Nhlrc2 is crucial during mouse gastrulation

Summary The loss of NHL repeat containing 2 (Nhlrc2) leads to early embryonic lethality in mice, but the exact timing is currently unknown. In this study, we determined the time of lethality for Nhlrc2 knockout (KO), C57BL/6NCrl‐Nhlrc2 tm1a(KOMP)Wtsi/Oulu, embryos and the in situ expression pattern of Nhlrc2 based on LacZ reporter gene expression during this period. Nhlrc2 KO preimplantation mouse embryos developed normally after in vitro fertilization. Embryonic stem (ES) cells established from KO blastocysts proliferated normally despite a complete loss of the NHLRC2 protein. Nhlrc2 KO embryos from timed matings implanted and were indistinguishable from their wildtype littermates on embryonic day (E) 6.5. On E7.5, Nhlrc2 KO embryo development was arrested, and on E8.5, only 6% of the genotyped embryos were homozygous for the Nhlrc2 tm1a(KOMP)Wtsi allele. Nhlrc2 KO E8.5 embryos showed limited embryonic or extraembryonic tissue differentiation and remained at the cylinder stage. Nhlrc2 expression was ubiquitous but strongest in the epiblast/ectoderm and extraembryonic ectoderm on E6.5 and E7.5. NHLRC2 is essential for early postimplantation development, and its loss leads to failed gastrulation and amniotic folding in mice. Future studies on the evolutionarily conserved NHLRC2 will provide new insights into the molecular pathways involved in the early steps of postimplantation development.


Summary
The loss of NHL repeat containing 2 (Nhlrc2) leads to early embryonic lethality in mice, but the exact timing is currently unknown. In this study, we determined the time of lethality for Nhlrc2 knockout (KO), C57BL/6NCrl-Nhlrc2 tm1a(KOMP)Wtsi /Oulu, embryos and the in situ expression pattern of Nhlrc2 based on LacZ reporter gene expression during this period. Nhlrc2 KO preimplantation mouse embryos developed normally after in vitro fertilization. Embryonic stem (ES) cells established from KO blastocysts proliferated normally despite a complete loss of the NHLRC2 protein.
Nhlrc2 KO embryos from timed matings implanted and were indistinguishable from their wildtype littermates on embryonic day (E) 6.5. On E7.5, Nhlrc2 KO embryo development was arrested, and on E8.5, only 6% of the genotyped embryos were homozygous for the Nhlrc2 tm1a(KOMP)Wtsi allele. Nhlrc2 KO E8.5 embryos showed limited embryonic or extraembryonic tissue differentiation and remained at the cylinder stage. Nhlrc2 expression was ubiquitous but strongest in the epiblast/ectoderm and extraembryonic ectoderm on E6.5 and E7.5. NHLRC2 is essential for early postimplantation development, and its loss leads to failed gastrulation and amniotic folding in mice. Future studies on the evolutionarily conserved NHLRC2 will provide new insights into the molecular pathways involved in the early steps of postimplantation development.

K E Y W O R D S
amniotic folding, ectoderm, embryonic stem cells, epiblast, gastrulation, Nhlrc2

