The Prion-like protein Shadoo is involved in mouse embryonic and mammary development and differentiation.

Shadoo belongs to the prion protein family, an evolutionary conserved and extensively studied family due to the implication of PrP in Transmissible Spongiform Encephalopathies. However, the biological function of these genes remains poorly understood. While Sprn-knockdown experiments suggested an involvement of Shadoo during mouse embryonic development, Sprn-knockout experiments in 129Pas/C57BL/6J or 129Pas/FVB/NCr mice did not confirm it. In the present study, we analyzed the impact of Sprn gene invalidation in a pure FVB/NJ genetic background, using a zinc finger nuclease approach. The in-depth analysis of the derived knockout transgenic mice revealed a significant increase in embryonic lethality at early post-implantation stages, a growth retardation of young Sprn-knockout pups fed by wild type mice and a lactation defect of Sprn-knockout females. Histological and transcriptional analyses of knockout E7.5 embryos, E14.5 placentas and G7.5 mammary glands revealed specific roles of the Shadoo protein in mouse early embryogenesis, tissue development and differentiation with a potential antagonist action between PrP and Shadoo. This study thus highlights the entanglement between the proteins of the prion family.

divergent observations, sometimes attributed to genetic compensation in mutated animals 13 , questioned the involvement of Shadoo during embryogenesis.
In this study, we report the establishment, through injection of a Zinc Finger Nuclease (ZFN), of FVB/NJ Sprn-knockout (Sprn 0/0 ) mice. Reproductive performances and transcriptomic analyses of these mice demonstrate that Shadoo is involved in embryonic development and more generally in the development and differentiation of several tissues. Furthermore, our data suggest antagonist actions of PrP and Shadoo, at least during embryonic stages.

Results
Generation of Sprn 0/0 mice. The ZFN target site is located at the start of the Sprn ORF (Fig. 1). ZFN mRNA was injected into 1 cell-fertilized FVB/NJ mouse eggs that were transferred into pseudo-pregnant mice. Tail-DNA analysis of the 29 resulting pups by PCR amplification of a ~400 bp genomic region surrounding the ZFN-target site and sequencing analysis allowed identifying 9 transgenic founder mice, carrying small deletions ranging from 1 to 13 bp. All founder mice were heterozygous for the mutated allele.
Offspring from two founders (17 and 05) were used to generate Sprn 0/0 mice. These founders were chosen as they transmitted a mutated allele (deletion of 1 nucleotide associated with the mutation of an adjacent nucleotide and deletion of 10 nucleotides, respectively), inducing a frameshift leading to an early premature stop codon. Translation of the corresponding mRNA is predicted to produce truncated nonsense proteins (Fig. 2) lacking all Shadoo domains (Fig. 1).
Sprn expression analysis was investigated in Sprn 0/0 mouse brains. Sprn-transcript levels seemed unaffected, as judged by semi-quantitative RT-PCR experiments (data not shown). This observation was expected since the ORF is located within the last exon of the gene and the mutated transcript should thus escape nonsense mediated decay 15 . Microarray transcriptomic analysis of other tissues led to similar observations as Sprn was not detected as differentially expressed between Sprn 0/0 and FVB/NJ mice (see below). The entire ORF was reverse-transcribed from brain RNAs of WT and Sprn 0/0 mice. No alternative splicing was detected and Sprn ORF from Sprn 0/0 mouse mRNAs carried the expected nonsense mutation (Fig. S1A). Of note is the lack of downstream in frame ATG codon close to the nonsense mutation, avoiding translation re-initiation (Fig. 2). Brain western blots analyses did not allow detecting Shadoo in WT or Sprn 0/0 brains, despite using several commercially (Abgent # AP4754b, PA5 24688, ab175070) or laboratory homemade (1909 S0, S4 et S5, J74 et J740 from CEA, France and 06sh1 from University Alberta, Canada) antibodies and detection techniques (ECL, IF, data not shown). Nevertheless, RNA investigations clearly demonstrate the knockout genetic status of these mice. increased lethality of Sprn 0/0 embryos. We next investigated whether the reduced litter size in Sprn 0/0 mice was associated with increased embryonic lethality. Pregnant females from Sprn 0/0 × Sprn 0/0 or WT × WT crossings were analyzed at G7.5 and G14.5. At G7.5, no resorption was observed in FVB/NJ mice (0/76 implantation sites (IS)). In contrast, a statistically significant 12.7% resorption rate was detected in Sprn 0/0 females (7/55 IS, p < 0.05). Empty decidua suggested that resorptions occurred early after implantation. At G14.5, this significant difference remained with a resorption rate of 19.7% in Sprn 0/0 mice (14/71 IS) compared to 6.2% in WT mice (10/162 IS). Thus, the reduced litter size observed in Sprn 0/0 × Sprn 0/0 crossings results from an increased fragility of the embryos at early post-implantation, a key period of mammalian fetal development.
