SMAD1 Is Dispensable for CDX2 Induction but Required for the Repression of Ectopic Small-Intestinal Gene Expression in Human-Pluripotent-Stem-Cell-Derived Colonic Organoids

Round 1

Reviewer 1 Report

Qu and colleagues present a detailed study of the regulation of BMP signaling in colonic development by employing organoids derived from human pluripotent stem cells with particular emphasis on the role of Smad1 as a regulator of BMP signaling in gut organogenesis. They use culturing techniques established during their previous work to differentiate pluripotent stem cells into colonic organoids, and subsequently use a combination of  techniques to test their hypothesis.

Study of SMAD1 regulation of mid/hindgut differentiation using human colonic organoids from hPSCs subverts a gap in in vivo studies as Smad1^-/- mice die by embryonic day 10.5 before the differentiation of intestine and colon. As such this study aims to fill a gap in current knowledge and is relevant to the research field. The study is well-designed and presented with logical investigative steps that are appropriately described.

While the overall quality of the presented study is high there are some questions:

1.       Figure 1c shows a weak SMAD1 band for the C6 clone and possibly a faint band for the C18 clone. Are all subsequent experiments carried out using the C18 clone? The figure legend and methods mention that C18 was the clone sent for sequencing but it should be explicitly clarified that C18 was the clone used for experimentation if that is the case.

2.       Additionally, have the authors confirmed that the cell lines at the end of the study represent those at the start of the study, i.e. there has been no overgrowth by residual WT cells?

3.       Were growth rates of WT and SMAD^-/- cells similar?

4.       Is it reasonable to expect organoids derived from both cell lines to reach comparable developmental stages at the same time?

5.       Redundancy between R-SMAD proteins has been described previously in multiple organs and pathways e.g., TGFβ signaling (Orvis et al, Biol Reprod. 2008 78(6):994-1001) and BMP signaling (Wei et al, Developmental Biology 2014 289(10):6604-18). The discussion of the current study should include consideration of Smad redundancy in the literature. 

6.       Figure 4c should refer to E-cadherin rather than its gene, CDH1.

7.       There is inconsistency in the abbreviations used for the pluripotent stem cells throughout the manuscript: iPSCs, IPSCs, hPSCs.

The standard of the written English is very high with a few minor corrections required:

Line 25 should read “examination of HCOs”,

Line 33 should read “in terms of endoderm-derived organs”,

Line 80 CRISPR is misspelt,

Line 97 should read “generate a CDX2+ monolayer”,

Line 106 green is misspelt,

Line 199 should read “studies in mouse”,

Line 306 should read "For the clump organoid”.

Author Response

Qu and colleagues present a detailed study of the regulation of BMP signaling in colonic development by employing organoids derived from human pluripotent stem cells with particular emphasis on the role of Smad1 as a regulator of BMP signaling in gut organogenesis. They use culturing techniques established during their previous work to differentiate pluripotent stem cells into colonic organoids, and subsequently use a combination of  techniques to test their hypothesis.

Study of SMAD1 regulation of mid/hindgut differentiation using human colonic organoids from hPSCs subverts a gap in in vivo studies as Smad1-/- mice die by embryonic day 10.5 before the differentiation of intestine and colon. As such this study aims to fill a gap in current knowledge and is relevant to the research field. The study is well-designed and presented with logical investigative steps that are appropriately described.

While the overall quality of the presented study is high there are some questions:

1. Figure 1c shows a weak SMAD1 band for the C6 clone and possibly a faint band for the C18 clone. Are all subsequent experiments carried out using the C18 clone? The figure legend and methods mention that C18 was the clone sent for sequencing but it should be explicitly clarified that C18 was the clone used for experimentation if that is the case. We apologize for this oversight. We did indeed use the C18 clone throughout the manuscript and we have updated the manuscript to reflect this.

2. Additionally, have the authors confirmed that the cell lines at the end of the study represent those at the start of the study, i.e. there has been no overgrowth by residual WT cells? We thank the reviewer for this important point. Our sequencing data shows that there is no detectable WT sequence in the SMAD1KO cells. However, the reviewer does make a good point that WT cells could be present below the limit of detection and have a selective advantage over SMAD1-/- cells. To rule out this possibility, we now provide PCR/Restriction enzyme analysis of WT and SMAD1 deficient 35 day old organoids (See Supplementary figure 1). It is clear that the SMAD1 KO HCO do not contain WT sequence.

3. Were growth rates of WT and SMAD-/- cells similar? Yes. Therefore, as mentioned above, WT cells are likely not present or do not outcompete SMAD1-/- cells.

4. Is it reasonable to expect organoids derived from both cell lines to reach comparable developmental stages at the same time? This is a good question but one that is difficult to answer. There are not extensive datasets with stage specific markers and even less so for the K3 cell line which we used for experiments. This is a relevant point since there have been reports of human fetal colon expressing small intestinal markers. We therefore have added this possibility to our discussion. We do see overlapping transcriptomes between WT and SMAD1-/- cells and expression of maturation markers such as CDH17, MUC2 and CHGA did not vary.