| INTRODUCTION
The NHL repeat containing 2 (Nhlrc2) gene encodes an evolutionarily conserved NHLRC2 protein consisting of a thioredoxin-like, NHLrepeat β-propeller, and β-stranded domains (Biterova, Ignatyev, Uusimaa, Hinttala, & Ruddock, 2018). Variants in this gene lead to a neurodegenerative and multiorgan disease called FINCA (OMIM #618278) (Brodsky et al., 2020;Rapp et al., 2021;Uusimaa et al., 2018). FINCA patients and mice harboring a similar genotype develop normally in utero and are healthy at birth, despite having only a low amount of affected NHLRC2 (Hiltunen et al., 2020;Uusimaa et al., 2018). However, several studies have implied an important, but still unknown, role of NHLRC2 in embryonic development across species. In bovine, p.Val311Ala substitution in the β-propeller domain of NHLRC2 has been reported to cause neural tube-related developmental malformations in Angus cattle (Denholm, 2017), and bovine oocytes with greater developmental competence have been reported to have higher Nhlrc2 expression (Nemcova et al., 2016). Nhlrc2 has Siri Lehtonen and Reetta Hinttala contributed equally to this study. also been associated with embryo-maternal crosstalk in ewes and bovine (Xiao et al., 2021;Q. Yang et al., 2020). The complete loss of Nhlrc2 leads to early embryonic lethality in mice (Delhotal, 2016;Perez-Garcia et al., 2018;Uusimaa et al., 2018), indicating its essential role in embryonic development. Only very few remnants of Nhlrc2 KO trophoblast giant cells have been detected on embryonic day (E) 9.5 (Perez-Garcia et al., 2018) but the exact time of lethality of the embryo is unknown.
The molecular function of NHLRC2 is still relatively poorly understood, but it has been connected with a variety of molecular pathways, including reactive oxygen species-induced apoptosis in colon cancer cells (Nishi et al., 2017); vesicle transport, cytoskeleton organization, and endothelial to mesenchymal transition in fibroblasts ; phagocytosis (Haney et al., 2018;Yeung et al., 2019) and actin dynamics (Haney et al., 2018) in human macrophages; and RNA metabolism in mouse neurons and hippocampus (Hiltunen et al., 2020). All these pathways are relevant to embryonic development, and a more detailed understanding of the early developmental defects arising from the loss of NHLRC2 can reveal new information on the early embryo development and function of NHLRC2.
In the current study, we determined the precise time of lethality for embryos lacking NHLRC2 to the time of gastrulation and describe the in situ expression pattern of Nhlrc2 on E6.5 and E7.5. We also found that NHLRC2 is not essential for embryonic stem (ES) cells in vitro. The data presented here provide new information on the essential role of NHLRC2 during early postimplantation period of embryonic development.

| RESULTS
2.1 | The preimplantation development of Nhlrc2 KO embryos is unaffected As part of our previous publication , we observed that morula-stage (E2.5) homozygous C57BL/6NCrl-  (Skarnes et al., 2011;The International Mouse Knockout Consortium, 2007) were absent after heterozygous breeding. Our recent studies have confirmed that the Nhlrc2 tm1a(KOMP)Wtsi allele indeed leads to complete loss of full-length Nhlrc2 without any leakage (Hiltunen et al., 2020) and can be used as an Nhlrc2 KO allele to study its effect during embryogenesis. We performed in vitro fertilization (IVF) with heterozygous gametes and out of the 96 embryos obtained, 82 (85.4%) developed to the blastocyst stage and attached to gelatinized plates, forming a colony in the following 10-day culture in vitro. This proportion was close to the 86.4% of embryos that formed a colony in the control IVF using wildtype (wt) C57BL/6NCrl gametes. Next, we designed the genotyping protocol to enable identification of all three genotypes from one polymerase chain reaction (PCR) product utilizing a SacI digestion site present in the targeting construct ( Figure 1). The genotype was determined for 60 IVF-derived embryos, which, revealed a normal Mendelian ratio of Nhlrc2 KO embryos (Table 1). Previously, the morulae were obtained from in vivo fertilization (IVO) on E2.5, and we decided to re-genotype 15 of them with the new genotyping protocol to confirm that the IVF did not promote the development of the Nhlrc2 KO embryos. The genotyping revealed that 20% of the IVO embryos were homozygous Nhlrc2 KO embryos (Table 1), contrary to our previous report. This was further verified by Sanger sequencing.
These results indicate that, against our previous data, preimplantation development is not compromised by the loss of Nhlrc2 in mice.

