Combined deletions of IHH and NHEJ1 cause chondrodystrophy and embryonic lethality in the Creeper chicken

The Creeper (Cp) chicken is characterized by chondrodystrophy in Cp/+ heterozygotes and embryonic lethality in Cp/Cp homozygotes. However, the genes underlying the phenotypes have not been fully known. Here, we show that a 25 kb deletion on chromosome 7, which contains the Indian hedgehog (IHH) and non-homologous end-joining factor 1 (NHEJ1) genes, is responsible for the Cp trait in Japanese bantam chickens. IHH is essential for chondrocyte maturation and is downregulated in the Cp/+ embryos and completely lost in the Cp/Cp embryos. This indicates that chondrodystrophy is caused by the loss of IHH and that chondrocyte maturation is delayed in Cp/+ heterozygotes. The Cp/Cp homozygotes exhibit impaired DNA double-strand break (DSB) repair due to the loss of NHEJ1, resulting in DSB accumulation in the vascular and nervous systems, which leads to apoptosis and early embryonic death.

This fascinating study by Kinoshita et al describes elegant characterisation and functional validation of a genetic deletion involving both the Ihh and NHEJ1 genes in an unconventional model, the creeper chicken. The data generated are generally novel, of high standard and will be of interest to a broad audience ranging including those studying chondrodystrophy, developmental biologists, and those studying genomic stability. However a few major concerns need to be addressed: 1) The second result section claims "Osteoblast differentiation is inhibited in the Cp mutant." The data provided in this section relates to abnormal shape and size of various long bones and the lack of mineralised tissue formation. None of this directly assesses osteoblast differentiation, for example by detecting transcripts of osteoblast differentiation markers or differentiating osteoblasts to form mineralised nodules in vitro. Failure of osteoblast proliferation or activity could equally explain these phenotypes. Osteoblast differentiation should be directly assessed.
2) The conclusion in the same section that "chondrocyte proliferation and subsequent differentiation are completely inhibited" (repeated in the discussion) is similarly not supported by direct evidence, for example histological analysis of chondrocyte proliferation. The authors subsequently show absence of the hypertrophic chondrocyte marker COL10, so an amendment of this conclusion may be sufficient.
3) Localisation of IHH and NHEJ1 in normal embryos shown in Figure 3 is exceedingly week. The NHEJ1 sense and anti-sense probes appear to have produced comparable staining in the neural tube in the images provided. These localisations need to be improved or replaced with protein-level analyses. 4) Interpretation of nick-end labelling (TUNEL assay) indicating double strand breaks is confusing as this assay is primarily used to detect apoptotic cells. It is very surprising that no apoptotic cells were observed in the +/+ embryos given this process is common during development (e.g. over the closed neural tube). γH2AX staining of cp/cp versus +/+ embryonic tissue is needed to confirm that the increase in TUNEL staining is not simply due to an increase in apoptosis in dyeing embryos. 5) Quantification and statistical comparison of imaging data (e.g. % TUNEL positive cells within rescue and control regions in Figure 4C, γH2AX staining in Figure 5g, etc) is expected throughout to demonstrate reproducibility and variability.
-Please make sure all abbreviations and gene names are full explained (e.g. GSP, ALAD, etc). -Lines 219-223 of the discussion seem to assume chondrocyte to osteoblast trans-differentiation as the main mode of long bone ossification ("differentiation pathway from chondrocytes to osteoblasts"). Without having directly assayed this as well as the canonical mode of endochondral ossification through blood vessel ingression, this discussion should be removed. -A schematic comparison of the known NHEJ pathway similarities/differences between humans, mice and chickens would be helpful.
-Lines 285-286 claim the Cp model "highlights the importance of NHEJ function I normal neurogenesis". The production of neurones has not been assessed (even if reduced, apoptosis of the neuroepithelium before the onset of neurogenesis would be likely to limit neuron production). This conclusion should be removed.