Transcriptomic analysis of E7.5 Sprn 0/0 embryos. To get deeper insight into the biological consequences of Sprn knockout in embryonic development, comparative microarray transcriptomic analysis of WT and Sprn 0/0 embryos was performed at E7.5. Following raw data normalization, differentially expressed genes (DEG) with a p-value < 0.05 and a fold change (fc) <−2 or >2 were analyzed. Overall, 147 DEG were identified, 86 upregulated and 61 downregulated, of which 29 were small nucleolar RNAs (Table S1). These small RNA genes regulate cell proliferation, migration and invasion by inducing Epithelial-Mesenchymal transition (EMT) 16,17 . Notably, a large set of small nucleolar RNAs with a conserved H/ACA box (snoRA) were downregulated. Such downregulation was found to be associated with cellular apoptosis 17 . Further, gene set enrichment analysis (GSEA) identified downregulation of genes involved in dorsal spinal cord development (enrichment score (ESc) 3.37, false discovery rate (FDR) 0.006), mitochondrial translational termination and elongation (ESc 2.7, FDR 3.08 × 10 −5 ), constituent of ribosomes (ESc 3.47, FDR 4.5 × 10 −17 ) and proteasome core complex (ESc 2.6, FDR 4.2 × 10 −5 ). Identification of these biological pathways revealed that Sprn 0/0 E7.5 embryos suffer from developmental defects.  ; developmental disorder, cell-to-cell signaling and interaction, immunological and/or hematological disease, organism survival and cell growth and proliferation ( Table 2). It suggested that Shadoo deficiency alters cell proliferation and differentiation resulting in abnormal early development. Potential related master regulator genes were identified in further analyses of the DEG, but at reduced fc (note S1). Collectively, these transcriptomic analyses suggest that Sprn knockout induces alteration of EMT pathways, impacting cell fate determination and inducing embryonic lethality soon after the implantation.

Potential opposite transcriptomic alterations between Sprn and Prnp knockout E7.5 embryos.
Some pathways identified above are reminiscent of those described for PrP modulation of EMT 18,19 . PrP regulates embryonic stem cell pluripotency and differentiation, and Shadoo and PrP share biological properties 9 although recent experiments suggest that it could depend of the model system used 20,21 . As we previously performed transcriptomic analysis of E7.5 Prnp 0/0 embryos 11 , we searched for potential shared DEG between Prnp 0/0 and Sprn 0/0 embryos. Two master regulators of pluripotency were differentially expressed in E7.5 Prnp 0/0 embryos, Oct4 and TGFβr2. Oct4 was overexpressed and TGFβr2 downregulated in Prnp 0/0 embryos 11 , i.e. oppositely to what is observed in Sprn 0/0 (note S1). Furthermore, out of the 147 DEG observed in Sprn 0/0 embryos, 12 were also differentially expressed in Prnp 0/0 embryos 11 but in an opposite sense, those indicated in Table 2 alongside Hspb7 and Igsf11 (Table S1). Importantly, Prnp was not differentially expressed in Sprn 0/0 embryo, excluding this observation be an indirect consequence of Prnp overexpression.
