当前位置: X-MOL 学术Circulation › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Defining the Fetal Gene Program at Single-Cell Resolution in Pediatric Dilated Cardiomyopathy
Circulation ( IF 37.8 ) Pub Date : 2022-10-03 , DOI: 10.1161/circulationaha.121.057763
Neda R Mehdiabadi 1, 2, 3 , Choon Boon Sim 1, 2 , Belinda Phipson 1, 4, 5, 6 , Ravi K R Kalathur 1, 2 , Yuliangzi Sun 7 , Celine J Vivien 1, 2 , Menno Ter Huurne 1, 2 , Adam T Piers 1, 2 , James E Hudson 8 , Alicia Oshlack 1, 4, 5, 9 , Robert G Weintraub 2, 6, 10 , Igor E Konstantinov 2, 6, 11 , Nathan J Palpant 7 , David A Elliott 1, 2, 3, 6, 12 , Enzo R Porrello 1, 2, 12, 13
Affiliation  

A central dogma in cardiac biology is that the postnatal heart adopts a fetal-like transcriptional state in response to stress.1 Consequently, the so-called “fetal gene program” is frequently used as a surrogate marker of adverse cardiac remodeling and heart failure in preclinical models.2 Fetal gene reactivation in heart failure is traditionally studied in cardiomyocytes; however, the extent to which the fetal gene program is recapitulated in other cardiac cell types is unknown. Here we define the human fetal gene program at single-cell resolution in dilated cardiomyopathy (DCM), a common cause of heart failure in children and adults.


Single-nucleus RNA sequencing profiles of apical left ventricle tissue from fetal (19–20 weeks), nondiseased (ND; 4–14 years), and early-onset DCM samples (5–10 years) were captured (Figure [A]; sample processing and data analysis based on previous work3). Methodological details, sample details, source data (GSE185100), code, and quality control data including removal of doublets and ambient RNA are available at https://www.HeartExplorer.org. Informed consent was obtained from all human subjects with approval from the Royal Children’s Hospital Human Research Ethics Committee (HREC 38192 and HREC 36358).


Figure. snRNA-seq analysis of fetal, pediatric nondiseased (ND) and pediatric dilated cardiomypathy (DCM) hearts. A, Summary of experimental design showing total number of biological samples and nuclei sequenced, as well as the median number of genes and unique molecular identifiers (UMI) detected per nucleus. B, Heatmap showing expression of cell type–specific marker genes across broad cell types. C, Uniform Manifold Approximation and Projection (UMAP) plot of nuclei showing distinct clusters of cardiac cell types across fetal, nondiseased (ND), and dilated cardiomyopathy (DCM) samples. D, Bar plot of cell type proportions for each group. E, Bar plot showing the number of DE genes in each cell type that are upregulated (FDR <0.05, log2FC >1.5, red) or downregulated (FDR <0.05, log2FC < –1.5, blue) for DCM versus ND. FDR <0.05 with the correction procedure of Benjamini-Hochberg was used. F, Bar plot showing the number of reactivated fetal genes in each cell type that are upregulated (FDR <0.05, log2FC >1.5, red) or downregulated (FDR < 0.05, log2FC < –1.5, blue). FDR <0.05 with the correction procedure of Benjamini-Hochberg was used. G, Venn diagram showing the fetal gene program (FGP) in cardiomyocytes (CM; left) and heatmap illustrating the expression of the FGP in CM for each group (fetal, ND, DCM) including gene ontology terms for significantly regulated biological processes and associated genes with |log2FC| >2 (right). H, Venn diagram showing the FGP in fibroblasts (Fib; left) and heatmap illustrating the expression of the FGP in Fib for each group (fetal, ND, DCM) including gene ontology terms for significantly regulated biological processes and associated genes with |log2FC| >2 (right). I, Top, Venn diagram showing upregulated fetal genes that are shared between CM and cardiac Fib in DCM including gene ontology terms and associated genes for a significantly regulated biological process. Bottom, Venn diagram showing downregulated fetal genes that are shared between CM and cardiac Fib in DCM including gene ontology terms and associated genes for a significantly regulated biological process. J, UMAP plot showing 6 distinct subclusters of CM (CM1–CM6) from 3 groups (fetal, ND, DCM; left) and heatmap showing expression of the FGP across CM subclusters (right). K, UMAP plot showing 6 distinct subclusters of cardiac Fib (Fib1–Fib6) from 3 groups (fetal, ND, DCM; left) and heatmap showing expression of the FGP across Fib subclusters (right). CM (Prlf) indicates cardiomyocyte (proliferative); Endo, endothelial; Er, erythroid; FC, fold change; FDR, false discovery rate; pericyte; and Smc, smooth muscle cell.


