Cooperation of LIM domain‐binding 2 (LDB2) with EGR in the pathogenesis of schizophrenia

Abstract Genomic defects with large effect size can help elucidate unknown pathologic architecture of mental disorders. We previously reported on a patient with schizophrenia and a balanced translocation between chromosomes 4 and 13 and found that the breakpoint within chromosome 4 is located near the LDB2 gene. We show here that Ldb2 knockout (KO) mice displayed multiple deficits relevant to mental disorders. In particular, Ldb2 KO mice exhibited deficits in the fear‐conditioning paradigm. Analysis of the amygdala suggested that dysregulation of synaptic activities controlled by the immediate early gene Arc is involved in the phenotypes. We show that LDB2 forms protein complexes with known transcription factors. Consistently, ChIP‐seq analyses indicated that LDB2 binds to > 10,000 genomic sites in human neurospheres. We found that many of those sites, including the promoter region of ARC, are occupied by EGR transcription factors. Our previous study showed an association of the EGR family genes with schizophrenia. Collectively, the findings suggest that dysregulation in the gene expression controlled by the LDB2‐EGR axis underlies a pathogenesis of subset of mental disorders.

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. Depending on transfer agreements, referee reports obtained elsewhere may or may not be included in this compilation. Referee reports are anonymous unless the Referee chooses to sign their reports.)

