Spatial Sorting Enables Comprehensive Characterization of Liver Zonation
Total Page：16
File Type：pdf, Size：1020Kb
ARTICLES https://doi.org/10.1038/s42255-019-0109-9 Spatial sorting enables comprehensive characterization of liver zonation Shani Ben-Moshe1,3, Yonatan Shapira1,3, Andreas E. Moor 1,2, Rita Manco1, Tamar Veg1, Keren Bahar Halpern1 and Shalev Itzkovitz 1* The mammalian liver is composed of repeating hexagonal units termed lobules. Spatially resolved single-cell transcriptomics has revealed that about half of hepatocyte genes are differentially expressed across the lobule, yet technical limitations have impeded reconstructing similar global spatial maps of other hepatocyte features. Here, we show how zonated surface markers can be used to sort hepatocytes from defined lobule zones with high spatial resolution. We apply transcriptomics, microRNA (miRNA) array measurements and mass spectrometry proteomics to reconstruct spatial atlases of multiple zon- ated features. We demonstrate that protein zonation largely overlaps with messenger RNA zonation, with the periportal HNF4α as an exception. We identify zonation of miRNAs, such as miR-122, and inverse zonation of miRNAs and their hepa- tocyte target genes, highlighting potential regulation of gene expression levels through zonated mRNA degradation. Among the targets, we find the pericentral Wingless-related integration site (Wnt) receptors Fzd7 and Fzd8 and the periportal Wnt inhibitors Tcf7l1 and Ctnnbip1. Our approach facilitates reconstructing spatial atlases of multiple cellular features in the liver and other structured tissues. he mammalian liver is a structured organ, consisting of measurements would broaden our understanding of the regulation repeating hexagonally shaped units termed ‘lobules’ (Fig. 1a). of liver zonation and could be used to model liver metabolic func- In mice, each lobule consists of around 9–12 concentric lay- tion more precisely. T 1 ers of hepatocytes . Blood flowing from portal nodes at the corner In this study, we developed an approach termed ‘spatial sorting’, of the lobules towards draining central veins generates gradients which uses surface markers with discordant zonation profiles to isolate of oxygen, nutrients and hormones along the lobule radial axis. very large amounts of hepatocytes from defined lobule layers (Fig. 1b). Additionally, Wnt morphogens are secreted by endothelial cells We used these for high-throughput profiling of mRNAs, miRNAs surrounding the central veins, resulting in a graded morphoge- and proteins (Fig. 1c), revealing previously unknown features of liver netic field2. This graded microenvironment gives rise to spatial zonation. These include a comprehensive proteomic zonation map heterogeneity in gene expression among hepatocytes residing and the identification of zonated miRNA with discordantly zonated at different lobule layers, a phenomenon that has been termed target genes. Our approach can be readily applied to profile other cel- ‘liver zonation’3,4,5,6. lular features of hepatocytes and other cell types in health and disease. We have recently used spatially resolved single-cell transcrip- tomics to uncover the global zonation patterns of hepatocyte gene Results expression7. We found that around half of all genes expressed in Spatial sorting enables isolating bulk hepatocyte populations hepatocytes are zonated, with specific functional specialization that from different lobule layers. We used our recently reconstructed seems to match the zonated microenvironment. This global zona- mRNA zonation map7 to identify zonated surface markers with a tion suggests that similar spatial heterogeneity of hepatocytes may large dynamic range in expression, spanning several radial lobule also exist for other cellular features, including proteins, metabolites layers (Fig. 1a and Supplementary Fig. 1a). We argued that the com- and regulatory molecules such as miRNAs. However, achieving bined staining of two inversely zonated surface proteins would be similar global zonation maps for cellular features beyond mRNA informative for inferring the lobule positions of single hepatocytes has encountered technical difficulties. (Fig. 1b), which would facilitate cell sorting of many cells according Immunohistochemistry enables the measurement of protein lev- to their spatial origin (Fig. 1b,c). CD73, encoded by the gene Nt5e, els with high spatial resolution but it is low-throughput and often is an enzyme that converts mononucleotides to nucleosides and limited by lack of availability of antibodies. Laser capture microdis- exhibits pericentral zonation7. E-cadherin, a cell–cell adhesion gly- section and digitonin perfusion enable extracting large numbers of coprotein encoded by Cdh1, exhibits periportal zonation12 (Fig. 2a). periportally or pericentrally enriched