On the basis of known marker genes,3 single-nucleus RNA sequencing analysis revealed 7 cell clusters across fetal, ND, and DCM samples including cardiomyocytes, fibroblasts, immune cells, smooth muscle cells, endothelial cells, neurons, and pericytes (Figure [B and C]). Postnatal cardiac maturation was associated with a significant increase in the relative proportion of fibroblasts and immune cells, accompanied by a significant decrease in the proportion of cardiomyocytes (Figure [D]; Benjamini-Hochberg false discovery rate <0.05 using propeller function in speckle R package). No statistically significant shifts in cellular composition were observed between DCM and ND hearts (Figure [D]). However, DCM samples were characterized by lower average expression levels of immune markers, suggesting that dysregulation of the immune system may be a feature of pediatric DCM.


To identify transcriptional pathways perturbed in DCM, we used pseudo-bulk profiling to define differentially expressed gene sets across shared cell clusters (Figure [E through H]). DCM-related gene expression alterations were predominantly observed in cardiomyocytes, fibroblasts, and immune cells (Figure [E]). To define the fetal gene program, we identified differentially expressed genes that were developmentally regulated between fetal and ND (fetal versus ND, false discovery rate <0.05, log2fold change < –1.5 or >1.5). This gene set was intersected with DCM-regulated differentially expressed genes (DCM versus ND, false discovery rate <0.05, log2fold change < –1.5 or >1.5) across all cell clusters, which showed reactivation of fetal genes predominantly in cardiomyocytes and fibroblasts (Figure [F]). Only a small fraction (~8%) of the fetal gene program was reinstated in cardiomyocytes in DCM (163 of 2079 developmentally regulated genes; Figure [G]). The fetal gene program in DCM cardiomyocytes was involved in upregulation of genes involved in muscle development and downregulation of genes implicated in the innate immune response (Figure [G]). Similarly, in fibroblasts, ~8% of the fetal transcriptome was re-engaged in DCM (139 of 1658 developmentally regulated genes (Figure [H]). The transcriptional response of DCM fibroblasts was associated with upregulation of a muscle differentiation program and downregulation of oxidative metabolic processes and the cell cycle (Figure [H]). It is interesting that among the most highly upregulated (log2fold change >2) fetal genes in DCM fibroblasts was MEOX1 (Figure [H]), a transcription factor recently identified as a central regulator of myofibroblast activation in the heart.4 In addition, a small subset of fetal genes (13 in total) overlapped between cardiomyocytes and fibroblasts in DCM (Figure [I]), suggesting this developmental transcriptional network is governed by shared regulatory mechanisms in both cell types.


Single-cell transcriptomics uniquely permits analysis of transcriptional heterogeneity across cell populations. We thus investigated whether the fetal gene program is broadly re-engaged in cardiomyocytes and fibroblasts or restricted to specific cell subpopulations in DCM. Subclusters of cardiomyocytes (Figure [J]) and fibroblasts (Figure [K]) were identified as previously described.3 The fetal gene program was ubiquitously engaged across all subclusters of cardiomyocytes (Figure [J]) or fibroblasts (Figure [K]) in DCM, suggesting that this transcriptional program is not restricted to a specialized subpopulation of cells in the disease state.


