Periodic Reporting for period 2 - LSO (Liver Spatial Omics)
Période du rapport: 2020-05-01 au 2021-10-31
The mammalian liver is a heterogeneous, yet highly structured organ, which performs diverse functions to maintain organismal homeostasis. Hepatocytes operate in repeating hexagonally shaped units termed lobules that are polarized by centripetal blood flow and morphogens. This polarized microenvironment facilitates optimal function by localizing specific processes to distinct lobule layers, a phenomenon known as ‘liver zonation’. Our lab discovered that about 50% of hepatocyte genes are zonated. This surprisingly broad spatial heterogeneity raises a fundamental question - do hepatocytes form a uniform population that differs due to spatially graded inputs or are hepatocytes at different zones rather distinct cell types? An additional important question is whether zonation can be observed for other cellular features beyond mRNA. Reconstructing Omics maps of the liver and consequently also other zonated mammalian tissues can enable a refined view of the cell types that make up our body and how local niches dictate their states.
To tackle these questions we are developing techniques for sorting massive amounts of hepatocytes from defined tissue coordinates with high spatial resolution using zonated surface markers. We have been performing deep and comprehensive profiling of the hepatocyte genome, methylome, epigenome, transcriptome, proteome and other features at each zone to characterize liver zonation at all relevant cellular scales.
The objectives of the project are:
1) Develop methodologies for spatial sorting of hepatocytes and other zonated mammalian cell types.
2) Reconstruct Omics maps of the liver and other tissues in physiological and pathological states.
3) Infer modes of zonated post-transcriptional gene regulation based on our omics zonation atlases.
To tackle these questions we are developing techniques for sorting massive amounts of hepatocytes from defined tissue coordinates with high spatial resolution using zonated surface markers. We have been performing deep and comprehensive profiling of the hepatocyte genome, methylome, epigenome, transcriptome, proteome and other features at each zone to characterize liver zonation at all relevant cellular scales.
The objectives of the project are:
1) Develop methodologies for spatial sorting of hepatocytes and other zonated mammalian cell types.
2) Reconstruct Omics maps of the liver and other tissues in physiological and pathological states.
3) Infer modes of zonated post-transcriptional gene regulation based on our omics zonation atlases.
We have previously uncovered broad zonation patterns of both parenchymal cells and non-parenchymal cells at the transcriptional level along the liver lobule axis. An outstanding question that guided us in the current ERC grant is whether cells at different tissue locations are inherently the same cell types, and exhibit transcriptional differences because they are exposed to spatially varying levels of morphogens, oxygen and nutrients, or rather whether they are epigenetically distinct cell types. For example, if I were to “pluck out” a periportal hepatocyte and insert it into the pericentral zone, would it behave as a pericentral hepatocyte or rather would remember its periportal identity. Addressing these questions requires measuring the zonal features of cells that go beyond mRNA, including DNA and chromatin methylations. Moreover, an outstanding question is whether proteins exhibit similar zonation patterns to their encoding mRNAs.
Given these fundamental questions in tissue biology, we set out to develop a generic approach that we termed ‘spatial sorting’ to measure any cellular feature along the tissues’ recurring axes. Spatial sorting is based on the usage of transcriptomics-based zonation atlases to identify zonated surface markers. Upon tissue dissociation, cells are incubated with a cocktail of spatially informative zonated surface markers, and gated so that hundreds of thousands of cells can be obtained in a spatially stratified manner (Ben-Moshe et al., Nature Metabolism 2019).
In the liver, we applied transcriptomics, microRNA (miRNA) array measurements and mass spectrometry proteomics on these spatially sorted populations to reconstruct spatial atlases of multiple zonated features. We demonstrated that protein zonation largely overlaps with messenger RNA zonation, with the periportal HNF4α as an exception. We identified zonation of miRNAs, such as miR-122, and inverse zonation of miRNAs and their hepatocyte target genes, highlighting potential regulation of gene expression levels through zonated mRNA degradation. Among the targets, we found the pericentral Wingless-related integration site (Wnt) receptors Fzd7 and Fzd8 and the periportal Wnt inhibitors Tcf7l1 and Ctnnbip1.
