Aging-regulated anti-apoptotic long non-coding RNA Sarrah augments recovery from acute myocardial infarction.

Long non-coding RNAs (lncRNAs) contribute to cardiac (patho)physiology. Aging is the major risk factor for cardiovascular disease with cardiomyocyte apoptosis as one underlying cause. Here, we report the identification of the aging-regulated lncRNA Sarrah (ENSMUST00000140003) that is anti-apoptotic in cardiomyocytes. Importantly, loss of SARRAH (OXCT1-AS1) in human engineered heart tissue results in impaired contractile force development. SARRAH directly binds to the promoters of genes downregulated after SARRAH silencing via RNA-DNA triple helix formation and cardiomyocytes lacking the triple helix forming domain of Sarrah show an increase in apoptosis. One of the direct SARRAH targets is NRF2, and restoration of NRF2 levels after SARRAH silencing partially rescues the reduction in cell viability. Overexpression of Sarrah in mice shows better recovery of cardiac contractile function after AMI compared to control mice. In summary, we identified the anti-apoptotic evolutionary conserved lncRNA Sarrah, which is downregulated by aging, as a regulator of cardiomyocyte survival.


4.
A more vexing question is whether Sarrah downregulation (either by Gapmers or using the Cispr model described later in the manuscript) is sufficient to evoke excessive apoptosis in wildtype mice and contributes to signs of cardiac senescence and/or apoptosis. If no additional signs of senescence are found, then the original term "age-related" was possibly prematurely chosen (this could be counter argumented by selecting a different mouse model, see comment #3 above), and suggests Sarrah is only involved in cellular death pathways. 5. More specifically on the execution of mentioned study in Fig.1, can the authors confirm that Sarrah overexpression was restricted to the heart and if so, only to heart muscle cells (Fig.1d)? If so, the endothelial component (Fig.1i) can only indirectly be affected in this study. The rescue in LVEF was relatively mild (Fig.1e), even when data are expressed as delta values. A more accurate physiological assessment could have been derived from MRI studies, invasive hemodynamics and/or including Doppler assessments. As presented, parameters of wall thickness and movements and ventricular volumes remain obscure. The quantification of the infarct size is suboptimal and should be derived from a 3D geometry of serially sectioned hearts. Fig.1g, h: quantification of TUNEL positive events and proliferating cells should be expressed per number of nuclei, not mm2 of histological sections. By what mechanism does Sarrah influence cardiomyocyte proliferation? 6. In the co-culture experiments with cardiomyocytes, a convincing approach would be to overexpress Sarrah and observe an increase in the proliferative capacity of HUVECs in the EHTorganoids. Do the authors think that Sarrah has a cell autonomous effect in endothelial cells or is the prevailing mechanism a form of communication between endothelial cells and cardiomyocytes, in which case conditioned media and the exosomal content could be of interest to study. 3. Line 112: "ontogenic" instead of "ontogenous"; 4. Figure 1a: specify that the red dot corresponds to Sarrah-siRNA; 5. Figure 2f: change "rest" with "non-cardiomyocytes".
Reviewer #2 (Remarks to the Author): The paper by Trembinski et al. evaluates the cardiac lncRNA transcriptome during aging. Via comparing RNA samples from young vs. old mouse heart cardiomyocytes, the authors identified >5000 expressed lncRNAs, of which 76 were selected for further validation. A siRNA screen, in combination with an apoptotic cell-based assay, finally identified Sarrah, a lncRNA located in the Oxct1 gene locus. Specifically, Sarrah knockdown increases apoptosis in cardiomyocytes. Next, the authors investigate the role of Sarrah in vivo in a mouse model of myocardial infarction, and demonstrate that Sarrah overexpression is protective in the stressed cardiomyocyte population. Several other experimental conditions are studied and confirm detrimental effects of Sarrah silencing on cardiomyocyte behavior. The proposed molecular mechanisms associated with Sarrah action could include formation of triple RNA-DNA helices at target gene promoters. This paper is interesting in its initial approach to identify aged-related lncRNAs in cardiomyocytes. The study is however quite confusing on several aspects. Sarrah is identified as expressed in the aged mouse cardiomyocyte population, is associated with apoptosis in the HL-1 cell line, and finally is demonstrated to produce beneficial effects in cardiomyocytes after myocardial infarction in vivo. These beneficial effects result from a role of Sarrah in apoptosis, in cardiomyocyte contractility, in regulating fibrosis and possibly in controlling heart perfusion via affecting endothelial cell function in the damaged heart. In addition, the authors take advantage in their study of mouse models of aging and myocardial infarction, of a rat model of HFpEF, and then of mouse HL-1 cells as well as human and rat cardiomyocytes in vitro. At the end, if Sarrah appears to be protective in all these different instances/pathophysiological situations, its exact role is not determined. Moreover, the molecular mechanisms responsible for these pleiotropic effects are not identified. If cardiomyocyte apoptosis is postulated to be at the center of Sarrah action, this should be studied in more details and the molecular mechanisms should be clearly described.
