The Eleanor ncRNAs activate the topological domain of the ESR1 locus to balance against apoptosis

MCF7 cells acquire estrogen-independent proliferation after long-term estrogen deprivation (LTED), which recapitulates endocrine therapy resistance. LTED cells can become primed for apoptosis, but the underlying mechanism is largely unknown. We previously reported that Eleanor non-coding RNAs (ncRNAs) upregulate the ESR1 gene in LTED cells. Here, we show that Eleanors delineate the topologically associating domain (TAD) of the ESR1 locus in the active nuclear compartment of LTED cells. The TAD interacts with another transcriptionally active TAD, which is 42.9 Mb away from ESR1 and contains a gene encoding the apoptotic transcription factor FOXO3. Inhibition of a promoter-associated Eleanor suppresses all genes inside the Eleanor TAD and the long-range interaction between the two TADs, but keeps FOXO3 active to facilitate apoptosis in LTED cells. These data indicate a role of ncRNAs in chromatin domain regulation, which may underlie the apoptosis-prone nature of therapy-resistant breast cancer cells and could be good therapeutic targets.

4. Page 7, top. "To investigate the significance of the Eleanor TAD, we measured the transcriptional activities of genes both within and outside the Eleanor TAD". To emphasize the point that the location of the TAD boundaries the authors have chosen is arbitrary and it depends on the resolution of the Hi-C data, I'm enclosing a picture of the same region using Hi-C data from Rao et al in GM12878 cells containing 4.5 billion reads instead of a couple hundred. The differences are obvious, although one could argue that this is due to the different cell type. However, the same region looks very similar in HCT116 cells at the same resolution. Is the ESR1 gene expressed in GM12878 cells? The domains the authors claim to be a TAD looks considerably different, with several CTCF loops that presumably would interfere with enhancer-promoter interactions and with the co-regulation the authors claim takes place within this domain. Are these CTCF sites present in MCF7 cells? If so, they presumably would form the same CTCF loops. Where are the different co-regulated genes present with respect to these loops? Are they all inside the loops or are some outside? 5. Page 7, bottom. "Upon further analysis of our 4C-Seq data, we observed intra-chromosomal ESR1 promoter interactions along the entirety of chromosome 6 in MCF7 and LTED cells". It is surprising that authors observe such long-range interactions in the 4C data. Based on the description in the Methods, it appears that authors only subtracted the input from the experimental with no distance normalization. Authors should use a standard 4C analysis pipeline such as that described by de Laat and collaborators in PMID 22961246.
6. Page 8, top. "However, there was a tendency for the 4C peaks to coincide with A compartments throughout the chromosome, suggesting that the ESR1 promoter region in the Eleanor TAD interacted with genes in the A compartments of multiple regions of chromosome 6. Consistently, ESR1 promoter interacting regions showed moderate coincidence with genes that were upregulated in LTED cells (red bars below 4C-Seq profiles in Fig. 2a)". Please explain in quantitative terms what "tendency" and "moderate" means. If the authors want to make the argument that regions that interact with the ESR1 gene represent upregulated genes, it may be better to plot the RNA-seq data. Also, it would help the reader asses the overlap between upregulated genes and 4C interactions if regions of interest are shown at a higher resolution, perhaps in the supplemental figures. 7. Page 9, bottom. "we inhibited Eleanors with resveratrol in LTED cells and performed 3C-qPCR analysis. We found that the ESR1 promoter and the b site were dissociated after resveratrol treatment (Fig. 2f). It should be noted that the transcriptional levels of the b site genes were mostly unchanged". It seems from Figure 2G that transcription of SNX3 and LACE1 changed quite a bit. How long was the resveratrol treatment? Is it possible that the RNAs for the other genes are more stable and that's why change was not detected? 