We have also applied the spatial sorting approach to the small intestine - another zonated metabolic tissue. We developed a spatial sorting surface marker cocktail that we optimized to isolate six spatially stratified enterocyte populations along the intestinal villus axis. We measured both mRNAs and proteins along these villus zones. Surprisingly, unlike the liver, we found that around 40% of the genes had anti-correlated zonation profiles of mRNAs and their encoded proteins along the villus axis (Harnik et al., under revision). Since space and time are analogous along the villus, we developed a Bayesian approach to infer translation and protein degradation rates from the combined zonation profiles. We showed that the discordances between mRNAs and proteins are not borne out of zone-dependent post-transcriptional rates, but can rather be attributed to variable protein lifetimes.
In addition to the liver and intestine, we have examined whether the mammalian pancreas exhibits zonated expression patterns. We applied a unique protocol for single molecule transcript imaging we developed for the pancreas, to explore potential zonation patterns of the exocrine pancreatic acinar cells. We identified spatial heterogeneity of acinar cell gene expression (Egozi et al., Cell Reports 2020). We found that peri-islet acinar cells exhibit a distinct molecular signature in db/db diabetic mice that includes upregulation of trypsin family genes and elevated mTOR activity. This zonated expression program seems to be induced by CCK, a hormone that is usually secreted only by enteroendocrine cells in the gut but that is up-regulated in islet cells of expanding islets. Elevated peri-islet trypsin secretion and its potential secretion into the basal acinar cell sides, that is towards the islet, could facilitate the islet expansion observed in this model via modulation of the islet capsule matrix components. Our study highlighted a molecular axis of communication between the pancreatic exocrine and endocrine compartments that may be relevant to islet expansion.
The liver is not only spatially heterogeneous; it is also subject to extensive temporal regulation, orchestrated by the interplay of the circadian clock, systemic signals and feeding rhythms. To explore the interplay between gene regulation in space and time we extended our study and performed scRNAseq of 20,000 hepatocytes from 10 mice sacrificed at four time points along the day. We used our landmark-reconstruction approach to obtain the temporally varying zonation profiles of hepatocyte genes on a global scale (Droin et al., Nature Metabolism 2021). In collaboration with Felix Naef from EPFL, we developed a mixed-effect mathematical model to assess the joint impact of space and time on hepatocyte zonation and found that most hepatocyte genes show a multiplicative space-time effect. We found that the circadian clock machinery is largely non-zonated and that zonation profiles are varying at different time points by a multiplicative zone-independent factor. We also identified circadian expression of Wnt ligands, secreted from pericentral non-parenchymal cells (NPCs), potentially accounting for the circadian rhythms of hepatocyte Wnt target genes.
Given these fundamental questions in tissue biology, we set out to develop a generic approach that we termed ‘spatial sorting’ to measure any cellular feature along the tissues’ recurring axes. Spatial sorting is based on the usage of transcriptomics-based zonation atlases to identify zonated surface markers. Upon tissue dissociation, cells are incubated with a cocktail of spatially informative zonated surface markers, and gated so that hundreds of thousands of cells can be obtained in a spatially stratified manner (Ben-Moshe et al., Nature Metabolism 2019).
In the liver, we applied transcriptomics, microRNA (miRNA) array measurements and mass spectrometry proteomics on these spatially sorted populations to reconstruct spatial atlases of multiple zonated features. We demonstrated that protein zonation largely overlaps with messenger RNA zonation, with the periportal HNF4α as an exception. We identified zonation of miRNAs, such as miR-122, and inverse zonation of miRNAs and their hepatocyte target genes, highlighting potential regulation of gene expression levels through zonated mRNA degradation. Among the targets, we found the pericentral Wingless-related integration site (Wnt) receptors Fzd7 and Fzd8 and the periportal Wnt inhibitors Tcf7l1 and Ctnnbip1.
We have also applied the spatial sorting approach to the small intestine - another zonated metabolic tissue. We developed a spatial sorting surface marker cocktail that we optimized to isolate six spatially stratified enterocyte populations along the intestinal villus axis. We measured both mRNAs and proteins along these villus zones. Surprisingly, unlike the liver, we found that around 40% of the genes had anti-correlated zonation profiles of mRNAs and their encoded proteins along the villus axis (Harnik et al., under revision). Since space and time are analogous along the villus, we developed a Bayesian approach to infer translation and protein degradation rates from the combined zonation profiles. We showed that the discordances between mRNAs and proteins are not borne out of zone-dependent post-transcriptional rates, but can rather be attributed to variable protein lifetimes.