Major points 1. It is not clear how was Sarrah identified. The authors initially identified 5439 annotated lncRNAs as expressed in the young and old mouse hearts. How many lncRNAs among these 5439 were differentially expressed between the young and aged hearts? How were the 76 selected lncRNAs chosen for further study? Were these 76 lncRNAs differentially expressed? Please provide the list of the 76 genes together with the list of the adjacent protein coding genes (PCGs). Were these PCGs related to apoptosis, other pathways? 2. Sarrah expression does not seems to be significantly downregulated in old mouse cardiomyocytes ( Figure 1C). Sarrah is also not changed in old vs. young non-myocyte cells. However, Sarrah is significantly downregulated in the old mouse heart as compared to the young mouse heart ( Figure  1D). It doesn't seem to add up with the assumption that Sarrah is predominantly expressed in cardiomyocytes in vivo.
3. Sarrah knockdown in cardiomyocytes affects their contractility. How can the authors exclude that this effect does not occur secondary of induced apoptosis? What is the postulated mechanism linking Sarrah and contractility? Do some Sarrah target genes encode sarcomeric proteins, calcium handling proteins, etc.
4. Sarrah is presented as a conserved transcript but data supporting this statement are not shown. Please provide evidence for sequence/syntenic conservation between mouse, rat and human Sarrah. Figure 1B, as mentioned in page 6, bottom, does not provided this information.
5. Moreover, Sarrah is presented as a hypoxia-responsive gene. What is the evidence for this?
6. The formation of triple helices at target gene promoters is proposed to play a role in Sarrah action. The authors mention enrichment of triple helix prone motifs at promoters of genes modulated after Sarrah silencing in human cardiomyocytes as assessed by using the Triplex Domain Finder software. This should be shown and the list of genes should be provided. 7. The Triplex Domain Finder analysis is performed presumably using the human sequence (please confirmed). How conserved is this putative mechanism? The authors mention little conservation in Sarrah sequences from different species. Does mouse Sarrah for instance contain a region predicted to form triple helices? What are the sequences that are deleted in mouse Sarrah in experiments depicted in Figure 4H.
8. Why considering only downregulated genes after Sarrah silencing for the Triplex Domain Finder analysis (Suppl. Figure 6A)? Sarrah might exert repressive action on target genes, which could be alleviated following Sarrah knockdown. Upregulated genes should not be excluded a priori of the analysis. In any case, it would important to compare up and downregulated genes, and to determine whether a disequilibrium is observed between these two gene sets in terms of percentage of genes containing triple helix prone motifs. 9. Is there overlap in the genes, which are downregulated following Sarrah silencing in human vs. mouse cardiomyocytes, and which contain a triple helix prone motif in their promoters?
10. Sarrah seems to exert trans-regulatory function on gene expression. The proposed mechanism via formation of triple helices is interesting but has not been previously described. Instead, lncRNAs have been extensively described as being able to partner with proteins, in particular chromatin modifiers. This mechanism seems to be quite relevant in the context of the present study. It is not clear therefore why the authors exclude this possibility. The authors could consider a RNA pulldown approach to identify Sarrah-bound protein. Along this vein, deletion in Sarrah sequences might as well alter binding of important protein partners.
11. Figure 4I. Overexpression of mouse Sarrah lacking the triple helix domain increases apoptosis in mouse cardiomyocytes. This is surprising. Overexpression of WT Sarrah should decrease apoptosis (This control should be included in the experiment), and possibly delta Sarrah should not produce any effects on cardiomyocyte apoptosis but there are no obvious reasons for the deleted transcript to increase apoptosis.
12. The target loci of Sarrah action should be investigated using a ChIRP approach.
Minor points 1. Often, the legends, and even the methods are not clear. For instance, given the variety of models used in the paper, the authors should carefully indicate in the legend which cell types and which species are studied in the various experimental approaches described in the paper. In particular, the authors should clearly indicated when a cell line is used and when they rely on primary cells. Several key methods are not described, for instance the Triplex Domain Finder analysis.
2. The protocol using total Histone 3 immunoprecipitation to detect potential association of Sarrah with chromatin ( Fig. 4C) should include a step in which precipitated material is treated with DNAse to avoid possible amplification of Sarrah-encoding DNA.

Manuscript:
Sarrah is an Aging-Regulated Anti-Apoptotic Long Non-Coding RNA in Cardiomyocytes that Augments Recovery from Acute Myocardial Infarction; Trembinski et al.