| Mouse ES cells are viable without Nhlrc2
As Nhlrc2 KO preimplantation embryos formed normal colonies in vitro, we next established four wt and four homozygous Nhlrc2 tm1a (KOMP)Wtsi ES cell lines from separate IVF-derived embryos using the 2i method (Nichols, Silva, Roode, & Smith, 2009). The Nhlrc2 tm1a (KOMP)Wtsi ES cells showed no NHLRC2 in immunoblotting, confirming   Figure 3a). On E7.5, the homozygous Nhlrc2 KO embryos were still present in the expected ratio according to Mendelian inheritance, but they were developmentally arrested or even degenerated ( Figure 3a and Table 1). The developing amnion and chorion were not visible in the Nhlrc2 KO embryos through stereomicroscope examination. On E8.5, we obtained only two homozygous Nhlrc2 KO embryos out of the 34 isolated embryos (5.9%) ( Table 1). In addition, four deciduae with only small remnants of an embryo were collected but could not be reliably genotyped. The development of the two confirmed Nhlrc2 KO embryos was severely arrested, and their gross morphology resembled that of an E6.5 rather than an E8.5 embryo ( Figure 3a).
Although the Nhlrc2 KO embryos grew in size to approximately that of an E7.5 embryo, no defined internal structures could be detected.
The genotyped Nhlrc2 KO embryos were distinguishable from their wt littermates from E7.5 onwards, based on their poor appearance. We then selected similar poorly surviving, presumably Nhlrc2 KO embryos, for further structural analysis with their wt littermates.
We examined the ultrastructure of E7.5 embryos using transmission electron microscopy (TEM) and the morphology of E7.5 and E.8.5 embryos from hematoxylin eosin-stained histological sections. The Developing amnion and allantois were also strongly stained on E7.5 (Figure 5a,b).

| DISCUSSION
Our results show that loss of NHLRC2 leads to the termination of embryonic development after E6.5 in mice and its loss is detrimental to amniotic folding and the completion of gastrulation. We also show that Nhlrc2 is widely expressed in the gastrulating embryo, with the strongest expression in the embryonic and extraembryonic ectoderms.
However, Nhlrc2 is not essential for mouse ES cells and their spontaneous differentiation.
The halt in development happened considerably later than we expected in the basis of our previous results . Previous attempts to establish cell lines with complete loss of Nhlrc2 in THP-1 cells (Yeung et al., 2019) and human fibroblasts  have yielded only knockdown cells, highlighting the essentiality of NHLRC2 in differentiated cells. Here, the establishment of Nhlrc2 tm1a (KOMP)Wtsi ES cell clones provided an important access to Nhlrc2 null control and enabled further optimization of the genotyping protocol. The new improved protocol, described in the current study, identifies all three genotypes from one PCR amplified product and thus requires less DNA template. This is crucial when genotyping such small samples.
The Nhlrc2 KO mouse embryos showed no normal ectoderm and mesoderm cell layer organization or amniotic folding. This could be an indication of a defective Nodal, bone morphogenic protein, transforming growth factor beta (TGF-β), or Wnt signaling (Tam & Loebel, 2007), as impaired mesoderm induction can result in loss of internal embryonic or extraembryonic structure development. For example, embryos lacking Smad2 or Smad4, a TGF-β mediator, have defective mesoderm induction and show some similarity to the phenotype described here (Sirard et al., 1998;Weinstein et al., 1998;. However, in contrast to Nhlrc2 KO embryos, these mutants also showed a defective egg cylinder formation not evident in Nhlrc2 KO embryos. The thioredoxin-like domain, also found in NHLRC2, is characteristic to oxidoreductases such as thioredoxin. Conversely, thioredoxin KO mice die shortly after implantation before E6.5, and their inner cell masses fail to proliferate in vitro (Matsui et al., 1996). The observation that NHLRC2 is not essential for mouse ES cells, raises the question of whether NHLRC2 may be critical for cell-to-cell communication or cell migration during embryogenesis, which are not important in a monolayer culture. Decrease in NHLRC2 has been found to affect cytoskeleton organization and endocytic vesicle trafficking pathway in vitro (Haney et al., 2018;Paakkola et al., 2018;Yeung et al., 2019). Actomyosin and microtubule cytoskeletal systems are essential for the various cell behaviors needed for gastrulation movements (Solnica-Krezel & Sepich, 2012). Endocytic pathway is required for signal amplification and termination for spatiotemporal regulation of cell movements during gastrulation (Wada & Sun-Wada, 2013). Moreover, mutations leading to decreased cell proliferation can result in failed mesoderm induction and developmental arrest on E6.5 (Power & Tam, 1993;Tam & Behringer, 1997). For example, the KO of Hnrnpc, a gene whose expression we have previously found to be affected in the neuronal precursor cells of the NHLRC2 deficient FINCA mouse model (Hiltunen et al., 2020), also leads to an inability to develop past the cylinder stage but does not affect the viability of ES cells (Williamson, Banik-Maiti, DeGregori, & Ruley, 2000).

