Ace2 safeguards embryonic hematopoietic stem and progenitor cell production by restraining Nlrp3-mediated pyroptosis

Significance

The emergence of hematopoietic stem and progenitor cells (HSPCs) during embryogenesis is a finely tuned process reliant on a delicate balance of inflammatory cues. While prior studies have identified proinflammatory pathways essential for HSPC development, how endothelial cells regulate and restrain inflammation to protect the hematopoietic niche remains largely unknown. Here, we uncover a conserved and endothelial-intrinsic role for angiotensin-converting enzyme 2 (ACE2) in safeguarding embryonic hematopoiesis. Using zebrafish and mouse models, we show that ACE2 suppresses Nlrp3–Caspase-1-dependent pyroptosis through its enzymatic conversion of Angiotensin II to the anti-inflammatory peptide angiotensin-(1–7). These findings uncover the developmental role of ACE2 in maintaining hematopoietic niche integrity and highlight inflammation control as essential for early blood stem cell formation.

Abstract

During vertebrate embryogenesis, hematopoietic stem and progenitor cells (HSPCs) originate from hemogenic endothelium (HE) in the dorsal aorta through endothelial-to-hematopoietic transition (EHT). While basal inflammation is essential for this process, excessive immune activation disrupts HSPC emergence. Here, we identify angiotensin-converting enzyme 2 (Ace2), a key component of renin–angiotensin system, as a crucial anti-inflammatory regulator of embryonic hematopoiesis in zebrafish and mice. Loss of Ace2 impairs HE specification and reduces nascent HSPC production. Mechanistically, transcriptomic profiling reveals that ace2 deficiency leads to aberrant activation of NLR family pyrin domain containing 3 (Nlrp3) signaling and pyroptosis in vascular endothelial cells. Importantly, pharmacological inhibition of Nlrp3 or Caspase-1 restores HSPC emergence upon ace2 deficiency, consistent with treatment with exogenous angiotensin-(1–7) [Ang-(1–7)], a downstream product of Ace2 enzymatic activity. Moreover, Ace2 knockdown in mouse embryos phenocopies the defects in zebrafish, demonstrating evolutionary conservation of ACE2 in developmental hematopoiesis in mammals. Together, our findings uncover an essential role for ACE2 in maintaining a permissive inflammatory environment for HSPC development and suggest therapeutic potential for targeting the ACE2/Ang-(1–7)/Nlrp3-pyroptosis axis in inflammatory hematopoietic disorders.

Hematopoietic stem and progenitor cells (HSPCs) are one of the adult stem cells with self-renewal and multidifferentiation capacity that sustain lifelong blood production in vertebrates (1, 2). During embryogenesis, HSPCs originate from a specialized subset of endothelial cells known as hemogenic endothelium (HE), through a tightly orchestrated process called endothelial-to-hematopoietic transition (EHT), which occurs within the aorta-gonad-mesonephros (AGM) region across vertebrates. Following their specification, nascent HSPCs migrate to a transient hematopoietic niche, such as the fetal liver in mammals or the caudal hematopoietic tissue (CHT) in zebrafish, where they undergo maturation and clonal expansion. Ultimately, HSPCs home to the bone marrow in mammals or the kidney marrow in zebrafish to establish and maintain adult hematopoietic homeostasis (3–9). Among embryonic hematopoietic sites, the arterial endothelium of the dorsal aorta in the AGM serves as a primary hematopoietic site and is uniquely endowed with the capacity to generate definitive HSPCs (10–12). However, the molecular mechanisms by which endothelial cells coordinate signaling cascades to initiate and sustain HSPC development remain incompletely understood. Thus, exploring the precise spatiotemporal regulation of HSPC emergence is essential for establishing a functional hematopoietic system.

Angiotensin converting enzyme 2 (ACE2) is a monocarboxypeptidase broadly expressed in vascular endothelium, smooth muscle, and epithelial cells across multiple organs in humans, including the heart, kidney, testis, gastrointestinal tract, brain, and lung (13–15). It functions as a key component of the renin–angiotensin system (RAS) by catalyzing the conversion of angiotensin II (Ang II) to angiotensin 1–7 [Ang-(1–7)], thereby regulating vascular tone, fluid homeostasis, and cardiovascular physiology (13, 16–18). Beyond its physiological roles, ACE2 is also known as the cellular entry receptor for the severe acute respiratory syndrome coronaviruses, SARS-CoV and SARS-CoV-2 (16, 19). Upon infection, SARS-CoV-2 hijacks ACE2-expressing epithelial cells in the lung, leading to robust viral replication and the release of inflammatory cytokines (19, 20). These cytokines activate surrounding myeloid cells, amplifying the inflammatory response and contributing to the cytokine storm, multiorgan dysfunction, and mortality in severe COVID-19 cases (21–25). While ACE2 has been extensively studied in cardiovascular, renal, and infectious diseases, its physiological functions in hematopoiesis remain unexplored.

Here, we identify Ace2 as a critical regulator of embryonic HSPC development by suppressing endothelial inflammation. Using zebrafish and mice as the model, we demonstrate that ace2 deficiency impairs HSPC emergence by activating Nlrp3 signaling and Caspase-1-dependent pyroptosis in vascular endothelial cells. Mechanistically, Ace2-mediated conversion of Ang II to Ang-(1–7) restrains proinflammatory signaling and preserves the hemogenic potential of the AGM microenvironment. These findings uncover a conserved, endothelial-intrinsic anti-inflammatory mechanism that safeguards hematopoietic development, highlighting Ace2 as a crucial modulator of vascular homeostasis during HSPC ontogeny.

Results

Ace2 Is Required for Embryonic HSPC Development.

To investigate the role of Ace2 in embryonic hematopoiesis, we first examined its expression dynamics during embryogenesis in zebrafish. Whole mount in situ hybridization (WISH) revealed that ace2 is a maternally expressed gene and becomes broadly expressed from 24 h postfertilization (hpf) to 4 d postfertilization (dpf) in zebrafish embryos (SI Appendix, Fig. S1A). Quantitative PCR (qPCR) revealed that ace2 transcripts are most abundant in endothelial cells/HE (kdrl⁺), while lower but detectable level are also present in HSPCs (CD41⁺gata1⁻), myeloid (lyz⁺) and erythroid (gata1⁺) populations (Fig. 1A). In contrast, nonendothelial and hematopoietic cells, such as muscle (myhz2⁺) and neuron (HuC⁺) populations, exhibited relatively lower expression levels (Fig. 1A). Moreover, transverse section of WISH embryos at 22 hpf revealed that ace2 is relatively enriched in the ventral wall of the dorsal aorta (DA) (SI Appendix, Fig. S1A). Together, these findings indicated that ace2 is preferentially expressed in endothelial cells, suggesting a potential role in definitive hematopoiesis.

