The EMBO Journal Peer Review Process File - EMBO-2014-90441
Manuscript EMBO-2014-90441
miR-290/371-Mbd2-Myc Circuit Regulates Glycolytic Metabolism to Promote Pluripotency Yang Cao, Wenting Guo, Shengya Tian, Xiaoping He, Xiwen Wang, Xiaomeng Liu, Kaili Gu2, Xiaoyu Ma, De Huang, Lan Hu, Yongping Cai, Huafeng Zhang, Yangming Wang and Ping Gao Corresponding author: Ping Gao, University of Science and Technology of China
Review timeline:
Submission date: Editorial Decision: Revision received: Accepted:
30 October 2014 25 November 2014 19 December 2014 22 December 2012
Transaction Report: (Note: With the exception of the correction of typographical or spelling errors that could be a source of ambiguity, letters and reports are not edited. The original formatting of letters and referee reports may not be reflected in this compilation.)
Editor: Thomas Schwarz-Romond
1st Editorial Decision
25 November 2014
Thank you very much for submitting your study reporting a miRNA-triggered Mbd2-Myc circuit in the control of glycolysis and reprogramming for publication in The EMBO Journal. Rather encouraging comments from two expert scientists are enclosed for your information and further perusal below. Both recognize the overall value and confirm general interest in your results. Ref#2 suggests slight improvements to crystallize the main and general advance from your study (re points background and literature presentation). S/he further emphasizes a few control experiments to improve confidence in data reliability and broaden their significance (major points 2, as well as complementary GoF experiments in point3). I am thus certain that you see yourself in a strong position to respond to all these in a relatively timely and constructive manner. I am thus delighted to invite formal submission of a revised and final study to your earliest convenience, please. Please do not hesitate to get in touch, in case I can be of further assistance/to discuss feasibility, amount and timeline of necessary revisions.
© EMBO
1
The EMBO Journal Peer Review Process File - EMBO-2014-90441
REFEREE REPORTS:
Referee #1: The manuscript "miR-290/371-Mbd2-Myc Circuit Regulates Glycolytic Metabolism to Promote Pluripotency" by Gao and colleagues reports the identification of a novel axis through which members of the miR-290 (miR-371 in humans) cluster of ES-specific miRNA control glycolysis and promote reprogramming. By performing an extensive and elegant series of genetic experiments in ESCs, the authors propose that members of the miR-290 cluster (in particular those with the "AAGUGCU" seed) indirectly control the expression of rate limiting glycolytic enzymes Ldha and Pkm2. The authors propose that these miRNAs directly repress Mbd2, which in turn leads to de-repression of Myc, which induces Ldha and Pkm2, thus promoting increased anaerobic glycolysis and reduced oxygen consumption. The manuscript is well written, the relevant literature is correctly cited, the experiments are generally well described and include the appropriate controls. The results are very convincing and generally support the model proposed by the authors. Overall this is a remarkable paper that will be of substantial interested to the general readership of Embo J. My only concern is that the model proposed by the authors seems to suggest that the ONLY way through which miR-294 controls glycolysis and promotes reprogramming is via Mbd2. This is unlikely to be case. For example, looking at the nice RNAseq dataset provided by the authors, it appears that miR-294 OE has a much stronger effect on Myc and Pkm expression compared to Mbd2 knockdown. This despite the fact that, as expected, miR-294 OE is less effective in reducing Mbd2 levels compared to siMbd2. It is therefore likely that additional miR-294 targets contribute to this phenotype and this should be explicitly discussed in the manuscript.
Referee #2: Review of "miR-290/371-Mbd2-Myc Circuit Regulates Glycolytic Metabolism to Promote Pluripotency " by Cao, et al. The authors use their well established DGCR8 knockout model to investigate the role of microRNAs in the metabolism of mouse embryonic stem cells. They find, first, that the DGRC8 knockout cells, which lack all microRNAs, produce reduced markers of glycolysis as compared to ESCs containing normal levels of microRNAs. The authors then query the usual suspects of microRNAs in ESCs and find that the well studied family of functionally equivalent microRNAs, the so-called "ESCCs", are responsible for stimulating glycolysis in ESCs. The authors then conduct an impressive epistatic analysis to identify the pathway by which the "ESCCs" regulate glycolysis namely through direct inhibition of MDB2, which causes Myc up-regulation, which causes PKM2 and LDHA up-regulation. Generally, the authors' methods are sound and convincing, and their conclusions are of interest to the fields of stem cell biology, somatic cell reprogramming and microRNA biology. Major concerns, discussed in more detail below, include a lack of transparency of quality control metrics for much of the data, incomplete representation of earlier and relevant literature, and a few key experiments to fully flesh out the epistatic analysis. These concerns should be easily addressed, and, if so, I recommend this manuscript for publication.
