Periodic Reporting for period 1 - StellateZEN (Unravelling the crosstalk between Stellate cells and Kupffer cells and its role in maintaining Stellate cell quiescence)
Reporting period: 2022-06-01 to 2024-12-31
Chronic liver diseases account for two million deaths a year, half of which can be attributed to liver fibrosis. The later stage of fibrosis, cirrhosis, leads to serious complications such as liver failure, portal hypertension, and represents a risk factor for developing hepatocellular carcinoma. Chronic liver disease can have multiple causes: alcohol, viral hepatitis, metabolic-associated fatty liver disease or drug induced liver injury. Independent of the etiology, activated hepatic stellate cells are the major source of excessive extracellular-matrix deposition that drives fibrosis. An increasing body of evidence suggest that upon resolution of liver inflammation, activated stellate cells can either undergo senescence, apoptosis, or revert to quiescence (i.e. homeostasis, unactivated). Hence, understanding molecular mechanisms driving and maintaining stellate cell quiescence is key and could potentially lead to novel fibrosis treatments.
Stellate cells are liver stromal cells that are in close contact with hepatocytes, liver sinusoidal endothelial cells and resident macrophages of the liver, the Kupffer cells (KCs). Recent studies have highlighted the important crosstalk between KCs and stellate cells. Our group showed that at steady state, they are always paired with one another. Moreover, upon KC depletion, stellate cells attract monocytes and promote their differentiation into KCs, highlighting the interconnectivity of these cells. A crucial part of this process is the BMP9/ALK1 signaling axis. In the liver stellate cells are the only source of BMP9 whilst KCs express the receptor ALK1. Importantly, while quiescent stellate cells are always paired with KCs, several studies have reported that profibrogenic activated stellate cells are instead paired with inflammatory macrophages. This suggests that stellate cell micro-environment forms a niche providing signals that imprint their identity and/or function.
In this project, our overall objective is to decipher stellate cell niche to understand which signals are important to drive and maintain their quiescent identity. This study aims at investigating the crosstalk between stellate cells and KCs in the liver.
Stellate cells are liver stromal cells that are in close contact with hepatocytes, liver sinusoidal endothelial cells and resident macrophages of the liver, the Kupffer cells (KCs). Recent studies have highlighted the important crosstalk between KCs and stellate cells. Our group showed that at steady state, they are always paired with one another. Moreover, upon KC depletion, stellate cells attract monocytes and promote their differentiation into KCs, highlighting the interconnectivity of these cells. A crucial part of this process is the BMP9/ALK1 signaling axis. In the liver stellate cells are the only source of BMP9 whilst KCs express the receptor ALK1. Importantly, while quiescent stellate cells are always paired with KCs, several studies have reported that profibrogenic activated stellate cells are instead paired with inflammatory macrophages. This suggests that stellate cell micro-environment forms a niche providing signals that imprint their identity and/or function.
In this project, our overall objective is to decipher stellate cell niche to understand which signals are important to drive and maintain their quiescent identity. This study aims at investigating the crosstalk between stellate cells and KCs in the liver.
In the mouse model CD64cre x ALK1-flox/flox, thereafter referred to as ALK1-KO, ALK1 signaling is abrogated in macrophages resulting in a drastic loss of KCs in the liver that are replaced with immature macrophages. The remaining KCs are clustered in islands throughout the tissue, thus the ALK1-KO liver contains both stellate cells still in contact with KCs (paired) and stellate cells no longer in contact with KCs (unpaired).
First, we performed single-nuclei transcriptomics to characterize stellate cells that are paired or unpaired with KCs. However, our preliminary transcriptomics analysis showed that stellate cells from the control and ALK1-KO transcriptomes were quite similar.
To verify if stellate cells in the ALK1-KO model were also behaving similarly to the control upon activation, I treated mice with Carbon tetrachloride (CCl4) to induce liver inflammation. There were no quantifiable differences between the paired and unpaired conditions, suggesting that unpairing stellate cells with KCs is not sufficient to disrupt their identity and that other signals from their niche are involved in maintaining their quiescence.
Since more signals seem to be necessary for stellate cell quiescent identity, we instead focused on directly disrupting stellate cells. LIM Homeobox 2 (LHX2) is a known stellate cell transcription factor that is down-regulated during stellate cell activation.
I characterized a new transgenic model in which we conditionally knocked-out Lhx2 (LRATcreERT2 x LHX2-flox/flox, thereafter referred to as LHX2-KO) specifically in stellate cells. Deletion of this gene in stellate cells has a drastic effect on the identity of stellate and other liver cells. Similarly to the ALK1-KO model, we observe islands of “escapee” cells where they maintain a normal phenotype.
Thus, we identified a new model that is relevant to answer our question, as it contains both paired and unpaired stellate cells, but also provides an environment where the cellular identity circuits are deeply perturbed.
We performed an in-depth transcriptomics analysis of the LHX2-KO model. Single cell transcriptomics showed that stellate cell identity was deeply perturbed, as they were displaying a more profibrotic (but not fully activated) gene profile. Escapee cells were transcriptionally close to their control counterparts, suggesting that the normal signaling circuits are maintained in the islands.
