10X单细胞空间联合分析揭示急性损伤后肝脏不同区域再生的时空程序

作者，追风少年i

7月1日，是2022年下半年的开始，单细胞空间持续进行时，细数一下发展，华大、新格元、寻因、墨卓、诺禾等等都在布局，从行业角度看，目前国内下游为主，标准的微笑曲线，上游实力与创新能力不足，同时也意味着发展潜力巨大。

不过呢，说实话，跟我关系不大，主要是看国内是不是真的有民族企业家，如果没有，微笑曲线可能波动一下都不可能。

好了，继续我们单细胞空间的分析内容，今天我们来分享一篇文献，The spatiotemporal program of zonal liver regeneration following acute injury,2022年6月发表于Cell Stem Cell，不用多说，我们来看看研究内容。

Summary of main events during APAP-induced liver damage and regeneration

肝脏是一个高度异质的器官。肝细胞和支持非实质细胞在被称为“肝小叶”的重复六边形解剖单元中活动。血液进入小叶的角落，称为“门脉节点”，并通过窦性通道流入引流的中央静脉 (CVs)。这种极化的血流与肝细胞的连续消耗和分泌相结合，产生氧气、营养物质和激素的梯度分布，导致高度异质的微环境，因此肝脏中不同位置细胞在基因表达上呈现显著差异，该现象称为“肝脏分区”，目前尚不清楚这种空间异质性如何影响肝脏病理和再生过程。

肝脏具有强大的再生能力。在急性剂量的药物作用下，试图排除这些外来物质的中心周围肝细胞被poisonous中间体淹没而死亡。剩下的肝组织进入再生模式，导致受损的小叶层迅速愈合和替换。分区再生涉及一系列相互协调的过程，对反应有严格的时空控制要求。死亡的肝细胞需要被有效清除，防止多种新抗原暴露以诱导适应性免疫系统。细胞外基质需要被快速构建以支持组织支架，但一旦新细胞形成就要分解，以防止持久的纤维化。最重要的是，起源于小叶区表达特征截然不同的肝细胞需要快速生成和重新编程，以接管中心周围肝细胞的关键功能。目前，在整个分区再生过程中，所有肝细胞类型的协调反应的空间和时间动力学仍亟待探索。

研究成果

A single-cell atlas of zonal regeneration after APAP

对小鼠不同时间段进行APAP处理，研究肝脏的变化过程，其中用到了bulk、单细胞以及空间技术表征肝脏图谱。其中bulk转录组分析了不同时间点的生物学通路变化、并且利用单细胞数据解卷积bulk数据分析细胞比例的变化，当然了，空间转录组分析细胞的位置变化与空间基因特征。

为了研究分区重生，他们在小鼠体内注射了300 mg/kg的APAP引发肝脏损伤，在不同时间点对肝脏进行采样，观察到大量的中心周围坏死在48小时后达到峰值。损伤的中心周围区域在72小时后出现收缩，出现新的中心周围肝细胞，在96h后完全取代坏死组织。作者在多个时间点进行批量和单细胞RNA-seq，并进行空间转录组学分析以揭示该重生过程的分子机制。bulk RNA-seq的主成分分析 (PCA) 显示，肝脏基因表达在6、24和48 小时出现差异，在96小时回归到对照水平。bulk RNA-seq数据的通路分析显示，早期免疫信号的诱导，在APAP后6和24小时达到峰值，增殖通路在APAP后48 h达到高峰。肝脏代谢功能，如谷胱甘肽和细胞色素P450代谢，以及静止的造血干细胞功能，在APAP后72 h重新出现。单细胞RNA-seq分析表明，注射APAP后48 h肝细胞丰度显著下降，与中心周围肝细胞大量坏死相一致，96小时后肝细胞丰度回归对照水平。相比之下，造血干细胞、单核细胞和巨噬细胞大量增加，在APAP注射后24-48小时达到峰值，然后下降到对照组水平。该分析显示出肝脏基因表达和多种肝细胞类型比例的重大变化，在损伤后4天就迅速恢复到控制水平。