© EMBO
2
The EMBO Journal Peer Review Process File - EMBO-2014-90441
Major Concerns: Background, Literature representation, and Novelty Several of the authors' claims of novelty are a bit strong given previous reports. Specifically: i) That the miR-302/290/372 microRNAs directly target MDB2 has been demonstrated at least twice by Lee, et al. 2013 and Subramanyam, et al 2011. Though the authors cite both these papers, it is not made clear in the manuscript that MDB2 is a previously validated target of these microRNAs. For example, figure 4E of this manuscript is a verification of the previously published figure 3C of the Lee et al, not a completely novel finding. ii) That knock-down of MDB2 enhances human somatic cell reprogramming has also been previously established by the same two papers mentioned above. The authors should acknowledge this in their introduction. However, the role of MDB2 knockdown in mouse reprogramming is, to this reviewer's knowledge, novel. iii) Previously, the miR-302/290/372 microRNAs were shown to inhibit a well-established inhibitor of glycolysis, the AMPK complex, and direct inhibition of this complex was shown to enhance reprogramming - Judson et al, 2013 & Faubert, et al, 2013. In this manuscript, the authors do perform additional experiments to flesh out this model - namely, i) they directly measure the effect of these microRNAs on glycolysis in ESCs and ii) they block miR-302/290/372 induced reprogramming using genes that activate glycolysis. These data are novel to this manuscript and constitute significant findings, but the previous observations cited above are directly relevant and should be discussed in the introduction. Despite these concerns on the stated novelty of these findings, it should be noted that this reviewer finds the authors' other data to be of sufficient interest and novelty for publication. Specifically, the identification of the MDB2-MYC-PKM2/LDHA pathway, the linking of each of these to glycolysis regulation, and the discovery that some 13% of miR-302/290/372 regulated genes are regulated indirectly through MDB2, are significant. 2) Transparency of quality controls in experiments. The analysis conducted, with few exceptions (see point 3 below), is thorough and convincing as shown. However, some key information is missing in order to perform a clear and comprehensive review. Please include the following additional information to supplement what is shown: i) Please include the number and description (technical / biological) of replicates for each experiment in the figure legends. ii) Many of the assays used are significantly effected by the health of the cells (for example glucose uptake, lactate production, somatic cell reprogramming), yet it is not made clear in the figure's current format whether the possibility of sick or dying cells was accounted for. If this quality control data already exists, please include it in supplemental figures. If it doesn't, the authors should consider re-blotting their Westerns for markers of apoptosis activation (cleaved PARP or CASP) or conducting a single experiment where miR-290, MBD2, PKM2, LDHA, and Myc are overexpressed and knocked down in the relevant cell types and cell viability and markers of apoptosis are assayed. iii) Many of the key experiments in this manuscript include glucose uptake, lactate production, and OCR assays. However, for each assay, the different conditions the authors test are assayed one at a time, in pair-wise fashion, and always compared to a normalized control. However, direct comparison of the different conditions is useful for interpreting the data used in their epistasis arguments. For example, the authors, at the bottom of page 9, attempt to compare across experiments the magnitude of glucose uptake and lactate production in PKM2 and LDHA overexpression conditions (S3C&D) to miR-290 over-expression conditions (2A&B) to make the point that PKM2 and LDHA over-expression is the down-stream mechanism of miR-290 over-expression, but the direct comparison is never shown. Conversely, in figure 4G, knockdown of MBD2 does not have as significant an affect on these processes as miR-290 over-expression (2A&b), despite the level of gene knock-down being greater - but can this comparison be made since the different conditions were tested across different experiments? Interpretation of these results would be made © EMBO
3
The EMBO Journal Peer Review Process File - EMBO-2014-90441
significantly easier if the authors conducted one series of assays across key conditions (as suggested in part ii, knockdown and over expression of miR-290, MBD2, PKM2, LDHA, and Myc side-byside). Alternatively, the authors could provide the raw, un-normalized data from the experiments portrayed here in a supplemental table. 3) Incomplete epistatic data If the pathway the authors have modeled is accurate, over-expression of LDHA and PKM2 would rescue the negative affect shMyc has on the miR-290 over expression-induced and siMBD2-induced increase in glycolysis. The experiments in 5J&L should be repeated with LDHA and PKM2 overexpression. Minor Concerns: a) Misinterpretation of 6G - directly before the discussion, the authors interpret the data in Fig 6G as PKM2 or LDHA knock-down "completely blocked reprogramming-promoting effects of miR-371 cluster". The data pretty clearly shows that shRNAs completely block all reprogramming, not just the additional microRNA-induced affect. This is fascinating data and should not be understated. b) It's interesting that knockdown of either PKM2 or LDHA blocks miR-290-induced glycolysis (Fig 2H&I), but over-expression of either phenocopies miR-290-induced glycolysis (Fig S3c&d). Can the authors comment on this potential discrepancy and how it would fit into their model? One would expect that if knock-down of only one of the two wouldn't block miR-290-induced glycolysis if over-expression of either induces glycolysis. What is the affect of PKM2 over-expression on LDHA expression and vice versa?
