicated in reducing friction and minimising adhesions. Out of the common markers studied in animals and stem cell models, BNC1 was seen to specifically define both populations of epicardial cells and the most sensitive and specific marker of epicardial cells across the atlas ( Supplementary Fig. 2 ). WT1 was expressed by the cells of the serous pericardium, including epicardium and controversially, TCF21 expression was not relatively high in any of the epicardial or pericardial cells and was preferentially expressed in fibroblast clusters ( Supplementary Fig. 2 ). In addition to TBX18 , the mesothelial epicardium was selective for SBSPON , and ITLN1 , while the EPDCs were selective for CPB1 and VEGFC ( Fig. 4B ). Interestingly, with evidence for VEGFC in driving angiogenesis 25 , these cells potentially guide coronary vessel formation as they invade the developing myocardium to form coronary pericytes and SMCs. Secondly, we found three clusters of cells describing either the parietal layer of the visceral pericardium, the fibrous pericardium, or a mixed cluster of pericardial cells with both visceral and fibrous markers ( Fig. 4A ). All three of these clusters expressed COL6A3 , COL12A1 and COL16A1 ( Fig. 4B ). We found that the parietal layer of the serous pericardium distinctly expressed TSHZ2 , AUTS2 and SETBP1 , while the fibrous pericardium expressed the chondrocytic marker ITM2A and procollagen peptidase enhancer PCOLCE ( Fig. 4B ). Interestingly, a strong expression of exosomal pseudo markers CD9 , CD63 and CD81 was seen in fibrous and visceral serous pericardial cells (epicardium) ( Fig. 4B ), suggesting that these pericardial layers may contribute to the exosome rich pericardial fluid 26 . We spatially validated our annotations with immunohistochemistry using antibodies for TSHZ2 on a section of intact pericardium revealing TSHZ2 expression in non-myocardial tissue adjacent to the pericardial cavity, confirming these cells as parietal serous pericardium ( Fig. 4C ). Furthermore, cell2location mapping of the visceral serous pericardium onto spatial transcriptomics 9 showed an intramyocardial enrichment of EPDCs ( Fig. 4D ). Overall, these results reveal new markers of the pericardium suggesting functional differences between the pericardial cells and allowing future researchers to distinguish between pericardial layers ( Fig. 4E ). Endocardium and valves Within the endocardium and the valve lineage, we’ve identified cells of the endocardial cushion, valve endothelial and valve interstitial cells, alongside endocardial cells ( Fig. 4F ). Interestingly, all the cells within this lineage expressed POSTN ( Fig. 4G ). NPR3 expression was specific to the endocardium, where other endocardial markers PCDH7 and SMOC1 were shared with the endocardial cushion ( Fig. 4G ). Endocardial cushion cells on the other hand had a strong expression of BMP4 and HAS2 , and also expressed PROX1 , COL26A1 , TMEM132D and TSPAN8 ( Fig. 4G ). In fact, TMEM132D and TSPAN8 emerged as specific markers of endocardial cushion cells across the atlas ( Supplementary Fig. 2 ). Endothelial cells of the valves express APCDD1 along with endocardial cushion cells, however, had specific expression of DKK2 , KCNJ2 and SULF1 within the endothelial lineage ( Fig. 4G ). Valve interstitial cells on the other hand expressed PTN , ID4 and PRRX1 . These cells also expressed TGF-β signalling inhibitor BAMBI , together with endocardial cushion cells ( Fig. 4G ). Overall, COL9A2 emerged as the most specific marker of valve interstitial cells across the whole dataset ( Supplementary Fig. 2 ). We were interested in cell-cell communication between myocardial layers and the epicardial or endocardial membranes the layers are nestled between. ADGRG6 signalling, associated with the integrity of the compact wall and the identity of trabeculated cardiomyocytes 27 , originated from both myocardial layers towards the epicardium and the endocardium ( Fig. 4I, J ). Conversely, signals directed at PLXNA4 , expressed by both compact and trabeculated cardiomyocytes, originated from both the epicardium and the endocardium ( Fig. 4I, J ). Discussion In this study, we have provided a reference atlas of the foetal heart at unprecedented scale and resolution, both in terms of numbers of cells represented as well as the temporal spread of samples. We have combined single-cell transcriptomics with spatial transcriptomics 9 to improve our understanding of the molecular signatures that govern the human heart and great vessel development through the first and second trimesters. As the utilisation of single-cell genomics in research grows exponentially and we uncover more nuanced cellular signatures, there is a growing demand for clear cell annotations and for ensuring consistent descriptions of the same biological entities 28 . Bioinformatic analysis can divide datasets into increasing numbers of fine-grained