Acute dose of APAP induces liver damage and regeneration

• 图注：(A) A schematic of the experimental design. Mice were injected with 300 mg APAP/1 kg body weight. Livers were harvested for bulk sequencing at 6, 24, 48, 72, 96 h, 1 week, and 1 month after injection (3 mice per time point), as well as saline injected controls at different time points (2 mice per time point). Livers were dissociated for single-cell sequencing at 24, 48, 72, 96 h, 1 week, and nontreated controls (2–4 mice for each time point, marked with black asterisks). Livers from 24, 48, and 72 h were also taken for Visium spatial transcriptomics (blue asterisks). Mouse injection illustration was created with BioRender.com.
  (B) Images of liver lobules at different time points following APAP injection. CV, central vein; PV, portal vein. Yellow dashed lines mark the borders of the damaged areas. Cell nuclei are stained with Dapi (blue). Cell membranes are stained with phalloidin (gray). Scale bars, 20 mm.
  (C) Uniform manifold approximation and projection (UMAP) visualization of the integrated data of all 23,944 cells from 6 time points (n = 28 mice). Cells are colored by their cell type annotation.
  (D) UMAP visualization of the integrated data. Cells are colored by the time following APAP injection.

  
  Analysis of Visium spatial transcriptomics dataset

• 图注：(A) H&E image of Visium slide of liver of 24h after APAP injection. White square marks the inset shown in B and C.
  (B) Close up look at the tissue shows distinct regions of pericentral damage and periportal undamaged zones. C - central vein, P - portal vein.
  (C) Overlay of the tissue with the Visium spot (spatial clusters). Yellow spots mark spots with fibrogenic regions, which were manually segmented (methods), and grey spots mark non-fibrogenic regions. C - central vein, P - portal vein
  (D) Projection of the spots overlaying the tissue, colored by their classification into fibrotic or non-fibrotic regions.
  (E) Projection of the spots overlaying the tissue, colored by their distances from the central veins. Each connected region of fibrotic spots was skeletonized (methods) to get to the “central core”. The distance of each spot to the nearest “central core” was then calculated. The “central cores” are marked in black x.
  (F) Projection of the spots overlaying the tissue, colored by their expression level of the hepatocyte pericentral gene Cyp2e1, log10 normalized. The “central cores” are marked in black x.
  (G) Projection of the spots overlaying the tissue, colored by their expression level of the hepatocyte periportal gene Cyp2f2, log10 normalized. The “central cores” are marked in black x.
  (H-J) Scatter plots of log2 of the periportal to pericentral expression ratios of zonated hepatocyte (H), HSC
  (I) and endothelial (J) specific genes found in the computationally inferred zonation patterns of the single cell transcriptomics and in the zonation pattern in spatial transcriptomics dataset (methods). Dots are colored by the Visium slide with which the single cell reconstruction was compared

肝细胞表现出区域性重编程和跨肝小叶的广泛增殖

中央周围和门静脉周围肝细胞在未受干扰的肝脏中表现出明显不同的基因表达。肝细胞单细胞数据包括来自对照、APAP 注射后 48 和 72 小时的 2,770 个细胞。使用空间转录组学数据集，确定了肝细胞标志性基因，这些基因在整个再生过程中在门静脉或中央分区。使用这些计算推断单个测序肝细胞沿肝小叶轴的坐标。随后，将每个时间点的细胞分成三个区域——中央周围、小叶中部和门静脉周围——并对它们的表达进行平均以获得动态肝细胞分区分布。正如预期的那样，具有周围中心特征的肝细胞在 48 小时内耗尽。分区坐标分布在 72 h 时接近损伤前模式。使用单分子荧光原位杂交 (smFISH) 来证明经典的分区基因，如中央 Glul 和 Cyp2e1 和门静脉 Ass1 在损伤后 96 小时呈现其损伤前分区模式。值得注意的是，虽然门静脉距离没有显着变化，但新形成的肝细胞表现出显着更高的倍性水平，这种增加在中央肝细胞中更为突出。因此，分析表明，肝细胞在急性肝损伤 4 天后恢复其区域分子特性。