8. Page 10, top. "Taken together, these data show that the chromatin interaction between ESR1 and FOXO3 loci requires Eleanors, and that the interaction plays a role in coupregulation of genes in the Eleanor TAD in LTED cells. Eleanor transcription is required to resist cell death in the transition from MCF7 to LTED and that the FOXO3 containing TAD somehow is permissive for Eleanor transcription". Perhaps I'm missing something related to this conclusion. My understanding is that the authors observe correlations among various events with no basis for casual inferences. First, does treatment with resveratrol affect Eleanors and this then affects transcription of coding genes? Or are the two events correlated without an established causal relationship? Second, the authors never show causality between interactions and transcription. Authors assume that interactions determine transcription, but it is equally possible, and perhaps more likely, that changes in transcription are determined by binding of protein complexes to the promoters and regulatory regions of genes, and that these proteins then mediate the interactions. Loss of interactions after resveratrol treatment could be a consequence of decreased transcription, which the authors argue it does not change (see #7), but Figure 2G appears to suggest that it does. Does RNA pol II remain unchanged after treatment? The TAD is not permissive for transcription, unless the authors are talking about specific CTCF loops, which do not necessarily correspond to TADs. 9. Page 11, top. "because most other coding (exons) and non-coding regions in the Eleanor TAD were transcribed in the same direction". What does "most" mean, 51% or 99%? Please be quantitative.
10. Page 11, bottom. "knockdown of ESR1 mRNA with RNAi or specific degradation of ER protein with a chemical compound ICI 782,780, had no effect on pa-Eleanor(S) expression". Evidence showing that ESR1 protein was depleted in these experiments should be shown. 11. Page 12, top. "To investigate the molecular function of pa-Eleanor(S), we knocked it down in LTED cells using Antisense LNA GapmeR (LNA) (Fig. 4a)". It is surprising that KD of pa-Eleanor had a modest effect when measured by qPCR ( Figure 4a) but a very strong effect when measured by FISH ( Figure 4C). Please comment on this discrepancy.
12. Page 12, bottom. "To investigate these, we first knocked down ESR1 mRNA in LTED cells and found that the transcriptional levels of genes in the Eleanor TAD, including pa-Eleanor(S) and the Eleanor RNA cloud, did not change ( Supplementary Fig. 5a-c). Therefore, Eleanors are specific ncRNAs that play a role in the activation of the chromatin domain. Consistent with this result, ICI 182,780 did not change the expression levels of either Eleanor TAD genes or pa-Eleanor(S)". Authors should show that the levels of ESR1 protein actually changed in these experiments. Changes in ESR1 RNA levels shown in supplemental Figure 5A are very small.
13. Page 13, bottom. "To investigate the function of pa-Eleanor(S) in gene regulation through chromatin structures, we performed 3C-qPCR analysis. We found that the interaction between the ESR1 promoter and the b site at 6q21 reduced with pa-Eleanor(S) knockdown in LTED cells (Fig.  5a), supporting the idea that the ncRNA was involved in the long-range chromatin interaction. In contrast, transcriptional levels of the genes at 6q21 showed little to-no change in the LTED cells with pa-Eleanor(S) knockdown (Fig. 5b)". This is a critical experiment and conclusion that should not rely on just one 3C experiment. Authors should perform 4C and examine all interactions between ESR1 and other sequences in the chromosome. Furthermore, gene expression changed significantly in at least one of the two experiments shown in Figure 5B. These changes in expression could account for the changes in interactions shown in Figure 5A.