Summary of Evaluation Criteria: Novelty The authors identify a previously unknown mechanism by which "ESCC" miRNAs increase glycolysis through direct inhibition of MBD2, thereby up-regulating cMyc and PKM2 and LDHA. This relationship is true both in the maintenance and establishment of pluripotent stem cells. Although not every step in this pathway is novel, most of it is, and the relationships are interesting. Further, the discovery that MBD2 knock-down accounts for a larger percentage of miR-290-cluster regulated targets is, in itself, novel and interesting. Broad biological significance This is a generally well conducted analysis of a biological mechanism directly linking a microRNA to an epigenetic modifier to a transcription factor to regulators of glycolysis. Investigators interested in any of these aspects should find this paper of interest. Importance to specific field MicroRNAs can target hundreds of different targets, but the downstream consequence of this coregulation is generally poorly explored. Here the authors take a microRNA family that has already been well-studied and identify a new set of significant down-stream consequences from a previously identified target. In doing so they highlight the need to consider both direct and indirect targets when assessing microRNA function, they solidify a novel role for this family in regulating an important cell process, and they identify two new genes that are strong regulators of somatic cell reprogramming - these findings should be of considerable interest to the microRNA, ESC, and iPS fields. Strong evidence for conclusions drawn Assuming the authors can provide the clarity and quality control data requested in point 2, the most
© EMBO
4
The EMBO Journal Peer Review Process File - EMBO-2014-90441
obvious evidence missing is that addressed by point 3. The rest of the experiments provide strong evidence for the conclusions drawn. References: Subramanyam, et al (2011) Nature Biotech. Multiple targets of miR-302 and miR-372 promote reprogramming of human fibroblasts to induced pluripotent stem cells. Lee et al (2013) Stem Cells. Epigenetic regulation of NANOG by miR-302 cluster-MBD2 completes induced pluripotent stem cell reprogramming. Judson, et al (2013) NSMB. MicroRNA-based discovery of barriers to dedifferentiation of fibroblasts to pluripotent stem cells Faubert, et al (2013) Cell Metabolism. AMPK Is a Negative Regulator of the Warburg Effect and Suppresses Tumor Growth In Vivo
1st Revision - authors' response
19 December 2014
(see next page)
© EMBO
5
Point-by-point response to the referees’ comments Editor’s Decision: Rather encouraging comments from two expert scientists are enclosed for your information and further perusal below. Both recognize the overall value and confirm general interest in your results. Ref#2 suggests slight improvements to crystallize the main and general advance from your study (repoints background and literature presentation). S/he further emphasizes a few control experiments to improve confidence in data reliability and broaden their significance (major points 2, as well as complementary GoF experiments in point3). I am thus certain that you see yourself in a strong position to respond to all these in a relatively timely and constructive manner. I am thus delighted to invite formal submission of a revised and final study to your earliest convenience, please. I formally do have to remind you that The EMBO Journal only allows one single round of revisions. Our Response: We are delighted upon Editor’s kind decision to allow us to revise this manuscript. Meanwhile, we are grateful for the encouraging and insightful comments from our expert referees. Accordingly, we have performed additional experiments and addressed all the comments raised by the referees, with a hope to make this manuscript better. Our point-by-point responses are appended below. For the editor and referees’ convenience, we have appended all the revised figures in this single file. And we thus labeled the figures from Figure R1 to Figure R5. Referees' comments: Referee #1: Comment 1-1: The manuscript "miR-290/371-Mbd2-Myc Circuit Regulates Glycolytic Metabolism to Promote Pluripotency" by Gao and colleagues reports the identification of a novel axis through which members of the miR-290 (miR-371 in humans) cluster of ES-specific miRNA control glycolysis and promote reprogramming. By performing an extensive and elegant series of genetic experiments in ESCs, the authors propose that members of the miR-290 cluster (in particular those with the "AAGUGCU" seed) indirectly control the expression of rate limiting glycolytic enzymes Ldha and Pkm2. The authors propose that these miRNAs directly repress Mbd2, which in turn leads to de-repression of Myc, which induces Ldha and Pkm2, thus promoting increased anaerobic glycolysis and reduced oxygen consumption.
