Genetic dissection identifies Necdin as a driver gene in a mouse model of paternal 15q duplications

Maternally inherited duplication of chromosome 15q11-q13 (Dup15q) is a pathogenic copy number variation (CNV) associated with autism spectrum disorder (ASD). Recently, paternally derived duplication has also been shown to contribute to the development of ASD. The molecular mechanism underlying paternal Dup15q remains unclear. Here, we conduct genetic and overexpression-based screening and identify Necdin (Ndn) as a driver gene for paternal Dup15q resulting in the development of ASD-like phenotypes in mice. An excess amount of Ndn results in enhanced spine formation and density as well as hyperexcitability of cortical pyramidal neurons. We generate 15q dupΔNdn mice with a normalized copy number of Ndn by excising its one copy from Dup15q mice using a CRISPR-Cas9 system. 15q dupΔNdn mice do not show ASD-like phenotypes and show dendritic spine dynamics and cortical excitatory-inhibitory balance similar to wild type animals. Our study provides an insight into the role of Ndn in paternal 15q duplication and a mouse model of paternal Dup15q syndrome.

Remarks to the Author: In this manuscript Tamada et al report new findings from analysis of a new line of mutant mouse model homologues to smaller region from AS imprinting center and UBE3A within the human 15q11-q13 region and other related studies. They conclude that the paternally expressed Ndn is a driver gene for the neurodevelopmental defect reported in the human paternal 15q11-q13 duplication syndrome.
Maternal duplication of human chromosome 15q11-q13 region (Dup15q) has been strongly implicated in autism spectrum disorder (ASD). While the pathogenic role of paternally derived duplication remains uncertain. A recent genetic study has implicated a significant contribution of paternal duplication to ASD/DD/MCA with penetrance rate of ~20%. While the total number of subjects analyzed is large in this report but the number of paternal Dup15q case is still relatively small (~10). So the conclusion regarding the role of paternal Dup15q remains somewhat questionable but is stronger than previous data in literature.
The same group has reported the mutant mice carrying the large 6Mb duplication that mimic the common human 15q11-q13 duplication before. Unexpected, the paternal but not maternal duplication display some autism-like behaviors. It remains unclear why maternal duplication does not contribute any significant defect in mice. In this study, the author has produced a new line of smaller duplication (upstream of AS-IC and downstream of Ube3a within the homologous region of 15q11-q13. The paternal duplication of this smaller region did not result in the similar behavioral or cellular phenotypes of that observed in the large 6Mb duplication. It is noted that the same analysis of the small maternal duplication was not included in manuscript and could be valuable. Because of these results, the focus of the study was then turned to the paternally expressed genes proximal to the small duplicated region. A series of experiments including in vivo rescue has been conducted. The results provided the support that Ndn is the critical gene to responsible for the autism phenotypes of the large 6 Mb paternal duplication in mice. Overall, authors should be complimented on the large number of experiments that are both in vitro and in vivo and at cellular, functional and behavioral levels. The conclusion is reasonably convincing based on the results presented. However, several issues may be worthwhile to discuss in revision Major concerns: 1. The reference #17 is cited as the major evidence supporting the pathogenicity of the paternal duplication of 15q11-q13. The large sample size was repeatedly mentioned. This statement may be misleading. The large population was screened for the 15q11-q13 CNV but the total number of cases for genotype and phenotype study of the paternal duplication is ~10 and small. In addition, the penetrance of paternal duplication is ~20% which indicate that many cases carrying the paternal duplications are not affected. This caveats of this report should be discussed in detail.