Altogether, these observations suggest that Shadoo and PrP exert antagonist functions in early embryonic development. It fits in with the above-mentioned observation that Sprn 0/0 and Prnp-overexpression 14 both induces smaller average litter sizes. However, these opposite differential gene expressions involved a limited number of genes and small variation levels, and we cannot formally exclude that this observation made across different experiments, might be coincidental. comparative analyses of Sprn 0/0 pups' growth curves and of Sprn 0/0 placentas. To assess whether pups body weight is affected in Sprn 0/0 mice, as previously reported 12 , we compared the growth curves of WT and Sprn 0/0 pups fed by WT lactating females. The body weight gain was lower for Sprn 0/0 versus WT pups (statistically significance between L3-L10, Fig. 3A). At variance with the previous data 12 , there was no difference according to the pup sex. Abnormal growth could have various origins, one being placental dysfunction as in Prnp 0/0 mice 14 . Placental dysfunction would explain that at birth, the weight of the Sprn 0/0 pups were significantly lighter (1.34 +/− 0.14 g for Sprn 0/0 pups (n = 45) versus 1.4 +/− 0.16 g for WT pups (n = 97), t-test p < 0.05), suggesting intra-uterine growth retardation. We thus analyzed whether Sprn knockout affects placental efficiency. Fetuses and placental tissues from E14.5 WT and Sprn 0/0 mice were weighted. No significant difference was observed between the two genotypes (Table 3). Histological examination of E14.5 and E17.5 placentas did not reveal structural or vascularization modification (data not shown). The relative proportions of the labyrinth layer, spongiotrophoblast and decidua zones were unaffected.
We next performed comparative transcriptomic analysis on E14.5 WT and Sprn 0/0 placentas. Following raw data normalization, DEG with a p-value < 0.05 and a fc < −2 or > 2 were selected. 133 genes were identified, of which 17 were downregulated (Table S2), including 12 small nuclear RNAs. Seven DEG were common between placentas and embryos, all of them behaved similarly between the two tissues; 4 small nuclear RNAs alongside Glp1r, LY96 and SLCO4C1 genes. This small number of shared DEG was expected due to transcriptomic divergences between organs besides early stages of development 22 . Validation of the transcriptional regulation of five genes, TBPA, Gli3, GC, PSG23, SerpinA1 (α1-antitrypsin), was tested by RT-qPCR. All results were concordant (Table S3). Furthermore, immunoblot analyses were performed to examine the expression levels of two proteins, SerpinA1 and Fibrinogen. The results confirmed the tendency observed at the transcriptomic levels, with on average a higher expression level of the two proteins in the knockout animals, which did not reach statistical significance (Fig. S2). IPA ® of the DEG identified two major associated networks (score > 35); lipid metabolism, molecular transport, small molecule biochemistry and tissue development, cellular movement (Table 4). Global gene activation  Table 2. Identified networks with differentially expressed genes in E7.5 embryos. Bold-faced genes: positively differentially expressed genes in Sprn0/0 embryos. Underlined genes: negatively differentially expressed genes in Sprn0/0 embryos. Yellow colored genes: genes also found differentially expressed in Prnp0/0 E7.5 embryos 11 . Of note, these genes were all differentially expressed in an opposite way between Sprn0/0 and Prnp0/0 embryos.
Altogether, these data suggest that at E14.5, Sprn 0/0 placentas underwent an adaptive biological process, involving LXR/RXR and FXR/RXR activation highlighted by the apo-lipoproteins and serpinA1 upregulations, to compensate earlier developmental dysfunctions.

Sprn-invalidation affects mouse lactation.
