SEARCH

SEARCH BY CITATION

Keywords:

  • Notch;
  • Val1744;
  • N1ICD;
  • cell fate specification;
  • feedback loop;
  • RBPJk;
  • mind bomb;
  • cardiovascular development;
  • hematopoiesis;
  • CNS development;
  • somitogenesis;
  • sense organs;
  • thymus;
  • pancreas

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. CONCLUSIONS
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Signaling through Notch receptors, which regulate cell fate decisions and embryonic patterning, requires ligand-induced receptor cleavage to generate the signaling active Notch intracellular domain (NICD). Here, we show an analysis at specific developmental stages of the distribution of active mouse Notch1. We use an antibody that recognizes N1ICD, and a highly sensitive staining technique. The earliest N1ICD expression was observed in the mesoderm and developing heart, where we detected expression in nascent endocardium, presumptive cardiac valves, and ventricular and atrial endocardium. During segmentation, N1ICD was restricted to the presomitic mesoderm. N1ICD expression was also evident in arterial endothelium, and in kidney and endodermal derivatives such as pancreas and thymus. Ectodermal N1ICD expression was found in central nervous system and sensory placodes. We found that Notch1 transcription and activity was severely reduced in zebrafish and mouse Notch pathway mutants, suggesting that vertebrate Notch1 expression is regulated by a positive feedback loop. Developmental Dynamics 236:2594–2614, 2007. © 2007 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. CONCLUSIONS
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Notch signaling regulates cell fate specification and tissue patterning through local cell interactions. Vertebrate Notch genes encode an evolutionarily conserved group of receptor proteins (Notch1–Notch4 in mammals), containing a large extracellular domain that interacts with membrane-bound DSL ligands (for Delta, Serrate and Lag2), expressed in neighboring cells. Upon ligand-receptor binding, ubiquitin ligases such as mind bomb (Itoh et al.,2003) trigger endocytosis of the ligand intracellular domain. This process leads to a conformational change in Notch (Parks et al.,2000) that permits two consecutive protease cleavage events that result in the generation and nuclear translocation of the biologically active Notch intracellular domain (NICD; reviewed in Kopan,2002). The last cleavage step, γ-secretase–dependent, occurs between amino acids glycine 1743 (G1743) and valine 1744 (V1744; Schroeter et al.,1998), located within the Notch transmembrane domain (De Strooper et al.,1999) and is critical for Notch function in vivo (Huppert et al.,2000).

In the nucleus NICD binds to the transcriptional repressor CSL (acronym of vertebrate RBPJK/CBF1, DrosophilaSu (H), and C. elegansLag1; Kopan,2002), and converts it into a transcriptional activator (reviewed by Mumm and Kopan,2000). The NICD/CSL target genes include those encoding basic helix-loop-helix (bHLH) transcription factors of the HES (Davis and Turner,2001) and Hey/HRT/Herp families (Iso et al.,2003).

The best understood effect of Notch signaling is the diversification of cell fates within an equivalence group, in which all cells have the potential to adopt two alternative fates. In this situation Notch acts by means of a mechanism termed lateral inhibition, whereby a single cell within an equivalence domain expresses high Delta levels and adopts the primary fate (i.e., neuroblast) and, by activating Notch, inhibits surrounding cells from adopting this fate. A Notch-mediated negative feedback loop down-regulates ligand expression in signal-receiving cells (Heitzler et al.,1996). Examples of lateral inhibition include neuroblasts segregation in Drosophila (Skeath and Thor,2003), vertebrate early neurogenesis (Chitnis et al.,1995; de la Pompa et al.,1997) and sensory hair cell formation in the vertebrate inner ear (Eddison et al.,2000). Notch signaling can also generate contiguous domains of cells with the same fate, an embryonic field. This signaling mechanism, termed lateral induction (Lewis,1998), occurs in flies during wing margin boundary formation (Panin et al.,1997) and in vertebrates during induction of proneural domains in the ear (Daudet and Lewis,2005), limb bud margin formation (reviewed in Irvine and Vogt,1997), somite boundary formation (reviewed in Lewis,1998), and cardiac valve development (Timmerman et al.,2004). In this case, Notch activation promotes ligand production by means of a positive feedback loop.

Functional studies indicate that Notch1 is likely the most relevant receptor of the pathway because of its involvement in a great variety of developmental processes (reviewed in Radtke and Raj,2003; Weinmaster and Kintner,2003; Radtke and Clevers,2005; Yoon and Gaiano,2005; Bolos et al.,2007). The Notch1 transcription pattern in embryonic and fetal tissues has been extensively characterized (Weinmaster et al.,1991; Del Amo et al.,1992; Reaume et al.,1992; and Supplementary Figure S1, which can be viewed at http://www.interscience.wiley.com/jpages/1058-8388/suppmat). However, Notch receptor activation depends on where ligands and receptor-modifier proteins such as fringe (Bruckner et al.,2000) are expressed; transcription thus does not necessarily reflect where Notch1 will be activated. Two methods have been used to detect Notch activity reliably. The first is based on the use of Notch reporter mice (Ohtsuka et al.,1999; Duncan et al.,2005; Souilhol et al.,2006; Vooijs et al.,2007), a sophisticated strategy that, nevertheless, has four important limitations, including low reporter sensitivity, lack of receptor specificity, excessive marker perdurance, which masks dynamic Notch activation (see Souilhol et al.,2006), and potential activation of the reporter by other signaling pathways. A second approach attempts to detect Notch activity by immunostaining. Thus, N1ICD has been detected at relatively late stages of kidney (Cheng et al.,2003), central nervous system (CNS; Tokunaga et al.,2004; Barsi et al.,2005), inner ear (Murata et al.,2006), and heart development (Watanabe et al.,2006), and retinal angiogenesis (Hellstrom et al.,2007; Hofmann and Luisa Iruela-Arispe,2007), although we still lack a detailed study on Notch1 activation pattern during embryogenesis.

We report the development of a simple, reliable and highly sensitive immunohistochemistry method to detect N1ICD activity using an antibody that recognizes the conserved V1744 amino acid exposed by γ-secretase cleavage (Schroeter et al.,1998), together with antigen retrieval and tyramide amplification of signal. We analyze Notch1 signaling activity in the mouse embryo during gastrulation and at specific stages of organogenesis. We show Notch1 activation in cardiac and arterial endothelium, presomitic mesoderm (PSM), endodermal derivatives, CNS, and sense organ development. Our results show a complex pattern of Notch1 activity in the embryo, including tissues for which severe developmental alterations have been reported. We find that both Notch1 expression and activity is greatly reduced in mouse RBPJk and in zebrafish mind bomb mutants, suggesting that, as described for its ligands, a positive feedback loop may regulate Notch1 transcription in vertebrates.

RESULTS AND DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. CONCLUSIONS
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Elegant work by Schroeter et al. demonstrated that minimal amounts of NICD are enough to activate a HES1-dependent reporter in cells, explaining why it has been so hard to detect N1ICD in the nucleus in vivo (Schroeter et al.,1998). Nevertheless, different reports have documented Notch1 activity during kidney (Cheng et al.,2003), CNS (Tokunaga et al.,2004; Barsi et al.,2005), inner ear (Murata et al.,2006), somite (Huppert et al.,2005; Morimoto et al.,2005), and heart (Watanabe et al.,2006) development, and retinal angiogenesis (Hellstrom et al.,2007; Hofmann and Luisa Iruela-Arispe,2007).

To determine the sites of Notch1 activation during development, we have developed a sensitive staining protocol using an antibody against the conserved N-terminal V1744 amino acid of N1ICD exposed by γ-secretase cleavage (Fig. 1A; Schroeter et al.,1998; De Strooper et al.,1999). This antibody is N1ICD-specific and does not recognize NICD produced from other Notch proteins such as Notch4, which is also cleaved at a V residue (Saxena et al.,2001; Huppert et al.,2005). Amino acid sequence comparison among Notch1 receptor homologues revealed that these G and V residues are evolutionarily conserved from C. elegans glp1 to human Notch1 (Fig. 1A; Supplementary File S1), although an additional amino acid is located between these two residues in some species (zebrafish Notch1b, Drosophila Notch, and C. elegans lin12; Fig. 1A), which may affect antibody recognition. In this regard, attempts to stain zebrafish embryos with V1744 antibody were unsuccessful (not shown), despite the presence of the conserved γ-secretase cleavage site in the zebrafish Notch1a homologue (Fig. 1A).