The manuscript is well written, the relevant literature is correctly cited, the experiments are generally well described and include the appropriate controls. The results are very convincing and generally support the model proposed by the authors. Overall this is a remarkable paper that will be of substantial interest to the general readership of Embo J. Our Response: We appreciate the encouraging comments from the referee. Comment 1-2: My only concern is that the model proposed by the authors seems to suggest that the ONLY way through which miR-294 controls glycolysis and promotes reprogramming is via Mbd2. This is unlikely to be case. For example, looking at the nice RNAseq dataset provided by the authors, it appears that miR-294 OE has a much stronger effect on Myc and Pkm expression compared to Mbd2 knockdown. This despite the fact that, as expected, miR-294 OE is less effective in reducing Mbd2 levels compared to siMbd2. It is therefore likely that additional miR-294 targets contribute to this phenotype and this should be explicitly discussed in the manuscript. Our Response: We agree with the referee that, while our results clearly establish Mbd2 as a major downstream target to facilitate miR-294-mediated metabolic changes, we could not rule out that miR-294 might also regulate metabolism through Mbd2-independent mechanisms. As the referee pointed out, the RNA-seq dataset reminds us that miR-294 OE has a stronger effect on Myc and Pkm2 expression compared to Mbd2 knockdown, suggesting additional miR-294 targets involved potentially to contribute to this phenotype. We appreciate the referee to point this out and have discussed it to soften our conclusion in the revised manuscript (in the third paragraph of the discussion). Referee #2: Comment 2-1: The authors use their well established DGCR8 knockout model to investigate the role of microRNAs in the metabolism of mouse embryonic stem cells. They find, first, that the DGRC8 knockout cells, which lack all microRNAs, produce reduced markers of glycolysis as compared to ESCs containing normal levels of microRNAs. The authors then query the usual suspects of microRNAs in ESCs and find that the well studied family of functionally equivalent microRNAs, the so-called "ESCCs", are responsible for stimulating glycolysis in ESCs. The authors then conduct an impressive epistatic analysis to identify the pathway by which the "ESCCs" regulate glycolysis - namely through direct inhibition of MDB2, which causes Myc up-regulation, which causes PKM2 and LDHA up-regulation. Generally, the authors' methods are sound and convincing, and their conclusions are of interest to the fields of stem cell biology, somatic cell reprogramming and microRNA biology. Major concerns, discussed in more detail below, include a lack of transparency of quality control metrics for much of the data, incomplete representation of earlier and
relevant literature, and a few key experiments to fully flesh out the epistatic analysis. These concerns should be easily addressed, and, if so, I recommend this manuscript for publication. Our Response: We are grateful to the referee for careful analysis of our data and insightful comments on this study. Accordingly, we have performed additional experiments and revised the manuscript to address all the comments raised. Comment 2-2: Several of the authors' claims of novelty are a bit strong given previous reports. Specifically: i) That the miR-302/290/372 microRNAs directly target MDB2 has been demonstrated at least twice by Lee, et al. 2013 and Subramanyam, et al 2011. Though the authors cite both these papers, it is not made clear in the manuscript that MDB2 is a previously validated target of these microRNAs. For example, figure 4E of this manuscript is a verification of the previously published figure 3C of the Lee et al, not a completely novel finding. ii) That knock-down of MDB2 enhances human somatic cell reprogramming has also been previously established by the same two papers mentioned above. The authors should acknowledge this in their introduction. However, the role of MDB2 knockdown in mouse reprogramming is, to this reviewer's knowledge, novel. Our Response: We agree with the referee that, while we cited the work by Lee, et al. 2013 and Subramanyam, et al 2011, our original manuscript failed somewhat to make it clear the reported regulation of MBD2 by miR-302/290/372 microRNAs and its role in reprogramming. Now we have discussed