6th Aug 2020 1st Editorial Decision
Thank you for the submission of your manuscript to EMBO Molecular Medicine. We have now received feedback from the three referees whom we asked to evaluat e your manuscript . As you will see from the report s below, the referees acknowledge the pot ent ial int erest of the st udy. However, they also raise subst ant ial concerns about your work, which should be convincingly addressed in a major revision of the present manuscript .
While referee #1 and referee #2 are more support ive of publicat ion, referee #3 is much more reserved and ment ions limit at ions that must be addressed for the st udy to be conclusive: addit ional cont rols and experiment s are required, the inconsist encies bet ween dat a and text need to be addressed, along wit h clarificat ions and det ails. Referee #3 is also concerned about several pot ent ial overst at ement s, which we would ask you to pay at tent ion to and address wit hin reason.
We would welcome the submission of a revised version wit hin three mont hs for furt her considerat ion. ***** Reviewer's comments ***** Referee #1 (Remarks for Author): Review of Cooperation of LIM domain binding 2 with EGR transcription factors to regulate gene expression networks relevant to mental disorders Ohnishi et al studied a role of LIM domain binding 2 (LDB2) gene encoding a putative transcription regulator in mental disorders. In the present study, they used different approaches, phenotypic analyses of Ldb2 mouse model; yeast two-hybrid, identification of direct genomic targets of Ldb2 by chromatin immunoprecipitation (IP) followed by sequencing (ChIP-seq). The present manuscript deserves publication with minor revisions.
Minor revision in particular to mouse model data Mouse model data: pp 9-15 and Figures 2-6. For non-specialist reader, I think that it can be useful to resume phenotypical traits that characterize bona fide mouse models of schizophrenia. These phenotypical traits include neuronal architecture, synaptic plasticity, circuitry deficits and behavioral paradigms dependent on learning such as spatial working memory (WM) performance using a T-maze delayed non-match to place (DNMTP) task; Pre-Pulse Inhibition (PPI) (see for instance Gogos et al., 2019; Mukai et al., 2019). This information can highlight translational phenotypes from mouse to humans.
Pages 11-14: Analysis of Arc expression as a function of a cognitive test. The authors generated Arc Venus tg/tg; Ldb2 KO (AV KO), and Arc Venus tg/tg; Ldb2 WT (AV WT) to monitor neuronal activities in the amygdaloid complex through the Venus fluorescence. They found that the increment of the Middle Venus positive neurons was significantly larger in the AV KO compared to AV WT. These results indicating that Ldb2 suppresses the Arc gene transcription responding to the tone shock paired stimuli in the LA neurons ( Figure 5). Synaptic activity is reduced as a function of Arc expression whatever the genotype ( Figure 6 and text page 14). The logic is to present the result ( Figure 6) before the analysis of the number of positive neurons ( Figure 5), in order to present first the relation between arc expression and synaptic activity and second the inverse relationship between the number of ARC expression and the level of Ldb2 protein.
Furthermore, I think that indication of regression in graphs of Figure 6 with a graph for each condition will increase the readability of the figure. Discussion Phenotypic traits of the mouse model need to be discussed in comparison to what found in bona fide mouse models Results of Figure 6 have to be discussed with the indication that the decrease in mEPSPs possibly reflect a change in the number of postsynaptic AMPA receptors. The authors present an extensive modeling study using both mouse and iPSC-derived neurospheres to examine the potential for a novel risk gene for schizophrenia. An alternative cellular model would be iPSC-derived neurons as well as iPSC-derived neural cells from the patient.
Referee #2 (Remarks for Author): Ohnishi and colleagues conducted a very extensive mouse study to investigate the potential role of LDB2 in brain function and behavior. In a prior study, this group identified a rare translocation t(4;13) (p16.1;q21.31), in a schizophrenia patient. While the breakpoint on chromosome 13 was devoid of nearby genes, the breakpoint at chromosome 4 localized at the upstream region of the LDB2 gene, in effect, rendering the latter a plausible candidate deserving further studies. Toward this end, the authors created a knockout (KO) Ldb2 mouse, examined gene expression brain, conducted multiple behavioral assays and showed phenotype rescue using neuropsychiatric drugs. Several protein interactors with LDB2 were detected following yeast two-hybrid assay in human embryonic brain cDNA library. ChIP-seq analysis in human iPSC-derived neurospheres showed enrichment of EGR target DNA elements. The authors state that these findings suggest the possible involvement of the LDB2-EGR axis in neuropsychiatric disorders. Comments This manuscript reports on a tour-de-force effort to use both animal and human cellular models to examine various functional aspects of LDB2, a gene located close to the breakpoint on 4p16, a part of a very rare chromosomal structural aberration in a schizophrenia patient. The findings presented are important and additional supportive future data would heighten the relevance of the LDB2-EGR system to risk for schizophrenia. Minor comments • If the parents of the proband were available, were they unaffected and did any of them carry the translocation?
• Are there any known GWAS hits on 4p16.1 & 13q21.31 for any neuropsychiatric disorder? • Did the schizophrenia genomic DNA carry any of the known pathogenic CNVs, e.g., on 15q, 16p or 22q? • How far is LDB2 from the 4p breakpoint? • Did the authors try to sequence the 2 bands on Western blot? • Why did the authors use iPSC-derived neurospheres instead of either iPSC-derived neural progenitor cells or neurons? • It would have been interesting also to use iPSC-derived neural cells from the patient. • Y-2-H tends to have problems with signal-to-noise ratio, any comments about this issue? • There are too many numbers in pages 12-14 that are distracting. Presenting these numbers in Tables would help. • Fig. 1A: Replace "dimeraization" with "dimerization" • Fig. 1D Use of KO mice to study the function of the gene (LDB2) is appropriate. However, to insist that the model is a mental disorder model is not appropriate. They need to modify their statement in the manuscript.
Referee #3 (Remarks for Author): Ohnishi et al. have investigated the expression pattern of Ldb2 in mice, effect of Ldb2 KO in mouse behaviours and Arc expression and identified direct interactors of LDB2 and potential indirect binding sites in DNA. They have postulated possible mechanisms of phenotypes seen in the KO mice and the action of LDB2. This is the first study generating and studying Ldb2 KO mouse and in that sense, this study is novel and provides data that could benefit others. However, the link between the schizophrenia and the gene is very weak as it is based on one case although de novo, hence linking the findings of this study to mental disorders, as was stated in several places throughout the manuscript by the authors, is not appropriate and over-stretching. Another point that has to be corrected is the interpretation of their data and the building of hypothesis and model. The manuscript shows either no supporting evidence or rather opposite evidence to build their hypothesis, especially around Arc expression and mEPSCs. Also it has to be noted that mEPSC recording is not a measure of synaptic plasticity in this experimental setting. Often the statements were not directly supported by the results, therefore, either new experimental data needs to be added to support their statements or they need to be modified to reflect the data. Detailed comments are below. Introduction 1. Do patient derived tissues (any somatic tissue? Biopsy?) show reduced expression of LDB2? In order to support the breakpoint is affecting the expression of LDB2, this data is needed.
First result section: Ldb2 is expressed in neurons of restricted regions in the brain 2. Has the lab-made LDB2 antibody tested for cross reactivity for LDB1? 3. Fig 1D doesn't necessarily show that Ldb2 is expressed only in neuron. Remove referring of Fig.1D in the first result section claiming this. 4. Fig 1D doesn't show that layer V has most prominent expression. Layer II/III also has strong expression too. Reflect this in the result. Proper quantification is needed with together with NeuN staining in all layers. 5. Fig 1E doesn Second result section: Ldb2 KO mice display pleiotropic behavioral abnormalities related to mental disorders 7. In the first line "dysregulated expression of LDB2" should be lack of LDB2 as the model they use is deficiency rather than dysregulation. 8. "Although Ldb2 showed a layer-specific expression in the cerebral cortex," This was not investigated fully. The authors need to quantify the % of neurons expressing LDB2 in each cortical layer to support this. Even in the representative picture in Fig 1D, it doesn't seem it has specific expression in layer V.  (Fig 2B) was by chance rather than consistent observation. 15. Fig4A Haloperidol graph has unnecessary stars above WT saline. Remove them. 16. Was the locomotion between WT and KO after saline injection significantly different? There is no significance recorded in the figure so it is assumed that were not different. Then again as commented above (14), the increased locomotion in open field test (Fig 2B) was by chance rather than consistent observation. 17. "Importantly, deficits in fear-conditioning test as well as the hyperlocomotive trait of the KO mice were ameliorated by the treatment with typical (haloperidol) and atypical (clozapine) antipsychotics (Fig. 4A, B)." This sentence is not correct. Fig 4B Haloperidol (Fig 2, 3). 19 22. "that Ldb2 of excitatory neurons suppressed the number of the Arc gene-expressing neurons in response to the tone-shock paired or unpaired stimuli in the LA." This is not correct. Fig5B LA graph shows a significant increase of Venus+ cells (Arc expressing cells) in WT paired group. This means that Ldb2 in excitatory neurons did not suppress the expression Arc. If that was the case, then the WT paired group would show no increased in venus + cells in comparison to naïve and unpaired WT groups. The conclusion from Fig5B LA graph needs re-writing to accurately reflect the result. Suggestion: Lack of Ldb2 expression in excitatory neurons in LA of amygdala promoted the expression of Arc upon stimuli during fear conditioning trials. 23. "These results indicate that Ldb2 suppresses the Arc--gene transcription responding to the tone--shock paired stimuli in the LA neurons." This statement is incorrect as well. See comment 2 above. 24. "(1) the Ldb2 protein in LA neurons suppresses Arc-gene expression during fear conditioning," is incorrect. It should be that lack of LDB2 protein in LA neurons increases Arc expression during fear conditioning. 25. Is there any functional differences between cells in LA and BAI? Arc expression level in BAI is same between WT and KO. Are Arc expressing neurons in BAT not contributing to fear learning? 26. How do you interpret the increased LA Arc expression in unpaired group of KO? What do you think is initiating Arc expression in unpaired and paired group of WT and KO? This needs explanation. 27. "(2) the Ldb2 KO mice fail to induce appropriate plasticity of synapses due to Arc overexpression, consequently leading to impaired tone-related fear conditioning ability" this is not correct conclusion. The patch clamp experiment (looking at synaptic activity) didn't show any differences between WT and KO although there was Arc expression differences between WT and KO. The authors didn't do any experiment investigating synaptic plasticity and the mEPSC measured by patch clamp is not measuring synaptic plasticity. The conclusion drawn here and the following hypothesis (Fig7) is not correct hence needs modification. 28. In Fig7, there are "miniature EPSP ↓" &" Fear Learning ↓". This is not reflecting their findings as KO mice DID NOT show reduced mEPSCs but DID show reduced fear learning.
Fourth result section: Neurogenesis is reduced in the hippocampus of Ldb2 KO mice 29. Supple Fig 6 legend has an error in animal numbers. Correct "three animals for each genotype" to six animals for each genotype. Also report how many sections were quantified per animal. Representative pictures should also be included. 34. What is the explanation of other GO terms listed in Table 2 for the function of LDB2 and EGR? 35. "disruption of the 'LDB2-EGR axis' has a role in the etiology of mental disorders" The results shown in the study does not directly support the etiology of mental disorders. The LDB2-EGR axis is a possible mechanism under the phenotypes seen in LDB2 KO mice, NOT mental disorders. This needs to be made clear.