thumbnail image

Figure 1. Notch1 activity in the gastrulating mouse embryo. A: Amino acid sequence comparison of various Notch1 homologues spanning the γ-secretase cleavage site. The blue arrowhead indicates the S3 cleavage site preceding V1744, the amino terminal amino acid of N1ICD. This S3 cleavage site is conserved in mouse, human, and chick Notch1; zebrafish Notch1a; and C. elegans glp1, but not in Drosophila Notch and zebrafish Notch1b, where it is shifted one amino acid (green arrowhead). Boxes mark conserved amino acid stretches; identical amino acids appear in red. TM, transmembrane region; NICD, Notch intracellular domain. B: Western blot of embryonic day (E) 9.5 and E13.5 mouse embryos probed with anti-N1ICD (V1744) antibody, showing a 110-kDa N1ICD band in different tissues (top). Western blot from several mammary tumor cell lines (bottom). Note that only the highly invasive MDA-MB231 line expressed N1ICD. SMC-3 is a 140-kDa cohesin used as loading control. C: Planes of sections shown in D–K. D,E: Nuclear N1ICD staining (red) in the intraembryonic mesoderm region (iem, arrow) of the E7.0 wild-type (wt) embryo. Note that headfold region (hf) and embryonic ectoderm (ee) are negative. F: Detail of allantois with developing vessel expressing N1ICD (arrowhead). G: At E7.5, N1ICD is expressed in primitive endocardium (ec), cells of the cephalic mesenchyme (cm), amnion (am), and allantois (al). H: Detail of headfold region showing N1ICD staining in primitive endocardium (ec). ne, neuroectoderm; cm, cephalic mesenchyme. I: Detail showing N1ICD expression in blood island (bli) precursors. J: At E7.7, N1ICD is expressed in endocardium, cephalic mesenchyme and amnion (am). K: Detail of J, showing N1ICD expression in endocardium and primordium of dorsal aorta (da). Nuclei are counterstained with 4′,6-diamidine-2-phenylidole-dihydrochloride (DAPI, blue). Scale bars = 150 μm in D,G,J, 60 μm in E,F,H,I,K.

Download figure to PowerPoint

Western blot analysis detected a specific 110-kDa N1ICD band in mouse embryonic and fetal tissues (Fig. 1B). N1ICD expression appeared stronger in PSM, aorta/genital ridge/mesonephros (AGM) region, CNS, liver, and lungs and weaker in heart and limbs (Fig. 1B). N1ICD expression was also observed in chicken tissues (data not shown). Western blot of human mammary gland tumor cell lines showed an N1ICD band exclusively in MDA-MB231 cells (Fig. 1B), a highly transformed and invasive cell line (Stylianou et al.,2006).

To confirm the specificity of the signal obtained with the V1744 antibody, we stained Notch1-deficient embryos, which showed no sign of staining (Supplementary Figure S2A–C). Developmental analysis of Notch1 mutants showed that they are viable until embryonic day (E) 10.5 (Swiatek et al.,1994; Conlon et al.,1995). Nevertheless, to discard the possibility that Notch1 mutants show no N1ICD staining because they were dying, we analyzed tissue organization by hematoxylin/eosin (H&E) staining (Supplementary Figure S3), and the general structure of the vasculature by CD31/PECAM1 staining (Supplementary Figure S4), and found both of them to be relatively normal. We also examined whether there was increased apoptosis in Notch1 mutants by terminal deoxynucleotidyl transferase–mediated deoxyuridinetriphosphate nick end-labeling (TUNEL) assay (Supplementary Figure S5) and found no difference, as previously demonstrated (Conlon et al.,1995). These data indicate that E9.5 Notch1 mutants show no sign of tissue degeneration. The lack of V1744 staining in Notch1 mutants demonstrates that they do not express detectable N1ICD. Lastly, to discard background signal due to secondary antibody, we incubated sections without primary antibody but with secondary antibody followed by avidin/biotin-horseradish peroxidase (HRP) and tyramide-Cyn3 amplification. Only weak and dispersed background non-nuclear staining was observed (Supplementary Figure S2D–F).

We have examined N1ICD expression in the gastrulating mouse embryo and in specific organ systems during development. We have detected different levels of Notch1 activity in derivatives of the three germ layers (summarized in Table 1 and Supplementary Figure S6), where Notch1 has been shown to be required. We have also observed reduced Notch1 expression and activity in mouse and zebrafish Notch pathway mutants, suggesting that vertebrate Notch1 transcription is regulated by a positive feedback loop.

Table 1. N1ICD Expression Levels in the Embryo and Correlation With Mutant Phenotypesa
TissueStageN1ICD activityNotch1 LOF or GOF phenotype
  • a

    Tissues were grouped in five classes based on intensity ranges. The first group includes tissues with intensity below 160. The second group includes intensities between 160 and 170. The third group includes intensities between 170 and 190. The fourth group includes intensities between 190 and 200. The fifth group includes intensities higher than 200. The results are from three to four embryos of each stage.

  • *

    Classification of this tissue attending to an average of all nuclei. Intensity analysis reveals marked regional differences in N1ICD expression levels for a given tissue (see Supplementary Figure S6).

Hematopoietic precursorsE7.5–E10.54Robert-Moreno et al.,2005
  Watanabe et al.,2006
EndocardiumE7.5–E14.52*Timmerman et al.,2004
Arterial endotheliumE7.5–E10.55Krebs et al.,2000
 E11.52 
 E13.5–E14.5 
Arterial smooth muscleE7.5–E10.5Not detected
 E11.52 
 E13.5–E14.53* 
Central nervous systemE8.5–E14.53Reviewed in Yoon and Gaiano,2005
Peripheral nervous systemE9.5–E14.53Gaiano et al.,2000
Inner earE9.5–E14.52Kiernan et al.,2005
Eye (neural retina)E11.51Jadhav et al.,2006; Yaron et al.,2006
 E13.5–E14.52 
NoseE11.5–E13.54Not detected
Presomitic mesodermE8.5–9.53Conlon et al.,1995
Urogenital tissuesE9.5–E14.53Cheng et al.,2003
Epithelial cellsE13.5–E14.54Vauclair et al.,2005
Pancreatic acinesE13.52Murtaugh et al.,2003
ThymocytesE13.5–E14.55Washburn et al.,1997

Mesodermal N1ICD Expression During Gastrulation

The earliest time at which we detect N1ICD expression is in the primitive streak stage embryo at around E7.0, when we found nuclear N1ICD staining in the nascent mesoderm (Fig. 1C–E), while no N1ICD signal is observed in ectoderm. At E7.5, N1ICD is expressed in developing vessels of the allantois (Fig. 1F). In the embryo proper, gastrulation is advanced and the mesoderm migrates forward to lie beneath the entire embryonic ectoderm. N1ICD staining is particularly strong in cardiac crescent cells that will give rise to the primitive endocardium—the inner lining of the heart—and in the amnion (Fig. 1G,H). It is also detected in a few cells of the cephalic mesenchyme (Fig. 1G). At E7.5–E8.0, N1ICD stains predominantly hematopoietic cells in the blood islands of the extraembryonic mesoderm (Fig. 1I). Within the embryo, N1ICD is expressed in scattered cells of the cephalic mesenchyme, the primordium of the dorsal aorta, and in endocardial cells (Fig. 1J,K). This early N1ICD expression demarcates mesodermal territories that will give rise to cardiovascular lineages and cephalic mesenchyme.

Notch1 Activity During Cardiac Valves and Chamber Development

At early stages of cardiogenesis, we detected N1ICD staining throughout the endocardium and as development proceeds, staining was confined to specific cardiac regions. Thus, in the E8.5 wild type (wt) embryo, N1ICD ventricular expression was restricted to endocardial cells adjacent to the myocardium, while distal endocardial cells expressed low N1ICD levels (Fig. 2A,B). The diagrams in Figure 2C indicate the approximate planes of sections shown in Figure 2D–T.

thumbnail image

Figure 2. Endocardial Notch1 activity precedes valve primordia formation and trabeculation. A: N1ICD expression in ventricular endocardium of an embryonic day (E) 8.5 wild-type (wt) embryo. B: Detail showing N1ICD expression in endocardial cells (arrows) adjacent to myocardium. The distal endocardial cells express low N1ICD levels (arrowhead). v, ventricle. C: Schematic diagrams depicting E8.5–E10.5 heart, showing section planes analyzed. Outflow tract (OFT, light blue), left (brown) and right (orange) ventricles, atrioventricular canal (AVC, yellow), atria (dark green), and arterial (red) and venous (dark blue) poles. D,E: N1ICD expression in endocardial cells of the AVC (arrows). Arrows in (E) point to AVC endocardium where N1ICD expression is highest. F,G: N1ICD expression in endocardium of the OFT (arrows in G) at E9.5. H: HRT2 mRNA expression in the OFT endocardium (arrow) at E9.5. I,L: HRT2 transcription in the endocardium of the AVC (arrows in L). Note HRT2 expression in myocardium (arrowheads in I, L). J,K: N1ICD expression in endocardium of the AVC (arrows in K). M: HRT2 expression is reduced in the AVC endocardium of RBPJk mutants (arrow) but not in myocardium (arrowhead). N,O: At E10.5, N1ICD expression persists in specific endocardial cells of the OFT (arrows), while neighboring cells are negative (arrowhead). P,Q: AVC endocardial cells express N1ICD (Q, arrow), but transformed mesenchyme cells do not (Q, arrowhead). RT: N1ICD expression in endocardium of atrium (at) and ventricles at E9.5. S,T: Details of right (rv) and left (lv) ventricles. Note predominant N1ICD expression in endocardium at the base of trabeculae (arrows) and reduced N1ICD expression in distal endocardium (arrowheads). Signal is stronger in the right ventricle. Scale bars = 30 μm in A,D,F,I,J,M,N,P,R, 100 μm in B,E,G,H,K,L,O,Q,S,T.

