Architecture and evolutionary conservation of Xenopus tropicalis osteoblast-specific regulatory regions shed light on bone diseases and early skeletal evolution
This article has been Reviewed by the following groups
Listed in
- Evaluated articles (Review Commons)
Abstract
Understanding the genetic mechanisms underpinning the differentiation of osteoblasts, the bone producing cells, has far reaching implications for skeletal diseases and evolution. To this end, it is crucial to characterize osteoblastic regulatory landscape in a diverse array of distantly-related vertebrate species. By comparing of the ATAC-seq profile of Xenopus tropicalis ( Xt ) osteoblasts to liver, heart and lung control tissues, we identified 524 promoters and 6,750 distal regions whose chromatin is specifically open in osteoblasts. Nucleotide composition, Gene Ontology, and RNA-Seq confirmed that the identified elements correspond to bona fide osteogenic transcriptional enhancers, and TFBS enrichment revealed a well-conserved regulatory logic with mammals. Amongst the 357 Xt osteoblast-specific enhancers aligning to homologous human loci, 127 map to regions annotated as enhancers. Phenotype predictions based on the genes neighbouring these conserved enhancers are tightly related to impaired skeletal development. In addition, six conserved enhancers are located at loci associated to craniosynostosis ( mx2 , tcf12 ), osteopoikilosis ( lemd3 ), osteopenia ( gorab ), skeletal dysplasia ( flnb ) and craniofacial abnormalities ( gpc4 ). From an evolutionary perspective, the elephant shark genome aligns to 53 Xt osteoblast-specific enhancers that are also conserved and annotated as enhancers in humans, revealing an ancestral osteogenic role for the ATOH8, IRX3, NFAT, NFIB and MEF2C transcription factors, as well as for the FGF, IHH and BMP/TGFb signalling pathways. As the absence of bone in sharks is a derived feature, we propose that, in this lineage, the osteogenic regulatory network has been maintained for its function in odontoblasts. Our data argues in favour of a common origin for dentine and bone, and provides a glimpse into the key regulatory elements and upstream activators that drove the formation of an ancient type of mineralized tissue in the vertebrates that inhabited the oceans more than 460 million years ago.
Author Summary
During animal embryogenesis, distinct type of tissues are formed and assembled, resulting in an integrated, functional organism. During this process, cells must make important decisions, which largely rely on an accurate use of their genetic material. Here, we have studied how the genome “knows” that it must participate to the formation of the bone tissue in a frog animal model. We therefore identified important genomic regions that are involved in driving the expression of genes involved in the formation of a mineralized skeleton. On the one hand, we show that some of these regions are also present in humans, and, therefore, skeletal pathologies could be studied in the frog model at a genetic level. On the other hand, we also identify regions that are present in the genome of a shark, which allows us to propose an evolutionary framework for the early evolutionary origin of the vertebrate skeleton.
Article activity feed
-
Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.
Learn more at Review Commons
Reply to the reviewers
We are sincerely thankful to all reviewers for their work and constructive comments that allowed us to improve the quality of the present manuscript. We are very pleased to announce that we were able to tackle all the raised concerns (except the reporter assays which is a focus of future research for our laboratory, see below), and would like to briefly mention here three major improvements:
- We have crossed our data in frog with available human ATAC-Seq datasets. We have followed a similar approach to the one we employed in Xenopus, by “subtracting” human osteoblastic ATAC-Seq peaks with human liver, heart and lung. This cross-species validation …
Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.
Learn more at Review Commons
Reply to the reviewers
We are sincerely thankful to all reviewers for their work and constructive comments that allowed us to improve the quality of the present manuscript. We are very pleased to announce that we were able to tackle all the raised concerns (except the reporter assays which is a focus of future research for our laboratory, see below), and would like to briefly mention here three major improvements:
- We have crossed our data in frog with available human ATAC-Seq datasets. We have followed a similar approach to the one we employed in Xenopus, by “subtracting” human osteoblastic ATAC-Seq peaks with human liver, heart and lung. This cross-species validation strategy led to the identification of osteoblast-specific NFRs in human that compare very well to the Xenopus osteoblastic regulatory landscape (new Fig 6).
- We have included ChIP-Seq data that was performed by Patricia Hanna, a former PhD student from our laboratory, in collaboration with Laurent Sachs and Nicolas Buisine (these three researchers were incorporated as new co-authors). We were planning to publish this ChIP-Seq separately but find that it contributes very well to this manuscript (modified Fig 4, new Fig 7).
- We have included in situ hybridization analyses on frog and shark performed by David Muñoz, a former PhD student from our laboratory, in collaboration with Melanie Debiais-Thibaud and Catherine Boisvert (these three researchers were incorporated as new co-authors). This data ends nicely the manuscript by providing a biological dimension and by strengthening our evolutionary model (See new Fig 7).