this in the revised manuscript (Figure 4A and 4E). Comment 2-3: iii) Previously, the miR-302/290/372 microRNAs were shown to inhibit a well-established inhibitor of glycolysis, the AMPK complex, and direct inhibition of this complex was shown to enhance reprogramming - Judson et al, 2013 & Faubert, et al, 2013. In this manuscript, the authors do perform additional experiments to flesh out this model - namely, i) they directly measure the effect of these microRNAs on glycolysis in ESCs and ii) they block miR-302/290/372 induced reprogramming using genes that activate glycolysis. These data are novel to this manuscript and constitute significant findings, but the previous observations cited above are directly relevant and should be discussed in the introduction. Our Response: We thank the referee for pointing this out. Now we have discussed this point in the Introduction part of the revised manuscript and cited the relevant publications. Comment 2-4: Despite these concerns on the stated novelty of these findings, it should be noted that this reviwer finds the authors' other data to be of sufficient
interest and novelty for publication. Specifically, the identification of the MDB2-MYC-PKM2/LDHA pathway, the linking of each of these to glycolysis regulation, and the discovery that some 13% of miR-302/290/372 regulated genes are regulated indirectly through MDB2, are significant. Our Response: We appreciate the referee for nice comments about novelty and significance of our study. Comment 2-5: Please include the number and description (technical / biological) of replicates for each experiment in the figure legends. Our Response: We have added these descriptions in our revised figure legends. Comment 2-6: ii) Many of the assays used are significantly affected by the health of the cells (for example glucose uptake, lactate production, somatic cell reprogramming), yet it is not made clear in the figure's current format whether the possibility of sick or dying cells was accounted for. If this quality control data already exists, please include it in supplemental figures. If it doesn't, the authors should consider re-blotting their Westerns for markers of apoptosis activation (cleaved PARP or CASP) or conducting a single experiment where miR-290, MBD2, PKM2, LDHA, and Myc are over-expressed and knocked down in the relevant cell types and cell viability and markers of apoptosis are assayed. Our Response: We agree with the reviewer that the healthy status of various cell lines is an important factor to be considered in many of our assays. As a matter of fact, we routinely check the cell status by observing cell morphology, proliferation rate and viability in our laboratories. Here we performed additional western blotting assay to measure cleaved caspase-3 and pro-caspase-3 protein levels in cells with miR-290, MBD2, PKM2, LDHA, and Myc over-expressed or knocked down in a single experiment (Figure R1A, or Figure S5E in the revised manuscript). As indicated by the levels of pro-caspase-3 and cleaved caspase-3, manipulation of these genes did not have marked effects on cell viability. These results were included in the revised manuscript as Fig S5E. It’s worthy to note that the antibody we used could well detect the cleaved (active) form of caspase-3 in the control cells treated with high percentage of DMSO (Figure R1B). In addition, we provided the cell morphology of these cell lines showing that the overall status of the cells is rather healthy (Figure R2). Comment 2-7: iii) Many of the key experiments in this manuscript include glucose uptake, lactate production, and OCR assays. However, for each assay, the different conditions the authors test are assayed one at a time, in pair-wise fashion, and always compared to a normalized control. However, direct comparison of the different conditions is useful for interpreting the data used in their epistasis arguments. For example, the authors, at the bottom of page 9, attempt to compare across experiments the magnitude of glucose uptake and lactate production in PKM2 and LDHA