Since Sprn appears to be involved in embryonic EMT and in cellular proliferation and tissue development and since PrP C is involved in the self-renewal of mammary adult stem cells 25 , we assessed whether Sprn knockout affects lactation. We compared the growth curves of WT pups fed by WT or Sprn 0/0 lactating females. Their weight gain was significantly lower when fed by Sprn 0/0 females between L2 and L10 (Fig. 3B). No significant difference was observed between WT and Sprn 0/0 milk protein concentrations at L7.5 (Table S4). Consistently, SDS-PAGE and western analyses did not reveal modification of the milk protein profiles (Fig. S3). Iron, zinc and copper concentrations were also similar between Sprn 0/0 and WT milks (data not shown). These data thus suggest that Sprn invalidation affects mouse lactation.   www.nature.com/scientificreports www.nature.com/scientificreports/ Sprn-invalidation affects milk fat globules volumes and morphology. Analysis of milk fat globules (MFGs) revealed a significant reduction of the medians volume of Sprn 0/0 MFGs (n = 3166) compared to that of FVB/NJ MGFs (n = 5643) (Fig. 4A, kruskal-Wallis chi-squared = 22.114, p-value = 2.57 10 −06 ). Analysis of the size distribution of the MFGs showed that the majority of MFGs were smaller in Sprn 0/0 milk with a significant increase (up to 30% of total MFGs) of very small MFGs (size between 0 and 100 µm 3 , Fig. 4Aa). Intriguingly, the detailed analysis of this size range (Fig. 4A,b,b') showed that more than 30% of the very small MFGs in Sprn 0/0 milk were <10 µm 3 (Fig. 4Ab). Moreover, while size distribution of MFGs from FVB/NJ mice between 0 and 100 µm 3 was in favor of the largest ones, those from Sprn 0/0 mice were evenly distributed (Fig. 4Ab'). To what extend this difference in MGF size distribution affects milk digestibility and pups' growth remain unknown. Nevertheless, these observations suggest that the growth of intracellular lipid droplets (CLDs), potentially linked to their apical transport and secretory mechanisms, could be altered in the absence of Shadoo. We thus investigated whether Shadoo absence affects MFGs morphological properties. MFGs were co-labeled with Bodipy 493/503 for neutral lipids and either acridine orange for cytoplasmic crescents, FM4-64 for membranes, wheat germ agglutinin for glycoproteins and filipin for free cholesterol, before analysis by immunofluorescence microscopy (Fig. 4B). While MFGs from Sprn 0/0 and FVB/NJ mice exhibited similar proportions of cytoplasmic crescents, those from Sprn 0/0 mice were rather labeled at their periphery, suggesting a more acidic environment (Fig. 4B, AO). MFGs membranous morphology and surface glycoproteins distribution was similar in both genotypes (Fig. 4B, FM4-64; WGA). However, MFGs from Sprn 0/0 mice appeared enriched in free cholesterol, a hallmark of membrane microdomains called rafts (Fig. 4B, filipin). Collectively, these results point to subtle modifications of MFGs membranes from Sprn 0/0 mice, which may impact their size distribution 26,27 . Sprn invalidation affects mouse mammary epithelial tissue development. As Sprn 0/0 MGFs morphological properties were altered, we investigated whether differences could be observed at the MG level. Confocal imaging on tissue sections were performed at three MG developmental stages: mid-gestation (G10) and full lactation (L10) in the presence or absence of pups. Key features such as lipid droplets morphology, apical polarity and organization of the myoepithelial cells were analyzed (Fig. 5). At these stages, the lipid droplets, while correctly associated with the raft specific marker GM1 ganglioside, were generally smaller in Sprn 0/0 mammary epithelial cells with a more disperse distribution. The myoepithelial cells morphology and general organization around forming or active alveoli was markedly disturbed in Sprn 0/0 mice, as assessed by glycoproteins (WGA) and actin (PHA) staining. These data thus suggest that the lack of shadoo disturbs the establishment of the apico-basal polarity of mammary epithelial cells and alveoli formation.
Sprn-invalidation affects mouse mammary proliferation. We next investigated the potential incidence of Sprn 0/0 invalidation on mammary gland development by whole mount analyses. The selected pictures showed in Fig. 6 are representative of the mammary morphology of at least 3 animals of both genotypes. In virgin, G7.5 and lactating MGs, some mammary primary canals were enlarged in Sprn 0/0 tissues. In gestating and lactating tissues, Sprn 0/0 MGs had less and only primary side branching compared to their wild-type counterparts and were overall less dense. This suggests that Sprn-knockout is associated with developmental defects.