Download figure to PowerPoint

During heart development, valve formation is initiated at approximately E9.0 in the mouse, when in response to regionalized myocardial signals, a subset of endocardial cells which overlie the atrioventricular canal (AVC) and outflow tract (OFT) regions, undergo epithelial-to-mesenchyme transition (EMT; Eisenberg and Markwald,1995). Once transformed, the mesenchymal cells proliferate and invade the cardiac jelly, a basement membrane produced by the myocardium. After EMT, endocardial cushion cells elongate and undergo remodeling to give rise to the mature valve leaflets (Person et al.,2005). At E9.0–E9.5, N1ICD expression in prevalve territories was highest in endocardial cells of the AVC (Fig. 2D,E) and OFT region (Fig. 2F,G). At E9.5, transcription of the Notch target HRT2 (Fig. 2H) coincided with N1ICD staining in the OFT endocardium (Fig. 2F,G). N1ICD was also detected in the AVC endocardium (Fig. 2J,K; Supplementary Figure S7 and Supplementary Movie S1), with a distribution similar to that of HRT2 (Fig. 2I,L). HRT2 transcription was reduced in the AVC canal endocardium but not in the myocardium of RBPJk mutants (Fig. 2M), suggesting that only HRT2 expression in the endocardium is dependent on Notch signaling. At E10.5, N1ICD expression persisted in specific endocardial cells of the OFT (Fig. 2N,O) and AVC (Fig. 2P,Q). In contrast, no N1ICD expression was detected in mesenchymal cells within the cushion (Fig. 2Q), indicating that cells undergoing EMT down-regulate Notch1. These results indicate that Notch1 activity labels a specific endocardial population that will form the heart valve primordium. Manipulation of Notch activity in mouse and zebrafish embryos has demonstrated the essential role of Notch in cardiac valve formation (Timmerman et al.,2004; Noseda et al.,2004; Beis et al.,2005; Kokubo et al.,2005; Watanabe et al.,2006). In mouse Notch1 and RBPJk mutants, production of the myocardial EMT-inducing signal transforming growth factor β2 is affected (Timmerman et al.,2004), suggesting that endocardial Notch1 activity is required for endocardium–myocardium communication during valve formation.

At E9.5, N1ICD was detected throughout the atrial endocardium (Fig. 2R). In the ventricles, N1ICD was expressed unequally in the endocardium of developing trabeculae, the highly organized sheets of cardiomyocytes that form muscular ridges lined by endocardial cells (Ben-Shachar et al.,1985). Trabeculae, the first indication of ventricular differentiation (Sedmera et al.,1997), form as a result of interactions between myocardium and endocardium at the end of cardiac looping. N1ICD expression was stronger in endocardium at the base of trabeculae and weaker in distal endocardium (Fig. 2S,T; Supplementary Figure S8 and Supplementary Movie S2). The relevance of Notch1 expression in ventricular endocardium has been shown by recent work demonstrating the role of Notch1 in trabeculation (Grego-Bessa et al.,2007). N1ICD staining appeared to be stronger in the right ventricle, suggesting that Notch signaling responds to cues from the different heart precursor lineages (Kelly,2005).

Figure 3A shows the approximate planes of sections shown in Figure 3B–S. At E11.5, the heart showed strong nuclear N1ICD staining in the OFT (Fig. 3B,C) and AVC endocardium (Fig. 3D,E). The trabecular endocardium of the left (Fig. 3B,F) and right ventricles (Fig. 3B,G) expressed high N1ICD levels. At E13.5, N1ICD staining persisted in the semilunar valve endocardium (Fig. 3H,I), but decreased in the arterial endothelium component (Fig. 3I, thick arrow) and was detected in smooth muscle cells of the arterial wall (Fig. 3I, arrowhead). N1ICD also stained the endocardial lining of the developing trabeculae in the left atrium (Fig. 3J,K), suggesting that Notch1 signaling may be involved in atrial development. In the mitral valve N1ICD is expressed in the endocardium but staining is reduced in the valve component facing the atrium (Fig. 3L,M). At E14.5, the restriction of N1ICD expression to the endocardial component of the semilunar valve persist (Fig. 3P,Q) and, in the tricuspid valve, N1ICD is expressed in the endocardium lining the valve (Fig. 3R,S). This expression pattern probably reflects the shortening of the conal/proximal OFT territory (Webb et al.,2003), rather than a true local restriction of Notch1 activity. At these stages, N1ICD endocardial staining was also prominent in the right ventricle (Fig. 3N). These data suggest that cardiac Notch1 activity responds to patterning signals that govern valve, ventricular, and atrial development. The recent association between NOTCH1 haploinsufficiency and aortic valve disease in humans (Garg et al.,2005) suggests that the generation of mice with specific Notch targeted mutations may be an excellent model to study cardiovascular disease. We have not detected N1ICD expression in myocardium at any stage examined; this finding is consistent with the observation that HRT2 myocardial transcription is unaffected in RBPJk (Fig. 2M) and Notch1 mutants (not shown), suggesting that HRT2 expression in this tissue does not depend on Notch.

thumbnail image

Figure 3. Endocardial Notch1 activity in cardiac valves, and atrial and trabecular development. A: Scheme showing the approximate plane of sections shown in B–S. B,C: General view of an embryonic day (E) 11.5 heart section at the level of the outflow tract (OFT). C: Nuclear N1ICD staining is particularly strong in the OFT endocardium (arrow). DG: General view of a section at the atrioventricular canal (AVC) level with strong endocardial N1ICD expression (E, arrows). Trabecular endocardium of the left (F) and right (G) ventricles expressed high N1ICD levels (arrows). H,I: At E13.5, N1ICD staining persists in endocardium of the semilunar valve (slv; I, arrow), but decreased in arterial endothelium (I, thick arrow), and was detected in smooth muscle cells of the arterial wall (I, arrowhead). J,K: Detail of N1ICD staining in endocardium lining the developing trabeculae in the left atrium (la; arrow in K). L: General view of an E13.5 heart at the AV valves level. M: Detail showing N1ICD expression in endocardium of the mitral valve (mitv; arrows) and reduced staining in the valve component facing the atrium (arrowhead). N: Detail of N1ICD endocardial staining in the right ventricle (arrow). O: N1ICD staining was also strong in the developing coronary vessels of the left ventricle (arrow). P–S: At E14.5, N1ICD was restricted to the endocardial component of the valves. P: Morphology of the semilunar valve. Q: Note N1ICD endocardial expression (arrow) and its exclusion from the arterial endothelium (arrowhead) of the semilunar valve. R: Morphology of the tricuspid valve (tricv). S: N1ICD expression in endocardial lining of the tricuspid valve (arrow). Scale bars = 250 μm in B,D, 200 μm in H,L, 100 μm in C,E–G, 80 μm in I–K,M–S.

Download figure to PowerPoint

N1ICD expression was also detected in the developing coronary vessels (Fig. 3O), as staining with CD31/PECAM and V1744 antibodies shows (Supplementary Figure S9A–F). As it occurs in the rest of the vasculature (Lawson et al.,2001), Notch activity in the coronaries may define the arterial cell fate (see below).

Notch1 Activity in Vascular Development

The vascular system is one of the first organs to function in vertebrates. It is composed of arteries and veins that are defined by the direction of blood flow and that differ anatomically and genetically from the earliest stages of angiogenesis (Wang et al.,1998a). Expression of Notch genes during vascular development and patterning is restricted to arterial vessels (Villa et al.,2001), and Notch1 appears to be the key receptor for transduction of Notch signals in endothelial cells (Krebs et al.,2000). Notch is involved in the establishment of arterial endothelial fate (reviewed in Gridley,2001). Different lines of evidence demonstrate that Notch represses the venous fate in developing arterial cells (Zhong et al.,2001; Lawson et al.,2002) and recent findings indicate that the COUP-TFII transcription factor represses Notch signaling in venous endothelium, suggesting that venous identity is also under genetic control and not derived by a default pathway (You et al.,2005).