First of all, we would like to thank the reviewers for their very helpful comments and overall their efforts, which helped us to substantially improve the manuscript and to more thoroughly describe the role of lncRNA Sarrah in the heart.

Response to reviewer #1
This manuscript deals with the discovery of an age-regulated long non-coding RNA termed "Sarrah" and deciphers its location, differential expression and function in the context of the senescent heart and in acute myocardial infarction. The study contains a plethora of techniques and models, but could benefit from readjustments in the study design and more focus to justify the claims and conclusions about this lncRNA. Major 1. Fig. 1c: The description of Sarrah expression in cardiac cells is too vague with a separation between cardiomyocytes versus stromal cells. A more informative separation minimally includes cardiomyocytes, fibroblasts/pericytes and endothelial cells and qRT-PCR or Taqman analyses.
We thank the reviewer for this helpful comment and agree that a seperation of stromal cells into fibroblasts, endothelial cells and leukocytes would be more informative. We have therefore isolated cardiomyocytes, fibroblasts and endothelial from mouse hearts and measured Sarrah levels in all fractions (reviewer fig. 1, fig. 1c, supplementary fig. 1e). To condense the paper, we have moved the original figure to the supplements. The results indicate that Sarrah is expressed at comparable levels in endothelial cells and cardiomyocytes, but lower in fibroblasts. In accordance with these new findings, we removed the term "cardiomyocyte-enriched" from the manuscript. In total heart lysates, in which the majority of the RNA comes from cardiomyocytes, Sarrah is significantly reduced with aging ( fig. 1d).
Reviewer fig. 1: 12 week old mice were used for cell isolation of cardiomyocytes (CM), endothelial cells (EC) and fibroblasts (FB)). RNA was isolated and levels of cell type-specific markers (Tnnt2 for CM, Cdh5 for EC) and Sarrah were determined by qRT-PCR.
2. Fig. 1e: In the context of this manuscript, it is unclear why the authors studied hypertensive rats, HFpEF rats and rats with HFpEF and high fat diet. The added value of these disease subtypes distracts from the main conclusion that Sarrah is age-related and influences apoptotic pathways, and raises the question if Sarrah functionally participates in the HFpEF syndrome, where apoptosis is seemingly of a minor role in the pathogenesis.
The reviewer made a valid point here. We included the data derived from a HFpEF rat model to emphasize a potential of Sarrah during aging as HFpEF and cardiomyocyte stiffness mainly occur in the elderly. We agree though that the data might distract the reader from the main story line of the manuscript, which is about anti-apoptotic effects of Sarrah during aging and after acute myocardial infarction (AMI), and therefore moved these data to the supplementary figures (supplementary fig. 1g). Fig. 2, the authors quickly continue with an AAV gene delivery study for Sarah in the setting of acute myocardial infarction in young mice, which is incongruent with its reported function in agerelated heart disease. A far more impressive approach would be to rescue the hearts from a (prematurely) aged mouse model. A therapeutic study aimed to resupplement Sarrah expression would be expected later in the manuscript.

In
The reviewer made a very interesting point here. In the six month revision period, we did not have enough time to apply for ethical permission to use a progeria model (our choice would have been Ku80-/-mice), obtain these mice, breed enough numbers and perform experiments with 8-weeklong overexpression of Sarrah. However, we have now included more mice in the AAV gene delivery study and additionally performed MRI analysis of cardiac function. Of note, the mice in the study are subjected to acute myocardial infarctio, which is age-related and of particular therapeutic relevance. Furthermore, we moved the in vivo part of the manuscript towards the end (new fig. 5), as the reviewer suggests. 4. A more vexing question is whether Sarrah downregulation (either by Gapmers or using the Crispr model described later in the manuscript) is sufficient to evoke excessive apoptosis in wildtype mice and contributes to signs of cardiac senescence and/or apoptosis. If no additional signs of senescence are found, then the original term "age-related" was possibly prematurely chosen (this could be counter argumented by selecting a different mouse model, see comment #3 above), and suggests Sarrah is only involved in cellular death pathways.
We thank the reviewer for the suggestion and share the notion that it would be indeed interesting to study the effects of Sarrah knockdown or knockout in vivo.
Therefore, we intraperitoneally injected GapmeRs into 12 week old male C57Bl6/J mice to obtain an in vivo Sarrah knockdown model (reviewer fig. 2). We first injected the GapmeR sequence we used for our in vitro studies (TTGGAAAGGTGAGCTG) at 20 mg/kg, but obtained only a slight knockdown in the heart atrium and no knockdown in the ventricle. Injecting three doses of the GapmeR did not improve the knockdown efficiency. Eventually, we injected another GapmeR sequence (CAACATGGACAAGGAC) at 20 mg/kg, but still did not achieve a reasonable knockdown in the mouse heart. Reviewer fig. 2: Sarrah levels in the heart atrium and ventricle, liver and skeletal muscle following different GapmeR injection protocols.