We hope that our responses match the quality criteria of Review Commons and of its affiliated journals, thank you very much once again and kind regards, Sylvain Marcellini
Point-by-point description of the revisions:
Reviewer #1 (Evidence, reproducibility and clarity (Required)): Summary: The paper investigates the genetic mechanisms driving osteoblast differentiation in Xenopus tropicalis, shedding light on bone diseases and early skeletal evolution. Through ATAC-seq analysis, the study identifies osteoblast-specific regulatory regions, confirming their role as osteogenic transcriptional enhancers. A substantial number of these enhancers are conserved in humans, potentially offering insights into skeletal disorders. Additionally, the research highlights an evolutionary perspective by revealing shared regulatory elements between Xenopus tropicalis and the elephant shark, suggesting an ancient origin for mineralized tissues in vertebrates.
Major comments:
Methodology of this paper is kinda vague and the paper seems to be fragmented and not logically organized in a linear fashion.
Reply: We have improved the methodology section. We provide the accession numbers for all raw sequencing datasets generated for this study have been submitted and linked to the NCBI BioProject database (page 27). The paper has been almost completely rewritten and the figures substantially modified. There are now less figure which contain more information presented in a friendly fashion. The logic of the paper is as follows: -Identification of enhancers and promoters (Figs 1 and 2) -Characterization of their nucleotide sequence and TFBSs (Fig 3) -Validation with RNA-Seq and ChIP-Seq (Fig 4) -Global sequence conservation (Fig 5) -Cross validation with ATAC-Seq in human (Fig 6) -Evolutionary model (Fig 7).
Authors could provide evidentiary support that the control tissues are non-mineralized (and exp tissues are) by simple calcein staining. Mineralization occurs during tadpole stage, and calcification of heart and lung tissue in amphibians is not well understood. This will strengthen the attestation of these tissues as controls and provide a useful diagram for exactly what tissues were used.
Reply: We have performed Alizarin reg staining on larval skull, liver, heart and lung and show that, like in mammals, only the calvaria is mineralized (see page 6 and new Supporting Information 1).
There appears to be no mention of osteocytes or other cell types. What measures were taken to ensure that osteoblasts are the principal cell type being described? The reference for bone tissue extraction refers to a cell culture technique in which it is likely no osteocytes would prevail.
Reply: This is an important point to clarify because osteoblasts and their osteocytic progeny harbour a completely different function, physiology and gene expression profile. Our laboratory has studied frog osteocytes in details (Fritz et al, 2018), and we have added the following sentence “Of note, this extraction procedure does not harvest osteocytes that lie embedded within the bone matrix, allowing us to exclusively study osteoblasts. As controls, we also included larval liver, heart and lung following the criteria that they are nonmineralized (Supporting information 1) and unrelated to skeletal tissues”. See page 6.
Minor comments:
Data on conservation of mentioned transcription factors could be easily added (NFAT, etc.)
Reply: We have performed extensive protein alignments showing broad conservation of the osteogenic transcription factors for which we detected binding site enrichment in osteoblast-specific enhancers (see page 10 and new Supporting Information 7).
The data presentation is poor, especially figure 2 and figure 4.
Reply: Following the reviewer’s advice these figures have been eliminated and replaced by Figures 2B and 2C, which, we believe, present the same information in a much clearer and friendly fashion.
Line 115-117: "By focusing on annotated Xt transcription start sites (TSSs), we found that the ATAC-Seq NFR and mononucleosome signals form two distinct clusters," it would be helpful to briefly explain the significance of these two clusters. What does it indicate about the regulatory regions associated with TSSs?
Reply: We have clarified this point by being more explicit: “The first cluster is composed of 5,949 promoters harbouring a robust NFR located immediately upstream of the TSS and flanked by two well-positioned nucleosomes (Fig 1B, left panel), likely corresponding to expressed genes. By contrast, the second cluster contains 16,947 promoters showing weak NFR and diffuse mononucleosome signals (Fig 1B, right panel), and is probably enriched in transcriptionally repressed genes or genes expressed at low levels”. See Page 6.
Line 133-139: When discussing hierarchical clustering and the similarity of NFR landscapes between different tissues, you could provide a sentence or two to speculate on the potential biological implications. For instance, why might heart and lung tissues exhibit more similarity in NFR landscapes compared to osteoblasts and liver?
Reply: This is an interesting point to raise because there is data in the literature supporting our findings. We have modified the following sentence on page 7: “Hierarchical clustering showed that the landscape of the NFRs from heart and lung are more similar to each other than to osteoblasts or liver, which is true both for TSS and non-TSS regions (Fig 1D) and which parallels data obtained in mouse [10]”. Our novel analysis with human ATAC-Seq data also leads to the same finding (Page 13): “Available human liver, heart and lung ATAC-Seq datasets were retrieved, and hierarchical clustering confirmed a higher similarity for heart and lung, and that the osteoblast sample substantially differs from the three other tissues (Supporting information 11), similarly to the situation in frog (Fig 1D) and mouse [10]”.