Fig. 1.

Ace2 is involved in the development of embryonic HSPCs in zebrafish. (A) Quantitative PCR (qPCR) showing the expression of ace2 in CD41⁺gata1⁻ hematopoietic stem and progenitor cells (HSPCs) of 2 dpf, kdrl⁺ endothelial cells/hemogenic endothelial cells (EC/HEs), lyz⁺ myeloid cells, gata1⁺ erythroid cells, mylz⁺ muscle cells, and HuC⁺ neuron cells of 24 hpf. (B) Whole mount in situ hybridization (WISH) showing the expression of runx1 in the aorta-gonad mesonephros (AGM) region at 36 hpf, cmyb at 3 dpf, pu.1 in the caudal hematopoietic tissue (CHT) region, and rag1 in the thymus at 4 dpf of ace2^+/+ and ace2^−/−. Statistical data showing the expression of runx1 in the AGM at 36 hpf, cmyb at 3 dpf, pu.1 and gata1 in the CHT region, and rag1 in the thymus at 4 dpf of ace2^+/+ and ace2^−/−. (C) Confocal imaging and statistical data of Tg(cmyb:eGFP/kdrl:mCherry) showing the number of cmyb⁺kdrl⁺ cells in the AGM and cmyb⁺kdrl⁻ in the CHT region at 36 hpf and 3 dpf of ace2^+/+ and ace2^−/−. [Scale bar, (B and C), 50 μm.] The results are represented as means ± SD. Student’s t test, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

To determine whether Ace2 is required for embryonic hematopoiesis, we first performed gene knockdown experiments by using the ATG-morpholino (aMO) to block the translation of ace2 transcripts. Western blot validated a substantial reduction of Ace2 protein levels in morphants (SI Appendix, Fig. S1B). While primitive erythroid and myeloid progenitor markers appeared unaffected (SI Appendix, Fig. S1C), definitive hematopoiesis was clearly compromised. Both WISH and qPCR revealed that the expression of nascent HSPC marker, cmyb, was reduced in the AGM and CHT of ace2 morphants at 36 hpf and 2 dpf (SI Appendix, Fig. S1 D–G). Similarly, myeloid lineage differentiation was also impaired, as assayed by reduced expression of pan-myeloid cell marker pu.1 by WISH and qPCR, lymphoid/myeloid progenitor cell marker coro1a, and neutrophil-specific marker lyz at 4.5 dpf by qPCR (SI Appendix, Fig. S1 F and G). Similar phenotypes were assessed and confirmed by live imaging in the double transgenic line Tg(mpo:GFP/lyz:DsRed) embryos at 4.5 dpf in the CHT (SI Appendix, Fig. S1H). To assess lymphoid development, we also examined the expression of ikaros and rag1, which specifically indicates lymphoid progenitor and differentiated T cells from HSPCs. WISH results showed that ikaros and rag1 expression was evidently reduced in the CHT and thymus in ace2 morphant, respectively (SI Appendix, Fig. S1F), confirmed by live imaging in Tg(rag2:DsRed) transgenic embryos at 4.5 dpf (SI Appendix, Fig. S1I). Importantly, the expression of thymic epithelial cell marker foxn1 remained unaffected (SI Appendix, Fig. S1F), suggesting that the impaired T lymphoid differentiation capability was due to HSPC-intrinsic defects rather than thymic niche abnormalities. To validate these findings, we employed a second morpholino, ace2 splice MO (spMO), which led to pre-mRNA splicing blockage and resulted in a dramatic reduction in ace2 proteins (SI Appendix, Fig. S2A). WISH and qPCR results showed a similar hematopoietic defect upon ace2 depletion (SI Appendix, Fig. S2 B and C).

To genetically confirm the role of Ace2 and corroborate the hematopoietic defects of ace2 morphants, we generated an ace2 null-allele knockout mutant (ace2^−/−) harboring a 5 bp deletion (CCGCA) via CRISPR-Cas9 technology (SI Appendix, Fig. S2 D–G). Western blot and qPCR confirmed a substantial reduction at the protein and transcript level in ace2 mutants (SI Appendix, Fig. S2 H and I), in which the residual protein likely resulted from the partial retention of maternal transcripts. Consistent with the observed hematopoietic defects in ace2-knockdown embryos, we found that the expression of marker genes of HSPCs in the trunk and HSPCs-differentiated lineages including myeloid- and lymphoid cells in the CHT and thymus was reduced significantly in ace2 mutants by WISH (Fig. 1B). Similarly, live imaging further displayed an evidently decreased number of cmyb⁺ HSPCs in ace2 mutant in the background of double transgenic line Tg(cmyb:eGFP/kdrl:mCherry) at 36 hpf and 3 dpf, compared to the wild-type siblings (ace2^+/+) (Fig. 1C). Altogether, these results reveal that ace2 plays an indispensable role in HSPC development.

Ace2 Is Essential for HSPC Development by Sustaining HE Specification.

Having established that Ace2 is required for embryonic hematopoiesis, we next sought to explore when and how these defects arose during development. To this end, time-series analysis of runx1 expression between 23hpf-30hpf was performed. WISH analysis revealed a marked reduction in runx1 expression in ace2 morphants throughout this window (SI Appendix, Fig. S3A), implicating impaired HE specification. These results were further corroborated using double transgenic line Tg(kdrl:mCherry/runx1:enGFP) embryos (Fig. 2A), in which live imaging confirmed a significant loss of runx1⁺ HE cells in the dorsal aorta of ace2-deficient embryos at 22 hpf and 28 hpf. To further substantiate the requirement of ace2 for HE specification, we examined additional HE markers. Both WISH and qPCR further showed significant downregulation of gata2b and gfi1aa at 22 hpf and 28 hpf in ace2^−/− and ace2 morphants compared to ace2^+/+ and controls (Fig. 2 B–E), reinforcing a defect at the level of HE specification.

Fig. 2.