Reviewer #2 (Remarks to the Author):
This is a lovely manuscript, well written and clearly presented that deserves to be published.
Below are the points to be addressed: Major 1) From this work, the conclusion that NHEJ1 deficiency "alone" causes embryonic lethality in the Creeper chicken can be suggested but not directly made. Strict demonstration would require a chicken with only the NHEJ1 defect.
Therefore the title of the manuscript should be "Combined deletions of IHH and NHEJ1… Minor 1) P 5 line 95. NHEF1…..NHEJ1 2) Would it be possible to perform clonogenic survival assays after DSB induction (IR or radiomimetic drugs) with the established primary cells to further support the NHEJ defect.
3) P 10 line 226 discovery of NHEJ1….it is a co-discovery by ref 27 and 28 4) P 10 Lines 236-237. The CTR of NHEJ1 contains the NLS but also the Ku binding motif. Recruitment and retention of XLF at DSBs in cells requires Ku interaction see and refer to the work of Charbonnier, Caldecott and Chen. The data in this manuscript clearly show that truncated XLF is not recruited to sites of damage but that it is still can be localized in the nucleus. 5) P 12 line 281. Remove "was" Reviewer #3 (Remarks to the Author): In this manuscript Kinoshita et al explore the genetic and developmental origin behind the Creeper chicken phenotype. They discover that the phenotype is linked to a double deletion/inversion impairing the functions of two genes: IHH and NHEJ1. While the implication of IHH in the phenotype is well described, the implication of NEJ1 is rather new. Using expression analysis and functional experiments they gather arguments indicating that the loss of function of NEJ1 is causing the early mortality phenotype of homozygous creeper, a part of the phenotype that is not easily explained by the loss of IHH. This manuscript addresses a classical genetic problem: identifying genes causing a well-characterized phenotype using chicken as a model system. Because the work complements and rectifies nicely what has been published on the problem (i.e. the implication of IHH in the Creeper phenotype), this manuscript is of importance for the field and should be published in Communications Biology. However several experiments are lacking and there are some flaws in the manuscript. I will list them hoping that the authors will be able to address them to improve their work and be able to publish it: Main concerns: 1. The morphogenetic abnormalities of the early stages of Creeper embryos (title of the first part of the results) are not well characterized. That is an important piece of result that should be in the main figures. The only information are in Suppl Fig 3G where we can see a Tunnel staining in a mutant but not control embryos and a descriptive table: hypoplasia of heart and brain and abnormal island formation should be better documented (mutants versus control embryos).
2. The description of the deletions are very informative and important to this manuscript, however to be sure that the same deletions are causing the initial Creeper phenotype it will be informative to double check that these deletions exist in the JB and/or the Chinese strains too.
Minor concerns: 1. There is no clear rationale (in the main) text for choosing the GSP strains to study the mutation. 2. There is no explanation on why ALAD gene is being investigated and not another gene. 3. A bit of introduction on DSB would be welcome for the readers. 4. The panels of the figures are very small and are difficult to see (in particular Fig. 3 and Fig. 4). 5. Sentence starting at line 205 (discussion) is confusing: the deletion is likely the same in the Chinese strain (see main concern 2). Are they potentially different genetic mutation causing the Creeper phenotype? This is unclear to me.

Responses to the comments of the reviewers
Reviewer #1 (Remarks to the Author): 1) The second result section claims "Osteoblast differentiation is inhibited in the Cp mutant." The data provided in this section relates to abnormal shape and size of various long bones and the lack of mineralised tissue formation. None of this directly assesses osteoblast differentiation, for example by detecting transcripts of osteoblast differentiation markers or differentiating osteoblasts to form mineralised nodules in vitro. Failure of osteoblast proliferation or activity could equally explain these phenotypes. Osteoblast differentiation should be directly assessed.