over-expression conditions (S3C&D) to miR-290 over-expression conditions (2A&B) to make the point that PKM2 and LDHA over-expression is the down-stream mechanism of miR-290 over-expression, but the direct comparison is never shown. Conversely, in figure 4G, knockdown of MBD2 does not have as significant an effect on these processes as miR-290 over-expression (2A&b), despite the level of gene knock-down being greater - but can this comparison be made since the different conditions were tested across different experiments? Interpretation of these results would be made significantly easier if the authors conducted one series of assays across key conditions (as suggested in part ii, knockdown and over expression of miR-290, MBD2, PKM2, LDHA, and Myc side-by-side). Alternatively, the authors could provide the raw, un-normalized data from the experiments portrayed here in a supplemental table. Our Response: This is a very good point. We agree with the referee that the direct comparison should be made ideally across key conditions, whilst, in reality, we tend to conduct pair-wise comparison because many of the assays need to consume a lot of cells, thus making it difficult to compare routinely all the cells together. Following the referee’s suggestion, here we have performed additional experiments to directly detect the glucose uptake and lactate production of all the cells with miR-290, MBD2, PKM2, LDHA and MYC manipulations in a single experiment and the results confirmed our observations described in the original manuscript (Figure R3). Comment 2-8: If the pathway the authors have modeled is accurate, over-expression of LDHA and PKM2 would rescue the negative affect shMyc has on the miR-290 over expression-induced and siMBD2-induced increase in glycolysis. The experiments in 5J&L should be repeated with LDHA and PKM2 over-expression. Our Response: Per reviewer’s suggestion, we have over-expressed Pkm2 or Ldha in Dgcr8-/- ESCs expressing shMyc and measured glucose uptake and lactate production levels (Figure R4). We observed partial but significant rescue effect of overexpressing these enzymes on glycolysis following Myc knockdown. These results, now included in the revised manuscript as Fig. S5D, reinforced our conclusion that Ldha and Pkm2 are downstream targets of miR290/Mbd2/Myc pathway in facilitating glycolysis in stem cells. Nevertheless, it is interesting but not surprising to note that overexpression of Pkm2 or Ldha could not fully replace Myc to stimulate glycolysis, suggesting potent regulatory functions of Myc via potential targets other than these two enzymes. Comment 2-9: Misinterpretation of 6G - directly before the discussion, the authors interpret the data in Fig 6G as PKM2 or LDHA knock-down "completely blocked reprogramming-promoting effects of miR-371 cluster". The data pretty clearly shows that shRNAs completely block all reprogramming, not just the additional microRNA-induced affect. This is fascinating data and should not be understated.
Our Response: This is an interesting point. In mouse reprogramming experiments, though addition of Pkm2 and Ldha shRNAs reduced the efficiency to a lower level than the control groups, we still obtained some iPS colonies. In contrast, Pkm2 and Ldha shRNAs almost completely blocked all reprogramming of human somatic cells, even with miR-371 cluster over-expression. This is really interesting and might be because that the reprogramming efficiency of human cells is much lower than the mouse ones, hence, removing one potent stimulator would represent an significant barrier for human somatic reprogramming. We thank the referee for the careful analysis and have corrected our interpretation of Fig. 6G in the revised manuscript. Comment 2-10: It's interesting that knockdown of either PKM2 or LDHA blocks miR-290-induced glycolysis (Fig 2H&I), but over-expression of either phenocopies miR-290-induced glycolysis (Fig S3c&d). Can the authors comment on this potential discrepancy and how it would fit into their model? One would expect that if knock-down of only one of the two wouldn't block miR-290-induced glycolysis if over-expression of either induces glycolysis. What is the affect of PKM2 over-expression on LDHA expression and vice versa? Our Response: Again, this is a very interesting point. First, since Pkm2 and Ldha both participate in glycolysis process but at different steps, it is easy to understand that, when either Pkm2 or Ldha is knocked down, we observed blockage of miR-290-induced glycolysis, confirming Pkm2 and Ldha as downstream effectors; Secondly, in regard to glycolytic promoting effect of expressing only one of the enzymes, we wouldn’t assume the expression of one enzyme could affect that of the other. In fact, we tested our assumption by western blot analysis, which showed that PKM2 over-expression did not affect that of