Reviewer #3 (Remarks to the Author): In this manuscript, the authors investigated the long-range chromatin TADs interaction and their consequence on balance of LTED MCF-7 cells growth and apoptosis. They found lncRNA pa-Eleanor (S) mediated the interaction of two TADs. Silence of pa-Eleanor inhibited the formation of Eleanor clouds, the binding of the b site of 6q21 with ESR1 promoter and promoted LTED cell apoptosis. And inhibition of pa-Eleanor mimics the effect of resveratrol treatment. In general, this manuscript is the extension of the previous study (Nat Commun 6, 6966 (2015)) and clarified the mechanism of LTED cells primed for apoptosis. However, the roles of 6q21 TAD in promoting co-upregulation of genes in the Eleanor TAD were still unknown. Some issues need to be addressed.
Major comments: 1. This manuscript investigated the balanced chromosomal interaction between genes for cell proliferation (ESR1) and death (FOXO3) in LTED cells, mediated by Eleanors. However, gene expression in Eleanors TAD didn't affect the activation of FOXO3 contained 6q21 TAD. Also, there were no direct evidence to prove the interaction between the b site of 6q21 with ESR1 promoter promoted the transcription of Eleanors and ESR1. It seems only pa-Eleanor regulated ESR1 expression determined LETD cell growth or apoptosis. The authors should provide more experiments to show the functional interaction of 6q21 on ESR1 promoter. 2. The group's previous paper found u-Eleanor is involved in transcriptional activation of the ESR1 locus and maintains downstream intragenic Eleanor in LTED cells. The pa-Eleanor reported in the present study diaplayed the similar functions. The authors should discuss the relationship between u-Eleanor and pa-Eleanor. And was it a general mechanism for Eleanor ncRNAS in regulating ESR1 expression? Minor comments: 1. For all the cell apoptosis experiments, the authors had better provide flow cytometry results stained by annexin V/PI. 2. For all the flow cytometry experiments showed in this manuscript, the authors should also provide the isotype-treatment curve for each group. 3. In figure 5a, there was interaction between c site with ESR1 promoter, which was not consistent with fig 2d, 2f. the authors should explain the results. 4. In sup fig5e, the basal cell viability was largely different. The experimental conditions should be optimized. We thank the reviewer for this important comment, because it made us aware that our previous description was insufficient, and therefore we have revised the manuscript. As pointed out, FOXO3 expression is up-regulated in LTED cells and it remains high after resveratrol treatment. This is one of the reasons why we propose that it function as a key regulator for apoptosis in LTED-RES cells.
Please note that we do not claim that FOXO3 may be a key regulator for apoptosis in MCF7 cells.
FOXO3 is a tumor suppressor and promotes apoptosis (Ekoff et al., 2007;Skurk et al., 2004). We have now stated this clearly in the text on page 11, line 5.
We agree that the percentage of vital cells is already high (67%) in the control knock down cells (siGL3), so the effect of siFOXO3 may not have been high enough in the previous Supplementary Fig. 3. Therefore, according to this reviewer's suggestion, we performed the apoptosis assay with Annexin V to clarify the role of FOXO3 in apoptosis. The results showed that siFOXO3 reduced apoptosis in LTED (2M) cells (new Supplementary Fig. 4d). The effect of siFOXO3 was statistically significant, as P= 1.82E-05, and more prominent in LTED (2M) than in MCF7 cells, as expected. Our additional CCK-8 assay also supported the fact that the FOXO3 knockdown rescued the proliferation of LTED (2M) cells (new Supplementary Fig. 4b). For further confirmation, we overexpressed FOXO3 in LTED cells, and found that it induced cell death (new Supplementary Fig. 4e-g). Altogether, these results clearly demonstrate that FOXO3 functions to induce apoptosis in breast cancer cell adaptation under estrogen deprivation. We presented these new data in the new Supplementary Fig. 4b, d-g, and described them in the revised manuscript on page 10, line 2.