clusters, which may not always reflect the underlying biology. In our atlas, the ability to discern subtle cellular signatures and spatially deconvolute the cell types that would traditionally be considered the same enabled us to annotate our cells primarily based on anatomical location or function, and with better clarity compared to previous attempts 1 – 8 . With this atlas, the ability now to separate similar cellular signatures from one another will allow researchers to better distinguish the molecular mechanisms governed by specific cell types in development and disease, such as DASMC-specific mechanisms in patent ductus arteriosus. Additionally, our results revealed previously undefined transcriptional states of known anatomical structures, including the cells of the pericardium. Collectively, this highly resolved cell type information could be used to construct fine-tuned cellular models. Our analysis revealed distinct characteristics of left and right identities of the chambers, where the atrial right-left signature emerged more readily distinguishable compared to the ventricular left-right signature. This observation is intriguing, given the atrial cardiomyocytes are thought to share the same precursor pool of cells, yet they show a clearer distinction compared left and right ventricular cardiomyocytes, which are believed to develop form different progenitors in first and second heart fields. PRRX1, which played a pivotal role in differentiating between the two ventricular subtypes, has been suggested to contribute to the right-left axis information during heart looping 29 , however its role in right ventricular cardiomyocyte identity has not been clear. Interestingly, PRRX1 did not surface as a marker of the right ventricular cardiomyocytes in the adult human heart 17 , implying a role for PRRX1 in cardiomyocytes in a specific developmental context. The atlas presented here also serves as a reference for differences between human and animal models. For example, several markers were used to mark the entire CCS in mice 16 , 30 . However, within the developing human heart, no marker was identified that could capture all the components of CCS in a sensitive and a specific manner. Another important difference was observed in the epicardium, with the absence of TCF21 in human epicardial cells. In zebrafish, tcf21 has been used as a lineage marker for epicardial cells in regenerative studies, which may have implications for translating zebrafish studies into human therapy 31 . However, the true scope and power of this atlas lie beyond this article. We introduce an opportunity for extended utilisation of our atlas, including the potential for comparative studies involving different animal models of heart development as well as benchmarking and improving current models of human tissue generation from pluripotent stem cells. Additionally, there exist opportunities to tackle the technical limitations encountered in our study, notably the application of spatial transcriptomics methods that lack single-cell resolution. Overall, the insights obtained from this atlas are anticipated to have broad applications in cardiovascular research and contribute to a deeper understanding of human heart development. Supplementary text: Other cells Myocardial fibroblasts Differentiating and characterising the specific subtypes of fibroblasts proved challenging due to their converging molecular signatures, despite their distinct anatomical locations or functions. To address this issue and improve the reliability of our annotations, we utilized spatial transcriptomics. Having identified great vessel and coronary vessel adventitial fibroblasts, we further characterised three distinct fibroblast clusters ( Supplementary Fig. 5A ) that although sharing molecular similarities, exhibited distinct localization patterns across the myocardium ( Supplementary Fig. 5B ). These clusters were labelled as subepicardial fibroblasts, myocardial interstitial fibroblasts and myofibroblasts, with myofibroblasts localizing to the innermost section of the myocardium. Subepicardial fibroblasts were enriched in SLIT3 , BNC2 , and BRINP3 , while myocardial interstitial fibroblasts were enriched in SCN7A , ROBO2 and CD34 ( Supplementary Fig. 5C ). A recent study on mice demonstrated that CD34 + fibroblasts are situated deeper within organs 1 , aligning with our observations and suggesting that CD34 expression in fibroblasts could help distinguish fibroblasts based on their zonation. Finally, myofibroblasts were enriched in IGFBP7 , APOE and MCAM , as well as notch signalling related proteins NOTCH3 and a downstream target HES4 ( Supplementary Fig. 5C ). Neural cells Fine-grain labels within neural cells were neuron progenitors, parasympathetic neurons, sympathetic neurons, and chromaffin cells for neuron mid-grain label, and Schwann cells and Schwann cell precursors for glia mid-grain