已显示带状肝细胞群在稳态以及不同的再生模型中表现出不同的增殖模式。为了探索 APAP 再生过程中肝细胞增殖的带状动态，分析了不同时间点增殖肝细胞的带状分布。为此，使用 smFISH 量化了表达增殖标记 Mki67 的肝细胞比例。在 APAP 注射后 32 小时，肝细胞在整个小叶轴上增殖，有轻微的门静脉周围偏差。增殖在 40 和 48 小时增加，显示出中心偏差，然后在 72 小时趋于平稳。值得注意的是，在整个再生过程中，离受损组织区域较远的小叶中区和门静脉区的肝细胞对 Mki67 呈阳性。在不同小叶区域的增殖和非增殖肝细胞之间的差异表达分析揭示了增殖肝细胞中肝细胞代谢功能的降低。最近在肝细胞中删除 ZNRF3 和 RNF43 后观察到增殖肝细胞中肝细胞代谢功能的类似下降。肝细胞在整个小叶轴上的广泛增殖可能有助于沿肝板产生增加的有丝分裂压力，这可能有助于快速替换受损区域的细胞。重要的是，这种有丝分裂压力将小叶中肝细胞带入中央区，需要重新编程它们的转录状态。

Hepatocytes from all zones proliferate and re-establish zonation

• 图注：(A) Spearman correlation distances between hepatocyte-specific genes for pairs of control samples (n = 4) and mice at different time points (n = 3 mice per time
  point) after APAP injection. Horizontal line denotes median distance of control-control pairs. White dots are median distances for each time point.
  (B) UMAP visualization of hepatocytes (n = 2,770 cells), colored by time after APAP injection.
  (C) UMAP visualization of hepatocytes colored by the expression of the centrally zonated gene Cyp2e1.
  (D) UMAP visualization of hepatocytes colored by the expression of the periportal gene Cyp2f2.
  (E) UMAP of hepatocytes colored by their inferred lobule spatial coordinate, ranging from CV—cells closest to central vein, to PV—cells closest to portal vein.
  (F) Distributions of lobule spatial coordinates of hepatocytes at each time point.
  (G and H) smFISH of a liver lobule showing 3 zonated genes: pericentral Glul (red), Cyp2e1 (blue), and periportal Ass1 (green). CV, central vein; PV, portal vein.
  Scale bars, 20 mm. Shown are examples of a control lobule (G) and a lobule 96 h after APAP (H).
  (I) The fraction of proliferating hepatocytes expressing Mki67+ transcripts out of all hepatocytes located in either pericentral (red), midlobular (yellow), or periportal
  (green) zones. The analysis was performed on 2 mice from each time point, at least 5 lobules per mouse were quantified. Significance levels calculated using
  Kruskal-Wallis tests. White dots represent the median fraction.
  (J) smFISH of a liver lobule 72 h following APAP administration showing Mki67 single transcripts (gray dots). Dashed white line marks the damage border, arrows
  point to representative Mki67+ proliferating cells. Cell nuclei are stained with Dapi (blue) and membranes are stained with phalloidin (red). Scale bars, 20 mm.