The single-cell transcriptomic data sets provided here permit the identification of cell type–specific gene dysregulation in DCM. In contrast with a previous single-cell transcriptional analysis of pediatric DCM,5 the current study included fetal and nondiseased control tissues, thus enabling precise cellular delineation of the fetal gene programs reactivated in DCM. Surprisingly, the phenomenon of fetal gene reactivation in DCM does not appear to be restricted to cardiomyocytes because a similar proportion of genes (<10%) adopts a fetal-like expression pattern in both cardiomyocytes and fibroblasts. This work provides a multicellular framework to identify the critical gene expression networks that underpin DCM disease pathogenesis in children. It will be important to determine whether these transcriptional mechanisms are also operative in other forms of heart failure including DCM in adults.


The authors gratefully acknowledge technical support and advice for single-nucleus RNA sequencing from the Translational Genomics Unit at the Murdoch Children’s Research Institute. The authors thank M. Burton from the Murdoch Children’s Research Institute for technical assistance and advice for fluorescence-activated cell sorting. Human fetal tissues were procured from Advanced Bioscience Resources. ND and DCM samples were obtained from the Melbourne Children’s Heart Tissue Bank. Informed consent was obtained from all human subjects, and studies were conducted in accordance with the approved protocols and guidelines from the National Health and Medical Research Council of Australia. E.R.P., D.A.E., N.R.M., and C.B.S. conceptualized the project. N.R.M., Y.S., C.B.S., B.P., C.J.V., M.t.H., A.O., N.J.P., D.A.E., and E.R.P. designed the experiments. C.B.S. performed sample preparation for sequencing. N.R.M., Y.S., C.B.S., R.K.R.K., and B.P. analyzed sequencing data. A.T.P., R.G.W., I.E.K., and J.E.H. contributed to patient recruitment, consenting, and biobanking. N.R.M., Y.S., N.J.P., D.A.E., and E.R.P. analyzed and interpreted data. E.R.P., D.A.E., and I.E.K. obtained funding and managed the project. N.R.M., D.A.E., and E.R.P. wrote the article. All authors reviewed and edited the article.


The authors acknowledge grant and fellowship support from the National Health and Medical Research Council of Australia (E.R.P., I.E.K., D.A.E.), the Heart Foundation of Australia (N.J.P), the Snow Medical Research Foundation (J.E.H.), the Stafford Fox Medical Research Foundation (E.R.P, D.A.E.), and the Royal Children’s Hospital Foundation (E.R.P., D.A.E., I.E.K., R.G.W.). The Novo Nordisk Foundation Center for Stem Cell Medicine (E.R.P., D.A.E.) is supported by Novo Nordisk Foundation grants (NNF21CC0073729). The Murdoch Children’s Research Institute is supported by the Victorian Government’s Operational Infrastructure Support Program.


Disclosures Drs Porrello and Hudson are cofounders, are scientific advisors, and hold equity in Dynomics, a biotechnology company focused on the development of heart failure therapeutics. The other authors report no conflicts.


Circulation is available at www.ahajournals.org/journal/circ


For Sources of Funding and Disclosures, see page 1107–1108.




中文翻译:

在小儿扩张型心肌病中以单细胞分辨率定义胎儿基因程序

心脏生物学的一个中心法则是,出生后的心脏在应对压力时采用类似胎儿的转录状态。1因此,所谓的“胎儿基因程序”经常被用作临床前模型中不良心脏重塑和心力衰竭的替代标志物。2传统上在心肌细胞中研究心力衰竭中的胎儿基因再激活;然而,胎儿基因程序在其他心脏细胞类型中的重现程度尚不清楚。在这里,我们在扩张型心肌病 (DCM) 中以单细胞分辨率定义人类胎儿基因程序,这是儿童和成人心力衰竭的常见原因。


捕获了来自胎儿(19-20 周)、未患病(ND;4-14 岁)和早发性 DCM 样本(5-10 岁)的左心室顶端组织的单核 RNA 测序图谱(图 [A];样本处理和数据分析基于以前的工作3)。方法学细节、样品细节、源数据 (GSE185100)、代码和质量控制数据,包括去除双峰和环境 RNA,请访问 https://www.HeartExplorer.org。经皇家儿童医院人类研究伦理委员会(HREC 38192 和 HREC 36358)批准,所有人类受试者均获得知情同意。