It is not convincing how Eleanor is linked to FOXO3-mediated apoptosis upon ER deprivation.
Considering intimate and paradoxical linkage between ER and FOXO3, as stated in Discussion session, it will be nice to find other genes to explain balance between proliferation and apoptosis in breast cancer cell physiology. Are there any other candidate target genes, direct or indirect, to explain estrogen deprivation adaptation and resveratrol treatment?
To address this reviewer's comment, we re-analyzed our previously published RNA-Seq data and  Table 3. We discuss these results in the revised text on page 11, line 13. Fig. 2g, but authors did not mention on SNX3 at all. Are there any phenotypic changes after SNX3 knockdown or overexpression?

What is the role of SNX3? It is significantly co-regulated with LTED and LTED-RES in
As the reviewer pointed out, SNX3 expression was co-regulated with ESR1 in LTED and LTED-RES cells. The SNX3 protein is involved in intracellular trafficking (Chiow et al., 2012;Haft et al., 1998).
To understand its function in endocrine therapy resistance in breast cancer, we knocked down SNX3, and found that it had no effect on cell proliferation. We show these data in the new Supplementary   Fig. 4h, i, and describe them in the text on page 10, line 8.

In previous paper published in 2015, authors identified u-Eleanor and demonstrated RNA cloud
upon LTED and LTED-RES. In this paper, expression of pa-Eleanor and its switch from the antisense to the sense orientation during LTED were shown in Fig. 3. Since it contains 5'UTR of ESR1 mRNA, it has to be more clearly demonstrate that it is not a part of mRNA.
In the revised text, we show the appropriate position of the TSS for the ESR1 mRNA variant 1 (new Fig. 4a), which starts downstream of pa-Eleanor (S) and is produced mostly in mammary cells (Consortium, 2012;Gannon et al., 2001;Reid et al., 2002). The 5'UTR of this ESR1 mRNA is not included in pa-Eleanor(S). For confirmation, we performed an RT-PCR analysis and did not detect a transcript that is contiguous from pa-Eleanor(S) to exon 1 of the ESR1 mRNA (new Fig. 4d, lane 7).
In contrast, we successfully detected the transcript from upstream of exon 1 of the ESR1 mRNA, corresponding to pa-Eleanor (new Fig. 4d In addition, we found that the cellular phenotype was different in each knockdown. With the pa-Eleanor(S) knockdown, the genes in the Eleanor TAD and the Eleanor cloud formation were repressed (new Fig. 5a-c), while they were not affected with the ESR1 mRNA knockdown (new Supplementary Fig. 8a-d). Altogether, we concluded that pa-Eleanor is independent of the mRNA.
More importantly, how the strand switch occurs has to be understood.
We agree that the mechanism for the transcriptional switching is important. It is possible that RNA polymerase II transcription in the opposite direction may interfere with the movement of RNA polymerase II toward the ESR1 gene in MCF7 cells, which could be relieved in LTED cells. This idea is shared with the transcriptional interference in yeast and other eukaryotes (Hao et al., 2016;Mazo et al., 2007), and is now described in the discussion section of the revised manuscript on page 18, line 12.

Overall, relevance of Eleanor in the regulation of ESR1 and neighboring genes are distinct from
long-range interacting genes. Also, the regulation modes by uEleanor and pa-Eleanor are not directly assessed in the manuscript. Considering such ambiguity, the model shown in Fig. 6 is very vague and did not clearly summarize the findings. Model diagram should be more clear.
According to this reviewer, we have revised the model diagram in the new Fig. 7 for clarity. We also made it clear that the relevance of Eleanor in the regulation of ESR1 and its neighboring genes is distinct from that in long-range interacting genes, and we have revised the discussion section of the text on page 16, line 13.
To understand the regulation modes of pa-Eleanor and u-Eleanor, we investigated the expression of pa-Eleanor and u-Eleanor upon the knockdown of each ncRNA. The u-Eleanor knockdown did not change the expression of pa-Eleanor (new Supplementary Fig. 7f). In contrast, the knockdown of pa-Eleanor(S) suppressed the expression level of u-Eleanor (new Supplementary   Fig. 7g). These results suggest that pa-Eleanor upregulates the ESR1 transcription through u-Eleanor, or may directly affect the expression of ESR1. These results are summarized in the new Supplementary Fig. 7h, and described in the revised text on page 14, line 4.
6. Introduction section is not good well written. General significance of ncRNA-mediated regulation on TAD and its application to cancer cell adaptation should be included to draw more attention.
Since the manuscript is the follow-up of previous paper in Nature Communications, introduction should be more interesting. Also, the title is not good representation of the findings.
According to the reviewer, we created a new paragraph stating the significance of the ncRNA-mediated regulation on TAD and its possible application to cancer cell adaptation on page 4, line 9 of the revised manuscript.
We also changed the title to "Eleanors activate the chromatin domain, with balancing cell death and proliferation in breast cancer adaptation" Reviewer #2 (Remarks to the Author): The manuscript describes experiments examining the relationship between expression of Eleanor ncRNAs, ESR1, and changes in 3D interactions. The manuscript is potentially interesting but authors need to address the issues raised below, some of which are substantial.

Introduction. Authors have a very rudimentary and superficial understanding of 3D organization.
Authors should read and cite manuscripts by Rao  We thank this reviewer for providing these important comments. We agree that the two publications (Rao et al., 2014;Rao et al., 2017) are very important, in which high resolution Hi-C was performed and exhibited contact matrices binned at 1-kb resolution, and novel principles of chromatin looping.
The reports identified loop domains that demarcate contact domains, which may correspond to those generally referred to as sub-TADs. In addition, recent reports employing single cell Hi-C analyses indicated the variabilities of TADs among cells (Bintu et al., 2018;Finn et al., 2019). Accordingly, we have rewritten our introduction in the context of these new findings on page 3, line 18.
As described below, we would like to clarify that we had analyzed the Eleanor TAD at a 40-kb resolution (Fig. 1b), in contrast to the 250-kb or lower resolution in the previous report (Lieberman-Aiden et al., 2009). In this revised manuscript, we further analyzed several Hi-C data, including those from the work by Rao et al. (Rao et al., 2014), and consistently detected the Eleanor TAD (new Supplementary Fig. 2c). Considering this and many other reports regarding TADs (Donaldson-Collier et al., 2019;Le Dily et al., 2019;Nuebler et al., 2018;Schwarzer et al., 2017;Szabo et al., 2018), it is beyond the scope of this manuscript to discuss whether TADs do not really exist or not.