We apologize that it has not been feasible to generate a knockdown model for Sarrah and hope that the reviewers will appreciate our efforts. Furthermore, we find it more therapeutically relevant to show that Sarrah augmentation induces cardiac function, which we confirmed in a second experiment (see also below). The reviewer states that the term "age-related" is prematurely chosen. We used the term "age-regulated" merely to denote that aging reduces the expression of Sarrah. 5. More specifically on the execution of mentioned study in Fig. 1, can the authors confirm that Sarrah overexpression was restricted to the heart and if so, only to heart muscle cells (Fig. 1d)? If so, the endothelial component (Fig. 1i) can only indirectly be affected in this study. The rescue in LVEF was relatively mild (Fig. 1e), even when data are expressed as delta values. A more accurate physiological assessment could have been derived from MRI studies, invasive hemodynamics and/or including Doppler assessments. As presented, parameters of wall thickness and movements and ventricular volumes remain obscure. The quantification of the infarct size is suboptimal and should be derived from a 3D geometry of serially sectioned hearts. Fig. 1g, h: quantification of TUNEL positive events and proliferating cells should be expressed per number of nuclei, not mm2 of histological sections. By what mechanism does Sarrah influence cardiomyocyte proliferation?
These are all very valid points and we have addressed them accordingly. Firstly, Sarrah overexpression was performed using well characterized AAV9, which have been repeatedly used for cardiomyocyte gene transfer. These vectors only transduce cardiomyocytes and do not target endothelial cells or fibroblasts adequately at the dose used in our study (6x10^11 viral genomes per mouse; [1]). Indeed the effects on endothelial cell proliferation are very likely paracrine. Coculture experiments in vitro confirm that Sarrah depletion in cardiomyocytes reduces proliferation of endothelial cells (supplementary fig 7c). Conversely, overexpression of Sarrah in cardiomyocytes induces endothelial cell proliferation (supplementary fig. 7b and reviewer fig. 6).
Secondly, at the request of the reviewer we have repeated the AMI experiment with GFP-control and Sarrah overexpressing mice, using the exact conditions as originally performed. Now we included cardiac MRI measurements to quantify cardiac function ( fig. 5c-h, supplementary fig. 6e,  g, h and reviewer fig. 4). This experiment showed that our echocardiography analysis of the original study was very comparable to MRI measurements. Furthermore, we now include additional parameters in our main figures ( fig. 5c-h) and supplemental figures (supplementary fig.  e, g, h). At the end of this experiment, we used the hearts to perform serial sectioning as requested by the reviewer (fig. 5l and reviewer fig. 5).

Reviewer figure 4:
a. Adeno-associated virus serotype 9 (AAV9)-green fluorescent protein (GFP) control virus particles or AAV9-cytomegalovirus (CMV)-Sarrah virus particles were injected intravenously into mice two weeks prior to AMI surgery. Cardiac function was analyzed by echocardiography at 1, 7 and 14 days and by MRI at 14 days after AMI surgery. b. Sarrah levels of mouse hearts after AMI were measured by qRT-PCR. (n = 15-16; SEM; * t-test p < 0.05). c. Wall motion score index (WMSI), a parameter of heart muscle movement at 16 individual heart sections, was assessed from echocardiographic measurements. A WMSI decrease indicates an improved recovery of cardiac contractile function. Delta WMSI values at day 14 refer to day 1 (n = 6-9; SEM; * t-test p < 0.05). d. Left ventricular wall thickness was assessed from magnetic resonance imaging (MRI) at day 14 after AMI surgery (n = 7-8; SEM; * t-test p < 0.05). e. Stroke volume was calculated as the difference between end diastolic and end systolic volumes, which were assessed from MRI at day 14 after AMI surgery (n = 7-8; SEM; ** t-test p < 0.01). f. Cardiac contractile function was assessed by MRI. Displayed are delta ejection fraction values on day 14 after AMI in comparison to day 1 (n = 7-8; SEM; ** t-test p < 0.01). g. Representative images from MRI on day 14 after AMI during end systole and end diastole from apical to basal heart segments. Lastly, we have quantified apoptosis and proliferation normalized to number of nuclei ( fig. 5i-j).
Our data indicate that cardiomyocyte proliferation is not affected by Sarrah. Sarrah overexpression does induce proliferation of endothelial cells in a paracrine manner (see also next point).
