Lung EC Markers, Assessed at Levels of Lineage, Anatomy, and Cell Type

We identified 12 563 ECs within the integrated lung dataset based on the canonical endothelial markers PECAM1 (CD31), CDH5 (VE-cadherin), CLDN5, and ERG (Figures ​(Figures2,2, ​,4,4, and ​and5).5). We identified 147 panendothelial marker genes, defined as genes significantly expressed in all vascular and lymphatic EC populations compared with all other cell types (Table III in the Data Supplement). Panendothelial markers include genes coding for proteins associated with endothelial-intrinsic structures like the glycocalyx (PODXL, ST6GALNAC3, GALNT18), caveolae (CAV1, CAV2, CAVIN1, CAVIN2), focal adhesion structures (ITGA5, ITGB1, LAMB2, AKT3, PXN, MYL12A, MYL12B, PRK2), and adherens junctions (AFDN, TJP1, PTPRM; Figure ​Figure22 and Table III in the Data Supplement). All EC subpopulations express common transcription factors (ERG, ETS1, FLI1, SOX18) and regulators of endothelial migration (ANXA3, SASH1, LDB2, PTK2, MEF2C, STARD13). Other panendothelial markers include angiopoietin receptor TIE1, subunits of the receptors for calcitonin gene-related peptide and adrenomedullin CALCRL and RAMP2/RAMP3, as well as general endothelial receptors ROBO4 and S1PR1 (Figure ​(Figure22 and Table III in the Data Supplement).

Open in a separate window

Figure 4.

Localization of aerocytes and general capillary endothelial cells (ECs). A, Representative serial immunofluorescent images of microvascular markers PRX and CA4 with positive red staining in capillaries (CA4 with off-target positive staining in macrophages) and negative staining of a larger central vessel. The general endothelial markers CLDN5 and PECAM1 in green stain the larger central vessels as well, in addition to the microvasculature. The third image shows an immunostain of the aerocyte-specific marker HPGD in green (white arrows), in addition to general microvascular marker PRX in red. Nuclei are counterstained throughout with DAPI (4′,6-diamidino-2-phenylindole). The white box highlights an aerocyte, which is shown at higher magnification in the first image of B. B, Representative immunostains of aerocytes with the specific marker HPGD in green (cytoplasmic and nuclear expression pattern), which colocalizes with the general microvascular marker PRX in red. Nuclei are counterstained throughout with DAPI. C, In situ RNA hybridization stains of markers specific to the capillary subpopulations with SOSTDC1 staining aerocyte ECs (arrows) and FCN3 staining general capillary ECs (arrows) with positive staining in red. The black box highlights an area shown to the right of it in high magnification. Conventional immunohistochemical images of shown markers can be found in Figure IV in the Data Supplement.

Open in a separate window

Figure 5.

Systemic–venous endothelial cells (ECs) localize to the bronchial vascular plexus and visceral pleura. Representative, serial immunofluorescent images of systemic–venous EC markers COL15A1 and VWA1, panvenous marker ACKR1 (all in red), panendothelial markers PECAM1 and CLDN5, and lymphatic marker PDPN (all in green) in (A) a bronchus with accompanying arteries and (B) visceral pleura. In (A), COL15A1, VWA1, and ACKR1 stain the small vessels of the bronchial vascular plexus, but not the accompanying arteries; PECAM1 and CLDN5 stain all vessels. In (B), COL15A1, VWA1, and ACKR1 stain the pleural vessels, but not the alveolar microvasculature; PECAM1 and CLDN5 stain all vessels. PDPN identifies bronchial (A) and pleural (B) lymphatic vessels. VWA1 exhibits off-target staining in smooth muscle cells, and ACKR1 in ciliated cells. Nuclei are counterstained throughout with DAPI (4′,6-diamidino-2-phenylindole). Conventional immunohistochemical images of shown markers can be found in Figure IV in the Data Supplement.