Discussion
36. "LDB and LDB2-mediated gene expression networks have an important role in the pathogenesis of schizophrenia and other mental disorders, including bipolar disorder" The current study did not show any direct evidence of this, hence should be modified to accurately describe their findings. 37. Please explain what "the face and construct validities" in this context. It is not clearly explained. 38. "We also presented evidence that Arc/ARC-mediated synaptogenesis is one of downstream targets for LDB2/Ldb2 in the amygdala," authors did not show any data on synaptogenesis. This was their guess. This needs to be made clear. 39. "Our study revealed that the Ldb2 deficiency results in an excess of the Arc expression in excitatory neurons in the amygdala, probably causing the suppression of mEPSCs" Ldb2 KO mice had same frequency and amplitude of mEPSCs (Supple Fig5), therefore, the increased Arc expression is not affecting mEPSCs at all. This needs correction. In addition, the increased Arc expression in KO compared to WT was only observed in LA not BAI in amygdala. This needs to be made clear. 40. One of the significant finding was reduced neurogenesis in DG in hippocampus hence it needs to be discussed. Also explain how neurogenesis could have been affected in KO when LDB2 is not expressed in DG in WT mice.

Response to the Reviewers
Response to Referee #1:

Our response:
Accrording to the suggestion, we have added some behavioral tests and summarized them in Supplementary Table (Appendix Table S1). In particular, we conducted the 8-arm radial maze and Banes maze tests for the working memory assessment, in addition to Y-maze test. However, no signs of working memory deficits in the KO animal were detected. We also evaluated the social behavioral function of KO mice using the encounter method, but again no signs of social behavioral deficit were seen. Please pay attention to that we only claimed that Ldb2 KO mice 'at least partly' fulfill the behavioral or neuronal criteria as a schizophrenia or bipolar model. Although we agree that investigating the deficits in neuronal architecture in KO mice are critical, we believe that it should be assessed in the future study.