6. In the co-culture experiments with cardiomyocytes, a convincing approach would be to overexpress Sarrah and observe an increase in the proliferative capacity of HUVECs in the EHTorganoids. Do the authors think that Sarrah has a cell autonomous effect in endothelial cells or is the prevailing mechanism a form of communication between endothelial cells and cardiomyocytes, in which case conditioned media and the exosomal content could be of interest to study.
Thank you for this interesting comment. We agree that an overexpression approach would be more convincing and therefore overexpressed Sarrah in human cardiomyocytes, co-cultured them with HUVECs for 48 hours and analyzed HUVEC proliferation by BrdU staining followed by flow cytometry (supplementary fig. 7b and reviewer fig. 6). Indeed, after co-culture with Sarrah overexpressing cardiomyocytes, more HUVECs in S-phase are observed. Since we now focus the manuscript on the effects of Sarrah on cardiomyocyte apoptosis, we have refrained from describing the effects of cardiomyocyte Sarrah on endothelial cells in more detail in the manuscript.
Reviewer figure 6: Human cardiomyocytes were transduced with lentiviral vectors for overexpression and co-cultured with HUVECs on a transwell after medium change. Proliferation of HUVECs was assessed by bromodeoxyuridine (BrdU) staining and flow cytometry analysis after two days of co-culturing (n = 3; SEM; * t-test p < 0.05).
7. Fig. 4: Some level of experimental validation of Sarrah downstream targets and their involvement in apoptosis are warranted.
We have identified 135 downstream target genes. We selected the 5 that we hypothesized to be most likely to confer the anti-apoptotic effects of Sarrah. However, knockdown of these 5 genes (GPC6, PDE3A, ITPR2, PARP8, SSBP2) did not significantly induce apoptosis (data not shown). We hypothesize that, in analogy to microRNA function, regulation of the majority of targets together confers the anti-apoptotic action of Sarrah.

Minor
1. Write the name of the lncRNA in italics.
We thank the reviewer for his or her attention to detail. We have implemented the changes.
Response to reviewer #2: The paper by Trembinski et al. evaluates the cardiac lncRNA transcriptome during aging. Via comparing RNA samples from young vs. old mouse heart cardiomyocytes, the authors identified >5000 expressed lncRNAs, of which 76 were selected for further validation. A siRNA screen, in combination with an apoptotic cell-based assay, finally identified Sarrah, a lncRNA located in the Oxct1 gene locus. Specifically, Sarrah knockdown increases apoptosis in cardiomyocytes. Next, the authors investigate the role of Sarrah in vivo in a mouse model of myocardial infarction, and demonstrate that Sarrah overexpression is protective in the stressed cardiomyocyte population. Several other experimental conditions are studied and confirm detrimental effects of Sarrah silencing on cardiomyocyte behavior. The proposed molecular mechanisms associated with Sarrah action could include formation of triple RNA-DNA helices at target gene promoters.
This paper is interesting in its initial approach to identify aged-related lncRNAs in cardiomyocytes. The study is however quite confusing on several aspects. Sarrah is identified as expressed in the aged mouse cardiomyocyte population, is associated with apoptosis in the HL-1 cell line, and finally is demonstrated to produce beneficial effects in cardiomyocytes after myocardial infarction in vivo. These beneficial effects result from a role of Sarrah in apoptosis, in cardiomyocyte contractility, in regulating fibrosis and possibly in controlling heart perfusion via affecting endothelial cell function in the damaged heart. In addition, the authors take advantage in their study of mouse models of aging and myocardial infarction, of a rat model of HFpEF, and then of mouse HL-1 cells as well as human and rat cardiomyocytes in vitro. At the end, if Sarrah appears to be protective in all these different instances/pathophysiological situations, its exact role is not determined. Moreover, the molecular mechanisms responsible for these pleiotropic effects are not identified. If cardiomyocyte apoptosis is postulated to be at the center of Sarrah action, this should be studied in more details and the molecular mechanisms should be clearly described. Major 1. It is not clear how was Sarrah identified. The authors initially identified 5439 annotated lncRNAs as expressed in the young and old mouse hearts. How many lncRNAs among these 5439 were differentially expressed between the young and aged hearts? How were the 76 selected lncRNAs chosen for further study? Were these 76 lncRNAs differentially expressed? Please provide the list of the 76 genes together with the list of the adjacent protein coding genes (PCGs). Were these PCGs related to apoptosis, other pathways?
Considering the comparison of aged cardiomyocyte fraction with young cardiomyocyte fraction, there are 91 differentially regulated lncRNAs (p < 0.05). We manually curated the list of cardiomyocyte-enriched lncRNAs and removed lncRNAs for which we did not find reliable reads when assessing the data in a genome viewer. At the request of the reviewer, we have prepared the list of 76 lncRNAs with adjacent protein-coding genes (PCGs). Of all PCGs on both sides of the lncRNAs we performed a Pubmed search to assess a potential role in apoptosis (see Excel file "76 lncRNA and PCGs"). However, we did not select Sarrah based on the genomic location. We used this list of 76 lncRNAs as a starting point for a functional assay to identify lncRNAs that regulate apoptosis in cardiomyocytes. This assay showed that out of these 76 lncRNAs, the most robust induction of apoptosis was achieved by silencing Sarrah.