1
We thank the reviewers for the careful review of our manuscript and for many useful suggestions to improve it. Our responses to the points raised are as follows:

<Reviewer #1>
This manuscript describes a really nice series of studies that point to over-dosage of the imprinted gene Necdin being a key contributor to the phenotype associated with paternal 15q duplication. The authors have previously characterised a mouse model of paternal 15q duplication, and here they use a number of approaches to narrow down the key genes involved. Firstly, they generate a second 15q duplication mouse model (1.5Mb 15q dup) that has a smaller duplicated region and show that this model does not display many of the phenotypes seen in the original, larger 15q duplication model. This suggest that the genes (Necdin,Mkrn3 and Magel2)  We appreciate the reviewer's positive comments. We understand the confusion of the naming of a mouse. We have made the new Supplementary Fig. 1, which indicates details of the duplicated region with 1.5 Mb, and we hope this new figure helps the reader's understanding. We have also included additional explanations to be made clearer in the revised manuscript. For 15q dupΔNdn mice, we modified the sentence to be clearer as follows: Line 246 (revised manuscript) "Therefore, we examined whether the normalization of the genomic copy number of Ndn in 15q dup mice is sufficient to restore the abnormal phenotypes" => "Therefore, we examined whether removing a single copy of Ndn from original 15q dup mice is sufficient to restore the abnormal phenotypes" I was also curious as to why the authors did not show the expression levels of Necdin is altered following electroporation? How can they be sure that the over-expression of Necdin is equivalent to that of the other genes they screened (Snrpn,Magel2 and Mkrn3)?
Afterall this could explain the difference in degree of dendritic spine phenotype change between the different genes ( Figure 3).
According to the reviewer's comments, we quantified the expression level of Ndn both in transfected and non-transfected cells induced by in utero electroporation followed by immunohistochemistry (new Supplementary Fig. 4). We observed an about 3-fold increased level in transfected cells than non-transfected cells. Since the reliable antibodies for MAGEL2, SNRPN, or MKRN3 were not available, we could not quantify them accurately. However, we thought the expression levels of these genes are similar because the other plasmids have the same promoter and vector.
Based on this result, we corrected the manuscript as follows:

Result
Line 183 (revised manuscript) "The overexpression level of Ndn was verified by immunohistochemistory, and it showed about 3-fold increase compared to non-transfected neurons ( Supplementary Fig. 4)."

Material and Methods
Line 658 (revised manuscript) "Briefly, E15.5 pregnant C57BL/6J mice were anesthetized by inhalation of isoflurane" => "Briefly, E15.5 pregnant C57BL/6J mice or ICR (only for Supplementary Fig. 4) were anesthetized by inhalation of isoflurane" Line 819 (revised manuscript) "For VGAT/VGLUT analysis (Fig. 5f)," Line 822 (revised manuscript) "For quantification of the NDN expression level ( Supplementary Fig. 4), the same confocal microscope with a 100x objective lens (digital zoom x 1.5) were used, and Z-stacked images were obtained with 0.5 μm step (1024 x 1024 pixels). Then, the Z-projected images were obtained with maximum intensity. By seeing the GFP and DAPI signal, the transfected or non-transfected nuclei were manually selected (N = 20 for each). Mean intensity/pixel of NDN in the nucleus was used to compare transfected and non-transfected cells."

Minor issues:
Italicise gene names throughout We corrected them throughout the manuscript. According to the reviewer's comment, we corrected as below: Line 44 (revised manuscript) "However, a recent genetic study with a large sample size of Dup15q has implicated a significant…" => "However, a recent genetic study has implicated a significant…" Line 87 (revised manuscript) "The large number of…" => "Most" L. 90: Consider removing 'To the contrary' and beginning the sentence with 'However,'.
The study cited was not contrary to the idea that maternal duplication is pathogenic (indeed this study still found that maternal 15q duplication was more penetrant), but as it was a much larger study it DID show that paternal duplications can be pathogenic for ASD (this is the point the authors are trying to make).
We agree with the reviewer's comment. We corrected as below: Line 91 (revised manuscript) "To the contrary, however, an investigation with a large sample size found individuals with paternally derived duplication also met the criteria for ASD." => " In contrast, an investigation found individuals with paternally derived duplication also met the criteria for ASD, although its penetrance was estimated at ~20% and the number of cases is still small 19 ."