30th Nov 2020 1st Authors' Response to Reviewers
We have refered to Appendix Table S1 in the "Ldb2 KO mice display pleiotropic behavioral abnormalities related to mental disorders" section, as follows:

Revised)
To explore the possibility that dysregulated expression of LDB2 had an impact on the manifestation of schizophrenia, we next set out to analyze phenotypes of Ldb2-deficient mice (Fig 2,   Fig EV3, Appendix Table S1) employing our behavioral test batteries.

Discussion
Phenotypic traits of the mouse model need to be discussed in comparison to what found in bona fide mouse models.

Our response:
Considering the complexity of the pathogenic mechanism of schizophrenia, we think it difficult to say which are the bona fide mouse models for the illness. However, as explained in the Introduction section, DISC1 is an established causal gene. Therefore, we have added statements where the behavioral phenotypes of Ldb2 KO mice were compared to those of Disc1-modiffied animals, to the Discussion section as follows: We have corrected the typo.
The authors present an extensive modeling study using both mouse and iPSC-derived neurospheres to examine the potential for a novel risk gene for schizophrenia. An alternative cellular model would be iPSC-derived neurons as well as iPSC-derived neural cells from the patient.

Our response:
Thank you for the comment. Ldb2 expression is prominet in neurospheres. Neurospheres are enrichecd in neural progenitors. As for iPS-dedived neurons, it was very diffifult for us to ensure sufficient cell number for ChIP analysis.

Comments
1. This manuscript reports on a tour-de-force effort to use both animal and human cellular models to examine various functional aspects of LDB2, a gene located close to the breakpoint on 4p16, a part of a very rare chromosomal structural aberration in a schizophrenia patient. The findings presented are important and additional supportive future data would heighten the relevance of the LDB2-EGR system to risk for schizophrenia.

Our response:
We would like to express our thanks to the positive comments.
Minor comments 2. If the parents of the proband were available, were they unaffected and did any of them carry the translocation?

Our response:
The proband's parents were unaffected and did not carry the chromomal abnormatlity, supporting the idea that the translocation was causal in the proband. Please find that these observations are explained in the Introduction section.  Due to the unavailability of genomic DNAs from the proband's parents, we were not able to determine whether these deletions were de novo or inherited from one of the proband's parents, who manifested no psychiatric symptoms. Although this makes it difficult to evaluate the pathogenic role of these deletions in the proband, no known genes are mapped to these two deleted regions.
Moreover, to our best knowledge, no reports have identified these regions as susceptibility loci for schizophrenia. Collectively, these observations do not favor to the idea these two deletions were pathogenic in the proband. We have reflected these observations in the Introduction section (2nd paragraph) as follws: By reanalyzing the NGS data (Horiuchi et al.) carefully, we found that the proband's genome harbors large deletions in two chromosomes: chromosome 6 (10,657,982-10,660,876; 2,885 bp) and chromosome X (55,702,373-55,709,904; 7,447 bp). Due to unavailability of genomic DNAs from the proband's patients, we were not able to determine whether these deletions were de novo or inherited from one of the proband's parents, who manifested no psychiatric symptoms. However, no known genes are mapped to these two deleted regions. Moreover, to our best knowledge, no reports have identified these regions as susceptibility loci for schizophrenia or bipolar disorder. 6. Did the authors try to sequence the 2 bands on Western blot?

Our response:
Although we have not conducted amino acid sequencing of the two proteins, our monoclonal and polyclonal anitibodies, both of which were directed against the N-terminal potion of mouse Ldb2, commonly reacted to the two proteins. In addtion, the two bands disappeared in western blot when brain lyasate from a KO animal was used. Based on these observations, we concluded that the two bands represent the Ldb2 gene products. 7. Why did the authors use iPSC-derived neurospheres instead of either iPSC-derived neural progenitor cells or neurons?

Our response:
The reason is that Ldb2 expression is prominet in neurospheres. Neurospheres are enrichecd in neural progenitors. As for iPS-dedived neurons, it was very diffifult for us to ensure suffiecient cell number for ChIP analysis.
8. It would have been interesting also to use iPSC-derived neural cells from the patient.

Our response:
We agree with the comment. We would like to analyze cellular phenotypes of the proband iPSCderived neural cells in the fuure study. In the 1 st paragraph of Discussion section, we have mentioned this as follws: Revised) In the future study, the cellular phenotypes of the these cells should be addressed.
9. Y-2-H tends to have problems with signal-to-noise ratio, any comments about this issue?