2. Sarrah expression does not seems to be significantly downregulated in old mouse cardiomyocytes ( Figure 1C). Sarrah is also not changed in old vs. young non-myocyte cells. However, Sarrah is significantly downregulated in the old mouse heart as compared to the young mouse heart ( Figure  1D). It doesn't seem to add up with the assumption that Sarrah is predominantly expressed in cardiomyocytes in vivo.

We thank the reviewer for the valid remark. To address this concern, we have isolated cardiomyocytes (CM), endothelial cells (EC) and fibroblasts (FB) from mouse hearts and measured levels of cell-specific markers and Sarrah (see also point #1 form reviewer #1). The results indicate that Sarrah is expressed at comparable levels in endothelial cells and cardiomyocytes, but lower in fibroblasts.
In accordance with these new findings, we removed the term "cardiomyocyteenriched" from the manuscript. In total heart lysates, in which the majority of the RNA comes from cardiomyocytes, Sarrah is significantly reduced with aging (reviewer fig. 7, fig. 1c, supplementary  fig. 1e). Reviewer figure 7: 12 week old mice were used for cell isolation of cardiomyocytes (CM), endothelial cells (EC) and fibroblasts (FB)). RNA was isolated and levels of cell type-specific markers (Tnnt2 for CM, Cdh5 for EC) and Sarrah were determined by qRT-PCR. this effect does not occur secondary of induced apoptosis? What is the postulated mechanism linking Sarrah and contractility? Do some Sarrah target genes encode sarcomeric proteins, calcium handling proteins, etc.
The rat cardiomyocyte contractility measurements were performed on individual viable beating cardiomyocytes, therefore we already select for cardiomyocytes that have not undergone apoptosis yet. Several Sarrah target genes indeed regulate contractility. For example phosphodiesterase 3a (encoded by the Pde3a gene) is known to regulate Ca 2+ handling via regulation of PKA signaling. We postulate that in vivo, the increased cardiac contractile function after Sarrah overexpression is achieved through a combination of the following mechanisms: (1) increased survival of cardiomyocytes, leading to a higher number of cells that contribute to contractile function, (2)

increased contractile function of individual cardiomyocytes and (3) paracrine stimulation of endothelial cell proliferation likely contributing to increased perfusion.
4. Sarrah is presented as a conserved transcript but data supporting this statement are not shown. Please provide evidence for sequence/syntenic conservation between mouse, rat and human Sarrah. Figure 1B, as mentioned in page 6, bottom, does not provided this information.
We apologize for the lack of clarity. Unlike protein coding genes, most lncRNA genes are poorly sequence-conserved, but rather locus-conserved [2]. Indeed Sarrah is locus-conserved, which we mention in the manuscript: " (fig. 1b,  supplementary fig. 1d, supplementary table 1a). Nevertheless, Sarrah sequence conservation between mouse and human is relatively high (supplementary fig. 1d and reviewer fig. 8), which provides further evidence for its conserved function and relevance.

"we searched for homologous transcripts in humans, pigs and rats using publicly available sequencing and annotation databases (http://genome.ucsc.edu/; 18-22) and found transcripts in the ontogenic loci with small stretches of conserved sequences
Reviewer figure 8: The sequences of human and mouse Sarrah are partially conserved. 5. Moreover, Sarrah is presented as a hypoxia-responsive gene. What is the evidence for this?
We have investigated whether Sarrah is regulated in a hypoxia-responsive manner in both mouse and human cardiomyocytes by culturing the cells at 1 % and 0.2 % oxygen, respectively (supplementary fig. 6a-b and reviewer fig. 9). Additionally, we treated the cells with 300 µM deferoxamine mesylate (DFO), a hypoxia-mimicking compound. All conditions yielded a significant reduction in Sarrah levels and have led to the conclusion that Sarrah is downregulated by hypoxia.
6. The formation of triple helices at target gene promoters is proposed to play a role in Sarrah action. The authors mention enrichment of triple helix prone motifs at promoters of genes modulated after Sarrah silencing in human cardiomyocytes as assessed by using the Triplex Domain Finder software. This should be shown and the list of genes should be provided.
We have specified the regions in mouse and human Sarrah forming triple helices for the reviewers (see supplementary table 3). Additionally, we have prepared a list with the complete results from the Triplex Domain Finder software which includes all target genes and exact RNA binding sites in promoter regions of target genes for both human and mouse Sarrah (see Excel file "Triplex domain finder_Sarrah results").