<Reviewer #2>
This manuscript reports the finding that Ndn contributes to ASD phenotypes and neuronal morphology deficits in pat Dup15q mice. Although the authors did a lot of work , following a strong rationale, the conclusions are somewhat overstated and the relevance to human disease is incredibly inflated. Overall, I worry that this paper will muddy the waters in a field that desperately needs to find the real genetic contributors to Dup15q syndrome, and this paper has little, if any, relevance to that, despite the strong statements and the title that would lead readers to think otherwise. My specific criticisms are detailed below.
We appreciate the reviewer's important suggestions. Although, in our original manuscript, we emphasized that Ube3a is a primary gene for Dup15q considering from human genetics studies, we have edited the whole manuscript, including the title, and also added "Limitations" in the Discussion section (revised manuscript line: 468-488) to avoid overstatement of our conclusion in our revised manuscript.
Title "Genetic dissection identifies Necdin as a driver gene in 15q duplication syndrome" => "Genetic dissection identifies Necdin as a driver gene in a mouse model of 15q duplication syndrome" Major issues: 1. The premise that the paternally expressed genes may contribute to the pathogenesis of the Dup15q syndrome is false. Dup15q syndrome is overwhelmingly caused by maternal duplications of 15q11-q13. Some individuals with maternal duplications inherited their CNV from an overtly neurotypical mother with the same CNV on her paternal allele.
Individuals with paternal 15q11-q13 duplications while perhaps not fully neurotypical have a different disorder than Dup15q syndrome (see Cook et al., 1997). Even in the Isles et al., 2016 paper referenced as evidence for paternal duplications being involved in ASD, out of 10 individuals with paternally-inherited duplications, none have ASD (see suppl. Table 1).
We again agree that Dup15q syndrome is predominantly caused by maternal duplication. Indeed, we emphasized a maternal duplication is a dominant form in Dup15q syndrome repeatedly throughout our original manuscript. According to the reviewer's comment, we overall corrected the expression of the manuscript to be more precise and add a "Limitation" section (revised manuscript line: 468-488). In the article of Isles et al., we found 2 individuals with ASD and 1 "Autistic traits" out of 10 paternal duplications (suppl.    We added "construct" and " with overexpression of a deletion construct "

Figure legend
Line 1233 (revised manuscript) We added, "each construct of".

Overexpression studies never quantify the amount of overexpression of Ndn per neuron.
Transfection often leads to grossly overexpressed gene products and this is not the same as an additional single genetic copy with its native regulatory elements.
Thank you for this suggestion. We quantified the expression level of Ndn induced by in utero electroporation (new Supplementary Fig. 4). As responded to reviewer 1, we observed an about 3-fold increased level in transfected cells than non-transfected cells. Since the reliable antibodies for MAGEL2, SNRPN, or MKRN3 were not available, we could not quantify them accurately. However, we thought the expression levels of these genes are similar because the other plasmids have the same promoter and vector. Regarding the reviewer's concern ("Transfection often leads to grossly overexpressed gene products and this is not the same as an additional single genetic copy with its native regulatory elements."), this is why we prepared and analyzed 15q dup∆Ndn mice. We had already described in the original manuscript as follows:

Line 245 (revised manuscript) and line 242 (original manuscript)
" It is possible that the overexpression of Ndn achieved using in utero electroporation with a strong artificial promoter may not reflect physiological conditions. Therefore, we examined whether the normalization of the genomic copy number of Ndn in 15q dup mice is sufficient to restore the abnormal phenotypes"

<Reviewer #3>
In this manuscript Tamada Fig. 3).  We thank this helpful comment. As the reviewers suggested, the study (Isles et al.,

2016) provided screening of a large number of people but not a number of cases. We have corrected this issue in the Introduction as follows:
Line 91 (revised manuscript) "To the contrary, however, an investigation with a large sample size found individuals with paternally derived duplication also met the criteria for ASD." => "In contrast, an investigation found … met the criteria for ASD, although its penetrance was estimated at ~20% and the number of cases is still small."