We examined Notch1 activation during vascular development, focusing on major blood vessels such as the aorta, vena cava, cardinal veins, intersomitic vessels and meningeal plexus. In the E8.5 wt embryo, a strong N1ICD signal was observed in the dorsal aorta (Fig. 4A,B), while vein cells were negative (Fig. 4B). The Delta4 ligand, crucial for arterial cell fate specification (Duarte et al.,2004; Krebs et al.,2004), was also transcribed in the dorsal aorta (Fig. 4C; Shutter et al.,2000). The endothelium connecting the dorsal aorta with the second arterial arch also expressed N1ICD (Fig. 4D). At E9.5, restriction of N1ICD expression to the dorsal aorta and exclusion from veins persisted (Fig. 4E). Likewise, Delta4 was transcribed in arterial endothelium (Fig. 4F), whereas Delta1 was found in venous endothelium (Fig. 4G). The target gene HRT2 was transcribed throughout the arterial endothelium (Fig. 4H). The EphrinB2/EphB4 signaling pathway is also required for arterial–venous specification (Wang et al.,1998a), and functional studies in zebrafish have placed EphrinB2 downstream of Notch during establishment of arterial identity (Lawson et al.,2002). EphrinB2 is transcribed in arterial endothelium of the E8.5–E9.5 embryo (Fig. 4I,K). In RBPJk mutants, EphrinB2 transcription in arteries was greatly reduced (Fig. 4J,L), supporting a functional link between both pathways.

thumbnail image

Figure 4. Notch1 activity in vascular development and hematopoiesis. A: General view of a cross-section of an embryonic day (E) 8.5 wild-type (wt) embryo. B: Detail showing N1ICD expression in endothelial cells of dorsal aorta (da, arrow), while cells of the primary head vein (phv, arrowhead) are negative. nt, neural tube. C: Detail showing predominant Delta4 transcription in dorsal aorta. D: Detail showing N1ICD staining in the endothelium (arrows) connecting the dorsal aorta with the second arterial arch (II). E: Detail of an E9.5 section showing N1ICD expression in dorsal aorta (arrow), and lack of expression in anterior cardinal vein (acv, arrowhead). F: Delta4 transcription in E9.5 arterial endothelium (arrows). G: Delta1 transcription in E9.5 venous endothelium (arrows). H: HRT2 transcription in E9.5 arterial endothelium (arrow). I,K: EphrinB2 transcription in arterial endothelium (arrows indicate dorsal aorta) of an E8.5–E9.5 wt embryo. J,L: EphrinB2 transcription is greatly reduced in arterial endothelium of RBPJk mutants. M,N: N1ICD expression in aorta endothelium of anterior AGM region at E9.5–E10.5 (arrows). M: The thick arrow in M indicates the urogenital ridge. O,P: N1ICD expression in endothelium and hematopoietic clusters in posterior AGM region at E9.5–E10.5. Note N1ICD staining in cluster cells closer to the aorta endothelium (thick arrow in O, P), but not in cells budding from the aorta (arrowheads in O, P). Q: At E11.5, N1ICD expression in arterial endothelium begins to disappear, while smooth muscle cells are stained (arrow) and endothelium of anterior cardinal vein (acv) is negative. da, dorsal aorta. R: In E13.5 embryos, N1ICD expression has disappeared from aorta endothelium (arrowhead), while the smooth muscle cells surrounding the artery express N1ICD (arrow). S: Smooth muscle cells showing nuclear N1ICD staining (green, arrow) and cytoplasmic SM-actin staining (red, thick arrow). The arterial endothelium is negative (arrowhead). T: Arterial endothelium stained with anti-CD31/PECAM antibody (red, arrowhead). U,V: N1ICD is expressed in the vasculature of the liver primordium (arrows) at E9.5–E10.5. X,Y: At E13.5, N1ICD stains the vasculature of the meningeal plexus (X, arrows) and arteries of pulmonary vasculature (Y, arrows). nt, neural tube. All are transverse sections. Scale bars = 200 μm in A, 35 μm in B–D,M–Q,U, 50 μm in E–L,Y, 100 μm in R, 45 μm in S,V,Y, 40 μm in T, 80 μm in X.

Download figure to PowerPoint

N1ICD expression in arterial endothelium also reflects its requirement for the establishment of definitive hematopoiesis (Kumano et al.,2003), in which Notch-GATA2 (Kumano et al.,2001; Robert-Moreno et al.,2005) and Notch-Runx1 pathways (Burns et al.,2005) regulate hematopoietic stem cell (HSC) generation and expansion. Definitive hematopoiesis in the mouse embryo originates from the aortic floor in the para-aortic splanchnopleura (P-Sp; Cumano et al.,2001) and the AGM region (Medvinsky and Dzierzak,1996), in close association with endothelial cells. We found N1ICD expression in the dorsal aorta at E9.5–E10.5 (Fig. 4M,N), in a pattern corresponding to that of Delta4, Jag1, and Jag2 transcription in the AGM (Robert-Moreno et al.,2005), suggesting that these ligands activate Notch1 in this region.

At E9.5, N1ICD was expressed in hematopoietic clusters of the dorsal aorta (Fig. 4O, thick arrow), where it is involved in determination of the hematopoietic lineage (Robert-Moreno et al.,2005). At E10.5, in parallel with the progression of hematopoietic differentiation (Duncan et al.,2005), N1ICD staining concentrated to the basal cell of the cluster (Fig. 4P, thick arrow), in which Notch1-mediated GATA2 activation occurs (Robert-Moreno et al.,2005). Thus, N1ICD expression in the AGM region identifies the HSC compartment. At E11.5, endothelial N1ICD expression in the dorsal aorta begins to diminish (Fig. 4Q), and by E13.5, only smooth muscle cells surrounding the endothelium are stained (Fig. 4R), as costaining with N1ICD and smooth muscle actin (SMA) antibodies revealed (Fig. 4S). The arterial endothelium surrounded by the N1ICD- and SMA-positive layer expresses CD31/PECAM (Fig. 4T and Supplementary Figure S9A,B). Although Notch1 transcription has been reported in smooth muscle cells of regenerating endothelium (Lindner et al.,2001; Alva and Iruela-Arispe,2004), this is the first report of Notch1 activity in vascular smooth muscle cells during endothelial development. In agreement with our finding, in vitro studies have reported the activation of smooth muscle α-actin by NICD/RBPJK (Noseda et al.,2006).

We also detected endothelial N1ICD expression in vessels of the liver primordium at E9.5 and E10.5 (Fig. 4U,V). At E13.5, N1ICD also stained the meningeal plexus vasculature (Fig. 4X) and arteries of pulmonary vasculature (Fig. 4Y). N1ICD expression thus delineates the arterial endothelium during embryogenesis.

Notch1 Activity During Somitogenesis and in Endoderm-Derived Organs

Somites.

Notch signaling is essential for somite formation (Conlon et al.,1995). Whole-mount in situ hybridization (WISH) studies showed strong Notch1 transcription in the PSM and somites (Del Amo et al.,1992; Reaume et al.,1992). We found strong N1ICD expression in PSM cells from E8.5 onwards (Fig. 5A,B), with a pattern similar to Delta1 (Fig. 5C; Bettenhausen et al.,1995). Recently, two N1ICD expression patterns have been described in the PSM of the E10.5 mouse embryo (Huppert et al.,2005; Morimoto et al.,2005); the first is characterized by N1ICD expression in a broad area around somites S-1 and S0. A second pattern is defined by N1ICD accumulation in an anterior region coinciding with the cleft-forming area between S-1 and S0, and a wide area of N1ICD accumulation in the mid-PSM (Huppert et al.,2005). Embryos stained at E9.5 showed no N1ICD staining in the somites (Fig. 5D,E), in contrast to Delta1, which is strongly expressed in the posterior somite half (Fig. 5F; Bettenhausen et al.,1995), where it presumably activates other Notch receptors. Sagittal sections of E9.5 embryos revealed two regions of N1ICD expression in the PSM. The first one spans the cleft-forming area between S-1 and S0 (Fig. 5G, thick arrow). The second one the posterior PSM (Fig. 5G, arrow). In between both domains, there is a region that shows no detectable staining (Fig. 5G, bracket). The pattern of N1ICD activity in PSM is coincident with that of the glycosyltransferase fringe (Morimoto et al.,2005), supporting the idea that Notch1 activation in the PSM reflects the cyclic expression of its modulator lunatic fringe (Dale et al.,2003).

thumbnail image

Figure 5. N1ICD expression in somitogenesis and in different organ systems. A–W: Somitogenesis (A–G), urogenital system (H–N), vibrissae (O), ear epithelium (P), esophagus (Q), pancreas (R,S), thymus (T–W). A,B: General view (A) and detail (B) of a transversal section at embryonic day (E) 8.5 showing the neural tube (nt), heart (h), presomitic mesoderm (PSM), and dorsal aorta (da) stained for N1ICD. B: Arrow points to a N1ICD-positive cell. The section level is similar to that shown in Figure 2A. C: Sectioned whole-mount in situ hybridization (WISH) showing Delta1 transcription in PSM. D: General view of the tail region of an embryonic day (E) 9.5 embryo. The arrow indicates the PSM. E: Detail showing N1ICD expression in the neural tube (arrow) and lack of expression in the somites (arrowheads). F: Delta1 transcription in the somites (s, arrows) and neural tube (nt, thick arrow) of an E9.5 embryo. G: Sagittal section of the tail region of a E9.5 embryo showing N1ICD expression in the posterior PSM (arrow) and around the cleft-forming area between S-1 and S0 (thick arrow), with almost no expression in the PSM between these two regions (bracket). H: Transverse section of the AGM region showing N1ICD expression in the nephric tubules (arrows), while the nephric ducts are negative (arrowheads). I: Sagittal section of the AGM region with expression in the nephric tubule (arrow) and no staining in nephric duct (arrowhead). Note no N1ICD expression in somites (thick arrowhead). J: Sectioned WISH showing Delta1 expression in nephric tubules (arrows) at E9.5. K: At E10.5, N1ICD is expressed in cells of the mesonephric tubule (arrow) and the mesonephric duct is negative (arrowhead). Note expression in the dorsal aorta (da). L: General view of the developing renal capsule at E13.5. M: Detail showing N1ICD expression restricted to metanephric vesicles (arrow), while metanephric ducts are negative (arrowheads). N: At E14.5, N1ICD expression localizes to primitive glomerulae (arrows). O: N1ICD expression in the hair plug (arrow) and dermal condensation primordia of vibrissa (thick arrow). P: N1ICD expression in ear epidermis (arrow). Q: N1ICD expression in esophagus epithelium (arrow) and lack of signal in trachea (arrowhead). R: General view of a transverse section of the pancreas at E13.5. S: Detail showing N1ICD expression in the pancreatic artery endothelium (arrow) and acini (thick arrow). T: General view of a transverse section of E13.5 thymus. U: Detail showing N1ICD expression in thymocytes (arrow) surrounded by epithelial cells (arrowhead) that do not express N1ICD. V: Detail of a thymic lobule at E14.5, showing N1ICD-stained thymocytes (arrow). W: Detail showing proliferating BrdU-positive cells in at E14.5. Scale bars = 30 μm in A,E,F,H,–K,P–R, 15 μm in B,C, 200 μm in D,G, 60 μm in L, 40 μm in M,N, 50 μm in O, 10 μm in S, 35 μm in T, 15 μm in U, 20 μm in V,W.