label ( Supplementary Fig. 6A ). PHOX2B emerged as the most sensitive and specific marker for all the neural cells throughout the atlas ( Supplementary Fig. 2 ). Neurons were distinguished from glia by several markers, including SYT1 , PCSK1N and PHOX2A ( Supplementary Fig. 6B ). Neuron precursors showed enriched expression of ASCL1 and BNC2 ( Supplementary Fig. 6B ). Sympathetic and parasympathetic neurons shared the expression of GATA3 , ISL1 and RTN1 ( Supplementary Fig. 6B ). Sympathetic neurons displayed an enriched expression of DBH and VSTM2L , while parasympathetic neurons exhibited an enriched expression of SV2C and SLC5A7 ( Supplementary Fig. 6B ). Both neuron types innervated around the nodal pacemaker cells ( Supplementary Fig. 6C ). Chromaffin cells displayed a distinct expression profile, characterised by the enriched expression of DLK1 , EPAS1 , KCNJ6 ( Supplementary Fig. 6B ) Several markers, including PTPRZ1 , CDH19 and ERBB3 , distinguished glia from neurons ( Supplementary Fig. 6D ). Schwann cell precursors displayed enrichment in SLITRK6 , IL1RAPL2 and COL20A1 , while Schwann cells exhibited an enriched expression of LAMB1, COL14A1 and TGFBR3 ( Supplementary Fig. 6D ). Leukocytes Leukocytes encompass a diverse array of subtypes, some of which can exhibit subtle signatures that make them challenging to distinguish. We identified 20 cell types comprising the leukocytes, which span the myeloid and lymphoid lineages (Supplementary Fig. 7A). Cardiac macrophages can exhibit a variety of functions, from remodelling to angiogenesis 2 , 3 , and distinguishing the subtypes of macrophages can have regenerative implications 4 . We identified 4 subtypes of macrophages, that had distinct expression profiles, and all also shared the expression of LYVE1 + macrophage-enriched markers. In particular, CX3CR1 + macrophages expressed C3 , CX3CR1 , HTRA1 , BHLHE41 and OLFML3 (Supplementary Fig. 7B). TIMD4 + macrophages expressed KCNMA1 , IL18 , ANTXR1 , ESRRG , SAMD4A , MMP9 and MYO5B (Supplementary Fig. 7B). LYVE1 + macrophages had an enrichment in F13A1 , LYVE1 , SELENOP and FOLR2 , while ATF3 + macrophages showed enrichment in ATF3 , NR4A3 and CXCL8 (Supplementary Fig. 7B). We identified leukocyte-associated cells at the lymphatic vessels and the lymph node, which were lymphatic endothelial cells and lymph node fibroblastic reticular cells. Upon mapping these cells on a Visium slide, we observed their localization within a cardiac lymph node present in the section (Supplementary Fig. 7D). Lymphatic endothelial cells expressed PROX1 , MMP7 , STAB2 , PARDH6 , DOCK5 , LYVE1 and STON2 , while shared the expression CCL21 with lymph node fibroblastic reticular node cells (Supplementary Fig. 7E). Lymph node fibroblastic reticular node cells also expressed SLC22A3 , OCA2 , XKR4 , GRIN2B and CCL19 (Supplementary Fig. 7E). Methods Ethics The embryonic and foetal heart samples corresponding to the donor IDs BRC2251, BRC2253, BRC2256, BRC2260, BRC2262, BRC2263, C82, C83, C85, C86, C87, C92, C94, C97, C98, C99, and C104 were provided from terminations of pregnancy from Cambridge University Hospitals NHS Foundation Trust under permission from NHS Research Ethical Committee (96/085). The donor IDs Hst33, Hst36, Hst39, Hst40, Hst41, Hst42, Hst44, and Hst45 were provided from the MRC/Wellcome Trust Human Developmental Biology Resource (University College London (UCL) site REC reference: 18/LO/0822; www.hdbr.org ). Sample names and that information are listed in Supplementary Table 1. Sample collection and processing The processing of samples with donor IDs of BRC2251, BRC2253, BRC2256, BRC2260, BRC2262 and BRC2263 were described previously 7 . Briefly, these samples were stored at 4 °C overnight in Hibernate-E medium (ThermoFisher Scientific). The next day, the apex and the base of the heart were dissected and separately dissociated using 6.6 mg/mL Bacillus Licheniformis protease (Merck), 5 mM CaCl2 (Merck), and 20 U/mL DNase I (NEB). The mixture was triturated on ice for 20 seconds every 5 minutes until clumps of tissue were no longer visible. The digestion was stopped with ice-cold 10% fetal bovine serum (FBS, ThermoFisher Scientific) in phosphate-buffered saline (PBS, ThermoFisher Scientific). Cells were then washed with 10% FBS, resuspended in 1 mL PBS and viability assessed using Trypan blue. Cells were submitted for 10x library preparation with v3.0 chemistry for 3’ single-cell sequencing on a NovaSeq 6000 (Illumina) at the Cancer Research UK Cambridge Institute. For the processing of donors corresponding to the donor IDs C86, C94, C97 and C99, samples were stored at 4 °C overnight in HyperThermasol preservation solution (Merck). Tissue was first minced in a tissue culture dish using scalpel. Minced tissue was digested with type IV collagenase (final concentration of 3 mg/mL; Worthington) in RPMI (Sigma-Aldrich) supplemented with 10% fetal bovine serum (FBS; Gibco), at 37°C for 30 min with intermittent agitation. Digested tissue was then passed through a 100-µm cell strainer, and cells were pelleted by centrifugation at 500g for 5 min at 4°C. Cells were then resuspended in 5 ml of red blood cell lysis buffer (eBioscience) and left for 5 min at room temperature. It was then topped up with a flow buffer (PBS containing 2% (v/v) FBS and 2 mM EDTA) to 45 ml and pelleted by centrifugation at 500g for 5 min at 4°C. The resuspended cell solution was filtered through a 70-μm cell strainer (Corning), and live cells were manually counted by Trypan blue exclusion. Dissociated cells were first incubated with 5uL of FcR blocker for 5 min at room temperature and stained with anti-CD45 antibody (BV785 anti-human CD45 antibody, BioLegend, 304048) and DAPI (Sigma-Aldrich, D9542) prior to sorting. DAPI was used at a final concentration of 2.8 µM, and all antibody solutions were used at a final concentration of 5 µl per 100 µl cell suspensions containing fewer than 5 million cells. DAPI-CD45+ and DAPI-CD45-populations were sorted by FACS using MA900 Multi-Application Cell Sorter (Sony) and its proprietary software (Cell Sorter Software v3.1.1). Sorted cells were loaded on the Chromium Controller (10x Genomics) with a targeted cell recovery of 5,000–10,000 per reaction. Single-cell cDNA synthesis, amplification, gene expression library was generated according to the manufacturer’s instructions of the Chromium Next GEM Single Cell 5’ Kit v2 (10x Genomics). Libraries were sequenced using NovaSeq 6000 (Illumina) at Wellcome Sanger Institute with a minimum depth of 20,000–30,000 read pairs per cell. Samples used for single nuclei isolation were flash-frozen (unembedded) or frozen in OCT and stored at −80 °C, or formalin-fixed and subsequently embedded in paraffin blocks. All tissues were stored and transported on ice at all times until freezing or tissue dissociation to minimise any transcriptional degradation. Single nuclei were obtained from flash-frozen tissues using sectioning and mechanical homogenization as previously described. 5-10 mm thickness frozen tissues were first sectioned with cryostat in a 50 μm thickness section. All sections from each sample were homogenised using a 7 ml glass Dounce tissue grinder set (Merck) with 8–10 strokes of a tight pestle (B) in homogenization buffer (250 mM sucrose, 25 mM KCl, 5 mM MgCl2, 10 mM Tris-HCl, 1 mM dithiothreitol (DTT), 1× protease inhibitor, 0.4 U μl−1 RNaseIn, 0.2 U μl−1 SUPERaseIn, 0.05% Triton X-100 in nuclease-free water). Homogenate was filtered through a 40-μm cell strainer (Corning). After centrifugation (500g, 5 min, 4 °C) the supernatant was removed and the pellet was resuspended in storage buffer (1× PBS, 4% bovine serum albumin (BSA), 0.2U μl−1 Protector RNaseIn). Nuclei were stained with 7-AAD Viability Staining Solution (BioLegend) and filtered through 20-μm cell strainer (CellTrics). Positive single nuclei were purified by fluorescent activated cell sorting (FACS) using MA900 Multi-Application Cell Sorter (Sony) and its proprietary software (Cell Sorter Software v3.1.1). Nuclei purification and integrity were verified under a microscope, and nuclei were manually counted by Trypan blue exclusion. Nuclei suspension was adjusted to 1000–3,000 nuclei per microlitre and loaded on the Chromium Controller (10x Genomics) with a targeted nuclei recovery of 5,000–10,000 per reaction. 3’ gene expression libraries. Libraries were sequenced using NovaSeq 6000 (Illumina) at Wellcome Sanger Institute with a minimum depth of 20,000–30,000 read pairs per nucleus. Read Mapping After sequencing, samples were demultiplexed and stored as CRAM files. Each sample of single-cell data was mapped to the human reference genome (GRCh38-2020-A) provided by 10x Genomics using the CellRanger software (CellRanger v.6.0.2 or v.6.1.0, or CellRanger ARC v.2.0.0) with default parameters. Part of the single-cell samples were mixed with different donors after the nuclei isolation for cost-efficient experimental design (Supplementary Table 1) and computationally demultiplexed (Souporcell, v.2.0, 32 ) based on genetic variation between the donors. Preprocessing and quality control For the single-cell transcriptome data, the CellBender algorithm (remove-background, v.0.2) 33 was applied to remove ambient and background RNA from each count matrix produced using the CellRanger pipeline. Downstream analysis was performed using the Scanpy package (v.1.7.1) 34 . Doublets with a score of >0.15 were removed using Scrublet 35 . We performed quality control and filtering based on the following settings: number of genes>500, total counts>1000, % mitochondrial<5 (nuclei), % mitochondrial<20 (cell), % ribosomal<5 (nuclei), % ribosomal<20 (cell), red blood cell score<1 (nuclei), red blood cell score<2 (cell). Data integration of single-cell data After pre-processing and removal of low-quality