We performed next-generation spatial transcriptomics (10X VisiumHD) which to date has the highest resolution for spatial gene detection, and our BioIT team is developing a new algorithm and analysis pipeline to investigate these data. Preliminary analysis shows that the islands contain a normal niche, for example with the expression of BMP9 being maintained.
We compared the differentially expressed genes between control and LHX2-KO, generating a list of genes that are dependent on LHX2 intracellular signalization. We use the NicheNet algorithm to predict ligand-receptor pairs, i.e. predicting what molecules from neighboring cells can be detected by stellate cells. We used the LHX2-dependent gene list to narrow our prediction to interactions involved in the regulation of LHX2-dependent genes.
Additionally, we performed Multi-ome analysis of LHX2-KO liver cells, which combines epigenetic regulation and gene expression. This analysis will be complementary to bridge all transcriptomics data to determine how the Lhx2 gene expression is controlled, with the intent to discover the niche signal(s) maintaining stellate cell quiescence.
First, we performed single-nuclei transcriptomics to characterize stellate cells that are paired or unpaired with KCs. However, our preliminary transcriptomics analysis showed that stellate cells from the control and ALK1-KO transcriptomes were quite similar.
To verify if stellate cells in the ALK1-KO model were also behaving similarly to the control upon activation, I treated mice with Carbon tetrachloride (CCl4) to induce liver inflammation. There were no quantifiable differences between the paired and unpaired conditions, suggesting that unpairing stellate cells with KCs is not sufficient to disrupt their identity and that other signals from their niche are involved in maintaining their quiescence.
Since more signals seem to be necessary for stellate cell quiescent identity, we instead focused on directly disrupting stellate cells. LIM Homeobox 2 (LHX2) is a known stellate cell transcription factor that is down-regulated during stellate cell activation.
I characterized a new transgenic model in which we conditionally knocked-out Lhx2 (LRATcreERT2 x LHX2-flox/flox, thereafter referred to as LHX2-KO) specifically in stellate cells. Deletion of this gene in stellate cells has a drastic effect on the identity of stellate and other liver cells. Similarly to the ALK1-KO model, we observe islands of “escapee” cells where they maintain a normal phenotype.
Thus, we identified a new model that is relevant to answer our question, as it contains both paired and unpaired stellate cells, but also provides an environment where the cellular identity circuits are deeply perturbed.
We performed an in-depth transcriptomics analysis of the LHX2-KO model. Single cell transcriptomics showed that stellate cell identity was deeply perturbed, as they were displaying a more profibrotic (but not fully activated) gene profile. Escapee cells were transcriptionally close to their control counterparts, suggesting that the normal signaling circuits are maintained in the islands.
We performed next-generation spatial transcriptomics (10X VisiumHD) which to date has the highest resolution for spatial gene detection, and our BioIT team is developing a new algorithm and analysis pipeline to investigate these data. Preliminary analysis shows that the islands contain a normal niche, for example with the expression of BMP9 being maintained.
We compared the differentially expressed genes between control and LHX2-KO, generating a list of genes that are dependent on LHX2 intracellular signalization. We use the NicheNet algorithm to predict ligand-receptor pairs, i.e. predicting what molecules from neighboring cells can be detected by stellate cells. We used the LHX2-dependent gene list to narrow our prediction to interactions involved in the regulation of LHX2-dependent genes.
Additionally, we performed Multi-ome analysis of LHX2-KO liver cells, which combines epigenetic regulation and gene expression. This analysis will be complementary to bridge all transcriptomics data to determine how the Lhx2 gene expression is controlled, with the intent to discover the niche signal(s) maintaining stellate cell quiescence.
The broad impact that the deletion of one gene in stellate cells has highlighted that liver cells function as a module, where the four cell types maintain each other identities and with stellate cells at the center. Most importantly, this phenotype takes place when stellate cells lose only part of their identity (upregulation of some profibrogenic genes) without being completely activated (no upregulation of genes typical of activated stellate cells) and without inflammatory stimuli. This paves the way to a new paradigm that identity and activation of stellate cells are uncoupled and probably controlled by two separate mechanisms. This is especially relevant to better understand the molecular mechanisms behind stellate cell activation during inflammatory conditions.
Our Bio-IT team is developing innovative algorithms and pipelines to analyze the data from the state-of-the-art transcriptomics techniques. Since our biological question proved to be more complex than anticipated, further research is needed to untangle the signaling circuits connecting the four liver cell types, and understand what effects are direct or indirect. Nevertheless, we have produced a very complete dataset that will help address these questions. Importantly, by integrating multi-ome analysis with our single-cell and spatial transcriptomics datasets, we seek to determine how LHX2 gene expression is controlled in stellate cells, and what cellular signals are behind maintaining stellate cell quiescent identity.
Our Bio-IT team is developing innovative algorithms and pipelines to analyze the data from the state-of-the-art transcriptomics techniques. Since our biological question proved to be more complex than anticipated, further research is needed to untangle the signaling circuits connecting the four liver cell types, and understand what effects are direct or indirect. Nevertheless, we have produced a very complete dataset that will help address these questions. Importantly, by integrating multi-ome analysis with our single-cell and spatial transcriptomics datasets, we seek to determine how LHX2 gene expression is controlled in stellate cells, and what cellular signals are behind maintaining stellate cell quiescent identity.