In figure 1 D, the level of the increased expression is Ube3a is less than other
paternally expressed genes but it is statistically significant based on the data presented.
In addition, because the maternal allelic expression of Ube3a is cell type specific, the total RNAs from the brain tissue with mixed cell types may underestimate the difference too.
According to the reviewer's comment, we examined the expression level of Ube3a in primary hippocampal neurons to avoid mixed cell types. The substantial increase of Ube3a expression in brain tissue was not seen in primary hippocampal neurons (new Fig. 1e). Therefore, as the reviewer suggested, the increased Ube3a expression in 1.5Mb patDp mice is probably due to the expression in glial cells.

A similar restriction enzyme map that illustrates the genomic Southern analysis of 1.5
Mb duplication mice described in Figure 1

may help reader and future reference
According to the reviewer's suggestion, we prepared the schema of a genomic map for generating 1.5 Mb duplication mice (new Supplementary Fig. 1). "decreased social interaction"

The interpretation and statement of E/I imbalance in text may be a bit premature because only the function of pyramidal neurons was measured and the E/I imbalance should be a net effect of E/I either at single cell or network level.
We agree with the reviewer's opinion that the E/I imbalance, often found in ASDs, is a net effect of excitatory and inhibitory neurons. We investigated E/I balance not only by electrophysiology but also by immunohistochemistry with anti-VGAT and VGLUT antibodies (Fig. 5f-h). As the reviewer indicated, our description regarding the overexpression of Ndn in our original manuscript was improper. In this revised manuscript, we avoided using "E/I balance" when we analyzed only pyramidal neurons by electrophysiology as below. Unless the phenotype is pathognomonic for human diseases, these evidence is not direct.
Thank you for the reviewer's comment. As we replied to the comment of reviewer #2-8, we have corrected the indicated points to be more precise and also added a Although this issue is not our goal in this study and currently we do not have direct evidence, we propose the following hypothesis described in a "Limitations" section in this revised manuscript (revised manuscript line: 468-488).
Our study includes several concerns as follows. This study is limited to the animal. Therefore, these results may not be direct evidence for explaining Dup15q etiology.
There seems to be a different allelic contribution rate between humans and mice. Line 53-"generate paternal15q dupDNdn mice with a normalized copy number of Ndn by excising its one copy from paternal dup15q mice using…"

Matthew
We added "paternal" in the sentences as the reviewer indicated and throughout the manuscript as following points, L73: multi-dimensional abnormalities in 15q dup mice.
L94: excluding the target gene from 15q dup mice.
=> excluding the target gene from paternal 15q dup mice.
L439: is the decreased 5-HT seen in 15q dup mice.
=> is the decreased 5-HT seen in paternal 15q dup mice.
L450: developing cortex is altered in 15q dup and recovered in 15q dupΔNdn mice.
=> developing cortex is altered in paternal 15q dup and recovered in paternal 15q dupΔNdn mice.

Minor text edits for accuracy as follows:
Line 37 -Necdin regulates formation and maturation of cortical dendritic spines. Add "in mice" Line 45 -"…study has implicated a significant contribution of paternal duplication to ASD." Remove "significant", as it's not clear how significant this is.
Line 46-"etiology of maternally derived duplication is known to be UBE3A because this gene is…". Change "known" to "thought", unfortunately, this isn't known, yet.
Line 469-" Our study includes several concerns as follows. This study is limited to the animal. Therefore, these results may not be direct evidence for explaining Dup15q etiology. There seems to be a different allelic contribution rate between humans and mice." " "Our study has some important limitations. First, this study was carried out in mouse models, and there seems to be different contributions from the two parental alleles to the phenotypes observed in humans and mice. Therefore, these results may not be directly relevant for explaining Dup15q etiology in humans." We corrected above 4 comments as the reviewer indicated.