Page 6, top. "approximately 700 kb chromatin domain".
What makes this a domain? Would "region" be more appropriate?
We changed the sentence to "approximately 700 kb chromatin region" on page 6, line 5.

It is unclear from this description how any of this was calculated. Were the "sharp transitions" in 4C signal calculated by an objective computational approach or just visually?
We again thank the reviewer for raising this important point. Unlike the directionality index or insulation square analysis for the Hi-C analysis, the positioning of the border in 4C-Seq is not well established. In the previous manuscript, we determined the 4C-Seq boundaries by visually inspection.
With this reviewer's comment, we made a large effort to determine them quantitatively. We applied the Tukey-Kramer method to the sequencing read numbers in the region of 151,000,000-153,000,000, with 100-kb resolution (new Supplementary Fig. 2a). In the 4C-Seq of MCF7 and LTED cells, we found that the regions of approximately 151,700,000 and 152,700,000 were statistically distinct from the neighboring regions (P<0.05), suggesting that they are the places where the sharp transitions occur. This observation was consistent among all four of the experiments using MCF7 and LTED cells (new Supplementary Fig. 2a). We found that the region inside the 4C-Seq, at approximately 151,900,000 and 152,200,000, might be distinct from the neighboring region, which suggests the possible presence of three sub-domains, but they were observed inconsistently (new Supplementary Fig. 2a). Due to this inconsistency, we did not further pursue the sub-TAD.

It appears that TAD boundaries were calculated by binning the Hi-C data into 250 kb bins.
This would position the boundaries at a resolution of +/-250 kb, which represents a very large region when making the arguments the authors want to state. An objective observer, asked to place the boundaries of the 4C-seq, would not place them where the authors have. The region where the

4C-seq becomes background is not at the borders the authors have marked as the TAD boundaries, but further outside. Alternatively, one could argue that based on visual inspection the 4C-seq suggests the existence of three domains within the region claimed to be a TAD.
We deeply apologize for the confusing sentence in the previous manuscript, "Positions of TAD boundaries and A/B compartments (250-kb bins) in MCF7 cell were also derived from the same study". It should have been the "positions of the TAD boundaries were determined by insulation square analysis using 40-kb bins, while the A/B compartment profiles were calculated at 250-kb resolutions (Barutcu et al., 2015)". Therefore, we have more precisely rewritten the method section of the manuscript on page 25, line 14.
In the revised manuscript, we also analyzed the Hi-C data of IMR90 (human fetal lung cells), HMEC (human mammary epithelial cells) (Rao et al., 2014), and T47D cells (human ER positive breast cancer cells) (Consortium, 2012), and the high resolution analyses revealed the contact domains or sub-TADs more clearly. The directional index revealed that the same regions form boundaries (new Supplementary Fig. 2b), regardless of whether they were analyzed with higher resolutions (10-kb for HMEC and IMR90, 40-kb for T47D, and 10-kb for GM12878) than that for MCF7 cells (Barutcu et al., 2015) (Fig. 1b). All showed the TAD boundaries at 151,700,000 and 152,700,000, and thus matched the 4C-Seq boundaries that we described above.
In conclusion, we have determined the 4C-Seq boundaries quantitatively and found that they matched the TAD boundaries consistently determined with the four different cell types.
These findings are described in the results section on page 6, line 17.

To emphasize the point that the location of the TAD boundaries the authors have chosen is arbitrary and it depends on the resolution of the Hi-C data, I'm enclosing a picture of the same region using
Hi-C data from Rao et al in GM12878 cells containing 4.5 billion reads instead of a couple hundred.
The differences are obvious, although one could argue that this is due to the different cell type.