7. The Triplex Domain Finder analysis is performed presumably using the human sequence (please confirmed). How conserved is this putative mechanism? The authors mention little conservation in Sarrah sequences from different species. Does mouse Sarrah for instance contain a region predicted to form triple helices? What are the sequences that are deleted in mouse Sarrah in experiments depicted in Figure 4H?
We confirm that the Triplex Domain Finder analysis was initially performed with human data, which identified the triple helix domain of Sarrah and 135 target genes with triple helix prone motifs in the human genome, all of them being downregulated in a microarray experiment with Sarrah-depleted human cardiomyocytes.
Subsequently, we repeated the analysis using mouse Sarrah and promoter sequences from the mouse genome. As in human, the algorithm detected a triple helix forming motif in the murine Sarrah sequence as well as an enrichment of triple helix prone motifs in downregulated, but not upregulated gene promoters (see next point for enrichment in up-vs. downregulated genes in human).

The sequence deleted in mouse cardiomyocytes (HL-1 cell line) in figure 3h and supplementary figures 5e-g is the triple helix forming region in mouse (sequences of sgRNAs used in the experiment are provided in supplementary table 5c).
Sarrah target genes showed a significant overlap between human and mouse (see concern #9; list provided, sheet "Overlapping Sarrah targets" in Excel file "Triplex domain finder_Sarrah results").
8. Why considering only downregulated genes after Sarrah silencing for the Triplex Domain Finder analysis (Suppl. Figure 6A)? Sarrah might exert repressive action on target genes, which could be alleviated following Sarrah knockdown. Upregulated genes should not be excluded a priori of the analysis. In any case, it would be important to compare up-and downregulated genes, and to determine whether a disequilibrium is observed between these two gene sets in terms of percentage of genes containing triple helix prone motifs.
We 9. Is there overlap in the genes, which are downregulated following Sarrah silencing in human vs. mouse cardiomyocytes, and which contain a triple helix prone motif in their promoters?
This is indeed an interesting and relevant question. We identified 165 mouse genes that were predicted to contain triple helix motifs for mouse Sarrah in their promoters. Out of the 135 human gene promoters forming triple helices with Sarrah, 54 (40 %) were also targeted by mouse Sarrah. This includes for example the promoter of GPC6 that we characterized in both species. The list of overlapping target genes is provided for the reviewers (sheet "Overlapping Sarrah targets" in Excel file "Triplex domain finder_Sarrah results"). Of note, even though a 40 % overlap may seem low, this is in the same order of magnitude as target genes of well-conserved microRNAs. For example, the overlap of miR-34a targets between mouse and human is 54 % (according to Targetscan v7.2).
10. Sarrah seems to exert trans-regulatory function on gene expression. The proposed mechanism via formation of triple helices is interesting but has not been previously described. Instead, lncRNAs have been extensively described as being able to partner with proteins, in particular chromatin modifiers. This mechanism seems to be quite relevant in the context of the present study. It is not clear therefore why the authors exclude this possibility. The authors could consider a RNA pulldown approach to identify Sarrah-bound protein. Along this vein, deletion in Sarrah sequences might as well alter binding of important protein partners.
We fully agree that an RNA pulldown approach would be a highly relevant means for further characterizing the mechanism of Sarrah function and have addressed the reviewer's concern as suggested. RNA pulldown of Sarrah from lysate of the mouse cardiomyocyte cell line HL-1 followed by mass spectrometry identified CRIP2, a transcription factor expressed mainly in the heart, as an interaction partner of Sarrah ( fig. 4b-c and reviewer fig. 10a-b), which is known to recruit the transcriptional co-activator p300 [3] that acetylates histone H3 on lysine 27 (H3K27ac) to activate transcription. We validated the physical interaction between CRIP2 and Sarrah, between p300 and Sarrah as well as between H3K27ac and Sarrah in RIP experiments ( fig. 5d and reviewer fig. 10c) and conclude that Sarrah facilitates gene transcription through recruiting CRIP2 and p300 to open chromatin loci. Of note, triple helix-mediated RNA-DNA interaction in cis has been shown for several lncRNAs (see http://dx.doi.org/10.1016/ j.chembiol.2016.09.011 for a recent review [4]), and is a mechanism to guide epigenetic modifiers like p300 to particular genomic regions.