Lung ECs separated into 2 broad populations in uniform manifold approximation and projection space (Figures ​(Figures1B,1B, ​B,2,2, and ​and55 and Table III in the Data Supplement): vascular ECs (10 469 cells from 66 participants) and lymphatic ECs (2094 cells from 64 participants). Lymphatic vessels, lined by lymphatic ECs, absorb interstitial fluid and carry it as lymph to a lymph node or collecting vein. Lymphatic ECs were identified on the basis of the expression of canonical lymphatic markers PROX1,¹⁵ LYVE1,¹⁵ FLT4,^15,16 and PDPN^16,17 (Figure ​(Figure22 and Table III in the Data Supplement). Immunofluorescent stains of lymphatic marker PDPN revealed lymphatic vessels in the distal lung (lymphatic capillaries) as well as surrounding large vascular vessels and bronchi (Figure ​(Figure55 and Figure IV in the Data Supplement). Lymphatic ECs express secreted proteins like CCL21¹⁶ mediating homing of T cells and semaphorins SEMA3A and SEMA3D,¹⁸ both necessary for lymphatic vessel maturation. Lymphatic marker genes comprise transcription factors TBX1, HOXD3, NR2F1, and NR2F2 and receptors KDR, GPR182, and TEK.

In contrast, blood vessels are lined by vascular ECs and carry blood from and to the heart. Panvascular markers (n=142) expressed across the remaining, nonlymphatic ECs include transmembrane glycoproteins such as ENG, PCDH17, CLEC14A, CLEC1A (C-type lectin domain family 1 member A), ESAM, and ITM2A; the cell surface receptors BMPR2, FLT1, ADGRL4, VIPR1, PLXNA2, FZD4, IL4R, and IL15RA; transporter and channel proteins such as SLCO2A1, SLCO4A1, and AQP1; as well as the transcription factors EPAS1, GATA2, FOXF1, and ETS2 (Figure ​(Figure22 and Table III in the Data Supplement).

Identification of Arterial, Venous, and Capillary ECs

To focus on vascular ECs, we generated a nonlymphatic EC dataset. To increase the power of our analysis, we included control vascular ECs from a sixth scRNAseq dataset¹⁴ (Leuvens Kankerinstituut, n=7 participants) of sorted, CD45^neg/CD31^pos lung ECs that brought the total number of vascular ECs to 15 142 from 73 participants. The Leuvens Kankerinstituut cohort had not been included in the initial lung dataset to avoid batch effects attributable to missing or selectively sorted non-ECs within this cohort.

We identified 3 broad vascular EC subtypes—arterial (n=2610 cells), capillary (n=7920 cells), and venous (n=4612)—based on expression of canonical marker genes (Figure ​(Figure33 and Table V in the Data Supplement). Arterial ECs were identified by EFNB2,^19–22 SOX17,^20,23 BMX,^19,24 SEMA3G,¹⁹ HEY1,^20,21 LTBP4,²⁵ FBLN5,²⁵ GJA5,²⁰ and GJA4.²⁰ Capillary ECs were identified by canonical markers CA4^16,26 and PRX,²⁷ as well as organism-wide capillary markers RGCC,¹⁶ SPARC,¹⁶ and SGK1.¹⁶ Venous ECs were identified through the canonical transcription factor NR2F2^20–22 (COUP-TFII [COUP transcription factor 2]), as well as VCAM1,^16,22 ACKR1²⁸ (Figure ​(Figure2),2), and SELP²⁸ expression (Figure ​(Figure3).3). Capillary and venous ECs separated further into 2 distinct populations each—venous ECs into pulmonary–venous and systemic–venous ECs and capillary ECs into aerocytes and general capillary ECs—for which a detailed characterization is given in the next 2 paragraphs (for makeup of vascular ECs, see Figure IIIA in the Data Supplement).

Open in a separate window

Figure 3.