Ace2 is required for HSPC development by sustaining HE specification. (A) Confocal imaging and statistical data of Tg(runx1:enGFP/kdrl:mCherry) showing the number of runx1⁺kdrl ⁺ cells in the AGM of ace2^+/+ and ace2^−/− at 22 hpf. (B) WISH showing the expression of gata2b and gfi1aa in the AGM of ace2^+/+ and ace2^−/− at 22 hpf. (C) WISH showing the expression of gata2b and gfi1aa in the AGM of control and ace2 morphants at 28 hpf. (D) qPCR showing the expression of gata2b and gfi1aa in ace2^+/+ and ace2^−/− at 22 hpf. (E) qPCR showing the expression of gata2b and gfi1aa in the AGM of control and ace2 morphants at 28 hpf. (F) WISH showing the expression of runx1 in the AGM of control, ace2 morphants, and ace2 morphants injected by zebrafish ace2 (zace2), mouse Ace2 (mAce2), and human ACE2 (hACE2) mRNA at 28 hpf. (G) Confocal imaging and statistical data of Tg(cmyb:eGFP/ kdrl:mCherry) showing the number of cmyb⁺kdrl⁺ cells in the AGM at 30 hpf of ace2^+/+, ace2^−/− and ace2^−/− injected by zace2, mAce2, and hACE2 mRNA at 30 hpf. [Scale bar, (A, B, D, and E), 50 μm.] The results are represented as means ± SD. Student’s t test, *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

To ascertain the gene-specificity of the observed phenotype, we conducted rescue experiments via mRNA overexpression. qPCR and western blot analysis confirmed robust Ace2 overexpression following mRNA injection (SI Appendix, Fig. S3 B and C). Notably, overexpression of zebrafish ace2 efficiently restored the number of runx1⁺HEs at 28 hpf and kdrl⁺cmyb⁺ HE/HSPCs at 30 hpf in both ace2 morphants and mutants (Fig. 2 F and G), demonstrating that the hematopoietic defects were specifically attributable to ace2 loss.

To determine whether this regulatory function is evolutionarily conserved, we performed parallel rescue assays using mouse Ace2 and human ACE2 mRNA (SI Appendix, Fig. S3 D–G). Remarkably, overexpression of either ortholog similarly restored HE number and rescued the kdrl⁺cmyb⁺ HE/HSPC population in ace2-deficient embryos (Fig. 2 F and G), indicating functional conservation of ACE2 across vertebrate species. Taken together, these findings demonstrate that ace2 exerts a gene-specific and evolutionarily conserved regulatory function in HE specification, an essential upstream event in definitive HSPC development.

Ace2 Sustains HE Specification in an Endothelial Cell-Dependent Manner.

Since definitive HSPCs arise from arterial endothelium, we next investigate whether vascular abnormalities underlie the impaired HE specification observed in ace2-deficient embryos. We examined the expression of pan-endothelial markers (etv2, cdh5, and kdrl) and arterial/venous identity genes (ephrinB2, dll4, deltaC/dlc, tbx20, and msr) at 24 hpf. Both WISH and qPCR revealed a significant downregulation of these vascular genes in ace2 mutants and morphants compared to their siblings and control (Fig. 3 A and B and SI Appendix, Fig. S4 A and B), indicating compromised vascular development and arterial specification. To further assess vascular integrity, we performed FACS on the trunk from Tg(kdrl:mCherry) embryos. This analysis showed a reduction in the number of kdrl⁺ endothelial cells in ace2 morphants (Fig. 3C), suggesting that Ace2 deficiency results in a modest but meaningful impairment of endothelial homeostasis rather than overt vessel loss or collapse.

Fig. 3.

HE defects are attributed to the dorsal aorta disorder. (A) WISH showing the expression of etv2, kdrl, and dll4 in the AGM at 24 hpf of ace2^+/+ and ace2^−/−. (B) qPCR showing the expression of etv2, cdh5, kdrl, ephrinB2, dll4, dlc, and tbx20 in ace2^+/+ and ace2^−/− at 24 hpf. (C) Flow cytometry and statistical analysis of Tg(fli1a:EGFP) showing a reduced number of fli1a⁺ ECs in ace2 morphants at 28 hpf. (D) WISH and statistical data showing the expression of runx1 in control, ace2 morphants, and ace2 morphants injected with fli1a-mis-ace2-EGFP at 36 hpf. (E) WISH and statistical data showing the expression of runx1 in control, ace2 morphants, and ace2 morphants injected with runx1-mis-ace2-EGFP at 36 hpf. (F) WISH and statistical data showing the expression of runx1 in control, ace2 morphants, and ace2 morphants injected with fmylz-mis-ace2-EGFP at 36 hpf. [Scale bar, (A, D, E, and F), 50 μm.] The results are represented as means ± SD. Student’s t test, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, and n.s., not significant.

To determine whether the reduction in HE was caused by changed in endothelial cell survival or proliferation, we performed terminal TUNEL and BrdU assays. A significant increase in TUNEL⁺ endothelial cells (TUNEL⁺ fli1a⁺) was observed in ace2 mutants as early as 22 hpf (SI Appendix, Fig. S4C). In contrast, the proportion of proliferating endothelial cells (BrdU⁺ fli1a⁺) in the AGM was comparable between ace2 mutants and siblings (SI Appendix, Fig. S4D), indicating that the observed HE defects are due to altered cell death but not proliferation.

To test whether Ace2 regulates HE specification in an endothelial cell-autonomous manner, we generated tissue-specific ace2 overexpression construct driven by 1) the endothelial cell-specific fli1a promoter, 2) HSPC-specific runx1-enhancer and 3) muscle cell-specific fmylz promoter. Successful overexpression was validated by confocal imaging and western blot (SI Appendix, Fig. S4 E–J). Notably, only endothelial cell-specific ace2 overexpression restored HE formation in ace2 morphants, while HSPC- and muscle-specific overexpression failed to rescue the defect (Fig. 3 D–F). Collectively, these findings indicate that Ace2 is required to preserve endothelial integrity and arterial identity, and functions in a cell-autonomous manner within endothelial cells to sustain HE specification during early hematopoietic development.

EC-Derived NLRP3 Signaling Mediates Ace2-Dependent Regulation of HE Specification.