Line 134: To enhance clarity, you might consider using phrases like "Figure 3A" and "Figure 3B" instead of "Compare Fig 3A and B" to directly refer to the figures in the text.
Reply: This has been corrected has we have deeply improved the figures. See “Globally, the Pearson correlation coefficient was much higher for TSS than non-TSS peaks (Fig 1D), a finding consistent with previous studies showing that, between distinct cell types, histone marks are largely invariable at promoters while they display highly context-dependent patterns at enhancers [6, 7].” on page 7.
Line 142-144: Please consider briefly explaining why you chose liver, heart, and lung tissues as controls. What specific characteristics or functions of these tissues make them suitable for this comparative analysis?
Reply: We now mention “As controls, we also included larval liver, heart and lung following the criteria that they are nonmineralized (Supporting information 1) and unrelated to skeletal tissues.” on page 6.
When discussing the potential function of osteoblastic enhancers in cartilaginous fish, you might briefly mention the role of cartilage in these organisms and how these enhancers may have evolved to regulate cartilage-related processes.
Reply: We agree with the reviewer that this is an exciting point which is of high interest for our laboratory (see for instance our review, Cervantes et al, 2017). However, as we feel that the manuscript is already quite long and has many references, we preferred not to discuss this point and to simply focus on the osteoblast/odontoblast aspect of skeletal evolution.
Ensure that the formatting of your methods section is consistent. For example, consistently use italics for software/tool names (e.g., "SAMtools") and follow a standard format for listing parameters or options used in software/tools.
Reply: We have corrected these points.
Reviewer #1 (Significance (Required)): The paper's significance lies in its elucidation of osteoblast-specific regulatory regions in Xenopus tropicalis. By characterizing these regions and connecting them to specific genes and pathways, the study advances our understanding of osteogenesis. Additionally, the identification of conserved elements across vertebrates provides insights into the deep evolutionary origins of skeletal features, offering a unique perspective on vertebrate evolution. However, one of the main limitations of the study is the lack of extensive experimental validation for the identified regulatory regions, leaving a gap in confirming their functionality.
Reply: Thank you very much again for your helpful and constructive comments. As a functional validation, at least from the chromatin perspective, we have incorporated ChIP-Seq data (Fig 4) with four key histone marks present at active promoters (H3K4me3), active enhancers (H3K4me1), and at active chromatin (H3K27Ac) and repressed chromatin (H3K27me3). This ChIP-Seq was already available in our laboratory (thereby explaining the incorporation of three new co-authors, Dr Hanna, Dr Sachs and Dr Buisine), but we were planning to incorporate it in a different manuscript. However, we feel that it is important to include it in the present paper. Another functional validation lies in the identification of 138 conserved osteogenic enhancers harbouring a NFR both in frog and human (Fig 6). We do not intend to incorporate reporter assays at this stage, as this is a future direction of research for our laboratory, together with CRISPR mutagenesis.
Reviewer #2 (Evidence, reproducibility and clarity (Required)): In this study, Hector Castillo and the coauthors conducted ATAC-seq and RNA-seq analyses across several cell types in Xenopus tropicalis (Xt) to identify regulatory elements specific to osteoblasts. They explored the evolutionary conservation of the osteoblast regulatory elements across species. Their research encompassed the identification of osteoblast-specific regulatory elements through cross-tissue analysis, offering comprehensive insights into tissue-specific regulatory elements. These insights included cell-type-specific chromatin accessibility, biological functions predicted by gene ontology analysis, and potential transcriptional regulators associated with these regions. The cross-species analysis unveiled partial conservation of osteoblast-specific regulatory regions between the Xt and the human genome, with the shared genomic regions being linked to osteoblast-related genes. Additionally, the enriched transcription factors were identified in these regions. The study further explored comparative analyses involving multiple species, providing evolutionary insights into the gene regulatory mechanisms underlying osteoblast identity and pathology.
Major comment All the cross-species analyses in this study were primarily based on sequence conservation. However, since human osteoblast ATAC-seq data, as well as ChIP-seq and Hi-C data, are publicly available (PMID: 35906483), conducting a direct comparative analysis between Xenopus tropicalis (Xt) osteoblast ATAC-seq and human osteoblast ATAC-seq could provide more concrete evidence regarding the conservation of chromatin-accessible regions between these two species. This additional analysis has the potential to significantly strengthen the conclusions drawn in the study.