Assessing fine-granular endothelial heterogeneity of the human lung by single-cell RNA sequencing. A, Uniform manifold approximation and projections of 15 142 vascular endothelial cells (ECs) from 73 control lungs of 6 cohorts. Each dot represents a single cell, and cells are labeled, from left to right, by cell type, cohort, sample location, and participant. In the uniform manifold approximation and projection colored by participants, each color represents a distinct participant. B, Heat map of marker genes of all 5 vascular EC populations. Each column represents the average expression value for 1 participant, grouped by cell type and cohort. All gene expression values are unity normalized from 0 to 1 across rows. BWH indicates Brigham and Women’s Hospital; LKI, Leuvens Kankerinstituut; TGen, Translational Genomics Research Institute; VIB, Leuven Vlaams Instituut voor Biotechnologie; and WSI, Wellcome Sanger Institute.

Beyond canonical marker concordance, vascular EC subpopulations were characterized on the basis of their distinct gene signatures associated with their respective cellular roles. Arterial ECs lining pulsating arteries and arterioles are exposed to relevant transmural pressure and shear stress. To withstand these forces, arterial ECs express genes encoding for tight and gap junction proteins such as CLDN10, GJA5, GJA4, and FBLIM1; extracellular matrix proteins that contribute to wall elasticity and strength such as FBLN5, FBLN2, MGP, BGN, LTBP4, LTBP1, and FN1; and the protease inhibitors SERPINE2²⁹ and CPAMD8 (Figure ​(Figure3B3B and Table V in the Data Supplement). Arterial ECs are active secretory cells and express signaling molecules such as CXCL12, EFBN2, SEMA3G,³⁰ VEGFA, and enzymes of the nitric oxide pathway involved in modulating vascular tonus like NOS1, PDE3A, and PDE4D. Maintenance of their arterial identity is accomplished through the expression of Wnt signaling modulator DKK2, the Notch ligand DLL4, and the transcription factors HEY1, SOX5, SOX17, HES4, and PRDM16 (Figure ​(Figure3B3B and Table V in the Data Supplement).

The capillary network surrounds millions of alveoli in the lungs to enable gas exchange. Capillary ECs express genes coding for enzymes related to gas exchange including CA4^16,26 and CYB5A; multiple receptors including ACVRL1/TMEM100, ADGRF5, ADGRL2, F2RL3, IFNGR1, VIPR1, and ADRB1; intracellular signaling molecules like ARHGAP6, IFI27, PREX1, PRKCE, SGK1,¹⁶ SH2D3C, and SORBS1; structural proteins PRX,²⁷ SPARC,¹⁶ EMP2, ITGA1, and SLC9A3R2; and transcription regulators AFF3 and MEIS1; and exhibit the highest major histocompatibility complex class I protein expression of all ECs (Figure ​(Figure3B3B and Table V in the Data Supplement).

Venules and veins return blood to the heart. In contrast to arteries, veins are thin-walled and subject to low shear forces, which makes in particular the postcapillary venules the primary location for leukocyte extravasation. To facilitate this, venous ECs express genes encoding for proteins associated with diapedesis of leukocytes like VCAM1,^16,22 SELP,²⁸ and SELE²⁸ (Figure ​(Figure3B3B and Table V in the Data Supplement) as well as ACKR1²⁸ for nonspecific transcytosis of chemokines. Venous ECs are further characterized by the major transcription factor NR2F2,^20–22 as well as genes coding for secreted (ADAMTS9, IGFBP7), nuclear (HDAC9, RORA), and membrane proteins (ACTN1, LDLRAD3, LDLRAD4, and LRRC1; Table V in the Data Supplement).

scRNAseq Characterization of 2 Previously Indistinguishable Capillary EC Types

scRNAseq reveals that the previously and consistently observed patchy or mosaic histologic staining patterns of vWF (von Willebrand factor), THBD (thrombomodulin), and EMCN (endomucin)^14,31–34 as well as EDN1 (endothelin 1) and EDNRB (endothelin receptor B1)³⁵ in lung capillaries can be attributed to 2 discernable capillary populations: a vWF^neg/EMCN^high/EDNRB^pos population and a vWF^pos/EMCN^low/EDN1^pos population (Figure ​(Figure3A3A and ​and3B).3B). Following nomenclature proposed by Gillich et al.,³⁶ we refer to these 2 microvasculature ECs as aerocytes and general capillary ECs.