Our response:
As the reviewer suggested, Y-2-H sometimes produces substantial noise signals. As shown in the original version, we confirmed the protein-protein interactions using coimmunoprecipitation assay.
To clarify this point, we have amended the Result section as follows: Original) ......translin-associated factor X (TSNAX), pruvate kinase, muscle-type (PKM2), and CD9 and/or sulfide quinone oxidoreductase (SQOR), were also identified ( Table 1). Next, we performed co-immunoprecipitation assay by focusing on the proteins that potentially act as transcription factors or transcription regulatory factors. Revised) ......translin-associated factor X (TSNAX), pruvate kinase, muscle-type (PKM2), and CD9 and/or sulfide quinone oxidoreductase (SQOR), were also identified ( Table 1). Since two-hybrid screening is known to sometimes produce false-positive clones, we validated the specificities of the poteintial interactions with co-immunoprecipitation assay, especially focusing on the proteins that potentially act as transcription factors or transcription regulatory factors.
10. There are too many numbers in pages 12-14 that are distracting. Presenting these numbers in Tables would help.

Our response:
According to this comment, we have moved some of numbers to figure legends. 11. Fig. 1A: Replace "dimeraization" with "dimerization"

Our response:
We have corrected it.
12. Fig. 1D, F & G: An improvement in color intensity would help.

Our response:
We have created new Fig.1 with improved color contrast. 13. Fig. 2: Indicate in the legend what blue and red lines represent.

Our response:
We have added information required to the legends. 14. Fig. 4A: Replace "venral" with "ventral"

Our response:
We have corrected it in new Fig. 6.

Our response:
We have modified them as follows:   Fig. 7: Include a legend.

Our response:
We have decided to remove this figure because we found no significant difference in mEPSC parameters.
17. Fig. 8: The legend can be made more descriptive. Fig. 8F legend is missing.

Our response:
We have amended it as follows: Results of coimmunoprecipitation between overexpressed LDB2 and interaction candidates are Use of KO mice to study the function of the gene (LDB2) is appropriate. However, to insist that the model is a mental disorder model is not appropriate. They need to modify their statement in the manuscript.

Our response:
Thank you for the comments. We have amended some parts of the manuscript as explained in the later parts of our response. . This is the first study generating and studying Ldb2 KO mouse and in that sense, this study is novel and provides data that could benefit others. However, the link between the schizophrenia and the gene is very weak as it is based on one case although de novo, hence linking the findings of this study to mental disorders, as was stated in several places throughout the manuscript by the authors, is not appropriate and over-stretching.
Another point that has to be corrected is the interpretation of their data and the building of hypothesis and model. The manuscript shows either no supporting evidence or rather opposite evidence to build their hypothesis, especially around Arc expression and mEPSCs. Also it has to be noted that mEPSC recording is not a measure of synaptic plasticity in this experimental setting.
Often the statements were not directly supported by the results, therefore, either new experimental data needs to be added to support their statements or they need to be modified to reflect the data.
Our response: Thank you for the insightful comments. We have rewritten some parts of the manuscript as explained in the later parts of our response. As for the analysis of mEPSCs and its interpretation, we have rewritten the manuscript thoroughly.  2. Has the lab-made LDB2 antibody tested for cross reactivity for LDB1?

Our response:
We have carefully designed the antigen to avoid crossreactivity of the antibody to Ldb1/LDB1 (Fig.   1A, B). In fact, the antigen region was the most divergent between the two proteins. In addtion, we have experimentally confirmed the specificity of our polyclonal antibody. There were essentially no bands in the brain tissue lysate from Ldb2 KO mice (Fig. EV3C), supporting that the polyclonal antibody is specific to Ldb2. As for our monoloconal antibody, we have confirmed the specificity using tissue sections form Ldb2 KO mice (Fig. EV3F).
3. Fig 1D doesn't necessarily show that Ldb2 is expressed only in neuron. Remove referring of

Our response to #3-6:
We have conducted additional experiments and reflected the results to new Fig. 1 as follws: These data clearly show that > 80% of Ldb2-positive are NeuN-positive. In addition, we have presented the staining pattern of Ldb2 in the cerebral cortex, hippocampus and amygadala from KO mice, evidencing the speficity of the antibody we have created (Fig EV3F).
We have amended the result section to explain these data (second result section: Ldb2 KO mice display pleiotropic behavioral abnormalities relevant to mental disorders).
7. In the first line 'dysregulated expression of LDB2' should be lack of LDB2 as the model they use is deficiency rather than dysregulation.

Our response:
We have removed 'To explore the possibility that dysregulated expression of LDB2 had an impact on the manifestation of schizophrenia,' from the sentence.
8. "Although Ldb2 showed a layer-specific expression in the cerebral cortex," This was not investigated fully. The authors need to quantify the % of neurons expressing LDB2 in each cortical layer to support this. Even in the representative picture in Fig 1D, it doesn't seem it has specific expression in layer V.

Our response:
Please refer to our response to #3-6 and new Fig. 1.   9. In Fig. 2, write the number of mice used in each test either in the figure itself or in the figure legend. Same goes for the rest of figures.

Our response:
We have added the information.
10. Details of the behavioural test protocol needs to be written in the supplementary information or method.