Although the above findings establish ace2 as an endothelial cell-autonomous regulator of HE specification and HSPC emergence, the underlying mechanism remained unclear. To investigate how the ace2 loss impairs HE development at the molecular level, we performed Smart-seq2 transcriptome profiling of FACS-sorted endothelial cells (kdrl:mCherry⁺) from the control and ace2 morphants at 24 hpf (Fig. 4A). Differential gene expression analysis revealed a clear separation between control and ace2-deficient endothelial cells (SI Appendix, Fig. S5A), identifying that 1,100 significantly upregulated genes, and 908 downregulated genes in ace2-deficient ECs (Fig. 4B). To identify the pathways perturbed by ace2 deficiency, we conducted KEGG pathway enrichment analysis, which revealed a marked downregulation of genes associated with the Notch signaling pathway, consistent with impaired HE specification (Fig. 4C). Notably, NOD-like receptor (NLR) signaling pathway exhibited a significantly increased activity in ace2-deficient endothelial cells (Fig. 4C). Although NLR signaling has been implicated in organ homeostasis and autoimmune disease (26–29); however, its role in developmental hematopoiesis remains largely unexplored. We first examined the expression of NLR family genes in ace2-deficient endothelial cells at 28 hpf by qPCR. Consistent with transcriptomic results, ace2 knockdown led to a marked upregulation of NLR-related genes, with nlrp3 showing particularly strong induction (Fig. 4D). Western blot analysis further confirmed elevated Nlrp3 protein levels in ace2-deficient embryos at 28 hpf (Fig. 4E), implicating NLRP3 signaling as a potential mediator of ace2-mediated hematopoietic defects.

Fig. 4.

EC-derived NLR signaling is responsible for the ace2-mediated regulation on HE specification. (A) Schematic representation of the strategies used for cell preparation and Smart-seq2 at 24 hpf. ECs were sorted from Tg(kdrl:mCherry) embryos. (B) Volcano plot showing DEGs between ECs from Smart-seq2 transcriptomic dataset in control and ace2 morphants. (C) KEGG enrichment analysis showing the upregulated and downregulated expression of genes in ace2-deficient ECs. (D) qPCR showing the expression of nlrp1, nlrp3, nod1, nod2, nlrc3, nlrc5, and pycad in control and ace2-deficient ECs at 28 hpf. (E) Western blot showing the protein level of Nlrp3 in control and ace2 morphants at 28 hpf. (F) WISH showing the expression of kdrl, dll4, and dlc in the AGM at 24 hpf of ace2^+/+nlrp3^+/+, ace2^−/− nlrp3^+/+, and ace2^−/− nlrp3^−/−. (G) qPCR showing the expression of etv2, cdh5, kdrl, ephrinB2, dll4, dlc, and tbx20 in ace2^+/+nlrp3^+/+, ace2^−/− nlrp3^+/+, and ace2^−/− nlrp3^−/− at 24 hpf. (H) WISH and statistical data showing the expression of runx1 in the AGM at 24 hpf of ace2^+/+nlrp3^+/+, ace2^−/− nlrp3^+/+, and ace2^−/− nlrp3^−/−. (I) qPCR showing the expression of runx1 at 24 hpf, 28 hpf, and 36 hpf in ace2^+/+nlrp3^+/+, ace2^−/− nlrp3^+/+, and ace2^−/− nlrp3^−/−. (J) WISH and statistical data showing the expression of cmyb in the AGM at 36 hpf of control, ace2 morphant, and nlrp3^−/− embryos injected with ace2 morpholino. (K) Confocal imaging and statistical data of Tg(cmyb:eGFP/kdrl:mCherry) showing the number of cmyb⁺kdrl⁺ cells in the AGM at 36 hpf of control, ace2 morphant, and nlrp3^−/− embryos injected with ace2 morpholino. (L) WISH and statistical data showing the expression of runx1 in the AGM at 36 hpf of ace2^+/+, ace2^−/−, ace2^−/− treated with MCC950 and ace2^−/− treated with UK5099. (M) WISH and statistical data showing the expression of cmyb in the AGM at 36 hpf of ace2^+/+, ace2^−/−, ace2^−/− treated with MCC950 and ace2^−/− treated with UK5099. (N) Confocal imaging and statistical data of Tg(cmyb:eGFP/kdrl:mCherry) showing the number of cmyb⁺kdrl⁺ cells in the AGM at 36 hpf of ace2^+/+, ace2^−/−, ace2^−/− treated with MCC950 and ace2^−/− treated with UK5099. [Scale bar, (F, H, and J–N), 50 μm.] The results are represented as means ± SD. Student’s t test, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, and n.s., not significant.

To determine whether nlrp3 upregulation is functionally responsible for the impaired HE specification in ace2-deficient embryos, we employed both genetic and pharmacological approaches. First, we generated a nlrp3 genetic mutant carrying an 11 bp deletion (GTTCTGACCAA) and 1 bp substitution (A to C) (SI Appendix, Fig. S5 B and C), and validated the loss of nlrp3 expression by qPCR and western blot (SI Appendix, Fig. S5 D and E). Strikingly, WISH analysis showed that the vascular gene expression (kdrl, dll4, and dlc) in the trunk region was largely rescued in ace2^−/−nlrp3^−/− double knockout embryos (Fig. 4F), which was further confirmed by qPCR (Fig. 4G). To further assess the rescue of HE specification, runx1 expression was examined from 24 hpf to 36 hpf. WISH and qPCR showed a restoration of HE gene expression in double knockout ace2^−/−nlrp3^−/− embryos (Fig. 4 H and I). WISH and confocal imaging confirmed that the reduction in the number of cmyb⁺ HSPC and kdrl⁺cmyb⁺ HE was efficiently rescued upon ace2 knockdown in nlrp3^−/− (Fig. 4 J and K). In parallel, we treated ace2 mutant embryos with two potent and specific NLRP3 inhibitors, MCC950 and UK5099 (30), during the HE specification window (22 hpf to 36 hpf). We observed a robust rescue effect of HE/HSPC number in AGM of ace2^−/− embryos following MCC950 treatment, as shown by WISH and live imaging (Fig. 4 L–N). Similarly, UK5099 treatment restored HE and HSPC numbers in ace2 morphants, as verified by WISH (Fig. 4 L–N). Taken together, these results suggest that NLRP3 signaling functions downstream of ace2 to regulate HE specification.

Loss of Ace2 Triggers NLRP3–Caspase-1-Dependent Pyroptosis and Impairs HSPC Formation.