Reply: We are thankful to the reviewer for this insightful comment that dramatically improved the scope of our work. We have indeed incorporated available ATAC-Seq experiments performed on human osteoblasts (SRR12933513 and SRR12933514), liver (SRR21927033 and SRR21927032), heart (SRR21927531 and SRR21927534) and lung (SRR21927095 and SRR21927098). This is explained on pages 13-14 (results), pages 19-20 (discussion) and pages 22-24 (methods). Hence, we have uncovered 138 conserved enhancers that display an osteoblast-specific NFR both in frog and human (see new Fig 6). As the reviewer states, we believe that our conclusions have been significantly strengthened, allowing us to reformulate the manuscript title which now vehiculates a more functional message. Also, thanks to this comment, we were able to propose a more attractive title for our work: “Cross-validation of conserved osteoblast-specific enhancers illuminates bone diseases and early skeletal evolution”.
Minor comment 1 The authors made use of "annotated human enhancers" in their study; however, the specific definition or source of this annotation was not provided in the manuscript. It is crucial that the authors clarify the criteria or source used for annotating human enhancers to ensure transparency and allow readers to better understand the basis of their analyses and conclusions.
Reply: The reviewer is correct. These “annotated human enhancers” have now completely been eliminated for the study and replaced by the analysis shown in Fig 6 (and see our reply to the previous comment).
Minor comment 2 In relation to the association studies conducted between Xenopus tropicalis (Xt) osteoblast enhancers and genes related to human bone diseases, it's important for the authors to express their statements with caution. While the putative target genes may be potentially regulated by shared regulatory elements between Xt and humans, there exists no direct evidence demonstrating that these regulatory regions are the causative factors behind these diseases. It's worth noting that there are several other open chromatin regions in proximity to these putative target genes. As a result, the shared genomic regions may or may not have a direct relationship with human diseases. To establish a substantial linkage, more in-depth analyses would be required to provide evidence of a pathological connection.
Reply: This is an important point, on page 14 we now state “While the osteoblast-specific regulatory regions reported here might not be directly involved in the aetiology of the aforementioned diseases, their identification considerably improves our understating of the transcriptional control of these genes”.
Minor comment 3 In lines 394 to 397, the authors assert that the enrichment of TWIST1/2 transcription factor binding sites (TFBS) at Xenopus tropicalis (Xt) osteogenic enhancers is a novel finding. However, this claim lacks clarity regarding the novelty of this discovery, given that they reference previous literature (reference 42) that has already demonstrated the involvement of TWIST1/2 in osteoblast differentiation. The authors should provide a more precise explanation of how their specific findings related to TWIST1/2 TFBS enrichment contribute to existing knowledge or differ from previous studies to clarify the novelty of their results.
Reply: We now provide a clearer explanation by mentioning “In this respect, the reported enrichment in TWIST1/2 TFBS (Fig 3 and Supporting information 5) represents the first evidence that TWIST proteins might control the timing of osteoblastic differentiation through binding to hundreds of osteogenic enhancers, a possibility that could be confirmed by ChIP-Seq” on page 19.
Minor comment 4 Depositing the NGS data, including ATAC-seq and RNA-seq datasets, in a public database would be a valuable contribution to the research community.
Reply: Yes, this data has now been made available, see pages 26-27: “Data Availability. The raw sequencing datasets generated for this study have been submitted and linked to the NCBI BioProject database with the following accession numbers: PRJNA1011469 (ATAC-seq), PRJNA1021677 (RNA-seq), and PRJNA1056467 (ChIP-seq)”.
Reviewer #2 (Significance (Required)): The comparative analysis of ATAC-seq among different cell types in Xenopus tropicalis (Xt) provides a broad perspective on cell-type-specific chromatin accessible regions, which is a notable strength of the study. It's worth highlighting that, as far as known, this study represents the first report of ATAC-seq in Xt osteoblasts. However, it's important to acknowledge that the overall message of the study is consistent with previous findings in mammals. For example, the observation that non-transcription start site (TSS) regions were more cell-type-specific, correlating with cell-type distinct gene expressions, aligns with findings in mammalian systems. Additionally, many of the osteoblast regulators predicted from the data are already known osteogenic factors in mammals. The cross-species analysis provides valuable insights into the evolutionary aspects of putative enhancers in osteoblasts. The study identifies conserved gene regulatory regions and putative transcription factors associated with these genomic regions, shedding light on their potential roles in gene regulation. Moreover, the identification of conserved regions possibly linked to human skeletal diseases is a noteworthy aspect of the research, showcasing its strengths. However, it's essential to acknowledge a potential limitation related to this aspect of the study: the analyses conducted so far have been descriptive, primarily focusing on DNA sequence conservation. Given that several osteoblast ATAC-seq datasets from different species are publicly available, a more direct comparison between the Xt dataset and these other datasets could provide a deeper understanding of enhancer conservation and evolution. This study offers valuable resources for researchers in the field of skeletal biology and evolution. The comprehensive analysis of osteoblast-specific regulatory elements in Xenopus tropicalis, along with insights into their conservation and potential roles in human skeletal diseases, provides a foundation for further investigations in this area. Additionally, the evolutionary insights offered by the cross-species analysis contribute to the growing body of knowledge in evo-devo studies, shedding light on the evolution of gene regulatory mechanisms related to osteoblast identity. These resources and insights can serve as a valuable reference and guide for future research endeavors in both bone biology and evolutionary developmental biology. This reviewer specializes in the study of gene regulatory mechanisms in skeletal development and metabolism, primarily utilizing mouse and human tissues.