5. Redundancy between R-SMAD proteins has been described previously in multiple organs and pathways e.g., TGFβ signaling (Orvis et al, Biol Reprod. 2008 78(6):994-1001) and BMP signaling (Wei et al, Developmental Biology 2014 289(10):6604-18). The discussion of the current study should include consideration of Smad redundancy in the literature. We appreciate the reviewer bringing this to our attention. We now include these references in the manuscript.

6. Figure 4c should refer to E-cadherin rather than its gene, CDH1. Although E-cadherin is more widely used, Cadherin 1 is a recognized protein name based on unitprot. https://www.uniprot.org/uniprotkb?query=cdh1&facets=model_organism%3A9606

7. There is inconsistency in the abbreviations used for the pluripotent stem cells throughout the manuscript: iPSCs, IPSCs, hPSCs. We have fixed the issues with IPSCs. hPSCs is used to denote human pluripotent stem cells which encompasses both embryonic stem cells and induced pluripotent stem cells.

he standard of the written English is very high with a few minor corrections required:

Line 25 should read “examination of HCOs”,

Line 33 should read “in terms of endoderm-derived organs”,

Line 80 CRISPR is misspelt,

Line 97 should read “generate a CDX2+ monolayer”,

Line 106 green is misspelt,

Line 199 should read “studies in mouse”,

Line 306 should read "For the clump organoid”.

We have fixed these issues.

Reviewer 2 Report

In the present manuscript, Jorge Munera and colleagues used SMAD1 KO human iPS cells and organoid culture systems to elucidate the role of SMAD1 in differentiating human iPS cells into colon cells. First, the authors assessed the impact of SMAD1 KO on mid/hindgut differentiation by combining BMPs or their inhibitor addition and SMAD1 KO iPS cells, but concluded from immunostaining and RNAseq results that posterior hindgut patterning was not affected. However, the authors confirmed from ToppFun analysis using RNAseq data that small intestinal marker genes were ectopically expressed in SMAD1 KO iPS cell-derived organoids after long-term culture over 35 days.

General comments:

The reviewer generally believes the findings will interest researchers in this field. However, there are numerous problems with the experimental methods and the data presented in this manuscript. Most problematic is the complete lack of analysis of SMAD5 and 9, which could be a source of redundancy in SMAD1. The reviewers believe that redundancy cannot be discussed without data on the expression transition of SMAD1/5/9 in the normal organoid differentiation system and changes in SMAD5/9 expression due to SMAD1 KO. The reviewers also consider that experiments assessing the effects of SMAD on cell differentiation should always examine SMAD phosphorylation levels and that the redundancy argument could be advanced by examining changes in SMAD5/9 phosphorylation levels by SMAD1 KOs.

Major Points:

1.     TGFbeta signaling is required to maintain undifferentiated human iPS cells and is often included in undifferentiated maintenance media. Are there any problems with the undifferentiated nature of the SMAD1 KO iPS cells generated by the authors?　The redundancy could be due to other SAMDs if there is no difference. The authors should also mention the role of SMAD1 in maintaining undifferentiation.

2.     It contributes to the differentiation of iPS cells into definitive endoderm, as the authors also add BMP4 to the culture medium in the early stages of differentiation. Therefore, the effect of SMAD1 KO on differentiation into endoderm should also be assessed before assessing differentiation into mid/hindgut tissue.

3.     The authors conclude from the analysis in Fig. 4D alone that increased expression of small intestinal genes by SMAD1 KO occurred in long-term cultures, but this conclusion needs to be assessed by RT-PCR analysis and antibody staining for individual genes.

Minor points:

1.               As there are no positive controls for SOX2 staining in Fig. 2, the reviewer cannot judge the results of this experiment. Experimental controls are needed, for example, employing the noggin addition method used to generate SOX2-positive cells in previous studies.

2.               In this experiment, DMH1 is used as the BMP inhibitor in Fig. 3CD and LDN in Fig. 3D What was the reason for changing the BMP inhibitor?

3.               In Figures 3A and 3B, it appears that the "BMP2" in the figures is actually a combination of EGF and BMP2. Are EGF also added to those listed as "BMP2" "BMP2 + DMH1," etc in Fig. 3CDE?

The English language in this paper is clear and well-structured. There are no noticeable linguistic errors, allowing for easy comprehension of the content.

Author Response

We thank the reviewer for their careful review of our manuscript.

In the present manuscript, Jorge Munera and colleagues used SMAD1 KO human iPS cells and organoid culture systems to elucidate the role of SMAD1 in differentiating human iPS cells into colon cells. First, the authors assessed the impact of SMAD1 KO on mid/hindgut differentiation by combining BMPs or their inhibitor addition and SMAD1 KO iPS cells, but concluded from immunostaining and RNAseq results that posterior hindgut patterning was not affected. However, the authors confirmed from ToppFun analysis using RNAseq data that small intestinal marker genes were ectopically expressed in SMAD1 KO iPS cell-derived organoids after long-term culture over 35 days.

