Embryonic liver developmental trajectory revealed by single-cell RNA sequencing in the Foxa2eGFP mouse

The liver and gallbladder are among the most important internal organs derived from the endoderm, yet the development of the liver and gallbladder in the early embryonic stages is not fully understood. Using a transgenic Foxa2eGFP reporter mouse line, we performed single-cell full-length mRNA sequencing on endodermal and hepatic cells isolated from ten embryonic stages, ranging from E7.5 to E15.5. We identified the embryonic liver developmental trajectory from gut endoderm to hepatoblasts and characterized the transcriptome of the hepatic lineage. More importantly, we identified liver primordium as the nascent hepatic progenitors with both gut and liver features and documented dynamic gene expression during the epithelial-hepatic transition (EHT) at the stage of liver specification during E9.5–11.5. We found six groups of genes switched on or off in the EHT process, including diverse transcripitional regulators that had not been previously known to be expressed during EHT. Moreover, we identified and revealed transcriptional profiling of gallbladder primordium at E9.5. The present data provides a high-resolution resource and critical insights for understanding the liver and gallbladder development.

In this work Mu and colleagues use single-cell RNA-seq (scRNA-seq) to identify the developmental trajectory of embryonic liver from gut endoderm to hepatoblasts and characterize the transcriptome of the hepatic lineage. They found both gut and liver features in the emerging liver primordium and dynamic gene expression changes during the epithelial-hepatic transition (EHT) at the stage of liver specification from E9.5-E11.5, with six groups of genes switched on or off in the EHT process. RXR signaling and transcription factors including Lzts1 were identified as potential regulators of EHT. Additionally, the authors identified the transcriptional profile of the gallbladder primordium at E9.5. Together the data provides a high-resolution map of gene expression changes during hepatic specification that should be useful for understanding the liver and gallbladder development.

Critique:
Overall this study is clearly described, technically well executes and does provide an important resource. However, the study is exclusively descriptive and no effort has been made to follow up predictions about for example RXR signaling or Lzts1 as regulators of EHT. This is a weakness that should be addressed in a revised version.
Reviewer #2 (Remarks to the Author): Mu and colleagues performed single-cell RNA sequencing on the FOXA2 positive cells, endogenously labeled with eGFP in the transgenic mice. Authors have isolated these eGFP-positive cells from different stages of the embryonic mice to understand the liver and gallbladder organogenesis at the single-cell resolution. In addition to the scientific advancements, the manuscript provides an enormous amount of genomics data, and therefore, could serve as a great resource for the community. Moreover, authors have appropriately taken all the necessary quality check measurements to ensure that all the claims are made from the good quality cells.
In order to further improve the quality of the manuscript, I have a few specific comments: 1. To ensure the high quality of the cells, authors have used various filtering criteria, including the minimum expression, the minimal number of expressed genes per cell, etc. It will be informative for the readers if authors can graphically depict these parameters, and could clearly indicate the total number of cells after each filtering criteria, further segregated into distinct developmental time points. This new figure will give an overview of the sanity checks and the information about the number of contributing cell-types at each developmental time-point. So far all this information is in the text. In this work Mu and colleagues use single-cell RNA-seq (scRNA-seq) to identify the developmental trajectory of embryonic liver from gut endoderm to hepatoblasts and characterize the transcriptome of the hepatic lineage. They found both gut and liver features in the emerging liver primordium and dynamic gene expression changes during the epithelial-hepatic transition (EHT) at the stage of liver specification from E9.5-E11.5, with six groups of genes switched on or off in the EHT process. RXR signaling and transcription factors including Lzts1 were identified as potential regulators of EHT. Additionally, the authors identified the transcriptional profile of the gallbladder primordium at E9.5. Together the data provides a high-resolution map of gene expression changes during hepatic specification that should be useful for understanding the liver and gallbladder development.

Point-by-Point Response to Reviewers' Comments:
Thanks for your comments. They are useful and provide an enhanced clarity to our readers! We appreciate your effort and time! In this rebuttal, the text in blue represents our response to your comments (in black) and the text (in red) denotes the changes made to the revised manuscript.