To consolidate these observations, we explored whether Shadoo invalidation impacts proliferation and/or apoptosis of mammary epithelial cells. MG sections of each genotype were labeled for Ki67, active caspase3 at mid-gestation and full lactation (Fig. S4). At mid-gestation both proliferating and apoptotic cells were detected in developing alveoli from FVB/NJ MG. Double labeling with TUNEL indicated that aCasp3-positive apoptotic cells differed from Ki67-positive proliferating cells. In Sprn 0/0 MGs, there were fewer proliferating Ki67-positive cells while apoptotic cells were found in similar or higher proportions than in WT MG. While apoptotic cells mostly localized near the forming lumen of alveoli in FVB/NJ, they appeared more dispersed in Sprn 0/0 MG. During lactation, few proliferating and apoptotic cells were observed at the periphery of the alveoli in the mammary tissue of both FVB/NJ and Sprn 0/0 mice. These results suggest that the absence of Shadoo generates an imbalance in favor of apoptosis in the growing mammary epithelial tissue, associated with a defect/delay in lumen formation.

Sprn-invalidation affects mouse mammary transcriptomic. Comparative transcriptomic analysis
was investigated on G7.5 WT and Sprn 0/0 MGs. Fifty-three DEG were identified with a p-value < 0.05 and fc < −2 or >2, 21 downregulated and 32 upregulated (Table S5) (Table 5). It suggests that in Sprn 0/0 , gestating mammary epithelial cells undergo, associated with an immune reaction, a growth inhibition correlating with the whole mount observations. Search for deregulation of known  NJ; a,c,e,g,i) or mutant (Sprn 0/0 , b,d,f,h,j) mice. It highlights defects in duct tree formation in 6-week-old virgin, in duct side-branching during early gestation (G7,5) and consequently in the density and organization of the mammary epithelium during lactation (L2: day 2 of lactation, 10: day 10 of lactation; +pups: in the presence of pups, −pups: in the absence of pups). Bar = 0.5 mm. Enlarged views of each developmental stage show the differences in mammary epithelial tissue organization between WT and Sprn 0/0 mice. Bar = 200 nm.
Overall, transcriptomic analysis of Sprn 0/0 gestating MG revealed the potential deregulations of pathways corroborating phenotypic and histologic analyses of this tissue. These results highlight a yet undescribed role of Shadoo in lactation.

Discussion
The biological role of the members of the prion family remains mostly unknown, despite intensive researches focusing on PrP due to its major implication in TSE and potential involvement in other more prevalent neurodegenerative diseases such as Alzheimer's disease. Less attention focused on the related Shadoo protein, probably resulting from its recent discovery and potentially limited implication in TSE 29,30 . The previously reported knockout of Sprn did not induce overt phenotypes beside a moderate, transient and sex-related, growth deficiency that could originate from a role of Shadoo in the control of feeding behavior, in association with its expression in hypothalamic neurons 12 .
A role of PrP in mouse embryonic development, around the implantation stage and the transition from anaerobic to aerobic metabolism, was suggested as well as in the development of the nervous system and of extra-embryonic tissues. Shadoo shares with PrP some biological properties 9 , although other biological roles are specific to this protein and/or even antagonist to PrP 20,21 . A biological redundancy between PrP and Shadoo was evocated following the observation that the knockdown of Sprn induces early embryonic lethality in Prnp 0/0 mouse transgenic lines 10,11 . This hypothesis was weakened by the absence of phenotype of double Prnp/Sprn knockout mice 12 . The present report reinvestigates the role of Shadoo in FVB/NJ embryonic development through the generation of knockout mice using a ZFN approach, allowing its assessment in a non-mixed genetic background. This study highlights that Sprn 0/0 mice suffer from subtle, sometimes transient, but significant alterations of developmental and differentiation processes that affect both embryonic and adult tissues (Table 6).
Sprn 0/0 mice were viable and fertile, but analysis of the transmission rate in heterozygous intercrosses and of the average litter size in knockout homozygous crossings revealed a non-Mendelian distribution of the genotypes and a reduced number of pups, respectively. Further investigations highlighted an abnormal lethality rate of Sprn 0/0 embryos at an early post implantation developmental stage, between E4.5 and E6.5. This suggested that the absence of Shadoo is not detrimental for embryonic implantation but weakens the survival potential of the embryos early post-implantation.