Aerocytes (n=2317) can be distinguished by their expression of genes encoding for the endothelin receptor EDNRB, the transcription factors TBX2 and FOXP2, the pattern recognition receptors CLEC4E and SPON2, and signaling mediators such as PRKG, CHRM2, S100A4, and EDA (Figure ​(Figure3B3B and Table V in the Data Supplement). Aerocytes are the only EC type expressing prostaglandin-degrading HPGD, which is responsible for the first-pass deactivation of most prostaglandins as they pass through the vascular bed of the lung.³⁷ Aerocytes specifically express SOSTDC1, antagonist to BMPR2, which itself is expressed in all ECs and its mutations are associated with pulmonary arterial hypertension. In contrast to all other ECs, aerocytes do not express major components of endothelial-specific Weibel-Palade bodies (vWF, SELP, EDN), storage granules essential for hemostasis and extravasation of leukocytes. Aerocytes are characterized by the absence of expression of genes otherwise expressed in all other ECs including THBD, CD93, PTPRB, ANO2, CRIM1, MECOM, LIFR, PALMD, GNA14, CYYR1, and LEPR. Aerocytes are more rare in the human lung compared with general capillary ECs and were profiled in the median per participant in a ratio of aerocytes to general capillary ECs of 0.48:1 (IQR, 0.26–1.16; Figure IIIB in the Data Supplement).

General capillary ECs (n=5603) are the most generic vascular cell type; they specifically express only 140 genes, compared with arterial (n=373) or venous (n=280) ECs or aerocytes (n=203). Nevertheless, the general capillary EC population can be identified by the expression of genes related to transcytosis of low-density lipoprotein cholesterol and other lipids such as GPIHBP1¹⁶ and CD36³³; innate immune response including FCN3, BTNL9, BTNL8, and CD14; and cytokine receptors such as IL7R and IL18R1 (Figure ​(Figure3B3B and Table V in the Data Supplement).

We localized both capillary EC types by immunofluorescent microscopy using CA4 (Carbonic anhydrase 4) and PRX (Periaxin) as common capillary markers and HPGD (15-hydroxyprostaglandin dehydrogenase) as an aerocyte-specific marker, and by in situ hybridization using the subtype-specific markers SOSTDC1 for aerocytes and FCN3 for general capillary ECs (Figure ​(Figure4).4). In the control lungs, both consistently localize to and intermingle in the endothelium of the alveolar wall.

Systemically Perfused Vessels of Bronchi and the Visceral Pleura Are Lined by a COL15A1-Positive EC Type

Venous ECs divide into COL15A1^pos and COL15A1^neg subpopulations. Immunofluorescent microscopy using the markers COL15A1 (Collagen alpha-1(XV) chain) and VWA1 (von Willebrand factor A domain-containing protein 1) revealed that COL15A1^pos ECs localized to systemically supplied vessels of the bronchial vascular plexus, as we previously reported,¹³ and of the visceral pleura (Figure ​(Figure5),5), whereas vessels supplied by the pulmonary circulation remained negative. In contrast, the panvenous marker ACKR1 (Atypical chemokine receptor 1) stained both pulmonary and systemic venules (Figure ​(Figure5).5). Furthermore, 4 scRNAseq samples from biopsies of the systemically supplied large airways contributed almost exclusively cells to the COL15A1^pos population (Figure ​(Figure3A).3A). On the basis of the immunofluorescent stains and the tissue source of the samples, we were able to assign 2 venous subpopulations to the systemic and pulmonary circulation: systemic–venous ECs (COL15A1^pos, n=2291) and pulmonary–venous ECs (COL15A1^neg, n= 2321). The integrated dataset allowed us to identify additional genes specific to systemic–venous ECs in addition to COL15A1, a multiplexin collagen linking basement membrane to underlying connective tissue. We also identified the extracellular matrix protein vWA1, as well as PLVAP, which encodes an EC-specific protein that forms the stomatal and fenestral diaphragms of blood vessels and regulates basal permeability, leukocyte migration, and angiogenesis (Figure ​(Figure3B3B and Table V in the Data Supplement). Systemic–venous ECs express genes that encode for specific sets of transcription factors, including EBF1, EBF3, MEOX1, MEOX2, and ZNF385D; the cell surface receptors TACR1, ROBO1, and CYSLTR1; the peptidases CPXM2 and MMP16; the phosphodiesterases PDE7B and PDE2A; as well as antagonist of fibroblast growth factor signaling SPRY1.