Critique:
Overall this study is clearly described, technically well executes and does provide an important resource. However, the study is exclusively descriptive and no effort has been made to follow up predictions about for example RXR signaling or Lzts1 as regulators of EHT. This is a weakness that should be addressed in a revised version.
Thanks for the critique. We appreciate the reviewer's point that by including details about Lzts1 and RXR in the abstract. It led to the suggestion that we had done extensive validation work on their function during liver specification; as such, we deleted that sentence from the abstract and modified the results to minimize implications about functions of the genes. On the other hand, we realized that our original abstract didn't sufficiently feature our exciting discovery of many new transcription factors that are induced at the time of liver specification, so we have added that point to the revised abstract. We feel that the revised paper provides better focus on the extensive new information that we have discovered, and thank the reviewer for raising their point.
Page 6 : "We found six groups of genes switched on or off in the EHT process, including diverse transcriptional regulators that had not been previously known to be expressed during EHT." Then, in the text page 17, we tweaked the header of the relevant section a bit, to soften the claims.
Page 17: "Significant transcription factors and RXR complex signaling dynamics during liver specification" For the predictions for LXR/RXR signaling, we did described some interpretations for the role of LXR/RXR during EHT in the previous manuscript (Page 18, cited below), but perhaps the explanation was too long and not sufficiently understandable. By Ingenuity pathway analysis (IPA), we found that LXR/RXR signaling was very highly upregulated based on the high expression of both upstream and downstream genes ( Figure S10D). Interestingly, by Motif analysis, we found that the promoters of Alb, C3, Apo and Serpin had putative RXRA elements ( Figure 3F). Together, Alb and Serpin served as both the ligands and the targets in LXR/RXR pathway. A positive-feedback loop of LXR/RXR pathway during EHT ( Figure 3G) is implied and able to explain that Alb and Serpin family genes were increased over 1,000-fold in a short time in liver primordium compared with the gut tube.
To validate the role of the LXR/RXR pathway, we analyzed the promoters of the differentially expressed genes between the gut tube and hepatoblasts by motif analysis. The promoters of 49 genes (including Alb, C3, Apo and Serpin family members) highly expressed in the hepatoblasts had putative RXRA elements (Table S8). The expression of these target genes increased in the liver primordium and peaked within the liver bud ( Figure 3F). Combined with IPA analysis, Alb and Serpin family served as both the ligands and the targets in the LXR/RXR pathway, which implies a positive-feedback loop during liver specification ( Figure 3G). " As a transcription factor, Lzts1 was barely not expressed in gut tube and started to be expressed in liver primordium by our finding, which was not reported previously. We validated the expression of Lzts1 in hepatoblasts at E11.5 by immunofluorescence to support our finding ( Figure 3E). We tried to explore more information related to Lzts1 during EHT. Unfortunately, no strong evidence was found about the connection between other genes and Lzts1 during EHT, either by pathway analysis, motif analysis or network analysis. This is might due to limited information about the function of Lzts1. Our data presents a preliminary discovery of expression of Lzts1 in hepatoblasts, and more investigation such as knock-out experiments is needed to reveal the role of Lzts1 during liver development in the future. Mu and colleagues performed single-cell RNA sequencing on the FOXA2 positive cells, endogenously labeled with eGFP in the transgenic mice. Authors have isolated these eGFP-positive cells from different stages of the embryonic mice to understand the liver and gallbladder organogenesis at the single-cell resolution. In addition to the scientific advancements, the manuscript provides an enormous amount of genomics data, and therefore, could serve as a great resource for the community. Moreover, authors have appropriately taken all the necessary quality check measurements to ensure that all the claims are made from the good quality cells.

