|Home | About | Journals | Submit | Contact Us | Français|
In mammals, two spatially and temporally distinct waves of fiber cell differentiation are crucial steps for normal lens development. In between these phases, an anterior growth zone forms in which progenitor cells migrate circumferentially, terminally exit the cell cycle and initiate differentiation at the lens equator. Much remains unknown about the molecular pathways orchestrating these processes. Previously, the Notch signal transduction pathway was shown to be critical for anterior lens progenitor cell growth and differentiation. However, the ligand or ligand(s) that direct these events are unknown. Using conditional gene targeting, we show that Jagged1 is required for lens fiber cell genesis, particularly that of secondary fiber cells. In the absence of Jagged1, the anterior growth and equatorial transition zones fail to develop fully, with only a handful of differentiated fiber cells present at birth. Adult Jagged1 conditional mutants completely lack lenses, along with severe anterior chamber deformities. Our data support the hypothesis that Jagged1-Notch signaling conveys a lateral inductive signal, which is indispensable for lens progenitor cell proliferation and differentiation.
A hierarchy of genes that regulate tissue morphogenesis, growth and two waves of fiber cell differentiation during ocular lens development. The vertebrate lens initiates from a placodal thickening of the surface ectoderm, adjacent to the optic vesicle. This lens placode invaginates into a pit, and then vesicle, within the space vacated by the forming optic cup (reviewed in Lang, 2004; McAvoy et al., 1999). As the lens vesicle forms, it separates from the surface ectoderm, which later gives rise to the cornea. Within the hollow lens vesicle, anterior-posterior compartmentalization occurs as posterior progenitor cells, closest to the forming retina, exit the cell cycle and differentiate as primary lens fibers. Thus, proliferative lens progenitor cells are progressively sequestered in the anterior vesicle, where they coalesce into an epithelial growth zone that persists into adulthood. Once formed, the anterior epithelial layer (AEL) produces secondary lens fibers by peripheral cell movement, through a germative zone, and into the transition zone at the lens equator. Cells in the transition zone become postmitotic, differentiate, move centrally and elongate as mature fiber cells.
The transcription factors Pax6, Sox1, Sox2, Prox1, Foxe3, Pitx3, AP2α, and Maf are critically required for lens formation, and constitute a partial lens regulatory gene network (reviewed in Cvekl and Duncan, 2007; Graw, 2003; Lang, 2004). Importantly, mutations in several of these human genes cause anophthalmia, microphthalmia, Peter’s Anomaly and/or aphakia, wherein the lens is absent or defective by birth (Ashery-Padan et al., 2000; Blixt et al., 2000; Glaser et al., 1994; Grimm et al., 1998; Kim et al., 1999; Medina-Martinez et al., 2005; Rieger et al., 2001; Semina et al., 1998; Semina et al., 1997). But, signal transduction pathways are equally important during lens development. For example FGF, BMP, Wnt and Notch signaling regulate key aspects of lens formation (Beebe et al., 2004; Cain et al., 2008; Faber et al., 2002; Jia et al., 2007; Lovicu and McAvoy, 2001; Ogino et al., 2008; Robinson, 2006; Rowan et al., 2008; Song et al., 2007; Zhao et al., 2008).
The Notch pathway functions in the lens, as it does throughout the body, to transduce cell contact-mediated communication. The ligands are of two types: Delta/Deltalike (DLL) or Jagged/Serrate/Lag2, which bind a Notch receptor via their extracellular DSL domain. There are four Notch receptor genes in mammals (reviewed in Gridley, 2003; Louvi and Artavanis-Tsakonas, 2006), and whether one or multiple Notch genes act during lens formation remains unclear. Nevertheless, upon ligand-receptor binding a series of proteolytic cleavages are triggered that release the intracellular domain (NotchIC), allowing it to form a nuclear complex with the Su(H)/Rbpj and mastermind/MAML transcription factors. This complex then activates the transcription of downstream effector genes (reviewed in Ilagan and Kopan, 2007; Kopan, 2002; Louvi and Artavanis-Tsakonas, 2006). Major targets of Notch signaling are the hairy-E(spl)/Hes transcriptional repressors (reviewed in Davis and Turner, 2001; Kageyama et al., 2007). While invertebrates (e.g. Drosophila), encode essentially one gene for each part of the Notch pathway, vertebrates have multiple paralogues for nearly every component. This allows additional levels of regulatory complexity during vertebrate development, such that different ligands may transduce information about distinct cellular processes to the same receptor, or one ligand may send multiple signals by binding and activating more than one receptor. Multiple ligands, receptors and downstream effectors simultaneously acting within a single cell or cell type can impose additional layers of signal information. Finally, the Notch pathway transduces two main types of signals, a classical lateral inhibition signal (Cabrera, 1990; Simpson, 1990), wherein adjacent cells compete for ligand expression, or a lateral induction signal (Eddison et al., 2000; Lewis, 1998), in which a group of cells exhibit cooperative ligand expression.