To elucidate how Ace2 deficiency triggers NLRP3 signaling activation and disrupts HSPC development, we investigated the downstream consequences of Nlrp3 activation. As a member of the NLR family, cytosolic NLRP3 forms an inflammasome complex that activates Caspase-1, leading to the cleavage and secretion of proinflammatory cytokines such as IL-1b and IL-8 and cleavage of Gasdermin E (GSDME) to induce pyroptosis, a lytic and proinflammatory form of programmed cell death (31–33). To determine whether pyroptosis contributes to the hematopoietic defects in ace2-deficient embryos, we performed gene set enrichment analysis (GSEA) using Smart-seq2 data. GSEA revealed a strong upregulation of inflammatory gene signatures—a hallmark of pyroptosis—in ace2-deficient endothelial cells (Fig. 5A). This was supported by qPCR analysis, which demonstrated significant upregulation of inflammatory-related genes (Fig. 5B). Additionally, both protein and transcript levels of key executors of pyroptosis, including cleaved Caspase-1, pro- and cleaved-IL-1β, and the transcript of Gsdmeb, were markedly elevated in ace2-deficient endothelial cells (Fig. 5 C and D), indicating a rubout activation of the pyroptotic pathway. To test the functional relevance of Caspase activity, we treated ace2-deficient embryos with Z-VAD-FMK (pan-Caspase inhibitor) and Belnacasan (Caspase-1-specific inhibitor). Both treatments significantly suppressed inflammatory gene upregulation (Fig. 5E) and restored HE and HSPC numbers, as assessed by WISH and live imaging (Fig. 5 F and G), highlighting Caspase-1 activation as a critical mediator of the hematopoietic defects. To determine whether the link between Ace2 and Caspase-1-mediated pyroptosis is mediated by NLRP3, we inhibited NLRP3 pharmacologically and genetically. Western blots and qPCR demonstrated that both chemical inhibition and nlrp3 knockout efficiently attenuated pyroptotic signaling in the ace2-deficient embryos (SI Appendix, Fig. S6 A–D).

Fig. 5.

NLRP3 drives Caspase-1-dependent pyroptosis that maintains HE survival. (A) Gene Set Enrichment Analysis showing the enrichment plots for the enriched pathways in the control EC versus ace2-deficient EC. (B) qPCR showing the expression of il1b, il4, il6, il10, il11b, il12a, and il34 in control and ace2-deficient ECs at 28 hpf. (C) Western blot showing the protein level of Nlrp3, Caspase-1, cleaved Caspase-1, IL-1β, and cleaved IL-1β in ace2^+/+ and ace2^−/− at 36 hpf. (D) qPCR results showing the expression of nlrp3, caspa, gsdmeb, and il-1β in ace2^+/+ and ace2^−/− at 36 hpf. (E) qPCR results showing the expression of il1b, il4, and il6 in control, ace2 morphants, ace2 morphants treated with Z-VAD-FMK and ace2 morphants treated with Belnacasan at 36 hpf. (F) WISH and statistical data showing the expression of runx1 in the AGM at 28 hpf of control, ace2 morphants, ace2 morphants treated with Z-VAD-FMK and ace2 morphants treated with Belnacasan. (G) Confocal imaging and statistical data of Tg(cmyb:eGFP/kdrl:mCherry) showing the number of cmyb⁺kdrl⁺ cells in the AGM at 36 hpf of ace2^+/+, ace2^−/−, ace2^−/− treated with Z-VAD-FMK and ace2^−/− treated with Belnacasan. (H) WISH and statistical data showing the expression of runx1 in ace2^+/+, ace2^−/−, ace2^−/− treated with Disulfiram, Necrostatin-1, 3MA, and Ferrostatin-1 at 24 hpf. (I) WISH and statistical data showing the expression of cmyb in ace2^+/+, ace2^−/−, ace2^−/− treated with Disulfiram, Necrostatin-1, 3MA, and Ferrostatin-1 at 36 hpf. (J) Confocal imaging and statistical data of Tg(cmyb:eGFP/kdrl:mCherry) showing the number of cmyb⁺kdrl⁺ cells in the AGM of control, ace2 morphants, and ace2 morphants treated with Disulfiram, Necrostatin-1, 3MA, and Ferrostatin-1 at 36 hpf. (K) qPCR showing the expression of il1b in the AGM of control, ace2 morphants, and ace2 morphants-treated with Disulfiram at 36 hpf. (L) qPCR showing the expression of nlrp3, caspa, gsdmeb, and il1b in ace2^+/+, ace2^−/−, ace2^−/−treated with Ang-(1–7) at 36 hpf. (M) WISH and statistical data showing the expression of cmyb in ace2^+/+, ace2^−/−, ace2^−/− treated with Ang-(1–7). [Scale bar, (F–J and M), 50 μm.] The results are represented as means ± SD. Student’s t test, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, and n.s., not significant.

Given that NLRP3 can be involved in multiple cell death modalities (34), we next assessed which pathway was primarily responsible for HSPC impairment. We administered embryos with inhibitors of various cell death pathways: 3-MA (autophagy inhibitor) (35), necrostatin-1 (necroptosis inhibitor) (36), ferrostatin-1 (ferroptosis inhibitor) (37), and Disulfiram (pyroptosis inhibitor). Among these, only Disulfiram effectively rescued the HE and HSPC phenotype in ace2-deficient embryos, as confirmed by WISH and confocal imaging (Fig. 5 H–J), and a significantly reduced il1b expression, as determined by qPCR (Fig. 5K). In contrast, the other inhibitors had no apparent effects on HE specification and HSPC development (Fig. 5 H–J), suggesting that pyroptosis, rather than apoptosis, autophagy, necroptosis, or ferroptosis, is the primary contributor to the HSPC impairment in the absence of Ace2. Taken together, these findings demonstrate that ace2 safeguards HE and HSPC development by restraining NLRP3–Caspase-1-dependent pyroptosis.

Ace2 Restricts NLRP3 Signaling Activation Through Angiotensin-(1–7).

We next explored how Ace2 enzymatic activity regulates NLRP3 signaling during embryonic hematopoiesis. As a monocarboxypeptidase within RAS, Ace2 hydrolyzes proinflammatory Ang II into Ang-(1–7), a peptide known for its anti-inflammatory properties (38). Previous studies have shown that Ang-(1–7) can attenuate NLRP3 inflammasome activation and limit inflammation in fibrotic diseases, including liver fibrosis (39). Based on these findings, we speculated that Ang-(1–7) mediates the anti-inflammatory effects of Ace2 by suppressing NLRP3-Caspase-1 signaling. To test this, we treated ace2 mutants with exogenous Ang-(1–7) from 28hpf to 36hpf. Strikingly, Ang-(1–7) supplementation significantly reduced nlrp3 transcript levels in ace2 mutants (Fig. 5L), accompanied by downregulation of caspa and inflammatory cytokines, such as il1b, as assessed by qPCR (Fig. 5L). Importantly, WISH analysis revealed a robust restoration of HSPC formation in the AGM of Ang-(1–7)-treated ace2 mutants (Fig. 5M), indicating a functional rescue of hematopoiesis. Collectively, these results support a model in which endothelial Ace2 maintains hematopoietic niche integrity by generating Ang-(1–7), which acts to suppress NLRP3–Caspase-1-mediated pyroptotic signaling. This enzymatic suppression of inflammatory stress is essential for preserving the endothelial microenvironment required for proper HSPC development.