Download figure to PowerPoint

Kidney.

Xenopus Notch1 mRNA is transcribed in the epithelial pronephric tubules and is required for cell fate determination during pronephros development (McLaughlin et al.,2000). Inhibition of γ-secretase activity in mouse metanephroi impairs podocyte and proximal tubule formation (Cheng et al.,2003); Notch2 inactivation leads to defective kidney development (McCright et al.,2001), although no specific role has been reported for Notch1 in nephrogenesis. In the developing mouse kidney, we observed N1ICD expression at E9.5 in the nephric tubule, but not in nephric duct (Fig. 5H,I); similarly to Delta1 (Fig. 5J) and Jag1 transcription (not shown) at this stage. At E10.5, N1ICD was expressed in mesonephric tubule cells (Fig. 5K). At E13.5, N1ICD expression was restricted to metanephric vesicles, while metanephric tubules and ducts were negative (Fig. 5L,M). At E14.5, we detected N1ICD expression in the comma and S-shaped bodies of kidney glomerulae (Fig. 5N).

Epithelia.

At E13.5, N1ICD was expressed in the hair plug and in the dermal condensation primordia of vibrissa (Fig. 5O), as well as in ear epithelium (Fig. 7P) and esophagus epithelia (Fig. 5Q). These patterns were consistent with the proposed role for Notch in hair follicle development and homeostasis (Vauclair et al.,2005) and epithelial self-renewal (Radtke and Clevers,2005).

thumbnail image

Figure 7. Notch1 activity in peripheral nervous system (PNS) and sense organ development. (A–F) PNS, (G–P) inner ear, (Q–T) eye, (U–X) nose. A: General view of the central nervous system (CNS) at embryonic day (E) 10.5, at the level of the trigeminal ganglion. IV, fourth ventricle. Note strong N1ICD signal in ventricular zone (arrow). B: Detail showing N1ICD expression in trigeminal ganglion (V, arrow) and lack of signal in the primary head vein (phv, arrow). C,D: N1ICD staining in trigeminal ganglion at E11.5 with detail in (D). E,F: General view of spinal cord at E13.5, showing N1ICD expression in the ventricular zone (arrow) and dorsal root ganglia (DRG, F), where N1ICD-expressing cells (arrow) and nonexpressing cells are mixed (F). G,H: At E9.0, the invaginating otic cup (otc, arrow) expresses N1ICD. I,J: At E10.5, the ventral region of the otic vesicle (arrow) and cells of the acoustic ganglion complex (VIII) express N1ICD (arrows). K,L: At E11.5, N1ICD expression in the inner ear is restricted to specific cells of the endolymphatic sac (es, arrow), but the semicircular canal (scc) is negative. M: At E13.5, neuroepithelial cells in the cochlear canal (cc) express N1ICD (arrow), but the semicircular canal (thick arrow) remains negative. N: Higher magnification showing positive N1ICD staining in sensory cells of cochlea (arrow) and semicircular canal (thick arrows). O: General view of inner ear at E14.5. P: Detail showing N1ICD expression in neuroepithelium of the cochlear canal (arrow). Q: At E11.5, N1ICD staining is weakly expressed in the neural retina (nr, arrow). Signal is also detected in the optic stalk (os, arrow), hyaloid artery (ha, arrow) and vasculature around the eye. l, lens. R,S: At E13.5–E14.5, N1ICD staining is observed throughout the distal neural retina (arrow and inset in S) and hyaloid plexus (hp), but is excluded from the central retinal region (arrowheads). T: BrdU staining is restricted to the neural retina (arrow) and lens epithelium (thick arrow) and is excluded from central neural retina and lens fibers (arrowheads). U: General view of nasal pit region at E11.5. V: Detail showing N1ICD expression in nasal epithelium (arrow) and in surrounding vessels (arrowhead). W,X: General view of primitive nasal cavity at E13.5 (W) and detail showing N1ICD staining in serous gland (X, arrows). Scale bars = 600 μm in A,K,N, 250 μm in B,C,G,S,T, 150 μm in D, 1.2 mm in E, 350 μm in F, 80 μm in H, 300 μm in I,L, 150 μm in J, 800 μm in M, 1 mm in O, 500 μm in P, 200 μm in Q,R, 600 μm in U, 400 μm in V,W, 180 μm in X.

Download figure to PowerPoint

Pancreas.

The mammalian pancreas is composed of endocrine (islets of Langerhans) and exocrine lineages (acini), both derived from a common progenitor pool in foregut endoderm (Gu et al.,2003). Functional studies indicate that Notch is required to prevent excessive precursor commitment to the endocrine lineage (Apelqvist et al.,1999). The phenotype caused by loss-of-function of Notch1 has not yet been studied, but gain-of-function analyses indicate that ectopic N1ICD expression prevents differentiation of both exocrine and endocrine lineages (Murtaugh et al.,2003). At E13.5, we detected N1ICD in the pancreatic artery and in the acini (Fig. 5R,S), the latter concurring with the proposed role for Notch in preventing endocrine differentiation.

Thymus.

Several lines of evidence link Notch signaling with cortical thymocyte differentiation (for a review, see Wu,2006). In mice, commitment to the T lineage occurs after thymic-seeding progenitors enter the thymus and interact with thymic stromal cells, triggering Notch activation (Harman et al.,2003). The relevance of Notch1 in this process is demonstrated by lack-of-function studies in mice, whereby Notch1 deletion in thymus impairs T-cell fate specification (Radtke et al.,1999), and by the early discovery that the human NOTCH1 gene is rearranged in T-cell lymphoblastic leukemia (Ellisen et al.,1991) and that N1ICD can induce T-cell neoplasms (Pear et al.,1996). N1ICD was expressed in developing thymocytes at E13.5 (Fig. 5T,U) but also in stromal cells, as double staining with pan-cytokeratin and V1744 antibodies indicate (Supplementary Figure S9G–I). At E14.5, N1ICD strongly stains the thymic lobules (Fig. 5V), where cell proliferation is very intense (Fig. 5W), suggesting that, at this stage, Notch1 may maintain developing thymic cells in a proliferative state.

Notch1 Activity During Nervous System Development

The role of Notch in neural development has been studied extensively in Drosophila and in vertebrates (reviewed in Artavanis-Tsakonas et al.,1999; Louvi and Artavanis-Tsakonas,2006). Loss-of-function studies in mice support the idea that Notch signaling inhibits neuron differentiation by lateral inhibition (reviewed in Yoon and Gaiano,2005). Other Notch functions in the mammalian nervous system such as neurite outgrowth and branching regulation (Redmond et al.,2000), learning and memory (Costa et al.,2003), glial differentiation (Dorsky et al.,1997; Wang et al.,1998b), and myelination (Givogri et al.,2002), are currently under intensive study. Initial CNS expression of N1ICD was detected in E8.0–E8.5 embryos, in scattered dorsal cells of the neural folds in the mid- and hindbrain regions (data not shown). At E9.5, N1ICD expression is detected throughout the hindbrain, particularly in the distal or abluminal region, and with less intensity in cells closest to the luminal side of the neural canal (Fig. 6A). Delta1 and Jag1 expression were more restricted and somewhat complementary; Delta1 was widely expressed in the hindbrain, particularly in the dorsal region (Fig. 6B; Bettenhausen et al.,1995; Lindsell et al.,1996). Jag1 was transcribed in two longitudinal stripes on either side of the midline (Fig. 6C; Lindsell et al.,1996), as described for the chick embryo (Myat et al.,1996). Proliferation analysis by bromodeoxyuridine (BrdU) incorporation into DNA indicates that BrdU-positive cells were located predominantly in the abluminal region of hindbrain (Fig. 6D), like N1ICD- and at difference with Jag1- and Delta1-positive cells (Fig. 6B,C). By E10.5, the pattern of N1ICD expression detected throughout the CNS neuroepithelium (see forebrain expression, Fig. 6E,F; and hindbrain, Fig. 6G) closely resembled that of BrdU staining (Fig. 6H). In E11.5 embryo brain, N1ICD was found at the roof of the telencephalon (Fig. 6I,J) as well as the lateral walls and roof of the diencephalon (Fig. 6I). In both cases, we observed N1ICD expression in nuclei close to the ventricular zone. At E11.5, in spinal cord of the mid-thoracic region, N1ICD was expressed in cells of the ventricular zone and in cells of the condensing dorsal root ganglia, while postmitotic ventral horn cells were N1ICD-negative (Fig. 6K). BrdU staining in the spinal cord (Fig. 6L) also resembled that of N1ICD. At E13.5, N1ICD expression in the brain marked cells near the neural canal of the diencephalon (Fig. 6M,N). In the spinal cord ventricular zone (Fig. 6O), the pattern was similar to BrdU staining (Fig. 6P), indicating that N1ICD-positive cells are proliferating.