However, the same region looks very similar in HCT116 cells at the same resolution. Is the ESR1 gene expressed in GM12878 cells?
We thank the reviewer for providing the Hi-C data. According to the comment, we analyzed the TAD boundaries at 40-kb bins for these Hi-C data of GM12878 cells (Rao et al., 2014) on the 3D Genome Browser. Alignments of the TAD profiles of the GM12878 (containing 4.5 billion reads) and MCF7 (0.15 billion reads) cells (Barutcu et al., 2015) showed that the TAD in GM12878 and the one in MCF7 cells (0.15 billion reads) shared most of the region (151,825,000-153,000,000 for GM12878 cells, and 151,750,000-152,750,000 for MCF7 cells). As the reviewer pointed out, the ESR1 gene is not expressed in both GM12878 and HCT116 cells (RNA-Seq of ENCODE). These data are presented in the new Supplementary Fig. 2c, and described in the revised manuscript on page 6, line 20.

The domains the authors claim to be a TAD looks considerably different, with several CTCF loops
that presumably would interfere with enhancer-promoter interactions and with the co-regulation the authors claim takes place within this domain. Are these CTCF sites present in MCF7 cells? If so, they presumably would form the same CTCF loops. Where are the different co-regulated genes present with respect to these loops? Are they all inside the loops or are some outside?
According to the reviewer, we now show the data for the CTCF ChIA-PET of MCF7 cells (Consortium, 2012) at the bottom of the new Supplementary Fig. 2c. These loops that were anchored with CTCF may well correspond to the CTCF loops that were described in Rao et al. (Rao et al., 2014). There are no CTCF loops that traverse beyond the Eleanor TAD. Please note that no enhancer elements governing the co-regulation of the genes in the Eleanor TAD have been described. Therefore, it may not be appropriate to speculate whether the CTCF loops would interfere with the enhancer-promoter interactions, interfere with the co-regulation of the genes within this domain, or represent local promoter-enhancer interactions.

Page 7, bottom. "Upon further analysis of our 4C-Seq data, we observed intra-chromosomal ESR1 promoter interactions along the entirety of chromosome 6 in MCF7 and LTED cells".
It is surprising that authors observe such long-range interactions in the 4C data. Based on the description in the Methods, it appears that authors only subtracted the input from the experimental with no distance normalization. Authors should use a standard 4C analysis pipeline such as that described by de Laat and collaborators in PMID 22961246. According to this reviewer, we re-analyzed our 4C-Seq data using the suggested standard pipeline described by de Laat and colleagues (van de Werken et al., 2012). In this method, the interactions were normalized against the distance from the bait. We again found interactions between the ESR1 promoter and 6q21 in LTED cells. These new data are presented in the new Supplementary Fig. 3a, labeled as 4C-Seq (normalized against distance).
6. Page 8, top. "However, there was a tendency for the 4C peaks to coincide with A compartments throughout the chromosome, suggesting that the ESR1 promoter region in the Eleanor TAD interacted with genes in the A compartments of multiple regions of chromosome 6. Consistently, Fig. 2a)".

Please explain in quantitative terms what "tendency" and "moderate" means.
We apologize for not describing the correlations in quantitative terms. We analyzed the overlap between the 4C peaks and the A compartments along chromosome 6 with a 250-kb bin resolution. In LTED and MCF7 cells, more than 60% of the regions showed the overlap. Similarly, we found that 60% of the regions showed the overlap between the 4C peaks and the genes upregulated in LTED cells. We show these new data in the new Supplementary Fig. 3b, c, and describe them in the revised manuscript on pages 8, lines 10 and 17.
If the authors want to make the argument that regions that interact with the ESR1 gene represent upregulated genes, it may be better to plot the RNA-seq data. Also, it would help the reader asses the overlap between upregulated genes and 4C interactions if regions of interest are shown at a higher resolution, perhaps in the supplemental figures.
We have plotted the genes upregulated in LTED cells as red bars below the A/B compartment graph in the new Fig. 2a and the new Supplementary Fig. 3a. The bars are derived from our previous RNA-Seq data (Tomita et al., 2015). The data suggested that the regions that interact with the ESR1 gene correlated with the genes upregulated in LTED cells. We also show an enlarged view in the new Supplementary Fig. 3a, to help the readers assess the overlap between upregulated genes and 4C interactions. Furthermore, we calculated the overlap between upregulated genes and 4C interactions. The results are presented in the new Supplementary Fig. 3c., and described in the text on page 8, line 17.