Whereas aerocytes are characterized by their ability to break down prostaglandins (HPGD) and their lack of hemostatic Weibel-Palade bodies, pulmonary–venous ECs—which are located downstream—express genes encoding for proteins associated with prostaglandin synthesis including PTGS1 and PTGIS, and with complement and coagulation cascades such as C7, PLAT, PROCR, THBD, C1R, and CLU (Figure ​(Figure3B3B and Table V in the Data Supplement). Pulmonary–venous ECs express the most specific pulmonary–venous gene CPE as well as Wnt modulator DKK3, which contrasts with the DKK2 expression in arterial ECs. Genes coding for structural proteins specific to pulmonary–venous ECs include EFEMP1 and CDH11. Overlapping with lymphatic ECs, but expressed at lower levels compared with them, pulmonary–venous ECs express factor V/Va binding and extracellular matrix protein MMRN1¹⁶ and PKHD1L1 (Figure ​(Figure3B3B and Table V in the Data Supplement).

EC Subpopulations’ Connectivity With Other Lung Cell Populations

Complex tissues like the lung have distinct cellular niches that are tightly regulated by intercellular communications. To study the role of ECs in tissue homeostasis within the human lung, we performed a connectomic analysis that mapped the averaged gene expression per cell type to known ligand–receptor interactions catalogued in the NicheNet database.^38,39 We focused our analysis on potential communication pathways between EC and non-EC cell types. Because of the complexity of the cell–cell signaling connectome, we can only present a limited fraction of our findings here (for the full connectome, see Table VI in the Data Supplement). All ligand–receptor interaction networks can be explored on www.LungEndothelialCellAtlas.com.

We first identified the intercellular signaling patterns wherein ECs are senders of ligands (Figure ​(Figure6A).6A). Arterial ECs are the most active EC subpopulation in this respect, expressing Notch ligands (DLL4, JAG1, and JAG2), which can be sensed by NOTCH3 in pericytes and smooth muscle cells (SMCs) as a part of a signaling axis that is crucial for arterial EC specification, as well as proliferation and differentiation of supporting pericytes and vascular SMCs.⁴⁰ Arterial ECs, and to a lesser extent venous and general capillary ECs, express EDN1, which can be sensed by pericytes, SMC, and alveolar fibroblasts through EDNRA; this is a vasoconstrictive axis⁴¹ implicated in pulmonary hypertension. Arterial ECs also secrete CXCL12, a chemokine that facilitates homing, migration, and survival of immune cells through binding to CXCR4, which is expressed by most lymphoid cells (T regulatory, T helper, T cytotoxic, B, and innate lymphoid cells) and dendritic cells (DCs; plasmacytoid DCs [pDCs], mature DCs, and Langerhans DCs).⁴² Lymphatic ECs secrete CCL21, which is sensed by pDCs through CXCR3 and by mature DCs, Langerhans DCs, and B cells through CCR7 (Figure ​(Figure6A),6A), representing a classic mechanism for guiding matured DCs to secondary lymphoid tissues.⁴³ Lymphatic EC, as well as AT1 (alveolar cell type 1) and pericytes, express PDGFA, with the corresponding receptor PDGFRB expressed in pericytes and SMCs. Lymphatic ECs were found to express the spatial guidance molecule SEMA3A that is critically important for lymphatic vessel maturation.¹⁸ Its corresponding receptors are expressed on alveolar and nonparenchymal fibroblasts and vascular ECs (NPR1); AT1, AT2, Club, basal, suprabasal, and mesothelial cells (PLXNA1/PLXNA2); and pDCs (PLXNA4; Figure ​Figure66A).

Open in a separate window

Figure 6.