thumbnail image

Figure 6. N1ICD expression and cell proliferation during central nervous system development. A–P: N1ICD expression (A,E,F,G,I–K,M–O), in situ hybridization (B,C), bromodeoxyuridine (BrdU) staining (D,H,L,P). A: At embryonic day (E) 9.5, N1ICD stains preferentially hindbrain cells in the abluminal region (arrow), and to a lesser extent those closest to the neural canal lumen (arrowhead). B,C: Delta1 (B) transcription is predominant in the dorsal region (arrow) and Jag1 (C) is transcribed in two stripes (arrows) on either side of the midline. D: At E9.5, BrdU staining in hindbrain is predominant in the abluminal region (arrow). E: An E10.5 embryo sectioned at the forebrain (fb) level. N1ICD staining is widespread in the neuroepithelium, except in the optic stalk (os, arrowhead). F: Detail showing N1ICD expression (arrow) in forebrain. G: In the hindbrain region, N1ICD is expressed in the neuroepithelium (arrow in inset) and is excluded from the cells closer to the lumen (arrowhead). H: BrdU staining in the hindbrain is relatively similar to N1ICD expression. I: General view at E11.5 of the forebrain region with telencephalic vesicles (tc; *) and diencephalon (dc). J: Detail of one vesicle. Note predominant N1ICD expression in nuclei closer to the ventricular zone (arrow) and its exclusion from the distal region (arrowhead). K: N1ICD expression in ependymal layer (arrow) of the spinal cord and condensing dorsal root ganglia (drg). No expression is detected in postmitotic ventral horn (vh) cells (arrowhead). L: BrdU staining at E11.5 reveals a pattern similar to that of N1ICD. M: General view of diencephalon at E13.5. N: Detail showing N1ICD staining in neuroepithelial cells closer to the neural canal (arrow), while distal cells are negative (arrowhead). O,P: In the spinal cord, N1ICD expression is restricted to the ventricular zone (arrow in O) in a pattern similar to BrdU staining (P). All are transverse sections. Scale bars = 300 μm in A–D, 500 μm in E–H,J–L,N–P, 0.8 mm in I; 1 mm in M.

Download figure to PowerPoint

During neurogenesis, Delta1 (Henrique et al.,1995) and Jag1 expression (Myat et al.,1996) identify prospective, nondividing neurons; in contrast, Notch1 transcription appears to define uncommitted, proliferating neural precursors (Del Amo et al.,1992). Experimental manipulation in Xenopus (Chitnis et al.,1995) and analysis of mouse targeted mutants (de la Pompa et al.,1997), have led to a model in which Delta1 (and Jag1 in complementary CNS domains) through their binding and activation of Notch, maintain a proliferative cell population from which differentiated postmitotic cells are produced. Lateral inhibition mediated by Delta1 or Jag1 prevents neuroepithelial cells from differentiating prematurely, acting as a negative feedback loop, limiting the amount of neural precursors cells that differentiate at any time point (Myat et al.,1996; Lewis,1998).

At E10.5–E11.5, N1ICD is expressed in the neuroepithelium of the fourth ventricle (Fig. 7A) and in peripheral nervous system (PNS) structures such as the developing trigeminal ganglion (V, Fig. 7A–D). The spinal cord at E13.5 shows strong N1ICD expression in the ventricular zone and dorsal root ganglia (Fig. 7E), where N1ICD-expressing cells intermingle with nonexpressing cells (Fig. 7F). Similarly to its role in the CNS, Notch activity in the PNS has been related to the inhibition of neurogenesis and maintenance of a progenitor cell population in the developing cranial ganglia (Ma et al.,1998; Wakamatsu et al.,2000).

Notch1 Activity During Ear, Eye, and Olfactory Epithelium Development

Inner ear.

The inner ear derives from the otic placode, a thickening of the epidermis adjacent to the hindbrain in the early embryo (Fekete,1996). This placode gives rise both to the inner ear epithelium, with its sensory patches, and to the sensory neurons that innervate these patches. The neuronal lineage segregates from the sensory epithelium lineage as the otic placode invaginates and forms first a cup and then a vesicle (Adam et al.,1998). Notch-mediated lateral inhibition regulates inner ear development, initially demarcating a prosensory region in the cochlear epithelium and then inhibiting progenitor cell differentiation to hair cells (Lanford et al.,1999; Zhang et al.,2000; Daudet and Lewis,2005). Of the four Notch receptors, only Notch1 is expressed during inner ear development (Lewis et al.,1998). RNA expression studies indicate that Notch1 is transcribed in the otic vesicle at E10.0 (Reaume et al.,1992). Anti-Notch1 staining showed Notch1 activity in the inner ear at fetal stages (Murata et al.,2006). We used the V1744 antibody to study the time-course of Notch1 activity in the embryonic ear. N1ICD was already expressed in the otic cup at E9.0 (Fig. 7G,H), presumably as a response to Jag1 (Myat et al.,1996; Lewis et al.,1998) and/or Delta1 (Morrison et al.,1999), both of which are expressed throughout the cup. At E10.5, the ventral lateral region (sensory epithelium) of the otic vesicle and neurons of the acoustic ganglion complex (VIII) express N1ICD (Fig. 7I,J), suggesting that Notch1 activity is required for the specification of otic vesicle cells that differentiate into nerve VIII ganglion neurons. In the inner ear, N1ICD is expressed at E11.5 in specific cells of the endolymphatic sac, whereas the semicircular canal is negative (Fig. 7K,L), and at E13.5, in neuroepithelial cells of the cochlear and semicircular canal (Fig. 7M,N). At E14.5, N1ICD is detected in the mid-basal region of the cochlear canal (Fig. 7O,P), indicative of inhibition of hair cell differentiation.

Eye.

Notch is a key regulator of retinal determination, controlling progenitor cell cycle exit, apoptosis, and differentiation (Austin et al.,1995; Dorsky et al.,1995; Henrique et al.,1997; Jadhav et al.,2006). Notch1 is expressed in the neural retina exclusively during the period of cell fate determination and differentiation (Lindsell et al.,1996). At E11.5, we detected N1ICD staining in the hyaloid artery and very weakly in the peripheral region of the neural retina (Fig. 7Q). The vasculature around the eye was also N1ICD-positive (Fig. 7Q). N1ICD expression in neural retina increased progressively at E13.5–E14.5, but was excluded from the central region (Fig. 7R,S). At E14.5, N1ICD was expressed predominantly in proliferating, BrdU-positive cells in the neural retina (Fig. 7S,T) and signal in the hyaloid plexus remained (Fig. 7S). N1ICD expression in neural retina coincided with that of Delta1 (Lindsell et al.,1996), suggesting that Delta-Notch interaction in this tissue would allow a progressive maturation of neuronal precursors. Analysis of retina-specific Notch1 mutants has shown that Notch1 not only maintains the progenitor state, but is also required to inhibit the photoreceptor fate (Jadhav et al.,2006; Yaron et al.,2006).

Olfactory epithelium.

From E11.5 onward, various Notch receptors are expressed in the olfactory epithelium (Mitsiadis et al.,2001). At E11.5, N1ICD expression was predominant in the basal layer of nasal epithelium (Fig. 7U,V), which contains neural progenitors, as well as in surrounding vasculature (Fig. 7V). At E14.5, we detected N1ICD in the epithelium of the primitive nasal cavity (Fig. 7W) and serous gland (Fig. 7X).