LDHA, and vice versa (Fig. R5). Therefore, a simple explanation we would provide here is that expression of only one enzyme will substantially increase the overall glycolysis flux, leading to enhanced glucose consumption and lactate production, as we have observed in this study. Comment 2-11: Summary of Evaluation Criteria: Novelty The authors identify a previously unknown mechanism by which "ESCC" miRNAs increase glycolysis through direct inhibition of MBD2, thereby up-regulating cMyc and PKM2 and LDHA. This relationship is true both in the maintenance and establishment of pluripotent stem cells. Although not every step in this pathway is novel, most of it is, and the relationships are interesting. Further, the discovery that MBD2 knock-down accounts for a larger percentage of miR-290-cluster regulated targets is, in itself, novel and interesting. Broad biological significance This is a generally well conducted analysis of a biological mechanism directly linking a microRNA to an epigenetic modifier to a transcription factor to regulators of
glycolysis. Investigators interested in any of these aspects should find this paper of interest. Importance to specific field MicroRNAs can target hundreds of different targets, but the downstream consequence of this co-regulation is generally poorly explored. Here the authors take a microRNA family that has already been well-studied and identify a new set of significant down-stream consequences from a previously identified target. In doing so they highlight the need to consider both direct and indirect targets when assessing microRNA function, they solidify a novel role for this family in regulating an important cell process, and they identify two new genes that are strong regulators of somatic cell reprogramming - these findings should be of considerable interest to the microRNA, ESC, and iPS fields. Strong evidence for conclusions drawn Assuming the authors can provide the clarity and quality control data requested in point 2, the most obvious evidence missing is that addressed by point 3. The rest of the experiments provide strong evidence for the conclusions drawn. Our Response: We appreciate the referee for the summary of novelty and significance of our study. For the convenience of the editor and all the referees, all the revised figures are appended below. Thank you for your kind consideration. All the major changes in the revised manuscript are marked as red characters.
Figure R1. Caspase-3 protein levels in different cell lines as indicated
A
Dgcr8-/-
43 KD
Pro- Caspase-3
34 KD 26 KD
Active Caspase-3 17 KD
55 KD 43 KD
ACTIN
B
DMSO blank
2%
5%
55KD 43KD 34KD
Pro- Caspase-3
26KD
17KD
Active Caspase-3
ACTIN
A. Western blotting showing caspase-3 protein levels in the indicated cell lines. B.
Validation of caspase-3 antibody using DMSO treated MDA-MB-435 cells.
Figure R2. Cell morphology of the indicated cell lines in culture WT ESC
miR-290-/-
Dgcr8-/-
Dgcr8-/- + miR-290 cluster
Dgcr8-/- + Mbd2
Dgcr8-/- + Myc
Dgcr8-/- + Pkm2
Dgcr8-/- + Ldha
Dgcr8-/- + NTC
Dgcr8-/- + shMbd2
Dgcr8-/- + shMyc
Dgcr8-/- + shPkm2
Dgcr8-/- + shLdha
Figure R3. Glucose uptake and lactate production of the indicated cell lines
Glucose level (nmol/μg)
A 0.016 0.014 0.012 0.01 0.008 0.006 0.004
0.002 0
Dgcr8-/- ESC
Lactate level (nmol/μg)
B 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
Dgcr8-/- ESC
A. Glucose uptake were measured in the indicated cell lines (n=3). B.
Lactate production were measured in the indicated cell lines (n=3).
Figure R4. Pkm2 and Ldha rescue the negative effect of shMyc on glycolysis
A
B *
1.2 *
1
Rel. Lactate level
Rel. Glucose level
1.2
0.8
0.6 0.4 0.2
*
1 0.8 0.6 0.4 0.2 0
0 shMyc Pkm2 Ldha
*
-
+ -
+ + -
+ +
shMyc Pkm2 Ldha
-
+ -
+ + -
+ +
A. Glucose uptake were measured in the indicated cell lines. Data were presented as mean (±SD). * P<0.05 as compared between indicated groups (n=3). B. Lactate production were measured in the indicated cell lines. Data were presented as mean (±SD). * P<0.05 as compared between indicated groups (n=3).
Figure R5. Pkm2 over-expression did not affect the protein level of Ldha, and vice versa
A
B Dgcr8-/-
Dgcr8-/EV
EV
Ldha
Pkm2 LDHA
PKM2 LDHA ACTIN
PKM2 ACTIN
A. Western blotting showing Ldha expression level upon over-expression of Pkm2 in Dgcr8-/- ESCs. ACTIN serves as loading control. B. Western blotting showing Pkm2 expression level upon over-expression of Ldha in Dgcr8-/- ESCs. ACTIN serves as loading control.