General comments:

The reviewer generally believes the findings will interest researchers in this field. However, there are numerous problems with the experimental methods and the data presented in this manuscript. Most problematic is the complete lack of analysis of SMAD5 and 9, which could be a source of redundancy in SMAD1. The reviewers believe that redundancy cannot be discussed without data on the expression transition of SMAD1/5/9 in the normal organoid differentiation system and changes in SMAD5/9 expression due to SMAD1 KO. The reviewers also consider that experiments assessing the effects of SMAD on cell differentiation should always examine SMAD phosphorylation levels and that the redundancy argument could be advanced by examining changes in SMAD5/9 phosphorylation levels by SMAD1 KOs. Although examination of SMAD phosphorylation is a way to examine BMP signaling, we did examine known SMAD1/5/9 mediated transcriptional targets such as ID1, ID3 and MSX2. Furthermore, there are no pSMAD5 or pSMAD9 specific antibodies that could be used.

Major Points:

1. TGFbeta signaling is required to maintain undifferentiated human iPS cells and is often included in undifferentiated maintenance media. Are there any problems with the undifferentiated nature of the SMAD1 KO iPS cells generated by the authors?　The redundancy could be due to other SAMDs if there is no difference. The authors should also mention the role of SMAD1 in maintaining undifferentiation. SMAD1 KO IPSCs generate both endoderm derived intestinal epithelium and mesoderm derived mesenchyme. They also can be differentiated into ectoderm as we do see some neuronal markers expressed in our BMP+LDN treated organoids. Furthermore, there is no reason to believe that SMAD1 deficiency would affect pluripotency since Smad1 KO mice gastrulate normally and survive until embryonic day 9-10. https://www.informatics.jax.org/marker/MGI:109452

2. It contributes to the differentiation of iPS cells into definitive endoderm, as the authors also add BMP4 to the culture medium in the early stages of differentiation. Therefore, the effect of SMAD1 KO on differentiation into endoderm should also be assessed before assessing differentiation into mid/hindgut tissue. See point above.

3. The authors conclude from the analysis in Fig. 4D alone that increased expression of small intestinal genes by SMAD1 KO occurred in long-term cultures, but this conclusion needs to be assessed by RT-PCR analysis and antibody staining for individual genes. We agree with the reviewer and we intended to perform antibody stainings on a panel of small intestinal genes. We ordered antibodies in July and due to supply chain issues, we have not yet received them.

Minor points:

1. As there are no positive controls for SOX2 staining in Fig. 2, the reviewer cannot judge the results of this experiment. Experimental controls are needed, for example, employing the noggin addition method used to generate SOX2-positive cells in previous studies. In figure 2C, at least 3 positive cells can be seen in the bottom right quadrant of the SMAD1-/- cells. These antibodies have been validated in previous studies (McCracken et al. Nature 2014, Nature 2017).

2. In this experiment, DMH1 is used as the BMP inhibitor in Fig. 3CD and LDN in Fig. 3D What was the reason for changing the BMP inhibitor? This was another supply chain issue.

3. In Figures 3A and 3B, it appears that the "BMP2" in the figures is actually a combination of EGF and BMP2. Are EGF also added to those listed as "BMP2" "BMP2 + DMH1," etc in Fig. 3CDE? Yes, those have EGF added also. Since “control” media has EGF, materials and methods) the BMP2 also has EGF and we state this in the methods.

Round 2

Reviewer 2 Report

Reviewing the revised manuscript, some of the previously raised concerns have not been adequately addressed.

 **Major point 3:** The authors mention an inability to evaluate due to the unavailability of antibodies. However, it was suggested that RT-PCR can be an alternative approach for this purpose. The authors are encouraged to consider using RT-PCR to address this point.

**Minor point 1:** A previous comment highlighted the necessity for positive control in the SOX2 staining experiment. In the revised manuscript, no positive control for the staining was provided. If the authors want to claim that there are no SOX2-positive cells in these cells, they should use RT-PCR or antibody staining to show that SOX2 expression is seen in positive controls in the same experiment but not in those cultured under the conditions shown in this paper. 

These issues must be thoroughly addressed in the subsequent revision.

Author Response

 **Major point 3:** The authors mention an inability to evaluate due to the unavailability of antibodies. However, it was suggested that RT-PCR can be an alternative approach for this purpose. The authors are encouraged to consider using RT-PCR to address this point. We now include RT-PCR analysis of 3 small intestinal genes that are upregulated in SMAD1-/- HCOs. For reference, we also include RNA-seq data on expression of these genes in different human organs to demonstrate that these are duodenum/small intestine specific.

**Minor point 1:** A previous comment highlighted the necessity for positive control in the SOX2 staining experiment. In the revised manuscript, no positive control for the staining was provided. If the authors want to claim that there are no SOX2-positive cells in these cells, they should use RT-PCR or antibody staining to show that SOX2 expression is seen in positive controls in the same experiment but not in those cultured under the conditions shown in this paper. We have included a figure for the reviewer of our validation of CDX2 and SOX2 antibodies. We also include single channels to show that SOX2 is expressed in a small number of cells in figure 1C.

These issues must be thoroughly addressed in the subsequent revision. We have thoroughly addressed the forementioned issues.