Absence of Shadoo transiently also affects the growth rate of the pups in this genetic background but independently from their sexes. This slightly differs from previous observations 12 . The difference could originate from the genetic backgrounds used in relation to the invalidation strategies. Although not formally exclusive with an abnormal feeding behavior, this transient phenotype could correlate with an abnormal biology of the placentas, as judged by our transcriptomic studies ( Table 6). Invalidation of Sprn led to placental defects, recalling observations made following Prnp-knockout 14 . However, in Sprn 0/0 mice, no histological placental defect were evidenced and the few genes described as DEG in Prnp 0/0 placentas were not transcriptionally affected in Sprn 0/0 . Our transcriptomic analysis rather suggests that Sprn 0/0 extra-embryonic tissues might undergo early developmental defects, highlighted by transcriptomic alteration in E7.5 embryos of genes affecting the formation of mouse extra-embryonic tissues such as Smad1. These early potential defects seem at least partially corrected during development, as exemplified by the activation of pathways at G14.5 supporting placentation and embryonic development. Studies of earlier developmental stages to assess placenta development should help validating this hypothesis.
We also evidenced an impact of the absence of Shadoo on MG physiology, objectivized by (i) growth curve analyses of FVB/NJ pups fed by Sprn 0/0 females, (ii) transcriptomic alterations of G7.5 mammary tissue, (iii) morphological development defects and alteration of milk fat globules synthesis that correlate with transcriptomic alterations, (iv) mammary cell/membrane alterations. Implication of the Prion protein family in the physiology of the MG per se has received little attention. Implication of PrP in mammosphere formation and repopulating activity was documented 25 , but its relationship with lactation performances remains controversial. Of note is the  www.nature.com/scientificreports www.nature.com/scientificreports/ absence of any mammary phenotype described so far for Prnp 0/0 mammals. In the present study, invalidation of Shadoo appears to induce various gestational/lactation mammary tissue and cellular defects that results in transient lactation deficit.
This phenotype exemplified Shadoo implication in the control of cellular proliferation and differentiation, also revealed at early post-implantation embryonic stages. Such an involvement in cell proliferation and differentiation was highlighted for PrP 31 . Implication of these two related proteins in similar biological functions alongside their expression in developing mammalian fetuses could substantiate the hypothesis of their biological redundancy. Surprisingly, our data rather suggest that Sprn and Prnp exert opposite effects on expression of master regulators of pluripotency and/or embryonic development, at least at E7.5. Thus, one could expect their co-invalidation to be neutral, as documented 12 . However, comparison of the data described here and for Sprn-knockdown/Prnp 0/0 embryos 10,11 highlights a developmental delay in the embryonic lethality between Sprn 0/0 and Sprn-knockdown/Prnp 0/0 embryos. Lethality of Sprn 0/0 embryos occurs early after implantation while that of Sprn-knockdown/Prnp 0/0 embryos was evidenced only at E10.5 10 , with E7.5 decidua carrying abnormal embryos 11 . It could suggest that Sprn-knockdown induces a fragility of the embryos at implantation that becomes detrimental to the further development of the Prnp 0/0 ones, between E8.5 and E9, when upturn of PrP RNA expression should occur 32 . Absence of such a scenario in Prnp 0/0 /Sprn 0/0 mice might result from genetic adaptation 13 .
In conclusion, this study provides evidence for a yet controversial role of the Prion-related protein Shadoo during mouse embryogenesis and more generally in the control of cellular proliferation and differentiation of various tissues, including adult ones such as the mammary epithelium. It also highlights a potential embryonic antagonist action of PrP and Shadoo, revisiting alongside recent papers the current view of a potential biological redundancy between these two proteins. Generation and establishment of Sprn-knockout mice. A preparation of mRNA encoding a ZFN targeting the start of the Sprn ORF was purchased from Sigma-Aldrich Corp (CompoZrTM ZFN design). The ZFN target site was 5′-ctgtgacagctgttccgccangggcggccgcggaggcg-3′. ZFN mRNA was injected at 2 ng/ μl into 1 cell-fertilized FVB/NJ mouse eggs. Surviving injected eggs were transferred into pseudo-pregnant recipient mice. Tail-DNA analysis of the 29 resulting live pups was performed by PCR using oligonucleotides 5′-gccggccttacgcgtactcttga-3′ and 5′-gcctattcccaagccctgatac-3′, using the Promega GoTaq G2 Flexi DNA polymerase kit. PCR conditions were made of 40 amplifications cycles; 94 °C-30 s, 60 °C-30 s and 72 °C-30 s. The amplified 400 bp long DNA genomic region surrounding the ZFN target site was sequenced to search for putative mutations.