The Conserved Role of ACE2 in Mammalian HSPC Development.

To determine whether the role of ACE2 in hematopoiesis is evolutionarily conserved in mammals, we first examined its expression in mouse embryos in the AGM of embryonic day (E) 10.5 mouse embryos. qPCR analysis revealed that Ace2 expression is enriched in endothelial cells (ECs, CD31⁺CD44^–Ter119^–), HEs (CD31⁺CD44⁺CD45^–Ter119^–), compared to Type 1 prehematopoietic stem cells (T1-pre-HSC, CD31⁺c-Kit⁺CD45^–Ter119^–) and Type 1 prehematopoietic stem cells (T2-pre-HSC, CD31⁺c-Kit⁺CD45⁺Ter119^–) isolated from the same region (40) (Fig. 6A and SI Appendix, Fig. S6E). This enrichment suggests a potential role for Ace2 in regulating the EHT and HSPC emergence in mice.

Fig. 6.

The conserved role of ACE2 in mammalian HSPC development. (A) qPCR showing the expression of Ace2 in T1 pre-HSC, T2 pre-HSC, EC, and HE from mouse AGM at E10.5. (B) Schematic illustration of Ace2 siRNA treatment for EHT induction in E10.5 mouse AGM HEs. (C) Morphological dynamics of CD31⁺CD44⁺CD45^–Ter119^– cells from E10.5 AGM following Ace2 knockdown and coculture on OP9-DL1 cells. (D) Morphological observation of 5-d cocultured CD31⁺CD44⁺CD45^–Ter119^– cells on OP9- DL1 cells showing the generation of round-shaped cells when Ace2 knockdown. (E and F) Flow cytometry and CFU analysis of 5-d cocultured CD31⁺CD44⁺CD45^– Ter119^– cells on OP9-DL1 cells showing the generation of CD45⁺ hematopoietic cells with hematopoietic colony-forming ability when Ace2 knockdown. [Scale bar, (C and D), 50 μm.] Biological replicates (A–F), n = 3. The results are represented as means ± SD. Student’s t test, *P < 0.05, **P < 0.01, ***P < 0.001, and n.s., not significant.

To functionally evaluate this role, we performed ex vivo short interfering RNAs (siRNA)-mediated loss-of-function analysis of Ace2 using E10.5 AGM-derived CD31⁺CD44⁺Ter119^– HEs, followed by time-lapse imaging, flow cytometry, and colony-forming units’ culture (CFU) assays (Fig. 6B and SI Appendix, Fig. S6F). In siControl (siCtrl)-treated AGM-derived HEs, we observed the expected flattened-to-round morphological transition and successful emergence of HSPCs. In contrast, siAce2-treated cells displayed impaired morphological conversion and defective HSPC generation (Fig. 6C). Eventually, Ace2 knockdown in E10.5 AGM-derived HEs markedly reduced the generation of round-shaped CD45⁺ hematopoietic cells (Fig. 6 D and E) and the formation of colonies (Fig. 6F). Together, these findings demonstrate that ACE2 is functionally required for EHT and definitive HSPC generation in the mammalian AGM, underscoring its evolutionarily conserved role in vertebrate hematopoiesis.

Discussion

In this study, we identify Ace2 as a critical regulator of embryonic hematopoiesis through its suppression of NLRP3-Caspase-1-dependent pyroptosis. As a key enzyme in the RAS pathway, Ace2 converts proinflammatory Angiotensin II into anti-inflammatory Angiotensin-(1–7), which in turn inhibits NLRP3 inflammasome activation. This prevents Caspase-1-mediated cleavage of Gasdermin D/E and the induction of pyroptosis—a lytic form of programmed cell death triggered by excessive inflammation. Loss of Ace2 leads to hyperactivation of pyroptotic signaling in endothelial cells, disrupting the HE niche and impairing HSPC emergence. Remarkably, exogenous Ang-(1–7) supplementation restores inflammatory balance and rescues HSPC development in ace2-deficient embryos. These findings uncover a previously unrecognized function of the ACE2/Ang-(1–7)-NLRP3 axis in preserving the vascular niche integrity and modulating inflammation tone required for hematopoietic development.

Well-controlled innate inflammatory signaling has emerged as a pivotal regulator of HSPC emergence during vertebrate embryogenesis. Accumulating evidence demonstrates that innate inflammatory cytokines—including tumor necrosis factor a (TNF-a), interferon (IFN), Toll-like receptor 4 (TLR4), and cGAS-STING-mediated inflammatory pathways—are essential for HSPC development in zebrafish, mice, and humans (41–45). These signals are largely derived from primitive myeloid cells, such as neutrophils and macrophages, that circulate through the embryonic vasculature and help shape the inflammatory microenvironment of the AGM (26–28, 41, 46, 47). Collectively, these studies suggest that the AGM constitutes an inflammatory microenvironment during definitive HSPC generation. However, the role of inflammation in hematopoiesis is double-edged: While acute or moderate inflammation can activate HSCs and protect them from stress-induced terminal differentiation in the adult bone marrow (29, 31, 32), chronic or excessive inflammation is detrimental—resulting in HSC exhaustion, myeloid skewing, and functional decline during aging (33, 48). Those observations underscore the existence of and importance of intrinsic protective mechanisms that fine-tune inflammatory responses to ensure proper HSPC emergence and maintenance. Despite progress in identifying proinflammatory cues required for HSPC production, our understanding of how endothelial cells orchestrate and restrain inflammatory cascades to sustain embryonic HSPC development remains incomplete. In this study, we identify a critical role for Ace2 in mitigating inflammatory stress. Specifically, loss of ace2 in endothelial cells leads to hyperactivation of intrinsic inflammatory signaling, triggering a self-protection mechanism—pyroptosis, a form of programmed cell death driven by excessive inflammation (49). In the context of embryonic hematopoiesis, such cell death disrupts vascular integrity and compromises HE specification. Importantly, these effects occur in the absence of exogenous infection, indicating that ACE2 functions as a homeostatic gatekeeper of inflammatory tone even under basal physiological conditions. (50). Although the COVID-19 pandemic has brought widespread attention to ACE2 as the viral entry receptor for coronaviruses, with its dysregulation linked to cytokine storm syndromes, our findings uncover a distinct developmental role for Ace2 in modulating immune responses independent of viral infection. The finding that ace2 deficiency alone is sufficient to provoke inflammatory toxicity and pyroptosis underscores the importance of tightly regulated inflammatory suppression in early vascular and hematopoietic development. Notably, our data (Fig. 5M) also showed that excessive supplementation of Ang-(1–7) leads to a reduction in cmyb expression, suggesting that when the Ang-(1–7)-mediated suppression of Nlrp3-driven inflammation is enhanced beyond an optimal level, the resulting overly dampened inflammatory state becomes detrimental to HSPC production. This observation further supports our conclusion that a balanced inflammatory tone is essential, where both insufficient and excessive inflammation can impair proper HSPC emergence and development. Therefore, a deeper understanding of these dynamic regulatory mechanisms may offer insights for developing stem cell–based therapeutic strategies.