| In vitro fertilization
IVF and embryo culture for 10 days on a gelatinized dish in an ES cell culture medium were performed at the Biocenter Oulu Transgenic and Phenotyping Core Facility (Oulu, Finland), as described previously (Hiltunen et al., 2020;Uusimaa et al., 2018).

| DNA extraction and PCR
After trypsinization, the ES cell pellets were lysed with 0.1 mg/ml proteinase K in lysis buffer (0.1 M Tris pH 8.5, 5 mM ethylenediaminetetraacetic acid, 0.2% sodium dodecyl sulfate [SDS], 0.2 M NaCl), and the DNA was isolated through ethanol precipitation.
PCR was used to amplify a region in Nhlrc2, where the insertion of the targeting construct resulted in a new SacI digestion site, with the primers shown in Table 2. Phire Hot Start II Polymerase (Thermo Fisher Scientific, Waltham, MA) and a Piko Thermal Cycler (Thermo Fisher Scientific, Vantaa, Finland) were used according to the manufacturer's instructions. Sanger sequencing was performed to validate the PCR product as described previously (Hiltunen et al., 2020). The PCR product from E6.5-8.5 embryos was precipitated with NaCl (4 M, 1:10) and ethanol before digestion overnight at À20 C or for 30 min at À70 C. The PCR product was digested using the SacI immunosurgery (Solter & Knowles, 1975), and the inner cell mass was transferred first to the feeders in 2i medium and then after the

| Immunoblotting
The protein isolation and immunoblotting protocol used has been described in detail previously (Hiltunen et al., 2020). In brief, ES cell pellets were solubilized with 1.5% dodecyl β-d-maltopyranoside Film, Tokyo, Japan) was used to detect bands after antibody incubations. Fiji software was used to determine the band intensities (Schindelin et al., 2012).

| Histology
Decidua were dissected on E7.5-8.5 and processed as described previously (Hiltunen et al., 2020 was used in the embedding of β-galactosidase-stained embryos to stabilize the X-gal reaction product (Masahira et al., 2005). Thin and semi-thin sections were cut throughout the embryo using a Leica Ultracut UCT microtome (Leica, Wetzlar, Germany), and toluidine blue-stained semi-thin sections were used to select thin sections. The thin sections were stained with uranyl acetate and lead citrate and examined using a Tecnai G2 Spirit 120 kV transmission electron microscope (FEI, Eindhoven, Netherlands). Images were captured using a Quemesa CCD camera (Olympus Soft Imaging Solutions GMBH, Münster, Germany).

| Spontaneous differentiation of mouse ES cells
A sample of 5 Â 10 6 ES cells were plated into two wells of an uncoated six-well plate at passage five. The medium was changed at Day 2. On Day 3, the formed embryoid bodies were transferred to a 6-cm plate coated with gelatin, and the medium was changed to a differentiation medium: complete basal medium (Millipore, Billerica, MA) with 10% FBS (Pan-Biotech, Aidenbach, Germany) and 100 μ/ml penicillin-streptomycin (Sigma-Aldrich, St. Louis, MO). The medium was changed every other day, and the cells were collected for RNA isolation after 16 days.

| β-Galactosidase staining
The embryos were stained using a β-galactosidase staining kit (Biovision, Milpitas, CA) according to the manufacturer's instruction.
The stained embryos were postfixed for 10 min with 4% paraformaldehyde and mounted in 1% low-melt agarose (ROTH, Karlsruhe, Germany). For histological sectioning, the embryos were embedded in paraffin after dehydration in ethanol and isopropanol and sectioned into 6-μm sections using a microtome (Microm, Walldorf, Germany). The embryos were counterstained in eosin for 1.5 min with a minimal xylene treatment to avoid the loss of the X-gal staining product.

DATA AVAILABILITY STATEMENT
The data generated or analyzed during this study are included in this published article. Further information and additional datasets are available from the corresponding author upon reasonable request.