In addition to the liver and intestine, we have examined whether the mammalian pancreas exhibits zonated expression patterns. We applied a unique protocol for single molecule transcript imaging we developed for the pancreas, to explore potential zonation patterns of the exocrine pancreatic acinar cells. We identified spatial heterogeneity of acinar cell gene expression (Egozi et al., Cell Reports 2020). We found that peri-islet acinar cells exhibit a distinct molecular signature in db/db diabetic mice that includes upregulation of trypsin family genes and elevated mTOR activity. This zonated expression program seems to be induced by CCK, a hormone that is usually secreted only by enteroendocrine cells in the gut but that is up-regulated in islet cells of expanding islets. Elevated peri-islet trypsin secretion and its potential secretion into the basal acinar cell sides, that is towards the islet, could facilitate the islet expansion observed in this model via modulation of the islet capsule matrix components. Our study highlighted a molecular axis of communication between the pancreatic exocrine and endocrine compartments that may be relevant to islet expansion.
The liver is not only spatially heterogeneous; it is also subject to extensive temporal regulation, orchestrated by the interplay of the circadian clock, systemic signals and feeding rhythms. To explore the interplay between gene regulation in space and time we extended our study and performed scRNAseq of 20,000 hepatocytes from 10 mice sacrificed at four time points along the day. We used our landmark-reconstruction approach to obtain the temporally varying zonation profiles of hepatocyte genes on a global scale (Droin et al., Nature Metabolism 2021). In collaboration with Felix Naef from EPFL, we developed a mixed-effect mathematical model to assess the joint impact of space and time on hepatocyte zonation and found that most hepatocyte genes show a multiplicative space-time effect. We found that the circadian clock machinery is largely non-zonated and that zonation profiles are varying at different time points by a multiplicative zone-independent factor. We also identified circadian expression of Wnt ligands, secreted from pericentral non-parenchymal cells (NPCs), potentially accounting for the circadian rhythms of hepatocyte Wnt target genes.
We plan to examine how omics features of hepatocytes change throughout the pathological process of drug-induced liver injury. The liver is a highly regenerative tissue. There are two distinct types of liver regeneration processes. One includes partial hepatectomy, whereby a part of the liver is removed, e.g. due to metastases or local tumor, or in a live organ donor, and the remaining part expands to replace the missing tissue. A second includes zonal regeneration, whereby hepatocytes in specific lobule zones are damaged, and are replaced by hepatocytes residing at different zones. We are particularly passionate to understand this second type of liver zonal regeneration, as this touches on the fundamental questions addressed by our project of what is the hepatocyte cell type. If hepatocytes in distinct zones carry not only differences in gene expression but also distinct epigenetic signatures, the incoming regenerating hepatocytes might be functionally compromised, unless they are able to reprogram to the new zonal epigenetic identity. In acetaminophen intoxication (APAP), the most common cause of acute liver damage, the most pericentral hepatocytes that express the xenobiotic enzyme Cyp2e1 necrotize. This gives rise to a dramatic regenerative response that involves activation of liver stellate cells into ECM-building fibroblasts, recruitment of immune cells, division of the remaining periportal and mid-lobular hepatocytes that create a mitotic pressure that pushes hepatocytes pericentraly, re-establishing hepatocyte mass. Within a week of damage the tissue exhibits almost complete histological remission. We are using scRNAseq, smFISH and spatial sorting to analyze in detail this regeneration process. Outstanding questions that guide us include: Do regenerating hepatocytes remember their epigenetic periportal/mid-lobular identity or rather completely reprogram and trans-differentiate into a pericentral state? Is there a division of labor whereby some stellate cells become activated whereas others proliferate to extend their numbers? What is the fate of activated stellate cells when the tissue is repaired – do they die, enter senescence or de-differentiate into the quiescent state? What are the niche signals emanating from zonal endothelial cells that regulate these cellular dynamics? Are zonated hepatocytes different in methylation patterns, and if they are is zonal regeneration accompanied by an effective reprogramming of hepatocytes? Addressing these questions can both reveal fundamental processes in liver biology and also identify potential avenues for fighting liver fibrosis.