Reply: Thank you very much again for your helpful and constructive comments.
Reviewer #3 (Evidence, reproducibility and clarity (Required)): Summary Starting with ATAC-seq on the Xenopus tropicalis (Xt) genome, the authors tried to identify regulatory regions, which were evolutionally conserved and critical for osteoblasts, through computational approaches. They obtained profiles of the nucleosome-free regions (NFRs), i.e., open chromatin regions, in bone, liver, heart, and lung of Xt. The NFRs contain TSS-associated regions (TSS regions) and non-TSS regions. They then identified tissue-specific NFRs. Tissue-specific NFRs were predominantly located in introns and intergenic regions, and the trend was more highlighted in non-TSS regions. Regarding osteoblast-specific NFRs, non-TSS regions were associated with genes related to osteoblasts. Osteoblast-specific TSS- and non-TSS regions were enriched with motifs of osteoblast-related transcription factors (TFs), including Smad, AP-1, TEAD, Runx2, Nfic, Twist, and Nfat. By integrating ATAC-seq data with RNA-seq data, they found that osteoblast-specific NFRs were associated with transcriptionally active genes. When inter-species conservation of the Xt tissue-specific NFRs was analyzed, osteoblast-specific ones were well conserved in human, chick, and Callorhinchus milii (elephant shark). The authors further identified human homologous regions to Xt osteoblast-specific NFRs, which were enriched with binding motifs of osteoblast-related TFs, proposing putative osteogenic enhancers associated with skeletal diseases. Lastly, they identified a set of Xt osteoblast-specific NFRs that were conserved with the human, chick, and elephant shark genomes. The putative target genes of NFRs are enriched with osteogenesis-related TFs. Based on these data, they propose that evolutionary origins of osteoblast and odontoblasts are common, given that elephant shark is a cartilaginous fish, where bone is absent but odontoblast is present.
Major comments A major critical concern on this work is that their findings and claim fully rely on bioinformatic analyses. Bioinformatic prediction should be verified by wet-type experiments. Otherwise, it is quite difficult to draw definitive conclusions. In particular, it remains to be verified if the "putative enhancers" that they computationally identified have actual enhancer activities in in-vivo contexts. ATAC-seq alone identifies open chromatin regions on the genome and is not enough to define the location of enhancers and their activities. The authors need to perform ChIP-seq for enhancer marks and reporter assays for enhancer activities, in order to verify their prediction on at least several key regions they propose.
Reply: We have taken very seriously the reviewer´s comments and have incorporated three major experimental validations that go beyond bioinformatic analyses: -ChIP-Seq data on 4 key histone marks, previously performed in our laboratory, performed on Xenopus primary osteoblasts (see Fig 4). -Available human ATAC-Seq data for osteoblasts and control tissues (see new Fig 6). -In situ hybridization on elephant shark dental plates (see new Fig 7). We therefore have deeply modified the whole manuscript and now propose a more attractive title for our work: “Cross-validation of conserved osteoblast-specific enhancers illuminates bone diseases and early skeletal evolution”. We were not able to incorporate Reporter assays because (i) these experiments are lengthy, (ii) the current manuscript is already quite extensive and (iii) this is a major future research focus of our laboratory.
Minor comments Line 146: Fig S2 is unlikely to be provided.
Reply: We would like to keep this data available for readers, former Fig S2 is now “Supporting Information 3”.
Lines 158 to 163 and Fig. 4: GO analysis was performed only on non-TSS peaks. What about TSS peaks?
Reply: We now state on page 8 “Due to the low number of regions, no significant results were obtained with lung-specific non-TSS ATAC-Seq peaks, or with any category of TSS”.
Line 269: In the text, the authors describe that 48 osteoblast-specific TSS peaks are aligned to corresponding regions on the human genome. However, Fig. S7 shows 46 peaks are aligned. Please double-check.
Reply: This discrepancy has now been corrected.
Lines 289 to 296, Figs. 8, and S11: Although TRPS1 appears in Fig. S11, the authors did not mention it in the main text and Fig. 8. Why is the gene specifically excluded from the explanation?