In the lens, Notch signaling has multiple roles. In the frog optic vesicle, Delta1 activates Notch in the lens placode, thereby triggering activation of the Foxe3 lens enhancer through binding of a NotchIC-Rbpj-MAML complex to DNA, adjacent to a site of Otx2 protein binding (Ogino et al., 2008). This synergistic activation of Foxe3 is crucial for lens vesicle formation and growth (Blixt et al., 2000). Meanwhile in the mouse, lens-specific deletion of Rbpj, or misexpression of the Notch1IC demonstrate that there are other, late Rbpj-dependent functions for Notch signaling, during primary fiber cell genesis, lens progenitor cell growth in the AEL and secondary fiber cell differentiation (Jia et al., 2007; Rowan et al., 2008). As primary fiber cells differentiate, one Notch ligand, Jagged 1 (Jag1), becomes localized posteriorly. Then, during secondary fiber formation Jag1 expression is confined to transition zone cells, and the anterior side of extended fiber cells, at the border with the AEL (Rowan et al., 2008). These data suggest Jag1 may act both in transition cells to regulate secondary fiber cell formation and in fiber cells to signal Notch-Rbpj-Hes1+ progenitor cells in the AEL (Rowan et al., 2008).
To investigate these ideas further, we directly tested the requirements for Jag1 during mammalian lens formation. Germline deleted Jag1 mutants die before lens formation is underway, but hemizygous mice have uncharacterized eye defects (Xue et al., 1999). Therefore, to determine when and how this ligand acts during lens development, we used a Cre-Lox strategy to selectively remove Jag1. We observed that although Jag1 is necessary for aspects of primary fiber cell formation, Jag1 plays a major role in the AEL and transition zone formation, since both compartments fail to become established in lens-specific deletion mutants. Without proper tissue compartmentalization, the lens is microphakic by birth, and contains only a handful of β-Crystallin+ cells. Jag1 conditionally mutant adult eyes lacked lenses, anterior chambers and pupillary openings. These data support a model that Jag1-Notch signaling is essential for lens growth and fiber cell differentiation. Furthermore, during lens development Jag1 appears to transduce multiple signals that are spatially and temporally restricted, and mediates distinct developmental events.
Jagged 1tm1JLew mice (Jag1CKO) were generated by Brooker and colleagues, maintained on a C57BL/6 background and genotyped as described (Brooker et al., 2006). Rbpjtm1Hon mice (RbpjCKO) were generated by Han et al., maintained on a 129/SvJ background and genotyped as described (Han et al., 2002). The abbreviation CKO indicates a “conditional knock-out” allele. Le-Cre mice were generated by Ashery-Padan et al., maintained on an FVB/N background and PCR genotyped as described (Ashery-Padan et al., 2000). Images of adult heads were captured with a Leica dissecting microscope and Optronics digital camera.
Embryonic and postnatal tissues were fixed in 4% paraformaldehyde/PBS for 1 hour at 4°C, processed through a sucrose/PBS series, cryoembedded and sectioned. Primary antibodies used were anti-BrdU (Serotec clone BU1/75 1:500), anti-cleaved PARP (Cell Signaling 1:500), anti-Cyclin D1 (Neomarkers SP4 1:100; or Santa Cruz 72-13G 1:500), anti-Cyclin D2 (Santa Cruz 34B1-3 1:200), anti-E cadherin (Zymed ECCD-2 1:500), anti-Foxe3 (gift from Peter Carlsson 1:1000), anti-β-Crystallin (gift from Richard Lang 1:8000), anti-γ-Crystallin (Santa Cruz 1:1000), anti-GFP (Molecular Probes 1:1000), anti-Hes1 (1:1000), anti-Jag1 (Santa Cruz 1:1000), anti-p27Kip1(BD Laboratories Clone 57 1:100), anti-p57Kip2 (Abcam 1:200), anti-Pax6 (gift from Grant Mastick 1:1000), anti-Prox1 (Covance 1:1000), anti-Pitx3 (gift from Marten Smidt 1:1000), anti-Six3 (gift from Guillermo Oliver 1:1000), anti-Sox1 (Affinity BioReagents 1:500), and anti-Sox2 (Chemicon 1:500), following (Lee et al., 2005). Secondary antibodies were directly conjugated to Alexa Fluor 488, Alexa Fluor 594 (Molecular Probes) or biotinylated (Jackson Immunologicals) and sequentially labeled with streptavidin Alexa 488 or 594 (Molecular Probes). Microscopic imaging was performed on a Zeiss fluorescent microscope with a Zeiss camera and Apotome deconvolution device. Whole-mount or cryosection in situ hybridization was performed as described (Brown et al., 1998) using a Jag1 digoxygenin-labeled antisense riboprobe. For S-phase analyses, BrdU (Sigma) was injected intraperitoneally as described in (Mastick and Andrews, 2001) and animals sacrificed 1.5 hours later for tissue processing that included 2N hydrochloric acid treatment of sections prior to antibody staining. Standard histology on paraffin embedded sections was also performed. Images were processed using Axiovision (v5.0) and Adobe Photoshop software (v7.0) and electronically adjusted for brightness, contrast and pseudocoloring.