Line 144 -158, Fig. 4a: In order to investigate osteoblast differentiation in the Cp mutant cells in vitro, we isolated calvarial cells of E15 embryos from the JB strain, cultured them for 16 days and then examined alkaline phosphatase activity, an osteoblast differentiation marker. The alkaline phosphatase activity was found in the wild-type (+/+) cells, whereas the activity was very weak in the Cp/Cp cells. Next, to confirm osteoblasts differentiation in vivo, we examined gene expression of a pre-osteoblast marker, RUNX2, and a mature osteoblast marker, osteopontin (OPN) in the hindlimb of the E15 embryos. RUNX2 and OPN were strongly expressed in the zeugopod region of the wild-type embryos; however, faint and no expression were found for RUNX2 and OPN, respectively, in the Cp/Cp embryos. Also in the digits, OPN expressed strongly in the wild-type but faint in the Cp/Cp embryos. These results clearly indicate that osteoblast differentiation is inhibited in the Cp/Cp embryos due to the loss of IHH expression.
2) The conclusion in the same section that "chondrocyte proliferation and subsequent differentiation are completely inhibited" (repeated in the discussion) is similarly not supported by direct evidence, for example histological analysis of chondrocyte proliferation. The authors subsequently show absence of the hypertrophic chondrocyte marker COL10, so an amendment of this conclusion may be sufficient. 3) Localisation of IHH and NHEJ1 in normal embryos shown in Figure 3 is exceedingly week.
The NHEJ1 sense and anti-sense probes appear to have produced comparable staining in the neural tube in the images provided. These localisations need to be improved or replaced with protein-level analyses.
We could not obtain IHH and NHEJ1 antibodies that react with chicken cells; therefore, we could not accomplish this experiment. However, we re-examined in situ hybridization of IHH and NHEJ1 with antisense and sense DIG probe, and we got the same result in which the sections stained with antisense probe had higher intensity compared to that stained with sense probe. Fig.   3a has been magnified to make it easy to see the signals.

4) Interpretation of nick-end labelling (TUNEL assay) indicating double strand breaks is
confusing as this assay is primarily used to detect apoptotic cells. It is very surprising that no apoptotic cells were observed in the +/+ embryos given this process is common during development (e.g. over the closed neural tube). γH2AX staining of cp/cp versus +/+ embryonic tissue is needed to confirm that the increase in TUNEL staining is not simply due to an increase in apoptosis in dyeing embryos.  -A schematic comparison of the known NHEJ pathway similarities/differences between humans, mice and chickens would be helpful.
This difference between birds and mammals has been unknown because there has been only one report on the chicken DSB repair pathway, which used the DT40 cell line. The HR pathway considered to be more important for the DSB repair in chicken DT40 cells than in mouse ES cells, suggesting that the roles of the two DSB repair pathways appear to be somewhat different between two species [Takata et al., EMBO J. 17, 4497-5508, 1998]. Further studies are needed to answer the reviewer's comment.
-Lines 285-286 claim the Cp model "highlights the importance of NHEJ function I normal neurogenesis". The production of neurones has not been assessed (even if reduced, apoptosis of the neuroepithelium before the onset of neurogenesis would be likely to limit neuron production).
This conclusion should be removed.
Line 336: The sentence has been removed following the suggestion.
Reviewer #2 (Remarks to the Author): Major 1) From this work, the conclusion that NHEJ1 deficiency "alone" causes embryonic lethality in the Creeper chicken can be suggested but not directly made. Strict demonstration would require a chicken with only the NHEJ1 defect. Therefore the title of the manuscript should be "Combined deletions of IHH and NHEJ.
Title: "Combined" has been added at the beginning of the title.
Minor 1) P 5 line 95. NHEF1…..NHEJ1 Line 101: "NHFJ1" has been corrected to "NHEJ1" 2) Would it be possible to perform clonogenic survival assays after DSB induction (IR or radiomimetic drugs) with the established primary cells to further support the NHEJ defect.