Page 9, bottom. "we inhibited Eleanors with resveratrol in LTED cells and performed 3C-qPCR
analysis. We found that the ESR1 promoter and the b site were dissociated after resveratrol treatment (Fig. 2f).
It should be noted that the transcriptional levels of the b site genes were mostly unchanged". It seems from Figure 2G that transcription of SNX3 and LACE1 changed quite a bit. How long was the resveratrol treatment? Is it possible that the RNAs for the other genes are more stable and that's why change was not detected?
To address the comments, first, we revised the text to describe the result, as follows. "It should be Next, we address the specific question. We treated the LTED cells with resveratrol for 24 hours.
Finally, to address the point raised by this reviewer, we measured the stabilities of RNAs for SNX3, LACE1, FOXO3 and others. We found that the stability of the FOXO3 RNA was lower than those of the LACE1 and SNX3 RNAs. Therefore, it is unlikely that the maintenance of FOXO3 transcription was due to its stability. These new results are presented in the new Supplementary Fig.   6a, and described in the revised manuscript on page 11, line 9. We thank the reviewer for pointing out the need for this clarification. We would like to make sure that we have shown that the resveratrol treatment affected Eleanors, but did not affect the transcription of the coding genes in 6q21. We demonstrated these data in the previous Fig. 2g (new   Fig. 3c). We have also described, "More importantly, the activation of the FOXO3 gene in the 6q21 domain is autonomous and not dependent on a long-range chromosomal interaction via Eleanors." on page 16, line 2.
Second, the authors never show causality between interactions and transcription. Authors assume that interactions determine transcription, but it is equally possible, and perhaps more likely, that changes in transcription are determined by binding of protein complexes to the promoters and regulatory regions of genes, and that these proteins then mediate the interactions. Loss of interactions after resveratrol treatment could be a consequence of decreased transcription, which the authors argue it does not change (see #7), but Figure 2G appears to suggest that it does.
We agree with the reviewer that it is important to mention the causality between interaction and transcription. We also agree that it is likely that the loss of interactions after the resveratrol treatment could be a consequence of the decreased transcription of the Eleanor TAD. We agree that once transcription has been activated by Eleanors, many transcription factors are gathered together, and then the long-range chromatin interaction may be established. Previous reports also showed that the enhancer region of Bcl11b (an essential gene for early T cell differentiation) was transcribed and then the chromatin interaction changed (Isoda et al., 2017). Therefore, we have revised the discussion section on page 16, line 21.
Does RNA pol II remain unchanged after treatment? The TAD is not permissive for transcription, unless the authors are talking about specific CTCF loops, which do not necessarily correspond to TADs.
According to the reviewer, we performed ChIP-qPCR, and found that the RNA polymerase II binding to the TSS region of the genes in the Eleanor TAD was reduced by the resveratrol treatment of LTED cells (new Supplementary Fig. 2e). In contrast, RNA polymerase II binding to the TSSs of SNX3 and FOXO3 in 6q21 did not change (new Supplementary Fig. 5d). Both are consistent with each gene's activity. We describe these findings in the revised manuscript on page 7, line 16 and page 11, line 10.

Page 11, top. "because most other coding (exons) and non-coding regions in the Eleanor TAD
were transcribed in the same direction".
We agree that a quantitative description is important. We found that most (90%) of the ESR1 region (chr6:152,128,814-152,424,408) was transcribed in the sense direction, based on a re-analysis of our stranded RNA-Seq data in MCF7 and LTED cells (Tomita et al., 2015), as shown in the graph below.
We now describe these data on page 12, line 11.
Sequencing reads mapped at the ESR1 gene. The numbers corresponding to sense and antisense RNA were counted and normalized against the total number of mapped reads.
10. Page 11, bottom. "knockdown of ESR1 mRNA with RNAi or specific degradation of ER protein with a chemical compound ICI 782,780, had no effect on pa-Eleanor(S) expression".
Evidence showing that ESR1 protein was depleted in these experiments should be shown.
According to the reviewer, we performed immunoblotting and the results showed that the ER protein levels were decreased by siESR1 and ICI 182,780 treatments in LTED cells. These new data are now presented in the new Supplementary Fig 8b, e, and described in the revised manuscript on page 13, line 1.
It is surprising that KD of pa-Eleanor had a modest effect when measured by qPCR (Figure 4a) but a very strong effect when measured by FISH ( Figure 4C). Please comment on this discrepancy.
We agree with the reviewer. This is consistent with our previous observation of the large effect of the u-Eleanor knockdown (Tomita et al., 2015). The result suggests that the level of pa-Eleanor may be critically regulated and a subtle reduction leads to large molecular and cellular alterations, including the Eleanor cloud disappearance (new Fig. 5c-e). We discuss this discrepancy in the revised text on page 13, line 20.