Interface hepatocytes up-regulate fetal programs and exhibit a mesenchymal shape（he interface between the damaged and nondamaged zones）

发现界面肝细胞表现出独特的表达特征，由在胎儿肝脏和肝细胞癌中表达但在成人肝细胞中不表达的基因组成。界面肝细胞不会简单地从门静脉周围/小叶中状态转变为中央周围状态，因为它们被推入中央周围区域。 相反，它们的细胞特性变化与胎儿基因的瞬时表达、蛋白质翻译和降解的升高以及细胞形态的改变有关。

Interface hepatocytes upregulate onco-fetal genes as they reprogram into pericentral hepatocytes

• 图注：(A and B) Expression levels of genes in pericentral hepatocytes 48 (A) or 72 h (B) after APAP injection plotted against their expression in control hepatocytes with
  matched distribution of lobule spatial coordinates. Gray dots represent all genes. Red/blue dots represent genes upregulated/downregulated respectively in the
  regenerating tissue, with mean expression level of above 5 3 10�6, at least 2-fold difference from matched control and FDR threshold of 0.01.
  (C) smFISH (top) and insets (bottom) of a liver lobule 48 h after APAP injection. Dashed white line delineate the damage border. CV, central vein; PV, portal vein.
  Nuclei and membranes are labeled with Dapi (blue) and phalloidin (green), respectively. Two representative interface hepatocytes outlined in orange shown in the
  insets (bottom), together with the smFISH labeling for mRNAs of Afp (left), Cdh17 (middle), and Spp1 (right). Scale bars, 20 mm. Laplacian of Gaussian filter was
  applied on the smFISH images.
  (D) smFISH of a liver lobule 48 h post-APAP injection for Apoa1 (red) and Actb (gray) mRNA. Nuclei are stained with Dapi (blue). CV, central vein. Scale bars, 20 mm.
  White dashed rectangle is the region displayed in (F).
  (E) GSEA normalized enrichment scores (NESs) of gene pathways significantly upregulated (red) or downregulated (blue) in interface cells. Dot size corresponds
  to number of pathway genes found in the dataset; opacity corresponds to false discovery rate (FDR). Genes sets used for the analysis were taken from Kyoto
  Encyclopedia of Genes and Genomes (KEGG) pathways dataset.
  (F) Magnification of the region marked in (D) showing interface (orange dashed lines) and noninterface (blue dashed lines) hepatocytes. Nuclei and membranes are
  labeled with Dapi (blue) and phalloidin (gray), respectively. Scale bars, 20 mm.
  (G) Quantification of circularity of interface (orange) and noninterface (blue) hepatocytes. White dots represent group circularity median. Gray boxes mark the 25–
  75 percentiles. Significance level was calculated using paired signrank test (n = 60 pairs of interface and adjacent noninterface hepatocytes, taken from 3 mice
  48 h after APAP injection and 3 mice 72 h after APAP injection).

HSCs exhibit spatial division of labor（空间分工）

成功的肝再生需要所有肝细胞类型的紧密协调反应。 HSC 是肝损伤反应和再生的关键参与者。分析强调了特定区域的 HSC 表达程序，这些程序可能促进急性 APAP 损伤后肝脏表现出的免疫募集、ECM 积累和分解的时间协调过程。

Hepatic stellate cells exhibit distinct expression programs in different zones

• 图注：(A) UMAP visualization of HSCs, colored by time after APAP administration. n = 2,311 cells, at least 2 mice per time point.
  (B) UMAP visualization of HSCs, colored by inferred zone
  (C–H) Temporal dynamics of selected genes in HSCs, stratified by zone. Lines denote the mean of the normalized expression over cells from the same zone and
  time point, patches denote the standard errors of the means (SE). For each gene, the mean and SE for pericentral HSCs (red), mid-lobule HSCs (yellow) and
  periportal HSCs (green) are presented. Selected genes belong to various processes: markers of quiescent HSCs (C), activated HSCs (D), ECM collagen genes (E),
  proliferation (F), centrally zonated immune modulators (G), and downregulated zonated immune modulators (H).

