Kidney single-cell transcriptomes predict spatial corticomedullary gene expression and tissue osmolality gradients [single nuclei]

Single-cell transcriptomics from dissociated organs lack information regarding the spatial origin of each cell, which limits their interpretation, particularly in complex and regionally heterogeneous tissues. This is relevant in the kidney, where cell types are exposed to a wide spectrum of cellular microenvironments along the corticomedullary axis, including steep gradients of extracellular osmolality and oxygen tension. Whether kidney single-cell transcriptomes can be exploited to predict spatial origins of cells and to provide physiological readouts of the cellular microenvironment is unknown. Here, we obtained single-cell transcriptomes of mouse kidney tissue from whole organs and from defined kidney zones (cortex, outer and inner medulla) and applied computational methods to reconstruct the spatial position of kidney tubule cells along the corticomedullary axis based on their transcriptomes. Our approach enabled a spatially resolved analysis of gene expression, showed a coordinated activation of osmolality-and hypoxia-associated genes towards the kidney medulla, and predicted that transcriptomes of a given cell type across different kidney zones change gradually rather than being clearly distinct in different anatomical zones. In genetically modified mice with a tubular concentration defect, spatial reconstruction of single-cell transcriptomics and osmogene expression quantitation accurately predicted reduced medullary osmolality. We conclude that our approach, which can be applied to any mouse whole kidney single-cell transcriptomic dataset, uncovers previously underappreciated information regarding spatial origin and microenvironment-dependent cellular states, adding improved readouts to existing and future datasets. The provided data comprise one sample of whole kidney single-nuclei data from a collecting duct-specific knockout of Grainyhead-like 2 and one control littermate.解离器官来源的单细胞转录组学（single-cell transcriptomics）数据缺乏每个细胞的空间起源（spatial origin）信息，这制约了其解读价值，在结构复杂且具有区域异质性的组织中尤为突出。这一点在肾脏中尤为关键：肾脏内的细胞类型沿着皮质髓质轴（corticomedullary axis）暴露于多样的细胞微环境中，伴随细胞外渗透压（extracellular osmolality）与氧张力（oxygen tension）的陡峭梯度。目前尚不明确是否可利用肾脏单细胞转录组预测细胞的空间起源，并获取细胞微环境的生理读数。本研究获取了全器官及特定肾脏区域（皮质、外髓质与内髓质）来源的小鼠肾脏单细胞转录组，并基于细胞转录组特征，通过计算方法（computational methods）重构肾小管细胞（kidney tubule cells）沿皮质髓质轴的空间位置。我们的方法实现了基因表达（gene expression）的空间分辨分析，揭示了向肾脏髓质方向渗透压相关与缺氧相关基因（hypoxia-associated genes）的协同激活，并预测：同一细胞类型的转录组在不同肾脏区域中呈渐进式变化，而非在不同解剖区域中呈现显著差异。在存在肾小管浓缩功能缺陷（tubular concentration defect）的基因工程小鼠（genetically modified mice）中，单细胞转录组空间重构与渗透压相关基因表达定量分析准确预测了髓质渗透压的降低。综上，本研究方法可推广至任意小鼠全肾脏单细胞转录组数据集，可揭示此前未被充分认知的细胞空间起源与微环境依赖的细胞状态相关信息，为现有及未来的数据集提供更优质的生理读数。本研究提供的数据包含1份集合管特异性敲除（collecting duct-specific knockout）Grainyhead样蛋白2（Grainyhead-like 2）的全肾脏单细胞核数据（single-nuclei data）样本，以及1份同窝对照样本。