Reply: This omission has now been corrected and now trps1 appears in Fig 6C, in Supporting Information 12, and is mentioned in the abstract and at pages 13-14 “Some cross-validated osteoblastic promoters and enhancers are located at loci of genes involved in skeletal diseases (See Supporting information 12 and Ref. [49]), such as osteoarthritis (adam12), osteoporosis (etv1), geroderma osteodysplasticum (gorab), keipert syndrome (gpc4), buschke-Ollendorff syndrome (lemd3), cleidocranial dysplasia (runx2) and trichorhinophalangeal syndrome type I (trps1).”.
Reviewer #3 (Significance (Required)):
- This work is potentially interesting, not just leading to identification of regulatory regions critical for osteoblast biology, but also providing evolutionary insight into bone development. However, as mentioned, lack of validation of bioinformatic prediction is a major weakness of this work. This work's concept would engage the interest in the field of bone development and skeletal transcriptional programs. However, the reviewer is not sure how much this work engages general interest.
- Expertise of the reviewer is mammalian skeletal development, particularly focusing on gene regulatory networks and epigenome during the process.
Reply: Thank you very much again for your helpful and constructive comments.
-
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #3
Evidence, reproducibility and clarity
Summary
Starting with ATAC-seq on the Xenopus tropicalis (Xt) genome, the authors tried to identify regulatory regions, which were evolutionally conserved and critical for osteoblasts, through computational approaches. They obtained profiles of the nucleosome-free regions (NFRs), i.e., open chromatin regions, in bone, liver, heart, and lung of Xt. The NFRs contain TSS-associated regions (TSS regions) and non-TSS regions. They then identified tissue-specific NFRs. Tissue-specific NFRs were predominantly located in introns and intergenic regions, and the trend was more highlighted in non-TSS regions. Regarding osteoblast-specific …
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #3
Evidence, reproducibility and clarity
Summary
Starting with ATAC-seq on the Xenopus tropicalis (Xt) genome, the authors tried to identify regulatory regions, which were evolutionally conserved and critical for osteoblasts, through computational approaches. They obtained profiles of the nucleosome-free regions (NFRs), i.e., open chromatin regions, in bone, liver, heart, and lung of Xt. The NFRs contain TSS-associated regions (TSS regions) and non-TSS regions. They then identified tissue-specific NFRs. Tissue-specific NFRs were predominantly located in introns and intergenic regions, and the trend was more highlighted in non-TSS regions. Regarding osteoblast-specific NFRs, non-TSS regions were associated with genes related to osteoblasts. Osteoblast-specific TSS- and non-TSS regions were enriched with motifs of osteoblast-related transcription factors (TFs), including Smad, AP-1, TEAD, Runx2, Nfic, Twist, and Nfat. By integrating ATAC-seq data with RNA-seq data, they found that osteoblast-specific NFRs were associated with transcriptionally active genes. When inter-species conservation of the Xt tissue-specific NFRs was analyzed, osteoblast-specific ones were well conserved in human, chick, and Callorhinchus milii (elephant shark). The authors further identified human homologous regions to Xt osteoblast-specific NFRs, which were enriched with binding motifs of osteoblast-related TFs, proposing putative osteogenic enhancers associated with skeletal diseases. Lastly, they identified a set of Xt osteoblast-specific NFRs that were conserved with the human, chick, and elephant shark genomes. The putative target genes of NFRs are enriched with osteogenesis-related TFs. Based on these data, they propose that evolutionary origins of osteoblast and odontoblasts are common, given that elephant shark is a cartilaginous fish, where bone is absent but odontoblast is present.
Major comments
A major critical concern on this work is that their findings and claim fully rely on bioinformatic analyses. Bioinformatic prediction should be verified by wet-type experiments. Otherwise, it is quite difficult to draw definitive conclusions. In particular, it remains to be verified if the "putative enhancers" that they computationally identified have actual enhancer activities in in-vivo contexts. ATAC-seq alone identifies open chromatin regions on the genome and is not enough to define the location of enhancers and their activities. The authors need to perform ChIP-seq for enhancer marks and reporter assays for enhancer activities, in order to verify their prediction on at least several key regions they propose.
Minor comments
Line 146: Fig S2 is unlikely to be provided. Lines 158 to 163 and Fig. 4: GO analysis was performed only on non-TSS peaks. What about TSS peaks? Line 269: In the text, the authors describe that 48 osteoblast-specific TSS peaks are aligned to corresponding regions on the human genome. However, Fig. S7 shows 46 peaks are aligned. Please double-check. Lines 289 to 296, Figs. 8, and S11: Although TRPS1 appears in Fig. S11, the authors did not mention it in the main text and Fig. 8. Why is the gene specifically excluded from the explanation?
Significance
- This work is potentially interesting, not just leading to identification of regulatory regions critical for osteoblast biology, but also providing evolutionary insight into bone development. However, as mentioned, lack of validation of bioinformatic prediction is a major weakness of this work. This work's concept would engage the interest in the field of bone development and skeletal transcriptional programs. However, the reviewer is not sure how much this work engages general interest.