Endothelial cells exhibit zonated cues along the regeneration process

Zonal endothelial cell populations differentially express genes involved in the regeneration process

• 图注：(A) UMAP visualization of endothelial cells, colored by time after APAP administration. n = 6,527 from 13 mice, at least 2 mice per time point.
  (B) UMAP visualization of endothelial cells, colored by expression level of the pericentral Wnt2 (top left), periportal Efnb2 (top right), sinusoidal endothelial cell
  marker Kit (bottom left) and vascular endothelial cell marker Vwf (bottom right).
  (C) UMAP visualization of endothelial cells, colored by inferred zone (STAR Methods). PC-LVECs, pericentral liver vascular endothelial cells; PC-LSECs, pericentral
  liver sinusoidal endothelial cells; mid-LSECs, mid-lobule liver sinusoidal endothelial cells; PP-LSECs, periportal liver sinusoidal endothelial cells; PPLVECs,
  periportal liver vascular endothelial cells.
  (D) Temporal dynamics of selected genes in endothelial cells, stratified by the zonated endothelial cell populations. Lines denote the mean of the normalized
  expression over cells from the same zone and time point, patches denote the standard errors of the means (SE). For each gene, the mean and SE for PC-LVEC
  (dark red), PC-LSEC (red), mid-LSEC (yellow), PP-LSEC (green), and PP-LVEC (blue) are presented. Lobule diagram highlights the different zonated endothelial
  cell subtypes with their respective colors.

Dynamics of myeloid cell subtypes along the regeneration process

骨髓细胞有助于急性肝损伤后的再生过程。与肝细胞和内皮细胞一样，活化的kuffer细胞与fetal kuffer细胞有显著相关性。 fetal kuffer细胞的共同趋势包括抗原呈递基因减少和脂质相关巨噬细胞基因增加。

Spatiotemporal patterns of myeloid cell gene expression

• 图注：(A) UMAP visualization of myeloid cell populations, colored by time after APAP administration. n = 9,338 cells from 15 mice, at least 2 mice per time point.
  (B) UMAP visualization of myeloid cell populations, colored by cell types.
  (C) The fractions of different myeloid cell subtypes at each time point.
  (D) UMAP visualization of myeloid cells, colored by expression levels of Mmp12.
  (E) smFISH scan (top) and zoomed-in insets (bottom) of a liver lobule 24 h after APAP injection. CV, central vein; PV, portal vein. Nuclei and membranes are labeled
  with Dapi (blue) and phalloidin (green), respectively. Pericentral (left), mid-lobule (middle), and periportal (right) KCs marked by dashed lines in the scan are
  enlarged in the insets (bottom), together with the smFISH for the Kupffer cell gene Marco (green) and Mmp12 (red) mRNAs, together with Dapi (blue). Scale bars,
  20 mm for top image, 10 mm for insets. Laplacian of Gaussian filter was applied on the smFISH images.

总的来说，该工作使用空间分辨单细胞RNA测序 (scRNA-seq) 来研究急性对乙酰氨基酚(APAP)处理后小鼠肝脏再生的动力学，发现肝细胞在肝小叶内增殖，产生有丝分裂压力，使坏死的中心区迅速再生。位于再生界面的部分肝细胞会在重新编程到中心周围状态过程中，短暂上调胎特异性基因。该研究有助于深刻理解肝分区再生的协同程序。

强调一个重点方法（空间为重）

Interaction analysis from spatial transcriptomics dataset

Ligand-receptor interactions were also investigated in the spatial-transcriptomics dataset. Ligand-receptor pairs in each Visium slide were analyzed. Spearman correlation of each ligand-receptor pair expression was calculated, either for across all slide spots, or for each of the three stratified lobule zones (discretized by growing distances from the central cores). In addition to correlations, interaction potential was also calculated as the mean product of the expression of the ligand and receptor over all spots, either across the whole slide, or across each zone.

好了，已经分享给大家了，生活很好，有你更好，百度文库出现了大量抄袭我的文章，对此我深表无奈，我写的文章，别人挂上去赚钱，抄袭可耻，挂到百度文库的人更可耻