cells8,9. However, these tech- We used immunofluorescence to validate the zonation of these two niques are limited in spatial resolution. Single-cell measurements of surface markers at the protein level (Fig. 2b,c). cellular features beyond mRNA are starting to emerge10,11; however, We perfused the livers of five mice fed ad libitum to dissoci- these technologies are less mature in tissues. A methodology that ate single cells and performed fluorescence-activated cell sorting enables massive isolation of pure cell types from defined layers with (FACS) of isolated hepatocytes stained with antibodies against CD73 high spatial resolution would enable generating organ-wide spa- labelled with allophycocyanin (APC) and E-cadherin labelled with tial atlases of key features, such as methylation patterns, chromatin phycoerythrin (PE). We filtered hepatocytes by size and selected conformations, miRNA content and proteomics. In the liver, such cells that were negative for the endothelial cell marker CD31 and 1Department of Molecular Cell Biology, Weizmann Institute of Science, Rehovot, Israel. 2Present address: Institute of Molecular Cancer Research, University of Zurich, Zurich, Switzerland. 3These authors contributed equally: Shani Ben-Moshe, Yonatan Shapira. *e-mail: [email protected] NatURE MetabOLISM | www.nature.com/natmetab ARTICLES NATURE METABOLISM a b c Spatial sorting Liver lobule Pericentral Spatial omics FACS gates miR1 miR2 miR3 Portal node Portal node Mid-lobule Periportal AAA FACS gates FACS gates AAA AAA Central vein Central vein Pericentral surface marker Pericentral Periportal Periportal surface marker mRNA Mass miR surface marker surface marker sequencing spectrometry microarray Fig. 1 | Spatial sorting approach for isolating large amounts of hepatocytes from distinct layers with high resolution. a, Identification of zonated surface markers. b, FACS enables defining gates that enrich for zonated hepatocytes according to their surface marker expression. c, Spatially sorted hepatocytes can be measured using multiple assays that require large input material, such as the RNA-seq, mass spectrometry and miRNA microarray applied in the current study. a mRNA b CD73/E-cadherin/DAPI Nt5e n Cdh1 1 Central vein 2 3 Portal 0.6 1 node mRNA expressio 0.2 (normalized to maximum) Central Portal vein node c Protein CD73 1 - Pericentral 2 - Mid-lobule 3 - Periportal 1 E-cadherin 0.6 0.2 Protein intensity (normalized to maximum) Central Portal vein node Lobule layer Fig. 2 | CD73 and E-cadherin are inversely zonated surface markers. a, CD73, encoded by Nt5e, and E-cadherin, encoded by Cdh1, are surface markers that are zonated at the mRNA level (data taken from Halpern et al.7). n = 1,415 cells from 3 mice. The lines show the sum-normalized mean of all cells; the shaded regions are ±s.e.m. b, CD73 and E-cadherin proteins are zonated. An example of a lobule stained by immunofluorescence with antibodies against CD73 (red) and E-cadherin (green) is shown. Blue, DAPI nuclear stain. Scale bar, 10 µm. The experiment was performed independently on three different mice. c, Quantification of immunofluorescence images (n = 8 lobules from three mice). The lines represent the mean intensity measured in the lobule layer; the shaded regions are ±s.e.m. across the eight lobules. the immune cell marker CD45, to avoid pairs of hepatocytes and hepatocytes from each gate. Contamination of NPC RNA was neg- non-parenchymal cells (NPCs)13. We further filtered out non-viable ligible and uniform across all eight isolated populations, validating cells and selected tetraploid hepatocytes using Hoechst staining our isolation and sorting approach (Supplementary Table 1). We (Fig. 3a and Supplementary Fig. 1b). Stratifying hepatocytes by compared the zonation profiles obtained via spatial sorting to our ploidy was important to obtain precise lobule localization spatially resolved single-cell RNA-seq map7. Zonation profiles were (Supplementary Fig. 1c,d). The selected hepatocytes displayed highly concordant (Fig. 3c and Supplementary Table 2), demon- strong anti-correlation in the fluorescence of CD73 and E-cadherin, strating the feasibility of our approach for isolating bulk hepatocytes as expected from the zonated expression patterns (Fig. 3b). with high spatial resolution. We defined eight gates based on the combined fluorescence of CD73 and E-cadherin (Fig. 3b). To ensure reproducibility, the Mass spectrometry proteomic measurements of spatially sorted gates were defined as the percentiles of the marginal expression lev- hepatocytes. We next applied spatial sorting to reconstruct the els of each surface marker, compared to the unstained control. To zonation patterns of the hepatocyte proteome. To this end, we validate that our defined gates represented sequential lobule layers, sorted 100,000 hepatocytes