The growth of the Cp/Cp fibroblast cells were very slow in culture, and the colony forming cell assay was very difficult due to the low ability of proliferation. So we have not done this assay.
3) P 10 line 226 discovery of NHEJ1….it is a co-discovery by ref 27 and 28 Line 271 -272: Following the suggestion, we have rewritten the sentence as follows (underlined). "Human NHEJ1, also known as Cernunnos or XLF, encoded by the NHEJ1 gene was discovered as an XRCC4-interacting protein by a yeast two-hybrid screening system and as the protein mutated in patients with growth retardation, microcephaly, and immunodeficiency. 37,38 ." 4) P 10 Lines 236-237. The CTR of NHEJ1 contains the NLS but also the Ku binding motif.
Recruitment and retention of XLF at DSBs in cells requires Ku interaction see and refer to the work of Charbonnier, Caldecott and Chen. The data in this manuscript clearly show that truncated XLF is not recruited to sites of damage but that it is still can be localized in the nucleus.
Line 294 -300: Following the suggestion by the Reviewer, we have added new sentences in the Discussion section citing three papers as follows. "The CTR of NHEJ1 contains the nuclear localization signal (NLS) but also the Ku binding motif, and recruitment and retention of NHEJ1 at DSBs in cells requires Ku interaction (Yano et al., 2008;Grundy et al., 2016;Nemoz et al., 2018). The present results clearly showed that the truncated NHEJ1 is not recruited to sites of damages but that it is still can be localized in the nucleus. Altogether, our findings strongly support that the Ku binding motif in the CTR of NHEJ1 is important for recruitment and retention of XLF at DSBs." 5) P 12 line 281. Remove "was" Line 332: "was" has been removed. We have showed the photographs that clearly demonstrate morphological abnormalities of the Cp/Cp embryos at E3 in Fig. 1b. We have added more data on morphologies of +/+, Cp/+, and Cp/Cp embryos in Supplementary Tables 2 and Table 3. 2. The description of the deletions are very informative and important to this manuscript, however to be sure that the same deletions are causing the initial Creeper phenotype it will be informative to double check that these deletions exist in the JB and/or the Chinese strains too. Minor concerns: 1. There is no clear rationale (in the main) text for choosing the GSP strains to study the mutation.
Line 52 -54, Line 74 -77: The GSP strain is the highly inbred strain derived from the Fayoumi breed, which shows quite low heterozygosity; its average heterozygosity is less than 0.01 for more than 50 microsatellite DNA markers (Nunome et al. Exp. Anim. 68, 177-193, 2019). We constructed a congenic strain of the Creeper allele that was introduced from the JB strain to characterize the Cp phenotype in homozygous genetic background. We have added the part that is deficient.
2. There is no explanation on why ALAD gene is being investigated and not another gene.
Line 140 -141: We studied expression pattern of ALAD in the extraembryonic blood vessels and blood island according to GEISYA web site (http://geisha.arizona.edu/geisha/). It was expressed in the blood island as in the web site. We used its expression as a marker of the blood island. We added this explanation in the result session.

A bit of introduction on DSB would be welcome for the readers.
L264 -269: We have added the introduction of DSBs with some references.
4. The panels of the figures are very small and are difficult to see (in particular Fig. 3 and Fig. 4).   Fig. 5a and 5c has been modified to magnify them. 5. Sentence starting at line 205 (discussion) is confusing: the deletion is likely the same in the Chinese strain (see main concern 2). Are they potentially different genetic mutation causing the Creeper phenotype? This is unclear to me.
L113 -115, L233 -237: As mentioned above, the mutation in the JB strain was different from that in the Chinese Xingyi bantam strain because no genetic fragments could be amplified with the former PCR primer that was used for detecting the mutation in the Chinese chicken.
However, the mutation in other Japanese native chicken breeds with the Creeper phenotype (Miyaji-dori, Jitokko) was the same