Although pyroptosis has been widely studied in infection and chronic inflammatory diseases, its impact on developmental processes is less well-understood. Here, we provide compelling evidence that reveals a previously unrecognized developmental role of Ace2 in safeguarding embryonic hematopoiesis through modulation of pyroptotic signaling. Specifically, our findings highlight how tightly regulated inflammatory signaling—particularly via the NLRP3–Caspase-1–GSDME axis—is essential for maintaining the endothelial microenvironment that supports HE specification and HSPC emergence. This work underscores the importance of the ACE2/Ang-(1–7) axis in restraining inflammatory cell death during early hematopoietic development, providing a mechanistic link between immune modulation, vascular integrity, and stem cell niche function.

Our in vivo functional studies were conducted in zebrafish; nevertheless, ex vivo AGM assays confirm that Ace2 deficiency impairs EHT and functional HSPC generation in the mouse. Future in vivo validation using conditional mouse knockout models will be essential to define the precise spatiotemporal and cell-autonomous roles of Ace2 within endothelial and hematopoietic compartments. In addition, siRNA-based loss-of-function assays cannot fully capture long-term effects on HSPC self-renewal and multilineage reconstitution potential, which will require transplantation-based functional studies in mammals. It will also be important to determine whether pharmacological modulation of the ACE2/Ang-(1–7)/NLRP3 axis can promote hematopoietic recovery under inflammatory or stress-induced conditions. Together, these findings establish an evolutionarily conserved role of ACE2 in hematopoiesis in vertebrates and highlight its potential therapeutic relevance in inflammation-associated hematopoietic disorders.

In summary, our findings highlight a conserved and previously unrecognized role of Ace2 in embryonic hematopoiesis, linking vascular inflammation control to stem cell development. Understanding how the ACE2/Ang-(1–7)/NLRP3 axis modulates hematopoietic niche opens broad avenues for therapeutic intervention in developmental hematopoietic disorders, stem cell dysfunction, and inflammatory diseases.

Materials and Methods

Ethics Statement.

All zebrafish experiments were conducted under the approval of the Animal Welfare Committee of Institute of Zoology, Chinese Academy of Sciences (license number: IOZ-IACUC-2021-017). The mice used in this study were bred under specific pathogen–free conditions and handled in accordance with institutional guidelines (animal license number: SYXK (Beijing) 2018-0021).

Zebrafish and Mice.

Zebrafish strains including AB, and transgenic line (Tg)(CD41:GFP) (51), Tg(gata1:DsRed) (52), Tg(lyz:DsRsd) (53), (runx1:enGFP) (54), Tg(cmyb:eGFP) (55), Tg(mpo:GFP) (56), Tg(rag2:DsRed) (57), Tg(kdrl:mCherry) (8), Tg(fli1a:EGFP) (58), Tg(HuC;Gal4-VP-16;UAS: mCherry-MA) (59) and Tg(myhz2:GFP)(60) were raised at 28 ^oC system water and staged as previously described. Wild-type C57BL/6 mice were purchased from SPF (Beijing) Biotechnology Co., Ltd. Embryonic day (E) 0 mice is the day that the vaginal plug was detected. This study was approved by the Ethical Review Committee in the Institute of Zoology, Chinese Academy of Sciences.

MOs, mRNA Synthesis, Plasmid Construction, and Microinjection.

The antisense morpholinos (MOs) of ace2 were purchased from GeneTools, and the MOs listed in SI Appendix, Table S1. ace2 aMO and ace2 spMO were injected at one-cell stage zebrafish embryos at the yolk/blastomere boundary. For the overexpression experiments in the whole embryos, mRNAs of zebrafish ace2, mouse Ace2, and human ACE2 were synthesized using the mMESSAGE mMACHINETM mRNA transcription synthesis kit (Invitrogen; AM1340) and cleaned up using the RNAclean Kit (TIANGEN; DP412), respectively. 150 pg of mRNA was injected into the one-cell stage embryos. For the EC, HSPC, and muscle-specific overexpression experiments, the full-length cDNA of ace2 was cloned and assembled into the pDestTol2pA2 vector with a fli1a and fmylz promotor, runx1 enhancer with a heatshock minimal promoter and an EGFP reporter to generate pDestTol2pA2-fli1a-mis-ace2-v2a-EGFP, pDestTol2pA2-runx1-mis-ace2-v2a-EGFP, and pDestTol2pA2-fmylz-mis-ace2-v2a-EGFP plasmid. The plasmid (50 pg) and Tol2 mRNA (25 pg) were coinjected into one-cell-stage zebrafish embryos at the yolk/blastomere boundary.

Generation of ace2 and nlrp3 Zebrafish Mutant Using CRISPR/Cas9.

The ace2 and nlrp3 mutants were generated by CRISPR/Cas9 as previously described (61). The guide RNAs of ace2 and nlrp3 mutants listed in SI Appendix, Table S1. The primers for genotyping are provided in SI Appendix, Table S2.

Whole Mount In Situ Hybridization.

The RNA probe targeting ace2 was first labeled by digoxigenin (DIG) with DIG RNA labeling mix (Roche, 11093274910) and then transcribed by T7 polymerase. WISH for zebrafish embryos was performed with probes, including ace2, runx1, cmyb, pu.1, ikaros, rag1, foxn1, gata2b, gfi1aa, etv2, dll4, kdrl, msr, and dlc as previously described (61). A Nikon SMZ 1500 microscope was used to collect figures. The primers for probe synthesis are provided in SI Appendix, Table S2.

TUNEL and BrdU Staining.

TUNEL staining was performed using In Situ Death Detection Kit, TMR red (Roche, 12156792910) as described previously (62). BrdU staining was performed as described previously (63). The fluorescence images were acquired by a Nikon confocal A1 laser microscope. The antibodies used for Western blotting are listed in SI Appendix, Table S3.