- Expertise of the reviewer is mammalian skeletal development, particularly focusing on gene regulatory networks and epigenome during the process.
-
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #2
Evidence, reproducibility and clarity
In this study, Hector Castillo and the coauthors conducted ATAC-seq and RNA-seq analyses across several cell types in Xenopus tropicalis (Xt) to identify regulatory elements specific to osteoblasts. They explored the evolutionary conservation of the osteoblast regulatory elements across species. Their research encompassed the identification of osteoblast-specific regulatory elements through cross-tissue analysis, offering comprehensive insights into tissue-specific regulatory elements. These insights included cell-type-specific chromatin accessibility, biological functions predicted by gene ontology analysis, and potential …
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #2
Evidence, reproducibility and clarity
In this study, Hector Castillo and the coauthors conducted ATAC-seq and RNA-seq analyses across several cell types in Xenopus tropicalis (Xt) to identify regulatory elements specific to osteoblasts. They explored the evolutionary conservation of the osteoblast regulatory elements across species. Their research encompassed the identification of osteoblast-specific regulatory elements through cross-tissue analysis, offering comprehensive insights into tissue-specific regulatory elements. These insights included cell-type-specific chromatin accessibility, biological functions predicted by gene ontology analysis, and potential transcriptional regulators associated with these regions.
The cross-species analysis unveiled partial conservation of osteoblast-specific regulatory regions between the Xt and the human genome, with the shared genomic regions being linked to osteoblast-related genes. Additionally, the enriched transcription factors were identified in these regions. The study further explored comparative analyses involving multiple species, providing evolutionary insights into the gene regulatory mechanisms underlying osteoblast identity and pathology.
Major comment
All the cross-species analyses in this study were primarily based on sequence conservation. However, since human osteoblast ATAC-seq data, as well as ChIP-seq and Hi-C data, are publicly available (PMID: 35906483), conducting a direct comparative analysis between Xenopus tropicalis (Xt) osteoblast ATAC-seq and human osteoblast ATAC-seq could provide more concrete evidence regarding the conservation of chromatin-accessible regions between these two species. This additional analysis has the potential to significantly strengthen the conclusions drawn in the study.
Minor comment 1
The authors made use of "annotated human enhancers" in their study; however, the specific definition or source of this annotation was not provided in the manuscript. It is crucial that the authors clarify the criteria or source used for annotating human enhancers to ensure transparency and allow readers to better understand the basis of their analyses and conclusions.
Minor comment 2
In relation to the association studies conducted between Xenopus tropicalis (Xt) osteoblast enhancers and genes related to human bone diseases, it's important for the authors to express their statements with caution. While the putative target genes may be potentially regulated by shared regulatory elements between Xt and humans, there exists no direct evidence demonstrating that these regulatory regions are the causative factors behind these diseases. It's worth noting that there are several other open chromatin regions in proximity to these putative target genes. As a result, the shared genomic regions may or may not have a direct relationship with human diseases. To establish a substantial linkage, more in-depth analyses would be required to provide evidence of a pathological connection.
Minor comment 3
In lines 394 to 397, the authors assert that the enrichment of TWIST1/2 transcription factor binding sites (TFBS) at Xenopus tropicalis (Xt) osteogenic enhancers is a novel finding. However, this claim lacks clarity regarding the novelty of this discovery, given that they reference previous literature (reference 42) that has already demonstrated the involvement of TWIST1/2 in osteoblast differentiation. The authors should provide a more precise explanation of how their specific findings related to TWIST1/2 TFBS enrichment contribute to existing knowledge or differ from previous studies to clarify the novelty of their results.
Minor comment 4
Depositing the NGS data, including ATAC-seq and RNA-seq datasets, in a public database would be a valuable contribution to the research community.
Significance
The comparative analysis of ATAC-seq among different cell types in Xenopus tropicalis (Xt) provides a broad perspective on cell-type-specific chromatin accessible regions, which is a notable strength of the study. It's worth highlighting that, as far as known, this study represents the first report of ATAC-seq in Xt osteoblasts. However, it's important to acknowledge that the overall message of the study is consistent with previous findings in mammals. For example, the observation that non-transcription start site (TSS) regions were more cell-type-specific, correlating with cell-type distinct gene expressions, aligns with findings in mammalian systems. Additionally, many of the osteoblast regulators predicted from the data are already known osteogenic factors in mammals.