Our response:
We have added information on the protocols for behavioral study in Supplementary Information. Fig 3A and B bar graphs in comparing WT/saline vs WT/MPD and WT/saline vs WT/MK-801.

Our response:
Those were not siginificant.

Our response:
We have corrected it.
13. Explain why the dose of MK-801 went up from 0.2 mg/kg in Fig 2B locomotion test into 0.5 mg/kg in 'popping' behaviour test. Did 0.2 mg/kg MK-801 not produce statistically significant result in popping behaviour? If so, this needs to be mentioned.

Our response:
Although we have conducted preliminary experiment in lower doses using a small number of animals, the effect of the drug was relatively weak, when compared to 0.5 mg/kg. Fig 2B open field test, KO mice are hyperactive. However, in Fig 3A and (Fig 2B) was by chance rather than consistent observation.

Our response:
In the home cage test, the activities in the day 1 were not different between KO and WT ( Fig. 2A).
However, the activities in the day 2 and thereafter were higher in KO than WT (Fig. 2A). Therefore, we believe that the KO mice intrinsically have a hyperactive trait and the data presented in Fig. 2B was not by chance. 15. Fig. 4A Haloperidol graph has unnecessary stars above WT saline. Remove them.

Our response:
We have removed the haloperidol data. Please refer to our response to #17. 16. Was the locomotion between WT and KO after saline injection significantly different? There is no significance recorded in the figure so it is assumed that were not different. Then again as commented above (14), the increased locomotion in open field test (Fig 2B) was by chance rather than consistent observation.

Our response:
In the open field test (Fig. 2B), hyperlocomotin was evident during the first 5 min. In contrast, culmitive locomotor activities in home cage in the day 1 were not different between KO and WT ( Fig. 4A). Therefore, the hyperlocomotin of the KO mice may be statistically masked when the locomotor activity is measured during more than 5 min in the novel place. We agree with the reviewer, in that a hyperlocomotive trait of KO mice is evident in some conditons. But the detailed mechanism remains elusive. 17. "Importantly, deficits in fear-conditioning test as well as the hyperlocomotive trait of the KO mice were ameliorated by the treatment with typical (haloperidol) and atypical (clozapine) antipsychotics (Fig. 4A, B)." This sentence is not correct. Fig 4B Haloperidol result doesn't show significant improvement in KO after haloperidol injection. Reflect this in the text. In addition, in both open field and fear conditioning test, both the WT and KO responded to drugs. Therefore, authors cannot justify the suitability of LDB2 KO as schizophrenia model based on this. If so, even the WT would've been considered as the disease model as they have responded to the drug.

Our response:
We were not able to determine the drug dose in which biological effects are seen in only KO animals. We have removed Fig. 4A and the haloperidol data from Fig. 4B, and revised the related description.
18. Add individual mouse data into the bar graphs in Fig 4 as was done in other figures (Fig 2, 3).

Our response:
We have modified the graphs.

Our response:
Please be aware that the basal activity in the light phase is much lesser than in the dark phase. In our opinion, this sometimes makes it diffucult to reproducibly detect relatively small diffrences. 20. For contextual test, there isn't haloperidol result. Why is that? Was it not performed or was it not significant?
Our response: It was not performed. We have removed all haloperidol result in the revised manuscript.
21. How many coronal sections were used for counting venus+ cells? Please specify.

Our response:
Thank you for the comment. We have now added the information in the text.

Our response:
Thank you for the comment. We have rewritten the text to reflect the result accurately.
23. "These results indicate that Ldb2 suppresses the Arc-gene transcription responding to the tone-shock paired stimuli in the LA neurons." This statement is incorrect as well. See comment 22 above.

Our response:
Thank you for the comment. We have corrected it.
24. "(1) the Ldb2 protein in LA neurons suppresses Arc-gene expression during fear conditioning," is incorrect. It should be that lack of LDB2 protein in LA neurons increases Arc expression during fear conditioning.

Our response:
We have rewritten the conclusion.

Our response:
Thank you for the constructive comment. The reason why Arc expression was increased in the unpaired group is unclear in the present study. One possibility is that, in the LA, excitability of excitatory neurons is lowered by Ldb2 and neurons will not fire unless the input of the foot shock (pain) and tone comes together. However, when the inputs of both pain and tone converge simultaneously to the neurons, they will fire and LTP occurs in those neurons, and subsequently, only the input of tone causes freezing behavior (cued-fear learning). On the other hand, the inhibitory mechanism is weakened in KO mice and the excitability of neurons is heightened nonspecifically, which makes it difficult to distinguish the sensory inputs and to establish the fear learning. In the BA, however, the inhibitory mechanism is weak from the beginning and neurons will fire nonspecifically in response to the input of tone or pain. Because this possibility is not supported by experimental results and, in that sense, too speculative, we do not describe it in the Discussion section. However, this issue is very important and we would like to pursue it in the future study.
27. "(2) the Ldb2 KO mice fail to induce appropriate plasticity of synapses due to Arc overexpression, consequently leading to impaired tone-related fear conditioning ability" this is not correct conclusion. The patch clamp experiment (looking at synaptic activity) didn't show any differences between WT and KO although there was Arc expression differences between WT and KO. The authors didn't do any experiment investigating synaptic plasticity and the mEPSC measured by patch clamp is not measuring synaptic plasticity. The conclusion drawn here and the following hypothesis (Fig7) is not correct hence needs modification.