Endothelial cell (EC)–focused intercellular communication. Visualization of a small subset of the connectomic analysis. Circos plots of top 75 edges by edge weight with EC subpopulations as (A) senders and (B) receivers. Edge thickness is proportional to edge weights. Edge color labels the source cell type. In both Circos plots, ligands occupy the lower semicircle and corresponding receptors the upper semicircle, and ligands and receptors are colored by the expressing cell type. The full results of the connectomic analysis can be found in Table VI. AM indicates alveolar macrophage; AT1/2, alveolar cell type 1/2; cDC1/2, classical dendritic cell type 1/2; cMono, classical monocyte; DC, dendritic cell; ILC, innate lymphoid cell; M, macrophage; ncMono, nonclassical monocyte; NK, natural killer; pDC, plasmacytoid dendritic cell; PNEC, pulmonary neuroendocrine cell; scRNAseq, single cell RNA sequencing; and SMC, smooth muscle cell.

We next identified selected cell–cell signaling patterns with ECs being on the receiving end (Figure ​(Figure6B).6B). Alveolar fibroblasts strongly express the morphogen SLIT2 and ANGPT1 (to a lesser extent also in pericytes and SMCs). Both SLIT2 and ANGPT1 can be sensed by all 5 EC subpopulations through the receptors ROBO4 and TEK/TIE1, respectively. Furthermore, alveolar fibroblasts are in a position to signal to aerocytes and lymphatic ECs through the VEGFD–KDR and VEGFD-FLT4 axes, respectively (Figure ​(Figure66B).

Not all ligands sensed by EC receptors are produced by resident lung cells. For example, the receptor VIPR1 is expressed by lung ECs but its main ligand VIP (vasoactive intestinal peptide) was not expressed by any cells profiled in the lungs. This suggests that the VIPR1 ligands targeting lung ECs are mainly derived from extrapulmonary organs. Therefore, alternate sources of ligands beyond those profiled in the present connectomic analysis may contribute to lung EC signaling and function.

Normal Lung Endothelial Subtypes and Pulmonary Hypertension

Whereas we focus on EC subpopulations in the normal lung, we also assessed the relevance of our findings to pulmonary hypertension. Using our integrated scRNAseq dataset, we attributed expression patterns of genes associated with mendelian genetic diseases of the lung to their cells of origin. BMPR2, known to be associated with pulmonary arterial hypertension, was expressed in all ECs, but SOX17 only in arterial ECs (see Results and Figure V in the Data Supplement). Reanalysis of a pulmonary arterial hypertension scRNAseq dataset⁴⁴ revealed that all vascular endothelial subpopulations could be observed in lungs of patients with pulmonary arterial hypertension (see Results and Figure VI in the Data Supplement). In contrast, scRNAseq revealed that commercially obtained, primary arterial, microvascular, and venous lung ECs lost their respective native lung phenotype (see Results and Figures VII and VIII in the Data Supplement).

For further details, see the Results in the Data Supplement.

Novel EC Populations Found in Humans Are Also Present in Mice

In mammals, the lung is the only organ in which gas exchange between blood and air takes place. Here, we aimed to explore whether cellular and transcriptomic principles of lung vascular biology are preserved and chose the most used mammalian model organism as comparator: the mouse. We applied our human workflow to mouse lung scRNAseq data to generate an integrated dataset of mouse lung cells from 18 control wild-type C57BL/6 mice from 6 cohorts^{14,16,39,45–47} (Figure IX in the Data Supplement and Table VII in the Data Supplement). All murine scRNAseq data had been generated from whole lung dissociations using 10×3′ scRNAseq kits based on v2 Chemistry. Only 2 out of 18 mice were female and the median age was 12.5 weeks (IQR, 10.75 – 34.0 weeks; for further details by cohort, see Table II in the Data Supplement). We identified 21 831 ECs within the integrated mouse lung dataset containing 57 974 single-cell transcriptomes (Figure ​(Figure7A7A and Figure IX in the Data Supplement). Subsetting the ECs from the mouse lung dataset for more specific subtype identification revealed the same EC subpopulations found in humans (Figure ​(Figure7A).7A). Regarding murine cell type marker genes in the full dataset and the EC subset, see Figures IX and X in the Data Supplement and Tables VIII and IX in the Data Supplement. Technical summaries of sample preprocessing are available in Table X in the Data Supplement.