cells and nuclei, single-cell transcriptomics data were integrated using scVI 10 , 11 , accounting for categorical covariates of donor, cell_or_nuclei, and kit_10x, as well as continuous covariates of total_counts, %mito, and %ribo. Neighbourhood identification, high-resolution leiden clustering, and further dimensional reduction using UMAP were performed based on the scVI latent space. Additionally, Force Atlas or Diffusion Map embeddings were utilised for visualisation 36 . Subsequently, clustering was performed employing the Leiden algorithm. Cells and nuclei of extra-cardiac origin were removed based on marker gene expressions: lung epithelial cells (IGFBP2, EPCAM, SOX2, and NKX2-1); hepatocytes (ALB, APOA1, TTR); hepatic stellate cells (CRHBP, FCN3, OIT3); parathyroid cells (PTH, MAFB, GATA3); and further erythrocytes (FTH1, HBA1, HBA2, HBB). The remaining data were integrated using scVI with the same variables, followed by the downstream processing. To annotate cell types, cell barcodes comprising major clusters were subsampled from the initial raw count matrix for subclustering analysis and processed separately through the same downstream analysis as described. Subclusters were annotated based on immunohistochemical and spatial transcriptomics validation, the anatomical location sequenced after targeted dissection, and literature evidence. Gene set scoring Gene set scoring was carried out using the sc.tl.score_genes() function in Scanpy, with the default settings. Cell cycle genes 37 , ribosomal genes, mitochondrial genes, genes related to hemoglobin, Y-chromosome genes were scored. Xist counts and the presence of Y chromosome gene expression was examined to determine the sex of the samples. Tip and stalk identity of capillary ECs were scored through a previously reported set of genes 38 . Cells with tip scores surpassing their stalk scores were identified as tip cells, while those with predominant stalk scores were annotated as stalk cells. cell2location mapping cell2location 39 was used for mapping the suspension dataset on spatial transcriptomics slides; for a detailed description, see 9 . Differential expression and gene set enrichment analysis Differential gene expression analysis was performed utilising the sc.tl.rank_genes_groups function within the Scanpy framework. Differentially expressed genes were identified applying a significance threshold of p-value 2. These identified gene lists were then subjected to exploration using Gene Ontology Biological Process 2023 40 through the GSEApy package 41 . Coronary endothelium continuum scFates 42 pipeline was employed to fit a curved trajectory across the embedded representation of the blood vessel endothelial cells. Genes that showed significant changes in expression over pseudotime were detected and their expression profiles were plotted. Among the coronary endothelial cells, the milestone linked to proliferating coronary endothelial cells was marked, and the associated features were plotted across the pseudotime trajectory. Niche analysis We used CellPhoneDB to infer cell-cell interactions between annotated clusters 43 . Expression matrices and the metadata relating cell-barcode to cluster were submitted to CellPhoneDB. To compute niche interactions, we utilised the statistical method provided by the cpdb statistical analysis method function, 1000 iterations and applied a p-value threshold of 0.01. We further filtered the interactions by selecting the interactions where both the ligand and receptor expressions were present in more than 40% of the cells within a particular cluster. Results were plotted using ktplots- py package 44 . Immunohistochemistry Primary human fetal heart tissue was fixed overnight in 4% PFA (Alfa Aesar) with gentle rocking at 4 C. The next day, the tissue was incubated in 30% sucrose (Sigma/Merck) in PBS overnight. Following day, hearts were embedded in OCT (Sakura Tek), frozen on dry ice, then sectioned onto slides in 10 μM thickness and kept in −80 until immunostaining. Cryosectioned heart slides were initially thawed for 10 minutes at room temperature and then rehydrated in a tris-buffered saline solution. The sections were surrounded with a hydrophobic pen (Abcam) and subjected to permeabilization for 10 minutes in a buffer containing 0.25% saponin (Alfa Aesar) in TBS. Following this, the permeabilization buffer was removed, and a blocking step was performed for 1 hour at room temperature using a blocking buffer, which was composed of 0.3M glycine (Merck), 10% goat serum (Merck), and 0.2% Tween-20 (Merck) in TBS. Subsequently, the blocking buffer was discarded, and the primary solution (Supplementary Table 2, in 10% goat serum and 0.2% Tween in TBS) was applied overnight at 4 °C degrees within a humidified chamber. The samples underwent three 5-minute washes with a washing buffer containing 0.2% Tween-20 in TBS. Secondary antibody staining (Supplementary