Our response:
Thank you for the comment. We agree with the reviewer and have rewritten the conclusion completely. Although there was no difference in mEPSC parameters between genotypes, the neuron that expresses more Arc shows smaller median amplitudes and lower frequency of mEPSCs regardless of genotypes. Therefore, it can safely be concluded that "(2) the amount of cellular Arc expression is inversely correlated with spontaneous synaptic activity in the LA".
28. In Fig7, there are "miniature EPSP ↓" &" Fear Learning ↓". This is not reflecting their findings as KO mice DID NOT show reduced mEPSCs but DID show reduced fear learning.

Our response:
Thank you for the comment. We have deleted Figure 7 because there was no difference in mEPSC parameters between genotypes and because this figure is somewhat speculative.
Fourth result section: Neurogenesis is reduced in the hippocampus of Ldb2 KO mice 29. Supple Fig 6 legend has an error in animal numbers. Correct "three animals for each genotype" to six animals for each genotype. Also report how many sections were quantified per animal.
Representative pictures should also be included.

Our response:
As the reviewer pointed out, we have used 3 animals/genotype, and both hemispheres were analyzed, resulting in n = 6 (2 x 3). We believe that many studies including ours have conducted this type of replication when brain tissue sections were analyzed. Unfortunately, we have not stored the

Our response:
We have explained the cell lines in the sixth Resutl section, and we have added a following sentence to the Method section: ChIP-seq data of the EGR1 were obtained using cancer-derived cell lines, ECC1 (endometrial cancer), HCT116 (colon cancer) and K562 (erythromyeloblastoid leukemia), and using an ES cell line H1ES from the NCBI SRR.
33. For the GO analysis, what was the background gene set? Use of correct background gene set is important as this will affect the enrichment result and interpretation of the result. There is no information on how GO analysis was performed in the method/supple method section.

Our response:
We are sorry for the confusion. As explained in the text, Appendix Table S1 lists gene names extracted from the common ChIP peaks between LDB2 and EGR1 in the ECC1 and K562 lines. The GO analysis was done using this gene set by comaring to all genes annotated. We have modified the descriptin in the result section as follows: Original) .... in the ECC1 and K562 lines were enriched in the upstream regions or 5' UTRs of target genes (Fig. 9E). A Gene Ontology analysis (http://geneontology.org) using the set of genes with overlapping ChIP-seq signals between the LDB2 and EGR1 (Supplementary Table 1), Revised) .... in the ECC1 and K562 lines were enriched in the upstream regions or 5' UTRs of target genes (Fig. 8E, Appendix Table S2). A Gene Ontology analysis (http://geneontology.org) using the genes listed in Appendix Table S1. Table 2 for the function of LDB2 and EGR?

Our response:
Currently, we have no reliable information on how other biological processes can be related to the phenotypes of Ldb2 KO mice in terms of the pathophysiology of schizophrenia.
35. "disruption of the 'LDB2-EGR axis' has a role in the etiology of mental disorders" The results shown in the study does not directly support the etiology of mental disorders. The LDB2-EGR axis is a possible mechanism under the phenotypes seen in Ldb2 KO mice, NOT mental disorders. This needs to be made clear.

Our response:
We have weaken our claim as follws: Revised) we speculate that LDB2/Ldb2 cooperates with the EGR family via unknown protein(s) to control expressions of numerous genes including ARC (Fig. 10), and that 'LDB2-EGR axis' may potentially underlie the etiology of subset of mental disorders. Discussion 36. "LDB and LDB2-mediated gene expression networks have an important role in the pathogenesis of schizophrenia and other mental disorders, including bipolar disorder" The current study did not show any direct evidence of this, hence should be modified to accurately describe their findings.

Our response:
We have modified the sentence as follws: Original) The current study provided multiple lines of experimental evidence that LDB and LDB2mediated gene expression networks have an important role in the pathogenesis of schizophrenia and other mental disorders, including bipolar disorder.

Revised)
Multiple observations in the current study favor to the idea that LDB and LDB2-mediated gene expression networks play a potential role in the pathogenesis of subset of schizophrenia and other mental disorders, including bipolar disorder.

37
. Please explain what "the face and construct validities" in this context. It is not clearly explained.