Open in a separate window

Figure 7.

Identification of conserved endothelial cell (EC) populations and conserved marker genes in humans and mice. A, Uniform manifold approximation and projection of all 57 974 mouse cells from 18 control mouse lungs colored by lineage membership and of the subset of 21 343 vascular ECs labeled by cell type, cohort, and animal. In the uniform manifold approximation and projection colored by participants, each color represents a distinct mouse. B, Violin plots of marker genes significantly expressed in systemic–venous ECs compared with all other vascular ECs (first to third row) and of homologues to human marker genes of systemic–venous ECs lacking specificity for murine systemic–venous ECs (fourth row). C, Heat map of conserved panendothelial, panvascular, and lymphatic EC marker genes. Each column represents the average expression value per cell type for ECs and per lineage for non-ECs. All gene expression values are unity normalized per species from 0 to 1 across rows. D, Heat map of conserved marker genes of 4 pulmonary EC populations. Each column represents the average expression value per cell type. All gene expression values are unity normalized per species from 0 to 1 across rows. Human genes are indicated by capital letters and mouse genes are indicated by small letters after the first letter, separated by a slash. Labels of mouse, but not human, cell types/lineages are given in italics.

We identified the previously undescribed systemic–venous EC type in mice as well based on the marker genes Col15a1 and Vwa1 (Figure ​(Figure7B7B and Figure XA in the Data Supplement and Table IX in the Data Supplement). Systemic–venous ECs are rare in mice and account for only 0.5% of all ECs (n=117 cells; Figure XB in the Data Supplement), probably because—unlike in humans—both the visceral pleura and intraparenchymal airways are supplied by the pulmonary rather than the systemic circulation.^48,49 As in humans, murine systemic–venous ECs express transcription factors Ebf1, Ebf3, Meox2, and Tshz2, phosphodiesterase Pde7b, and, overlapping with venous ECs, Selp and Ackr1. However, homologues to some human marker genes of systemic–venous EC show no specificity for murine systemic–venous ECs (Plvap, Spry1, Pkp4), a different specificity profile (Igfbp7), a very low expression (Tacr1, Robo1; Figure ​Figure5B),5B), or a homologue does not exist in mice (eg, ZNF385D).

Murine capillary ECs are characterized by the expression of canonical markers Car4,^16,26 Prx,²⁷ and Sgk1¹⁶ and split likewise in 2 subpopulations: aerocytes (n=4853 cells) and general capillary ECs (n=14 119 cells; Figure XA in the Data Supplement and Tables VIII and IX in the Data Supplement). Aerocytes were identified by the expression of Ednrb and Tbx2 and general capillary ECs by Gpihbp1 and Edn1 (Table IX in the Data Supplement). In the median per mouse, capillary ECs were profiled in a ratio of aerocytes to general capillary ECs of 0.24:1 (IQR, 0.15–0.39; Figure IIIB in the Data Supplement), indicating that aerocytes are less common in mice lung capillaries when compared with those of humans (ratio in humans of 0.48:1; IQR, 0.26–1.16). A breakdown of marker genes overlapping with humans for both populations is given in the following section. In contrast to what has observed in humans, murine aerocytes are also characterized by top marker genes like Igfbp7, Emp2, Rgs6, Abcc3, Ackr2, Fibin, and Chst1 and general capillary ECs by genes such as Glp1r, Adgrl3, Plcb1, and Hmcn1 (Table IX in the Data Supplement). Interesting patterns can be observed that distinguish both murine capillary cell types but that are only vaguely recognizable in humans: Apln is specifically expressed in aerocytes, whereas its receptor Aplnr is expressed in general capillary ECs. In contrast, Vegfa is expressed in general capillary ECs and its receptor Kdr on aerocytes (Table IX in the Data Supplement). In contrast to humans, major constituents of Weibel-Palade bodies—Vwf and Selp—are expressed neither in aerocytes nor in general capillary ECs.