数字。 胎儿、小儿无病 (ND) 和小儿扩张型心肌病 (DCM) 心脏的 snRNA-seq 分析。A,实验设计总结,显示生物样本和测序的细胞核总数,以及每个细胞核检测到的基因和唯一分子标识符 (UMI) 的中位数。B,显示细胞类型特异性标记基因在广泛细胞类型中的表达的热图。C,细胞核的均匀流形近似和投影 (UMAP) 图,显示胎儿、未患病 (ND) 和扩张型心肌病 (DCM) 样本中不同的心肌细胞类型簇。D,每组细胞类型比例的条形图。, 条形图显示 DCM 与 ND 的每种细胞类型中上调(FDR <0.05,log 2 FC >1.5,红色)或下调(FDR <0.05,log 2 FC < –1.5,蓝色)的 DE 基因数量。使用 Benjamini-Hochberg 校正程序的 FDR <0.05。F,条形图显示每种细胞类型中被上调(FDR <0.05,log 2 FC >1.5,红色)或下调(FDR < 0.05,log 2 FC < –1.5,蓝色)的重新激活胎儿基因的数量。使用 Benjamini-Hochberg 校正程序的 FDR <0.05。G,显示心肌细胞 (CM) 中胎儿基因程序 (FGP) 的维恩图;) 和热图说明每组(胎儿、ND、DCM)的 CM 中 FGP 的表达,包括显着调节的生物过程的基因本体术语和与 |log 2 FC|相关的基因 >2()。H,维恩图显示成纤维细胞中的 FGP(Fib;)和热图,说明每组(胎儿,ND,DCM)中 FGP 的表达,包括显着调节的生物过程的基因本体术语和与 |log 2相关的基因FC| >2()。_ _,维恩图显示了上调的胎儿基因,这些基因在 DCM 中的 CM 和心脏 Fib 之间共享,包括基因本体术语和显着调节的生物过程的相关基因。底部,维恩图显示了 DCM 中 CM 和心脏 Fib 之间共享的下调胎儿基因,包括基因本体术语和显着调节的生物过程的相关基因。J ,UMAP 图显示来自 3 个组(胎儿、ND、DCM;)的 6 个不同的 CM 亚群(CM1-CM6)和显示跨 CM 亚群的 FGP 表达的热图()。K,UMAP 图显示来自 3 组(胎儿、ND、DCM;)和热图显示 FGP 跨 Fib 子集群的表达()。CM(Prlf)表示心肌细胞(增殖性);内皮,内皮;呃,红细胞;FC,倍数变化;FDR,错误发现率;周细胞;和 Smc,平滑肌细胞。


基于已知的标记基因,3单核 RNA 测序分析揭示了胎儿、ND 和 DCM 样本中的 7 个细胞簇,包括心肌细胞、成纤维细胞、免疫细胞、平滑肌细胞、内皮细胞、神经元和周细胞(图 [B 和 C])。出生后心脏成熟与成纤维细胞和免疫细胞的相对比例显着增加有关,伴随着心肌细胞比例的显着下降(图[D];Benjamini-Hochberg在speckle R包中使用螺旋桨功能的错误发现率<0.05 )。在 DCM 和 ND 心脏之间未观察到细胞组成的统计学显着变化(图 [D])。然而,DCM 样本的特征是免疫标志物的平均表达水平较低,这表明免疫系统的失调可能是儿科 DCM 的一个特征。


为了识别 DCM 中受干扰的转录途径,我们使用伪批量分析来定义跨共享细胞簇的差异表达基因集(图 [E 到 H])。DCM 相关基因表达改变主要在心肌细胞、成纤维细胞和免疫细胞中观察到(图 [E])。为了定义胎儿基因程序,我们确定了在胎儿和 ND 之间发育调节的差异表达基因(胎儿与 ND,错误发现率 <0.05,log 2倍变化 <–1.5 或 >1.5)。该基因组与 DCM 调节的差异表达基因相交(DCM 与 ND,错误发现率 <0.05,log 2所有细胞簇的倍数变化 < –1.5 或 >1.5),这表明胎儿基因主要在心肌细胞和成纤维细胞中重新激活(图 [F])。在 DCM 中,只有一小部分 (~8%) 的胎儿基因程序在心肌细胞中恢复(2079 个发育调节基因中的 163 个;图 [G])。DCM 心肌细胞中的胎儿基因程序涉及肌肉发育相关基因的上调和先天免疫反应相关基因的下调(图 [G])。同样,在成纤维细胞中,约 8% 的胎儿转录组重新参与 DCM(1658 个发育调节基因中的 139 个(图 [H])。DCM 成纤维细胞的转录反应与肌肉分化程序的上调和氧化代谢过程和细胞周期(图[H])。2倍变化 >2) DCM 成纤维细胞中的胎儿基因是MEOX1(图 [H]),这是一种最近被确定为心脏中肌成纤维细胞活化的中枢调节因子的转录因子。4此外,在 DCM 中,一小部分胎儿基因(总共 13 个)在心肌细胞和成纤维细胞之间重叠(图 [I]),表明这种发育转录网络受两种细胞类型的共同调控机制控制。