Table 2) was carried out for 1 hour at room temperature. Following this, the slides were washed three times in the washing buffer before the application of DAPI-containing mounting medium (Vectashield). Finally, the sections were imaged using a Zeiss LSM 980 AiryScan microscope and analysed utilising Zeiss ZEN software. Data availability Open-access datasets will be available from ArrayExpress ( www.ebi.ac.uk/arrayexpress ) with accession numbers. Processed data of sc/snRNA-seq data are available for browsing gene expression and download from the Heart Cell Atlas ( https://www.heartcellatlas.org/foetal.html , currently password protected). We also developed a summary-level interactive data explorer on R shiny for this dataset accessible at https://sinha.stemcells.cam.ac.uk/ . A CellTypist model of cell type annotations was produced using a balanced sampling of variables in the healthy reference hearts and is available on the webportal. The human reference genome (GRCh38) used for read mapping is available from 10x Genomics ( https://support.10xgenomics.com/single-cell-gene-expression/software/release-notes/build ). Author contributions S.B., J.C., K.K. and V. R. K.-S. conceived and designed the experiments, performed the analysis, wrote the manuscript. M.C., J.C.M.L. and R.T. provided critical input into the design. H.D. heart collection, processing and sequencing. L.R., C.I.S., R.K., S.P. helped with single-nucleus RNA-seq data generation.S.Y.H. helped with structural annotation. X.H. helped with coordinating foetal tissue samples. L.G. heart collection, processing and sequencing, provided critical input into the design. S.A.T. and S.S. designed the experiments and wrote the manuscript. Competing interests S.A.T. is a scientific advisory board member of ForeSite Labs, Qiagen and Element Biosciences, and a co-founder and equity holder of TransitionBio and EnsoCell Therapeutics. S.S. is a co-founder and equity holder of ABS Biotechnologies. The remaining authors declare no competing interests. Supplementary tables Supp. Table1: Donor metadata Supp. Table 2: Antibody information Acknowledgements We thank the donors for granting access to the tissue samples. We also thank Roger Barker for his help in coordinating samples. We thank staff at the Wellcome Sanger Cytometry Core Facility, Cellular Genetics Informatics team and Core DNA Pipelines team for their support; A. Oszlanczi for her help on sample management; B. Çakır for his help on the Heart Cell Atlas web portal; and A. Wilk for administrative assistance.This work was made possible by a partnership between the Wellcome-MRC Cambridge Stem Cell Institute at University of Cambridge and Wellcome Sanger Institute. This project was made possible in part the Wellcome Trust Clinical PhD Fellowship to J.C.; the Overseas Research Fellowship of the Takeda Science Foundation to K.K; the Oxbridge BHF Centre for Regenerative Medicine (RM/17/2/33380) (V.K.S.); and BHF grants PG/17/24/32886 (L.G.) and RG/17/5/32936 (H.D.); This project has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie-Skłodowska-Curie grant agreement No. 101026233 (J.P.P.); Wellcome Trust (WT206194 to S.A.T); the Chan Zuckerberg Foundation (2021-237882 to S.A.T.); the British Heart Foundation (BHF) Senior Fellowship (FS/18/46/33663) (S.S. and L.G.). We also acknowledge core support from the Wellcome Trust, the Medical Research Council and the Wellcome Trust–Medical Research Council Cambridge Stem Cell Institute.This research was funded, in whole or in part, by the Wellcome Trust (grant no. 203151/Z/16/Z). Footnotes ↵ † Co-senior author References for Supplementary Text 1. ↵ Muhl , L. et al. Single-cell analysis uncovers fibroblast heterogeneity and criteria for fibroblast and mural cell identification and discrimination . Nat. Commun . 11 , 1 – 18 ( 2020 ). OpenUrl CrossRef PubMed 2. ↵ Dick , S. A. et al. Self-renewing resident cardiac macrophages limit adverse remodeling following myocardial infarction . Nat. Immunol . 20 , 29 – 39 ( 2019 ). OpenUrl CrossRef PubMed 3. ↵ Revelo , X. S. et al. Cardiac Resident Macrophages Prevent Fibrosis and Stimulate Angiogenesis . Circ. Res . 129 , 1086 – 1101 ( 2021 ). OpenUrl CrossRef 4. ↵ Vagnozzi , R. J. et al. An acute immune response underlies the benefit of cardiac stem cell therapy . Nature 577 , 405 – 409 ( 2019 ). OpenUrl References 1. ↵ Cui , Y. et al. Single-Cell Transcriptome Analysis Maps the Developmental Track of the Human Heart . Cell Rep . 26 , 1934 – 1950 .e5 ( 2019 ). OpenUrl CrossRef PubMed 2. Sahara , M. et al. Population and Single-Cell Analysis of Human Cardiogenesis Reveals Unique LGR5 Ventricular Progenitors in Embryonic Outflow Tract . Dev. Cell 48 , 475 – 490 .e7 ( 2019 ). OpenUrl CrossRef 3. Suryawanshi , H. et al. Cell atlas of the foetal human heart and implications for autoimmune-mediated congenital heart block . Cardiovasc. Res . 