Chemical Treatment.

Embryos were immersed in 10uM of MCC950 (MCE, HY-12815A), UK-5099 (MCE, HY-15475), Z-VAD(OMe)-FMK (MCE, HY-16658), Belnacasan (MCE, HY-13205), Disulfiram (MCE, HY-B0240), Necrostatin-1 (MCE, HY-15760), 3-Methyladenine (3-MA) (MCE, HY-19312), Ferrostatin-1 (MCE, HY-100579), or Talfirastide [Ang-(1–7)] (MCE, HY-12403) from 22 hpf to the timepoint that were analyzed in each experiments.

Small Interference RNA (siRNA) Transfection and CFU Assay.

For Ace2 siRNA knockdown assay, the siRNAs for Ace2 (NM_001130513.1) were designed and synthesized by GenePharma. CD44⁺CD31⁺ CD45^–Ter119^– cells were sorted from E10.5 AGMs, cultured in medium [α-MEM, 10% FBS, and cytokine cocktail (SCF, IL-3, and Flt3L)] and transfected with siRNAs using LipofectamineÔ. The following siRNA transduction and CFU were performed as previously described (40). The siRNA sequence is provided in SI Appendix, Table S4.

Confocal Microscopy and Time-Lapse Imaging.

For the confocal microscopy, zebrafish embryos were mounted in 1% low-melting agarose, and transgenic zebrafish embryos were imaged by a Nikon confocal A1 laser microscope with a 20× objective. The imaging was edited by ImageJ and Photoshop CS6. For the morphology change of mouse HE, CD44⁺CD31⁺CD45^–Ter119^–cells were isolated from E10.5 mouse AGM and then cocultured on OP9-DL1 cells. Their dynamics were subsequently imaged by time-lapse microscopy using the JuLI stage real-time cell history recorder (NanoEnTek).

Flow Cytometry.

For zebrafish embryos, the trunk region of (Tg)(CD41:GFP/gata1:DsRed) at 2 dpf, Tg(kdrl:mCherry) at 24 hpf, Tg(lyz:DsRsd) at 24 hpf, Tg(HuC;Gal4-VP-16;UAS: mCherry-MA) at 24 hpf, Tg(myhz2:GFP) at 24 hpf, control and ace2 morphants at 28 hpf with Tg(kdrl:mCherry) or Tg(fli1a:EGFP) background were collected and washed by PBST. After digesting into single-cell suspension using 0.5% trypsin, the reaction was stopped by adding fetal calf serum up to 10%. The cells were filtered through a 300 Mesh nylon cell strainer to make single-cell suspension and sorted using MoFlo XDP (Beckman Coulter). For mouse embryos, before or after siRNA transfected, single-cell suspension from the E10.5 mouse AGM was stained with antibodies for 30 min at 4 °C. 7-AAD (Biolegend, 420404) viability staining solution was used to exclude dead cells. Flow cytometry was performed using BD LSRFortessa™ X-20 (BD Biosciences), and the data were analyzed by FlowJo_V10. The antibodies used for Western blotting are listed in SI Appendix, Table S3.

Real-Time Quantitative PCR (qPCR).

For zebrafish and mouse embryos, total RNAs of flow-sorted cells were extracted by RNeasy MicroKit (QIAGEN, 74004) and then reversely transcribed and amplified as previously described (64). Total RNAs of the trunk region or tail region (20-40 embryos pooled for each sample) were extracted by TRIzol and reverse-transcribed using M-MLV Reverse Transcriptase. The cDNA was diluted five-fold to be used as the template for qPCR amplification (61). The qPCR was performed by the Bio-Rad system (CFX96), and data were analyzed by GraphPad Prism 8.0. Relative expression was normalized to β-actin using the formula 2^Δ[Ct(target)–Ct(β-actin)]. The primers for qPCR experiment are provided in SI Appendix, Table S2.

Western Blot.

Protein samples were extracted from zebrafish ace2^+/+, ace2^–/–, control embryos, ace2 morphants, zebrafish ace2 mRNA-, mouse Ace2 mRNA- or human ACE2 mRNA-overexpressed embryos, fli1a-mis-ace2-EGFP, runx1-mis-ace2-EGFP and fmylz-mis-ace2-EGFP injected embryos, MCC950-treated control embryos and ace2 morphants, nlrp3^+/+, nlrp3^−/−, nlrp3^−/− embryos-injected with ace2 morpholino at 36 hpf, as previously described. The antibodies used for Western blotting are listed in SI Appendix, Table S3.

Smart-seq2 and Data Processing.

5000 kdrl:mCherry⁺ cells of control and ace2 morphants were sorted per group from Tg(kdrl:mCherry) embryos at 24 hpf using MoFlo XDP (Beckman). We followed Smart-Seq2 steps to perform low-input cell transcriptome amplification (65). Libraries were sequenced to generate 150-bp paired-end reads on an Illumina Novaseq 6000 platform. Then, the original RNA-seq reads used Trimmomatic (v0.39) to filter low-quality reads and trim the linker sequence. Low-quality bases (usually at the 3′end) and adapter sequences can be removed at this preprocessing step. Reads were mapped back to the zebrafish genome (danRer11 assembly) using HISAT2 (version 2.1.0) (66). FPKM and read count for each gene were calculated using StringTie (version 1.3.3b) (67). Differential expressed gene were identified using DESeq2 (version 1.10.1 under R version 3.2.3) (68). Genes with an adjust P-value < 0.05 and absolute value of log₂ fold change > 1 were defined as differential expressed.

KEGG Enrichment.

KEGG enrichment analyses were performed using the enrichment analysis tool clusterProfiler(version 4.6.1) (69). The enriched KEGG pathway of ligand receptor genes in our data was obtained using function “enrichKEGG”, the “pvalueCutoff” was set as 0.05, and “qvalueCutoff” was set as 0.2.

GSEA.

GSEA (version 4.0.3) (70) was performed to compare gene expression signatures in ECs of control versus ace2 morphants. The annotated gene sets were selected from the Molecular Signatures Database (MSigDB version 7.5).

Statistical Analysis.

GraphPad Prism 6 was used to analyze the statistical data of WISH (71), qPCR, confocal imaging, flow cytometry, and CFU culture (63). Values are presented as the means ± SD. Student’s unpaired two-tailed t-test was applied for comparisons unless otherwise indicated.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

Smart-seq2 data have been deposited in GSA (GSA: CRA026453) (72). Study data are included in the article and/or SI Appendix.