单细胞转录组学独特地允许跨细胞群分析转录异质性。因此,我们研究了胎儿基因程序是否广泛地重新参与心肌细胞和成纤维细胞或仅限于 DCM 中的特定细胞亚群。如前所述鉴定心肌细胞亚群(图 [J])和成纤维细胞(图 [K])。3在 DCM 中,胎儿基因程序普遍存在于心肌细胞(图 [J])或成纤维细胞(图 [K])的所有亚群中,这表明该转录程序不限于处于疾病状态的特定细胞亚群。


此处提供的单细胞转录组数据集允许鉴定 DCM 中细胞类型特异性基因失调。与之前对儿科 DCM 的单细胞转录分析相比,5目前的研究包括胎儿和未患病的对照组织,从而能够对在 DCM 中重新激活的胎儿基因程序进行精确的细胞描述。令人惊讶的是,DCM 中胎儿基因重新激活的现象似乎并不局限于心肌细胞,因为相似比例的基因 (<10%) 在心肌细胞和成纤维细胞中都采用胎儿样表达模式。这项工作提供了一个多细胞框架来识别支持儿童 DCM 疾病发病机制的关键基因表达网络。重要的是确定这些转录机制是否也适用于其他形式的心力衰竭,包括成人的 DCM。


作者非常感谢默多克儿童研究所转化基因组学部门对单核 RNA 测序的技术支持和建议。作者感谢默多克儿童研究所的 M. Burton 为荧光激活细胞分选提供技术援助和建议。人类胎儿组织购自 Advanced Bioscience Resources。ND 和 DCM 样本来自墨尔本儿童心脏组织库。所有人类受试者都获得了知情同意,并且研究是根据澳大利亚国家健康和医学研究委员会批准的方案和指南进行的。ERP、DAE、NRM 和 CBS 对该项目进行了概念化。NRM、YS、CBS、BP、CJV、MtH、AO、NJP、DAE 和 ERP 设计了实验。C。BS 为测序进行样品制备。NRM、YS、CBS、RKRK 和 BP 分析了测序数据。ATP、RGW、IEK 和 JEH 为患者招募、同意和生物样本库做出了贡献。NRM、YS、NJP、DAE 和 ERP 分析和解释数据。ERP、DAE 和 IEK 获得了资金并管理了该项目。NRM、DAE 和 ERP 撰写了这篇文章。所有作者都对文章进行了审阅和编辑。


作者感谢澳大利亚国家健康和医学研究委员会(ERP、IEK、DAE)、澳大利亚心脏基金会(NJP)、雪地医学研究基金会(JEH)、斯塔福德福克斯医学研究基金会( ERP、DAE)和皇家儿童医院基金会(ERP、DAE、IEK、RGW)。诺和诺德基金会干细胞医学中心 (ERP, DAE) 由诺和诺德基金会 (NNF21CC0073729) 资助。默多克儿童研究所得到维多利亚州政府运营基础设施支持计划的支持。


Porrello和 Hudson 博士是联合创始人、科学顾问,并持有 Dynomics 的股权,这是一家专注于心力衰竭疗法开发的生物技术公司。其他作者报告没有冲突。


流通可在 www.ahajournals.org/journal/circ


有关资金来源和披露信息,请参见第 1107-1108 页。


更新日期:2022-10-04
down
wechat
bug