Notch-Dependent Positive Feedback Loop Regulates Notch1 Transcription

It has been proposed that, during Drosophila vein development (de Celis et al.,1997) and T-cell maturation (Deftos et al.,1998), a positive feedback loop regulates Notch transcription in cells in which Notch has been activated. We thus analyzed whether Notch1 activity was affected in RBPJk mutants. N1ICD expression was strongly reduced in the neural tube (Fig. 8A,B,D,E), in endocardium (Fig. 8A,C,D,F), and in the PSM (Fig. 8G,H). We previously showed that cardiac Notch1 transcription is reduced in E8.5 RBPJk mutants (Timmerman et al.,2004). We found a similar effect in the heart (compare Fig. 8I,J and 8K, L), PSM, and CNS of E9.5 RBPJk mutants (Fig. 8M,N). These observations were confirmed by semiquantitative reverse transcriptase-polymerase chain reaction (RT-PCR) analysis of E9.5 mutants (Fig. 8O). These data suggest that a NICD/RBPJK-dependent positive feedback loop regulates embryonic Notch1 transcription. To discard the possibility that the reduction in Notch1 transcription was due to a decline in viability of E9.5 RBPJk mutants, we examined tissue morphology by H&E staining that showed normal tissue organization in mutants (Supplementary Figure S3). We also examined the general vascular structure by CD31/PECAM staining, which was relatively normal (Supplementary Figure S4). Lastly, we examined whether there was increased apoptosis in E9.5 mutant embryos. TUNEL assays demonstrated no differences in apoptosis between wt and mutant embryos (Supplementary Figure S5). These data indicated that the reduced Notch1 transcription in E9.5 RBPJk mutants was not due to generalized arrest in developmental progression.

thumbnail image

Figure 8. Abrogation of Notch1-dependent signaling activity in mouse and zebrafish Notch pathway mutants. A–H: Embryonic day (E) 9.5 RBPJk mutants. A: General view of a transverse section stained with the anti-N1ICD (V1744) antibody. Section level is similar to that shown in Figure 2A. B: Detail showing the hindbrain (hb), hindgut (hg), and the collapsed dorsal aorta (da). Weak N1ICD expression is observed in the dorsal aorta (arrow), and no expression is detected in hb (arrowhead). C: Detail showing weak N1ICD expression in the outflow tract (OFT, large arrow) and right ventricle (rv, arrow) endocardium. D: General view of a transverse section at a more caudal level, similar to that shown in Figure 2D. E: Detail showing lack of N1ICD expression in hb (arrowhead). F: Detail showing very faint N1ICD expression in the atrioventricular canal (AVC, arrow) and left ventricle (lv, thick arrow) endocardium, where the majority of cells do not express N1ICD (arrowhead). G: General view of a transverse section at the level of the presomitic mesoderm (PSM). H: Detail showing very reduced N1ICD staining in PSM (arrows), da (thick arrow), and lack of staining in the neural tube (arrowhead). I–L: Notch1 Whole-mount in situ hybridization (WISH) analysis of E9.5 hearts. I,J: Notch1 is transcribed throughout the ventricular endocardium of the wild-type (wt) embryo (I); detail of left ventricle (lv, arrow in J). K,L: RBPJk mutants show reduced Notch1 transcription in ventricular endocardium (arrow in L). M: Notch1 transcription in E9.5 wt PSM and neural tube region (arrows in M) and marked reduction in E9.5 RBPJk mutants (N). O: Semiquantitative reverse transcriptase-polymerase chain reaction (RT-PCR) analysis of Notch1 transcription in E9.5 wt and RBPJk mutants. P–U: Notch1b WISH analysis of 48 hpf–72 hpf zebrafish embryos. P: Detail of the cardiac region of a 48 hpf wt embryo. The arrow indicates cardiac expression of Notch1b. Q,R: Detail of the cardiac region of 48 hpf mibtf191 (Q) and mibta52b (R) mutants showing reduced Notch1b transcription (arrows). S: Detail of the cardiac region of a 72 hpf wt embryo showing Notch1b expression restricted to valvular primodia (arrows). T,U: Detail showing very reduced cardiac Notch1b transcription in 72 hpf mibtf191 (T) and mibta52b (U) mutants (arrows). V: Semiquantitative RT-PCR analysis of Notch1b transcription in 72 hpf wt and mibta52b mutants. Scale bars = 100 μm in A,D,G,I,K,M,N, 40 μm in B,C,E,F,J,L, 30 μm in H, 125 μm in P–U.

Download figure to PowerPoint

As N1ICD binds to RBPJK in the nucleus, it was possible that N1ICD would be less stable in RBPJk mutants, and, therefore, only minimal amounts of the protein could be detected. Alternatively, as we have suggested, Notch1 signaling could regulate Notch1 transcription by a positive feedback loop. If this feedback loop exists, one might predict that it would involve genes acting downstream of Notch receptors and that upstream genes would not be involved. We, therefore, analyzed Notch1 transcription in zebrafish mutants for the ligand-modifying ubiquitin ligase mib, expecting that it would not be altered. We examined two mutants that drastically reduce wt mib function, mibta52b and mibtf91 (Itoh et al.,2003). We used WISH to analyze cardiac transcription of Notch1b—the zebrafish functional homologue of Notch1 (Walsh and Stainier,2001)—in both mib backgrounds. Figure 8P shows widespread cardiac Notch1b transcription in a 48 hours postfertilization (hpf) wt embryo. We observed reduced Notch1b transcription in 48 hpf mutants, which was moderate in mibtf91 (Fig. 8Q) and severe in mibta52b mutants (Fig. 8R). At 72 hpf, Notch1b expression was restricted to prevalve territories in the wt embryo (Fig. 8S and Walsh and Stainier,2001). In contrast, both mibtf91 (Fig. 8T) and mibta52b (Fig. 8U) mutants showed a severe decrease in Notch1b transcription at 72 hpf. Semiquantitative RT-PCR analysis of 72 hpf embryos indicated that Notch1b transcription was greatly reduced in mibta52b mutants (Fig. 8V), supporting our WISH results.

These data suggest that Notch1 signaling is required to sustain Notch1 transcription by a positive feedback loop involving both signaling and receiving cells, as mutations affecting Notch ligand-receptor interaction (mib) or Notch effector protein (RBPJk) lead to down-regulation of Notch1 transcription. After extensive analysis of the mouse and human Notch1 promoters we have not identified conserved RBPJK binding sites (data not shown), suggesting that this feedback loop may not be directly mediated by N1ICD/RBPJK.

CONCLUSIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. CONCLUSIONS
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

This study shows that our staining method to detect N1ICD using the V1744 antibody is a useful tool for the comparative analysis of Notch1 activity and function during mouse development. Both antibody staining and in vivo reporters are two fundamentally different approaches to study Notch activity and the information obtained is complementary: Antibody staining provides a snapshot view of the cells or tissues in which Notch is active. In vivo reporters give a retrospective or historical evaluation of Notch activity. A recent elegant study by Vooijs et al. (2007) reports the generation of a transgenic line in which N1ICD is replaced by the CRE recombinase, so that CRE activity is controlled by ligand-induced proteolysis of Notch1. This strain is combined with the CRE reporter strain Rosa26R (Soriano,1999) and Notch1 activation is visualized by β-galactosidase expression in N1::CREtg/+; R26Rtg/+ mice (Vooijs et al.,2007). In our V1744 staining, we use antigen retrieval coupled to tyramide amplification and obtain an excellent sensitivity, clearly superior to that attained so far with Notch reporters in vivo (Duncan et al.,2005; Souilhol et al.,2006; Vooijs et al.,2007). Our argument is based on the following observations: First, we detect N1ICD in all the tissues where Notch1 transcription has been described (Del Amo et al.,1992); second, we detect N1ICD in the heart, vasculature, PSM, and CNS at least 2 days earlier than with the use of Notch reporters (Duncan et al.,2005; Souilhol et al.,2006; Vooijs et al.,2007). Third, we detect Notch1 activity in derivatives of the three germ layers, including tissues with low levels of Notch1 activity (see Table 1 and Supplementary Figure S6). This finding is unlike with the use of Notch reporters so far, which allow for detection of Notch in tissues with relatively high activity levels (Duncan et al.,2005; Souilhol et al.,2006; Vooijs et al.,2007). We find that N1ICD expression levels do not necessarily correlate with a requirement of Notch1 function. For example, high N1ICD expression in vasculature and endocardium correlate with its essential role in arterial fate determination (Krebs et al.,2000) and cardiac development (Timmerman et al.,2004; Grego-Bessa et al.,2007). Lower Notch1 activity levels also reflect an important role of Notch, as in the inner ear (Lanford et al.,1999) or in the pancreas (Murtaugh et al.,2003). Notch1 activity overlaps with that of other Notch receptors in different tissues, such as the CNS (Lindsell et al.,1996), PSM (Williams et al.,1995), heart (Loomes et al.,2002), and vasculature (Krebs et al.,2000), although this finding does not suggest functional redundancy among receptors, as mutant analysis reveals (Yoon and Gaiano,2005; Bolos et al.,2007). Our staining method will be also a very useful tool to examine Notch1 activation in adult self-renewing tissues and in skeletal and cardiac muscle growth and repair.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. CONCLUSIONS
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Mice

Wild-type (wt) CD1 embryos were staged precisely by somite count and comparison with established references (Kaufmann,2001). Notch1 (Conlon et al.,1995) and RBPk (Oka et al.,1995) homozygous mutant embryos were genotyped by PCR analysis of yolk sacs. All experiments involving animals were performed according to institutional and European Union guidelines.

Zebrafish

The wt (AB), mibtf191, and mibta52b stocks were maintained under standard conditions. Mutant embryos can be distinguished morphologically, as they display severe trunk curvature (Lawson et al.,2001). Mutant embryos at 48–73 hpf were analyzed.