Reviewer figure 10: Identification and validation of protein interacting with Sarrah.
a. Sarrah pulldown was performed by adding 200 pmol of biotinylated scrambled oligo or two biotinylated Sarrah antisense oligos to HL-1 cell lysate, coupling to streptavidin beads and elution. Pulldown efficiency was determined by qRT-PCR of eluted samples (displayed as % input; n = 6; SEM; * t-test p < 0.05). b. Volcano plot showing all proteins identified by mass spectometry analysis that are enriched in Sarrah pulldown as compared to pulldown with a scrambled oligo. CRIP2, the first hit, is highlighted. c. RNA-immunoprecipitations with antibodies against CRIP2 (rabbit antibody), p300 (mouse antibody) and histone acetylation H3K27ac (rabbit antibody) were performed in primary human cardiomyocytes. Sarrah levels were measured by qRT-PCR (n = 4-8; SEM; * t-test p < 0.05; ** t-test p < 0.01; IgG: immunoglobulin G).
11. Figure 4I. Overexpression of mouse Sarrah lacking the triple helix domain increases apoptosis in mouse cardiomyocytes. This is surprising. Overexpression of WT Sarrah should decrease apoptosis (this control should be included in the experiment), and possibly delta Sarrah should not produce any effects on cardiomyocyte apoptosis but there are no obvious reasons for the deleted transcript to increase apoptosis.
We apologize for the confusion. The transcript containing the deletion, which we named Sarrah∆TH, is not overexpressed, but transcribed from the endogeous locus that we mutated using a CRISPR/Cas9-based approach. The deletion of the triple helix forming region of Sarrah is now described more precisely in the Materials and Methods section as well as in the figure legend. An increase in apoptosis with endogenously expressed Sarrah lacking its functional region corresponds to the effects observed after Sarrah knockdown ( fig. 3h, supplementary fig. 5e and  reviewer fig. 11).
Reviewer figure 11: Deletion of the endogenous Sarrah triple helix domain increases apoptosis.
Additionally, we included a caspase assay with Sarrah overexpressing human cardiomyocytes as suggested by the reviewer. The assay shows a profound decline of caspase-3/7 activity upon Sarrah overexpression and supports our in vivo findings that Sarrah overexpression has an antiapoptotic effect on cardiomyocytes ( fig. 1f and reviewer figure 12).
12. The target loci of Sarrah action should be investigated using a ChIRP approach.
We thank the reviewer for the suggestion. To address this remark, we first validated Sarrah association with genomic DNA by RIP using the S9.6 antibody [5] that recognizes RNA:DNA hybrids. Our results show that Sarrah is indeed enriched at these structures as compared to the U4 snRNA negative control ( fig. 3i and reviewer fig. 13a). In a second experiment, we performed ChIP using the same antibody and measured Sarrah target gene promoters in genomic DNA fragments associated with triple helices by qRT-PCR. All five target genes were, unlike the GAPDH promoter negative control, enriched over IgG control. Four of the genes were significantly enriched over IgG control after RNase H digestion, which degrades RNA-DNA-strcutures such as R loops without affecting triple helices [6] ( fig. 3j and reviewer fig. 13b). We therefore conclude that the majority of predicted Sarrah target genes forms triple helices in their promoters.
c. RNA-immunoprecipitation with the S9.6 anti-DNA-RNA-hybrid antibody was performed in crosslinked primary human cardiomyocytes. Levels of U4 snRNA as a negative control and Sarrah were measured by qRT-PCR (n = 3; SEM; * t-test p < 0.05; IgG: immunoglobulin G). d. RNA-immunoprecipitation with the S9.6 anti-DNA-RNA-hybrid antibody was performed in crosslinked primary human cardiomyocytes. Sonicated DNA fragments were used for qRT-PCR to quantify triplex formation in gene promoters.

Minor
1. Often, the legends, and even the methods are not clear. For instance, given the variety of models used in the paper, the authors should carefully indicate in the legend which cell types and which species are studied in the various experimental approaches described in the paper. In particular, the authors should clearly indicate when a cell line is used and when they rely on primary cells. Several key methods are not described, for instance the Triplex Domain Finder analysis.
We thank the reviewer for this feedback and have rewritten the figure legends and methods in order to make our work easier understandable for the reader.
2. The protocol using total Histone 3 immunoprecipitation to detect potential association of Sarrah with chromatin (Fig. 4C) should include a step in which precipitated material is treated with DNase to avoid possible amplification of Sarrah-encoding DNA.
We fully agree on this valid remark and are aware of the necessity to exclude amplification of DNA. A DNase digestion step was performed during all RNA extractions for this manuscript, which we now state in the Materials and Methods section. To additionally ensure that no DNA is precipitated and amplified during qRT-PCR, we included a -RT control of the input (the sample expected to contain the highest DNA amount) in our cDNA synthesis reaction to which no reverse transcriptase was added. Ct values of -RT control compared to Ct values of samples indicate that no DNA has been amplified and that detected Sarrah signals result from reversely transcribed RNA.