The cross-species analysis provides valuable insights into the evolutionary aspects of putative enhancers in osteoblasts. The study identifies conserved gene regulatory regions and putative transcription factors associated with these genomic regions, shedding light on their potential roles in gene regulation. Moreover, the identification of conserved regions possibly linked to human skeletal diseases is a noteworthy aspect of the research, showcasing its strengths. However, it's essential to acknowledge a potential limitation related to this aspect of the study: the analyses conducted so far have been descriptive, primarily focusing on DNA sequence conservation. Given that several osteoblast ATAC-seq datasets from different species are publicly available, a more direct comparison between the Xt dataset and these other datasets could provide a deeper understanding of enhancer conservation and evolution.
This study offers valuable resources for researchers in the field of skeletal biology and evolution. The comprehensive analysis of osteoblast-specific regulatory elements in Xenopus tropicalis, along with insights into their conservation and potential roles in human skeletal diseases, provides a foundation for further investigations in this area. Additionally, the evolutionary insights offered by the cross-species analysis contribute to the growing body of knowledge in evo-devo studies, shedding light on the evolution of gene regulatory mechanisms related to osteoblast identity. These resources and insights can serve as a valuable reference and guide for future research endeavors in both bone biology and evolutionary developmental biology.
This reviewer specializes in the study of gene regulatory mechanisms in skeletal development and metabolism, primarily utilizing mouse and human tissues.
-
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #1
Evidence, reproducibility and clarity
Summary:
The paper investigates the genetic mechanisms driving osteoblast differentiation in Xenopus tropicalis, shedding light on bone diseases and early skeletal evolution. Through ATAC-seq analysis, the study identifies osteoblast-specific regulatory regions, confirming their role as osteogenic transcriptional enhancers. A substantial number of these enhancers are conserved in humans, potentially offering insights into skeletal disorders. Additionally, the research highlights an evolutionary perspective by revealing shared regulatory elements between Xenopus tropicalis and the elephant shark, suggesting an ancient origin for …
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #1
Evidence, reproducibility and clarity
Summary:
The paper investigates the genetic mechanisms driving osteoblast differentiation in Xenopus tropicalis, shedding light on bone diseases and early skeletal evolution. Through ATAC-seq analysis, the study identifies osteoblast-specific regulatory regions, confirming their role as osteogenic transcriptional enhancers. A substantial number of these enhancers are conserved in humans, potentially offering insights into skeletal disorders. Additionally, the research highlights an evolutionary perspective by revealing shared regulatory elements between Xenopus tropicalis and the elephant shark, suggesting an ancient origin for mineralized tissues in vertebrates.
Major comments:
Methodology of this paper is kinda vague and the paper seems to be fragmented and not logically organized in a linear fashion. Authors could provide evidentiary support that the control tissues are non-mineralized (and exp tissues are) by simple calcein staining. Mineralization occurs during tadpole stage, and calcification of heart and lung tissue in amphibians is not well understood. This will strengthen the attestation of these tissues as controls and provide a useful diagram for exactly what tissues were used. There appears to be no mention of osteocytes or other cell types. What measures were taken to ensure that osteoblasts are the principal cell type being described? The reference for bone tissue extraction refers to a cell culture technique in which it is likely no osteocytes would prevail.
Minor comments:
Data on conservation of mentioned transcription factors could be easily added (NFAT, etc.) The data presentation is poor, especially figure 2 and figure 4. Line 115-117: "By focusing on annotated Xt transcription start sites (TSSs), we found that the ATAC-Seq NFR and mononucleosome signals form two distinct clusters," it would be helpful to briefly explain the significance of these two clusters. What does it indicate about the regulatory regions associated with TSSs? Line 133-139: When discussing hierarchical clustering and the similarity of NFR landscapes between different tissues, you could provide a sentence or two to speculate on the potential biological implications. For instance, why might heart and lung tissues exhibit more similarity in NFR landscapes compared to osteoblasts and liver? Line 134: To enhance clarity, you might consider using phrases like "Figure 3A" and "Figure 3B" instead of "Compare Fig 3A and B" to directly refer to the figures in the text. Line 142-144 :Please consider briefly explaining why you chose liver, heart, and lung tissues as controls. What specific characteristics or functions of these tissues make them suitable for this comparative analysis? When discussing the potential function of osteoblastic enhancers in cartilaginous fish, you might briefly mention the role of cartilage in these organisms and how these enhancers may have evolved to regulate cartilage-related processes. Ensure that the formatting of your methods section is consistent. For example, consistently use italics for software/tool names (e.g., "SAMtools") and follow a standard format for listing parameters or options used in software/tools.
Significance
The paper's significance lies in its elucidation of osteoblast-specific regulatory regions in Xenopus tropicalis. By characterizing these regions and connecting them to specific genes and pathways, the study advances our understanding of osteogenesis. Additionally, the identification of conserved elements across vertebrates provides insights into the deep evolutionary origins of skeletal features, offering a unique perspective on vertebrate evolution. However, one of the main limitations of the study is the lack of extensive experimental validation for the identified regulatory regions, leaving a gap in confirming their functionality.
-