116 , 1446 – 1457 ( 2019 ). OpenUrl 4. Asp , M. et al. A Spatiotemporal Organ-Wide Gene Expression and Cell Atlas of the Developing Human Heart . Cell 179 , 1647 – 1660 .e19 ( 2019 ). OpenUrl CrossRef PubMed 5. Sim , C. B. et al. Sex-Specific Control of Human Heart Maturation by the Progesterone Receptor . Circulation 143 , 1614 – 1628 ( 2021 ). OpenUrl CrossRef 6. Mehdiabadi , N. R. et al. Defining the Fetal Gene Program at Single-Cell Resolution in Pediatric Dilated Cardiomyopathy . Circulation 146 , 1105 – 1108 ( 2022 ). OpenUrl CrossRef 7. ↵ Knight-Schrijver , V. R. et al. A single-cell comparison of adult and fetal human epicardium defines the age-associated changes in epicardial activity . Nature Cardiovascular Research 1 , 1215 – 1229 ( 2022 ). OpenUrl 8. ↵ Ameen , M. et al. Integrative single-cell analysis of cardiogenesis identifies developmental trajectories and non-coding mutations in congenital heart disease . Cell 185 , 4937 – 4953 .e23 ( 2022 ). OpenUrl 9. ↵ Cranley , K. et al. Multiomic profiling of the developing human heart . In review . 10. ↵ Gayoso , A. et al. A Python library for probabilistic analysis of single-cell omics data . Nat. Biotechnol . 40 , 163 – 166 ( 2022 ). OpenUrl 11. ↵ Lopez , R. , Regier , J. , Cole , M. B. , Jordan , M. I. & Yosef , N . Deep generative modeling for single-cell transcriptomics . Nat. Methods 15 , 1053 – 1058 ( 2018 ). OpenUrl 12. ↵ Perez-Miguelsanz , J. et al. Early Appearance of Epicardial Adipose Tissue through Human Development . Nutrients 13 , ( 2021 ). 13. ↵ Franco , D. & Campione , M . The Role of Pitx2 during Cardiac Development: Linking Left–Right Signaling and Congenital Heart Diseases . Trends Cardiovasc. Med . 13 , 157 – 163 ( 2003 ). OpenUrl CrossRef PubMed Web of Science 14. ↵ Gore-Panter , S. R. et al. PANCR, the PITX2 Adjacent Noncoding RNA, Is Expressed in Human Left Atria and Regulates PITX2c Expression . Circ. Arrhythm. Electrophysiol . 9 , e003197 ( 2016 ). OpenUrl Abstract / FREE Full Text 15. ↵ Arita , Y. et al. Myocardium-derived angiopoietin-1 is essential for coronary vein formation in the developing heart . Nat. Commun . 5 , 1 – 14 ( 2014 ). OpenUrl CrossRef PubMed 16. ↵ Goodyer , W. R. et al. In vivo visualization and molecular targeting of the cardiac conduction system . J. Clin. Invest . 132 , ( 10 2022 ). OpenUrl CrossRef 17. ↵ Kanemaru , K. et al. Spatially resolved multiomics of human cardiac niches . Nature 619 , 801 – 810 ( 2023 ). OpenUrl CrossRef 18. ↵ Regan , C. P. , Manabe , I. & Owens , G. K . Development of a smooth muscle-targeted cre recombinase mouse reveals novel insights regarding smooth muscle myosin heavy chain promoter regulation . Circ. Res . 87 , 363 – 369 ( 2000 ). OpenUrl Abstract / FREE Full Text 19. ↵ Sinha , S. , Iyer , D. & Granata , A . Embryonic origins of human vascular smooth muscle cells: implications for in vitro modeling and clinical application . Cell. Mol. Life Sci . 71 , 2271 – 2288 ( 2014 ). OpenUrl CrossRef PubMed 20. ↵ Trimm , E. & Red-Horse , K . Vascular endothelial cell development and diversity . Nat. Rev. Cardiol . 20 , 197 – 210 ( 2023 ). OpenUrl 21. ↵ Su , T. et al. Single-cell analysis of early progenitor cells that build coronary arteries . Nature 559 , 356 – 362 ( 2018 ). OpenUrl CrossRef PubMed 22. ↵ Sissaoui , S. et al. Genomic Characterization of Endothelial Enhancers Reveals a Multifunctional Role for NR2F2 in Regulation of Arteriovenous Gene Expression . Circ. Res . 126 , 875 – 888 ( 2020 ). OpenUrl CrossRef 23. ↵ Segi , E. et al. Patent Ductus Arteriosus and Neonatal Death in Prostaglandin Receptor EP4-Deficient Mice . Biochem. Biophys. Res. Commun . 246 , 7 – 12 ( 1998 ). OpenUrl CrossRef PubMed Web of Science 24. ↵ Ye , H. et al. Interactions between Lubricin and Hyaluronic Acid Synergistically Enhance Antiadhesive Properties . ACS Appl. Mater. Interfaces 11 , 18090 – 18102 ( 2019 ). OpenUrl CrossRef PubMed 25. ↵ Chen , H. I. et al. The sinus venosus contributes to coronary vasculature through VEGFC-stimulated angiogenesis . Development 141 , 4500 – 4512 ( 2014 ). OpenUrl Abstract / FREE Full Text 26. ↵ Beltrami , C. et al. Human Pericardial Fluid Contains Exosomes Enriched with Cardiovascular-Expressed MicroRNAs and Promotes Therapeutic Angiogenesis . Mol. Ther . 25 , 679 – 693 ( 2017 ). OpenUrl CrossRef PubMed 27. ↵ Torregrosa-Carrión , R. et al. Adhesion G protein-coupled receptor Gpr126/Adgrg6 is essential for placental development . Sci Adv 7 , eabj5445 ( 2021 ). OpenUrl CrossRef 28. ↵ Osumi-Sutherland , D. et al. Cell type ontologies of the Human Cell Atlas . Nat. Cell Biol . 23 , 1129 – 1135 ( 2021 ). OpenUrl CrossRef 29. ↵ Ocaña , O. H. et al. A right-handed signalling pathway drives heart looping in vertebrates . Nature 549 , 86 – 90 ( 2017 ). OpenUrl CrossRef 30. ↵ Pallante , B. A. et al. Contactin-2 Expression in the Cardiac Purkinje Fiber Network . Circ. Arrhythm. Electrophysiol . 