Histology and In Situ Hybridization

H&E staining was performed by standard methods. WISH and sectioning were as described (de la Pompa et al.,1997). Details for probes will be provided on request.

BrdU Staining

BrdU labeling and staining were as described (Hakem et al.,1996), using mouse embryos fixed overnight in 70% ethanol.

N1ICD (V1744) Immunohistochemistry and Costainings

Embryos from timed matings were dissected and fixed for 2 hr (from E7.5–E11.5) or 4 hr (from E12.5–E14.5) in 4% paraformaldehyde at 4°C, dehydrated in a graded ethanol series, and then paraffin-embedded. Serial 7-μm sections were prepared and mounted on polylysine-treated slides. After deparaffinization and rehydration in an ethanol series, antigen was retrieved by boiling samples in a microwave oven for 10 min (twice, 5 min each) in sodium citrate (10 mM, pH 6.0) to avoid evaporation. Slides were allowed to cool to room temperature and were washed in distilled water (4 times, 3 min each) and phosphate buffered saline (PBS, 3 min). Endogenous peroxidase activity was quenched using 1% H2O2 in 100% methanol (40 min, room temperature) followed by 3 washes in PBS (5 min each) and 2 washes in 0.3% Triton/PBS (10 min each). For blocking, samples were incubated (1 hr) in histoblock solution (3% bovine serum albumin [BSA], 20 mM MgCl2, 0.3% Tween 20, 5% fetal bovine serum in PBS) before incubation with anti-cleaved Notch1 antibody (V1744 antibody, Cell Signaling; 1:100 dilution in histoblock solution; 4°C, overnight), followed by washing with PBS (2 washes, 5 min each) and 0.3% Triton X-100/PBS (2 washes, 5 min each). Slides were then incubated with biotinylated anti-rabbit IgG antibody (1:100 in 5% BSA in PBS; Vector Laboratories; 1 hr, room temperature), followed by two washes in PBS (5 min each) and two washes in 0.3% Triton/PBS (5 min each). Finally, the signal was amplified in two steps; first, avidin/biotin-HRP was added to the biotin of the secondary antibody (ABC kit, Vector Labs; 1 hr, room temperature). Tyramide-Cyn3 was then added to the HRP using a tyramide signal amplification (TSA)-cynanine 3 kit (NEL 744, Perkin Elmer). Incubation in TSA-Cyn3 was stage-dependent; thus, samples from E8.5 embryos or earlier stages were incubated at a 1:100 dilution (3.5 min) or 1:50 (3 min), whereas samples from E9.5 onward were incubated at 1:50 (3.5 min). Washes during amplification were done in 0.3% Triton/PBS (twice, 5 min each). Slides were mounted and nuclei counterstained using Vectashield with 4′,6-diamidine-2-phenylidole-dihydrochloride (DAPI, Vector Labs).

Costaining of N1ICD and SMA was done using the protocol described above for N1ICD, followed by incubation with an anti-SMA antibody (Sigma) conjugated with a cyn-3 fluorochrome (1:100 in Histoblock Solution). In this case, the TSA system used to detect N1ICD was TSA-fluorescein (Perkin Elmer).

Costaining of N1ICD and cytokeratin was done using the protocol described for N1ICD and incubating with TSA-fluorescein (1:50, PerkinElmer) to detect N1ICD. For cytokeratin staining, a pan-cytokeratin antibody against cytokeratins 4, 5, 6, 8, 10, 13, and 18 (1:100, C-2931, Sigma) was used, followed by anti-mouse cyn3 (1:100, Jackson) as secondary antibody.

For visualization, we used a Leica DMRB microscope with a Nikon DP70 camera and DP Controller software. Images were assembled using Adobe Photoshop. Confocal images were acquired with an Olympus Fluoview 1000 microscope.

CD31 Immunohistochemistry

Embryos were fixed in 4% paraformaldehyde for 2 hr at 4°C and processed as above with the exception of the sodium-citrate antigen retrieval step. Primary antibody (anti-mouse CD31, BD Pharmigen) was used 1:50 in Histoblock Solution, and secondary antibody (anti-mouse biotinylated, Vector) was used 1:100 in 5% BSA.

Cell Death (TUNEL) Assay

Apoptosis was analyzed by TUNEL staining using the In Situ Cell Death Detection Kit (Roche 1 684 817) according to manufacturer's instructions.

Western Blot

Protein pools from E9.5 and 13.5 mouse embryos and human mammary tumor cell lines (MCF7, HS57BT, BT594, MDA-MB468, MDA-MB231) were purified following a protocol for nuclear protein purification (Andrews and Faller,1991). Protein were separated by 8% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to Immobilon-P membranes (Millipore). Membranes were blocked with TBST/5% nonfat dry milk and incubated overnight with anti-cleaved Notch1 antibody (1:500, mouse or 1:1,000, human) in TBS/5% BSA. Anti–SMC-3 antibodies were used as loading control (1:1,000; Chemicon). HRP–anti-rabbit IgG (1:1,000, 1 hr; Dako Cytomation) was used as secondary antibody. Antibodies were diluted in TBST/5% nonfat dry milk or as indicated; washes were done using TBST.

Sequence Alignment

Sequences were aligned using the complete Notch1 amino acid sequences in PubMed. Clustal analysis was done at the EMBL-EBI Web site (http://www.ebi.ac.uk/clustalw/index.html).

RNA Isolation and Semiquantitative RT-PCR

E8.5–E9.5 wt and RBPJk mutant embryos were dissected in ice-cold PBS and RNA purified using Trizol (Invitrogen). cDNA synthesis was performed using a First Strand cDNA synthesis kit (Amersham), with 1 μg of total RNA per reaction. The sequences of Notch1 primers were 5′-CAATCAGGGCACCTGTGAGCCCACAT-3′ (forward) and 5′-TAGAGCGCTTGATTGGGTGCTTGCGC-3′ (reverse). The β-actin primers were 5′-GGACCTGGCTGGCCGGGACC-3′ (forward) and 5′-GCGGTGCACGATGGAGGGGC-3′(reverse). The PCR conditions for Notch1 were annealing at 60°C, 30 sec and extension at 72°C, 45 sec. For β-actin, the conditions were annealing at 62°C, 30 sec and extension at 72°C, 45 sec.

Total RNA was prepared from 72 hpf wt and mibta52b zebrafish mutant embryos using Trizol (Invitrogen). cDNA synthesis was performed as above with 1 μg of total RNA per reaction. The sequences of Notch1b primers were 5′-CCTGGGTATGGAATCTTGCGGTAT-3 (forward) and 5′-ATGGGCCAGACTCATCGTACTCCT-3 (reverse). The β-actin primers were 5′-CACTAGACGGCTGGAACGATGCTT-3 (forward) and 5′-GATCTCACTTGTGACGGCCTCGAT-3 (reverse). The PCR conditions for both sets of primers were annealing at 58°C, 30 sec and extension at 72°C, 30 sec. Amplified PCR products were cloned in the PCRII-TOPO vector (Invitrogen) and sequenced. PCR products were quantitated by phosphorimager analysis (Bio-Rad).

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. CONCLUSIONS
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

We thank J.M. Pérez-Pomares for critical reading of the manuscript, A.M. Pendas for help with the Notch1 sequence analysis, and C. Mark for editorial assistance. G.M. was supported by an I3P predoctoral fellowship from the Spanish National Research Council (CSIC). The Department of Immunology and Oncology was founded and is supported by the CSIC and by Pfizer.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. CONCLUSIONS
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. CONCLUSIONS
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

The Supplementary Material referred to in this article can be found at http://www.interscience.wiley.com/jpages/1058-8388/suppmat

FilenameFormatSizeDescription
DVDY21246Suppl.Fig.1.tif22291KSupporting Information file DVDY21246Suppl.Fig.1.tif
DVDY21246Suppl.Fig.2.tif20460KSupporting Information file DVDY21246Suppl.Fig.2.tif
DVDY21246Suppl.Fig.2A.tif3106KSupporting Information file DVDY21246Suppl.Fig.2A.tif
DVDY21246Suppl.Fig.3.tif5976KSupporting Information file DVDY21246Suppl.Fig.3.tif
DVDY21246Suppl.Fig.4.tif2583KSupporting Information file DVDY21246Suppl.Fig.4.tif
DVDY21246Suppl.Fig5.tif2197KSupporting Information file DVDY21246Suppl.Fig5.tif
DVDY21246Suppl.Fig.6.tif1036KSupporting Information file DVDY21246Suppl.Fig.6.tif
DVDY21246Suppl.Fig.7.tif1721KSupporting Information file DVDY21246Suppl.Fig.7.tif
DVDY21246Suppl.Fig.8.tif6384KSupporting Information file DVDY21246Suppl.Fig.8.tif
DVDY21246Suppl.Fig.9.tif3131KSupporting Information file DVDY21246Suppl.Fig.9.tif
DVDY21246Suppl.File1.doc2568KSupporting Information file DVDY21246Suppl.File1.doc

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.