Characterizing the Emergence of Liver and Gallbladder from the Embryonic Endoderm through Single-Cell RNA-Seq
Thanks for your comments. They are useful and provide an enhanced clarity to our readers! We appreciate your effort and time! In this rebuttal, the text in blue represents our response to your comments (in black) and the text (in red) denotes the changes made to the revised manuscript.
In order to further improve the quality of the manuscript, I have a few specific comments: 1. To ensure the high quality of the cells, authors have used various filtering criteria, including the minimum expression, the minimal number of expressed genes per cell, etc. It will be informative for the readers if authors can graphically depict these parameters, and could clearly indicate the total number of cells after each filtering criteria, further segregated into distinct developmental time points. This new figure will give an overview of the sanity checks and the information about the number of contributing cell-types at each developmental time-point. So far all this information is in the text.
Very good suggestion and absolutely! To make the manuscript more informative for the readers, in this revised manuscript, we have graphically depicted the parameters suggested by the reviewer 2, including minimal expressed gene number per cell, minimal mapped reads per cell, cell sub-clusters (Foxa2+/-) and cell number for each cluster at different developmental time points (Supplemental Fig. S7A of revised manuscript). 2. Authors have performed standard single-cell bioinformatics analysis, including DGE, pseudotemporal analysis, etc. Although the methods are explained well, but to really ensure the reproducibility of each figure panels, authors must provide the codes in GitHub, or at least for the reviewer's assessment.
Thank for this constructive comment! We have provided the source codes for single-cell bioinformatics analysis in our study and they can be accessed on Github (https://github.com/CellOmics-Yu/Mus_liver_development) now.
3. Importantly, although the authors have shown a good correlation between EGFP and FOXA2 gene across all selected genes, still it does not address the point that the transgenic mice indeed recapitulate the endogenous expression of FOXA2 (eGFP-/FOXA2+ population?). Therefore, since this piece of information is vital for all the downstream claims, the authors must quantitatively address this issue. The authors can perform the whole-mount two-color immunohistochemistry/in situ hybridization with anti-eGFP and in situ hybridization (probe for endogenous FOXA2). The authors should quantitatively represent the findings.
That is a very good point. I understand the concern that there might be some single-cells which were eGFP-/FOXA2+. However, we think this possibility is quite low. As eGFP was inserted to the exon3 of endogenous Foxa2 locus ( Figure 1A), this transgenic mice are able to isolate single-cells with the endogenous expression of Foxa2 by sorting eGFP fluorescence. These two genes were linked together, shared the same promoter of endogenous Foxa2, and should be co-expressed theoretically.
At the beginning, we did try to co-stain both FOXA2 and eGFP by immunofluorescence to show that the two proteins are co-expressed in hepatoblasts. We tried two eGFP antibody (Santa Cruz, sc-8334 and sc-9996), and unfortunately both antibodies failed to stain eGFP. One possible reason is that these two eGFP antibodies were both raised by full length GFP (amino acids 1-238) and N-terminal of eGFP was linked to Cterminal of FOXA2, which led to the failure of eGFP staining.
We have carefully calculated the correlation index between the expression of Foxa2 and eGFP quantitatively (Pearson r=0.95) ( Figure 1E). In Figure 1E, we can find that different single-cells have various expression of Foxa2/eGFP. However, in a particular single-cell, the expression of Foxa2 is generally equal to the expression of eGFP. Moreover, we can conclude there is basically no Foxa2+/eGFPsingle-cells, which confirmed that eGFP and Foxa2 were co-expressed. This way is more sensitive than in situ hybridization assay to quantify the expression level of Foxa2 and eGFP to draw the conclusion that eGFP and Foxa2 were co-expressed in single-cells.
4. Quantification of the co-labeling experiment ( Figure 1D) is missing.
Thanks for your timely suggestion! We have done three replicates (three slides) of the co-labeling experiments of DLK1 and FOXA2. We also quantified the cell number of FOXA2+/DLK1+ cells, FOXA2+/DLK1cells, and FOXA2-/DLK1+ cells by Venn plot (Figure S1E, below). The results quantitatively verified that FOXA2 and DLK1 were co-expressed. S1E, Quantification of co-labeling immunofluorescence assay of FOXA2 and DLK1 of three replicates at E12.5 by Venn plot.
In the current revison the authors have addressed some of the reviewers' concerns, but failed to address others.
Major point 1: Reviewer 1's concerns about this being a purely descriptive study with no efforts to experimentally test the predictions they make based on the scRNA-seq analyses is still a concern.
Major point 2: Reviewer 2's concern about co-expression of Foxa2 and eGFP and request for Foxa2/eGFP co-staining has apparently been attempted using two Santa Cruz antibodies that didn't detect the fusion protein. The authors suggest that the fusion protein context is preventing the antibodies from detecting eGFP. Were positive controls using FL eGFP included? Given that there is a plethora of anti-eGFP antibodies it should be possible to perform this analysis, which is critical as Reviewer 2 points out