3 , 186 – 194 ( 2010 ). OpenUrl Abstract / FREE Full Text 31. ↵ Kikuchi , K. et al. tcf21+ epicardial cells adopt non-myocardial fates during zebrafish heart development and regeneration . Development 138 , 2895 – 2902 ( 2011 ). OpenUrl Abstract / FREE Full Text 32. ↵ Heaton , H. et al. Souporcell: robust clustering of single-cell RNA-seq data by genotype without reference genotypes . Nat. Methods 17 , 615 – 620 ( 2020 ). OpenUrl CrossRef PubMed 33. ↵ Fleming , S. J. et al. Unsupervised removal of systematic background noise from droplet-based single-cell experiments using CellBender . Nat. Methods 20 , 1323 – 1335 ( 2023 ). OpenUrl CrossRef PubMed 34. ↵ Wolf , F. A. , Angerer , P. & Theis , F. J . SCANPY: large-scale single-cell gene expression data analysis . Genome Biol . 19 , 15 ( 2018 ). OpenUrl CrossRef PubMed 35. ↵ Wolock , S. L. , Lopez , R. & Klein , A. M . Scrublet: Computational Identification of Cell Doublets in Single-Cell Transcriptomic Data . Cell Systems 8 , 281 – 291 .e9 ( 2019 ). OpenUrl 36. ↵ Haghverdi , L. , Buettner , F. & Theis , F. J . Diffusion maps for high-dimensional single-cell analysis of differentiation data . Bioinformatics 31 , 2989 – 2998 ( 2015 ). OpenUrl CrossRef PubMed 37. ↵ Kowalczyk , M. S. et al. Single-cell RNA-seq reveals changes in cell cycle and differentiation programs upon aging of hematopoietic stem cells . Genome Res . 25 , 1860 – 1872 ( 2015 ). OpenUrl Abstract / FREE Full Text 38. ↵ del Toro , R. et al. Identification and functional analysis of endothelial tip cell-enriched genes . Blood 116 , 4025 – 4033 ( 2010 ). OpenUrl Abstract / FREE Full Text 39. ↵ Kleshchevnikov , V. et al. Cell2location maps fine-grained cell types in spatial transcriptomics . Nat. Biotechnol . 40 , 661 – 671 ( 2022 ). OpenUrl 40. ↵ Gene Ontology Consortium . The Gene Ontology (GO) database and informatics resource . Nucleic Acids Res . 32 , D258 – D261 ( 2004 ). OpenUrl CrossRef PubMed Web of Science 41. ↵ Fang , Z. , Liu , X. & Peltz , G . GSEApy: a comprehensive package for performing gene set enrichment analysis in Python . Bioinformatics 39 , btac757 ( 2022 ). OpenUrl 42. ↵ Faure , L. , Soldatov , R. , Kharchenko , P. V. & Adameyko , I . scFates: a scalable python package for advanced pseudotime and bifurcation analysis from single-cell data . Bioinformatics 39 , btac746 ( 2022 ). OpenUrl 43. ↵ Garcia-Alonso , L. et al. Single-cell roadmap of human gonadal development . Nature 607 , 540 – 547 ( 2022 ). OpenUrl CrossRef PubMed 44. ↵ Troul’e , K. et al. CellPhoneDB v5: inferring cell-cell communication from single-cell multiomics data . ( 2023 ). View the discussion thread. Back to top Previous Next Posted April 27, 2024. Download PDF Supplementary Material Email Thank you for your interest in spreading the word about bioRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following High-resolution atlas of the developing human heart and the great vessels Message Subject (Your Name) has forwarded a page to you from bioRxiv Message Body (Your Name) thought you would like to see this page from the bioRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share High-resolution atlas of the developing human heart and the great vessels Semih Bayraktar , James Cranley , Kazumasa Kanemaru , Vincent Knight-Schrijver , Maria Colzani , Hongorzul Davaapil , Jonathan Chuo Min Lee , Krzysztof Polanski , Laura Richardson , Claudia I. Semprich , Rakeshlal Kapuge , Monika Dabrowska , Ilaria Mulas , Shani Perera , Minal Patel , Siew Yen Ho , Xiaoling He , Richard Tyser , Laure Gambardella , Sarah A. Teichmann , Sanjay Sinha bioRxiv 2024.04.27.591127; doi: https://doi.org/10.1101/2024.04.27.591127 Share This Article: Copy Citation Tools High-resolution atlas of the developing human heart and the great vessels Semih Bayraktar , James Cranley , Kazumasa Kanemaru , Vincent Knight-Schrijver , Maria Colzani , Hongorzul Davaapil , Jonathan Chuo Min Lee , Krzysztof Polanski , Laura Richardson , Claudia I. Semprich , Rakeshlal Kapuge , Monika Dabrowska , Ilaria Mulas , Shani Perera , Minal Patel , Siew Yen Ho , Xiaoling He , Richard Tyser , Laure Gambardella , Sarah A. Teichmann , Sanjay Sinha bioRxiv 2024.04.27.591127; doi: https://doi.org/10.1101/2024.04.27.591127 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Developmental Biology Subject Areas All Articles Animal Behavior and Cognition (7644) Biochemistry (17725) Bioengineering (13915) Bioinformatics (42031) Biophysics (21486) Cancer Biology (18635) Cell Biology (25548) Clinical Trials (138) Developmental Biology (13397) Ecology (19937) Epidemiology (2067) Evolutionary Biology (24361) Genetics (15619) Genomics (22538) Immunology (17763) Microbiology (40468) Molecular Biology (17206) Neuroscience (88733) Paleontology (667) Pathology (2842) Pharmacology and Toxicology (4834) Physiology (7657) Plant Biology (15172) Scientific Communication and Education (2046) Synthetic Biology (4304) Systems Biology (9831) Zoology (2272)