Methods
Transgenic founder mice were crossed with FVB/NJ mice to establish transgenic lines. Intercrosses between heterozygous mice were used to derive transgenic knockout lines.
Reproductive performance analyses. All examined females were on their first pregnancy or lactation.
Intercrosses between heterozygous mice and between homozygous mice were performed and the number and genotype of pups recorded. To assess embryonic lethality, pregnant FVB/NJ or Sprn 0/0 females, following mattings with males of the same genotype, were euthanized at different gestation stages. Implantation points were recorded. Development and vitality of the embryos were observed under binocular microscope.
Comparative analysis of growth curves of FVB/NJ and Sprn 0/0 pups was performed, using FVB/NJ lactating females as foster mothers. Lactation performance was indirectly assessed by analysis of growth curves of FVB/ NJ pups fed either by FVB/NJ or Sprn 0/0 females. Six pups were placed under each lactating females and 3 females  Table 6. Summary of observed phenotypes and trancriptomic alterations in Sprn 0/0 mice.

MG whole mount and immunohistological analyses. Experiments involved analyses of at least 3 mice
per genotype and condition. Inguinal MGs (4 th , left) were excised, spread on glass slides and fixed in Carnoy's fixative (100% EtOH, chloroform, glacial acetic acid; 6∶3∶1) for 4 h at room temperature, washed in 70% EtOH for 15 min, rehydrated in water, and stained in carmine alum (2% carmine and 5% aluminum potassium sulfate in water) overnight. Tissues were then gradually dehydrated through serial ethanol baths and cleared in xylene overnight. MGs were kept in Permount Mounting Media (Fisher Scientific) until images were captured. For immunohistological analyses, mouse MGs were collected at different physiological stages and processed as described 26 . Tissue sections were labeled using Alexa488-conjugated WGA, rhodamine-conjugated phalloidin (Thermo Fisher), or filipin (Sigma-Aldrich) diluted in PBS for 30 min at 4 °C. Neutral lipids and nuclei were counterstained with BODIPY 493/503 or Nile Red and DAPI (4′-6-diamidino-2-phenylindole), respectively. Slides were mounted with Vectashield (Vector Laboratories LTD, United Kingdom) and stored at 4 °C until observation. For some experiments, Tissue sections or MFGs were also labeled for actin with rhodamine-conjugated phalloidin (Molecular Probes), GM1 ganglioside with Alexa 594-conjugated CTxB (Invitrogen), or free CH using filipin (Sigma-Aldrich). Each experiment was performed at least twice and included negative control without primary antibody. At least 20 images were randomly acquired per condition tested and further analyzed. All images were analyzed using ImageJ 1.47q software (http://rsb.info.nih.gov/ij/). transcriptomic analyses. Transcriptomic analyses were performed using Affymetrix MouseGene 2,OSt microarray on the Institute Cochin Genom'ic platform. Each analyzed embryonic RNA sample derived from the equal mixing of 6 individual total RNA preparations from E7.5 embryos. Different embryonic samples were used for each mixture. Similarly, each analyzed placenta RNA sample derived from the equal mixing of 6 individual total RNA preparations from E14.5 placentas. Different individual samples were used for each mixture. Each MG RNA sample derived from the equal mixing of 3 individual total RNA preparations from G7.5 MGs. Different individual samples were used for each mixture. Three biological replicates per tissue and genotype were performed. Differentially expressed genes between FVB/NJ and FVB/NJ Sprn-knockout tissues were identified following normalization of the raw data, with fold chance ratios and p values determined as described in the results section. Data were analyzed by Gene set enrichment analysis (GSEA, http://software.broadinstitute.org/ gsea/index.jsp). Differentially expressed genes were clustered and classified in pathways and networks by using Ingenuity (http://www.ingenuity.com/).