Our response:
We have explained them as follows: face validity: a similarity exists between behavioral phenotypes of a model animal and disease symptoms construct validity: a model animal can be generated based on a mechanistic theory of disease 38. "We also presented evidence that Arc/ARC-mediated synaptogenesis is one of downstream targets for LDB2/Ldb2 in the amygdala," authors did not show any data on synaptogenesis. This was their guess. This needs to be made clear.

Our response:
We have change the text as follows:

Original)
We also presented evidence that Arc/ARC-mediated synaptogenesis is one of downstream targets for LDB2/Ldb2 in the amygdala, Revised) Our ChIP-seq analyses suggested that binding sites of LDB2 and EGR proteins in human cells are concentrated to synaptogenesis-related genes including ARC. The expression of Arc was found to be modulated in the amygdala of Ldb2 KO mice. Therefore, it is tempting to speculate that synaptogenesis is one of the downstream targets for LDB2/Ldb2 in the amygdala, 39. "Our study revealed that the Ldb2 deficiency results in an excess of the Arc expression in excitatory neurons in the amygdala, probably causing the suppression of mEPSCs" Ldb2 KO mice had same frequency and amplitude of mEPSCs (Supple Fig5), therefore, the increased Arc expression is not affecting mEPSCs at all. This needs correction. In addition, the increased Arc expression in KO compared to WT was only observed in LA not BAI in amygdala. This needs to be made clear.

Our response:
Thank you for the comment. We agree with the reviewer. We have rewritten the statement as follows: "Our study revealed that the amount of cellular Arc expression is negatively correlated with synaptic activity in the LA, and that Ldb2 deficiency results in an excess of the Arc expression only in the LA, which probably causes the disturbance of auditory fear learning".
-We have now included the words "only in the LA" to emphasize that the increase of Arc expression was restricted to the LA. 40. One of the significant finding was reduced neurogenesis in DG in hippocampus hence it needs to be discussed. Also explain how neurogenesis could have been affected in KO when LDB2 is not expressed in DG in WT mice.

Our response:
Thank you for raising this critical point. We have currently no idea on the mechanism of reduced neurogenesis. For the readers' understanding, we have restructured the result section, pointing out that no clear expression of the Ldb2 protein was seen in the DG as follows: Revised) Neurogenesis in the hippocampus has a role in this learning. In addition, impairments in adult neurogenesis have been repeatedly reported in schizophrenia, affective disorders, and their model animals. We found that the neurogenesis in hippocampal dentate gyrus (DG) was remarkably decreased in the KO mice, although no clear expression of the Ldb2 protein was seen in the DG. Do the data meet the assumptions of the tests (e.g., normal distribution)? Describe any methods used to assess it.
Is there an estimate of variation within each group of data?

Reporting Checklist For Life Sciences Articles (Rev. June 2017)
This checklist is used to ensure good reporting standards and to improve the reproducibility of published results. These guidelines are consistent with the Principles and Guidelines for Reporting Preclinical Research issued by the NIH in 2014. Please follow the journal's authorship guidelines in preparing your manuscript.

B-Statistics and general methods
the assay(s) and method(s) used to carry out the reported observations and measurements an explicit mention of the biological and chemical entity(ies) that are being measured. an explicit mention of the biological and chemical entity(ies) that are altered/varied/perturbed in a controlled manner. a statement of how many times the experiment shown was independently replicated in the laboratory.
Any descriptions too long for the figure legend should be included in the methods section and/or with the source data.
In the pink boxes below, please ensure that the answers to the following questions are reported in the manuscript itself. Every question should be answered. If the question is not relevant to your research, please write NA (non applicable). We encourage you to include a specific subsection in the methods section for statistics, reagents, animal models and human subjects.

definitions of statistical methods and measures:
a description of the sample collection allowing the reader to understand whether the samples represent technical or biological replicates (including how many animals, litters, cultures, etc.).

The data shown in figures should satisfy the following conditions:
Source Data should be included to report the data underlying graphs. Please follow the guidelines set out in the author ship guidelines on Data Presentation.
Please fill out these boxes ê (Do not worry if you cannot see all your text once you press return) a specification of the experimental system investigated (eg cell line, species name).
At least 10 mice were used for behavioral analyses.
graphs include clearly labeled error bars for independent experiments and sample sizes. Unless justified, error bars should not be shown for technical replicates. if n< 5, the individual data points from each experiment should be plotted and any statistical test employed should be justified the exact sample size (n) for each experimental group/condition, given as a number, not a range; Each figure caption should contain the following information, for each panel where they are relevant:

Captions
At least 10 mice were used for behavioral analyses.
Not excluded.
The data anlayses were done in a genotype-blind format when possible No randamization was done.
When normal distribution was not seen, non-parametric tests have been performed.

Yes.
No randomization was used.
The data anlayses were done in a genotype-blind format when possible.
The data anlayses were done in a genotype-blind format when possible.

Data
the data were obtained and processed according to the field's best practice and are presented to reflect the results of the experiments in an accurate and unbiased manner. figure panels include only data points, measurements or observations that can be compared to each other in a scientifically meaningful way.