Page 12, bottom. "To investigate these, we first knocked down ESR1 mRNA in LTED cells and found that the transcriptional levels of genes in the Eleanor TAD, including pa-Eleanor(S) and the
Eleanor RNA cloud, did not change (Supplementary Fig. 5a-c) Figure 5A are very small.
According to the reviewer's comment, we performed immunoblotting, which revealed that the ER protein level was clearly decreased by siESR1 and ICI 182,780 treatments in LTED cells. These new data are shown in the new Supplementary Fig. 8b, e, and described on page 14, line 14.

Page 13, bottom. "To investigate the function of pa-Eleanor(S) in gene regulation through
chromatin structures, we performed 3C-qPCR analysis. We found that the interaction between the ESR1 promoter and the b site at 6q21 reduced with pa-Eleanor(S) knockdown in LTED cells (Fig.   5a), supporting the idea that the ncRNA was involved in the long-range chromatin interaction. In contrast, transcriptional levels of the genes at 6q21 showed little to-no change in the LTED cells with pa-Eleanor(S) knockdown (Fig. 5b)". This is a critical experiment and conclusion that should not rely on just one 3C experiment. Authors should perform 4C and examine all interactions between ESR1 and other sequences in the chromosome.
According to the reviewer's suggestion, we performed Hi-C, which is equivalent to 4C, but more suitable for a genome-wide analysis. We found that all interactions between ESR1 and other sequences in the chromosome did not dramatically change by the Eleanor inhibition with resveratrol (new Fig. 3a and new Supplementary Fig. 5a). However, a detailed examination suggested that the contact between the ESR1 promoter and FOXO3 became unstable with resveratrol treatment (new Fig. 3a and new Supplementary Fig. 5a, b). We present these data in the new Fig. 3a, the new Supplementary Fig. 5a, b, and describe them in the revised text on page 10, line 16. Figure 5B. These changes in expression could account for the changes in interactions shown in Figure 5A.

Furthermore, gene expression changed significantly in at least one of the two experiments shown in
We thank the reviewer for careful inspection and providing the interesting possibility. As the reviewer mentioned, the change in gene expression was not consistently observed (one of the two experiments). The change was small. It was reduced to 75%, while ESR1 mRNA was reduced to 15% in the same kind of experiment (new Fig. 5b). Therefore, we have to avoid proposing the suggested possibility.
Reviewer #3 (Remarks to the Author): In this manuscript, the authors investigated the long-range chromatin TADs interaction and their consequence on balance of LTED MCF-7 cells growth and apoptosis. They found lncRNA pa-Eleanor (S) mediated the interaction of two TADs. Silence of pa-Eleanor inhibited the formation of Eleanor clouds, the binding of the b site of 6q21 with ESR1 promoter and promoted LTED cell apoptosis. And inhibition of pa-Eleanor mimics the effect of resveratrol treatment. In general, this manuscript is the extension of the previous study (Nat Commun 6, 6966 (2015)) and clarified the mechanism of LTED cells primed for apoptosis. However, the roles of 6q21 TAD in promoting co-upregulation of genes in the Eleanor TAD were still unknown. Some issues need to be addressed.
Major comments: We thank the reviewer for this critical comment. As pointed out, the inhibition of gene expression in the Eleanor TAD did not affect the expression of the FOXO3 contained in 6q21 TAD. To investigate whether the interaction between the b site of 6q21 with the ESR1 promoter promoted the transcription of Eleanors and ESR1, we tried to introduce chromosomal interactions between the ESR1 promoter and 6q21 by overexpressing pa-Eleanor(S) in MCF7 and HeLa cells, to determine how they would affect the transcriptional levels of the genes in the Eleanor and 6q21 TAD in MCF7 cells. As shown below, we found that the overexpression did not change the expression of the genes in both chromatin domains. This may have happened, because the overexpression was exogenous and the pa-Eleanor(S) could not promote the chromosomal interaction. Consistently, the size of the