Tissue sections, separated by at least 60 μm, were antibody-stained and counted using Axiovision software. Three or more animals were analyzed per genotype and age and at least two independent sections through the center of the lens per animal quantified. Labeling indices were generated by dividing the number of antibody-positive cells by total DAPI-labeled nuclei, and Instat (v3.0) software used to perform ANOVA plus a Bonferroni posthoc test to determine p values.
Although basic expression of Jag1 in the developing vertebrate eye is known (Bao and Cepko, 1997; Jia et al., 2007), its expression pattern across the key stages of lens development is uncharacterized. Therefore, we examined Jag1 mRNA and protein expression from embryonic day 9.5 (E9.5) to postnatal day 3 (P3), by both in situ hybridization and anti-Jag1 staining (Fig. 1). From E9.5–10.5, Jag1 mRNA and protein are specifically expressed in lens placode (arrows in Figs. 1A,F), lens pit (not shown), and distal optic vesicle cells (arrowheads in Figs. 1A,F). When the lens vesicle pinches off from the surface ectoderm around E11, Jag1 mRNA and protein are abundant throughout vesicle cells (Fig. 1B,G). But once primary fiber differentiation commences between E11.5 to E12.5, Jag1 mRNA and protein become sequestered in posterior lens vesicle cells (Figs. 1B,C,G,H). Interestingly after E12.5, Jag1 mRNA and protein were no longer detectable in the peripheral optic cup.
In the E14.5 lens, Jag1 mRNA is abundant in equatorial cells (Fig 1D) once the AEL, transition zone, and fiber cell compartments are fully established. Likewise, Jag1 protein is observed in transition zone cell membranes, and along the anterior side of fiber cells, where they are in close contact with AEL progenitor cells (Fig 1I). These expression patterns for Jag1 mRNA and protein persist beyond birth (Figs 1E,J) to at least P3 (not shown). One consistent difference between Jag1 mRNA and protein patterns, is a broader protein domain that includes newly born secondary fiber cells (compare Figs. 1I,J to 1D,E), suggesting that either Jag1 mRNA is very tightly regulated, or Jag1 protein persists longer than the mRNA.
Previously, we used the Le-Cre driver (Ashery-Padan et al., 2000) to delete the nuclear Notch effector Rbpj during mouse lens development (Rowan et al., 2008). Loss of Rbpj causes adult eyes to be microphthalmic, with total loss of the pupillary opening and the anterior chamber, and dramatic reduction of lens tissue (Rowan et al., 2008). Using the same cre-lox deletion strategy (Fig. 2A), Le-Cre;Jag1CKO/CKO and Le-Cre; Jag1CKO/+ P21 mice were generated in expected Medelian ratios. Le-Cre;Jag1CKO/CKO mutants are devoid of fur around severely microphthalmic eyes that lack pupillary openings (Fig. 2D and data not shown). Heterozygotes have less severely reduced eyes (Fig. 2C). Histologic cross sections show that progressive loss of the lens correlates with reduction of Jag1 gene dosage (compare Figs. 2E–G; n = 3/genotype). Adult Le-Cre;Jag1CKO/CKO eyes are missing lenses and anterior chambers, accompanied by expansion of iris and/or ciliary body tissue (Fig. 2G). Upon close examination, small clumps of lens fiber-like cells are identifiable in some sections of Jag1 conditionally mutant eyes (arrows in Fig. 2G).
The adult phenotypes of Le-Cre;Jag1CKO/CKO and Le-Cre;RbpjCKO/CKO eyes are strikingly similar, but Jag1 conditional mutants are more severe (compare Figure 2G to Figure 1E of Rowan et al, 2008). To understand when lens formation goes awry in Le-Cre;Jag1CKO/CKO eyes and whether developmental defects are the same or different from those in Rbpj mutant lenses, we next analyzed P3 Jag1 conditional mutants. At this age, lens tissue could be discerned in sections through Le-Cre;Jag1CKO/CKO eyes, with a small number of β-Crystallin+ lens fibers (Figs. 3A–C). We also observed that Le-Cre;Jag1CKO/+ littermates have smaller sized lenses than controls. This loss of Crystallin+ fibers was also obvious at E18.5 (Supp Fig. 1A–C). To determine if the remnant lenses in Jag1 conditional mutants contain distinct AEL, transition zone and fiber cell regions, adjacent sections were colableled with anti-Ecadherin (Ecad) and anti-p57/Kip2, to mark the AEL and transition zone cells, respectively (Fig. 3D). At the boundary of these compartments, a small number of progenitor cells normally coexpress both markers (shown in yellow in the bracketed region in Fig. 3D). By contrast, Le-Cre;Jag1CKO/+ eyes have more peripheral cells coexpressing Ecad and p57/Kip2 (Fig. 3E). Strikingly, Le-Cre;Jag1CKO/CKO eyes lack both an AEL and transition zone, with only a few random cells expressing either marker (circled area of Fig. 3F). The breakdown of these peripheral compartments was also observed at E18.5 (Supp Fig. 2D–F). To confirm the loss of the AEL, we analyzed Foxe3 expression, which is expressed by lens progenitor cells from E9.5 to adulthood, where it regulates proliferation (Blixt et al., 2007; Blixt et al., 2000). We found E18.5-P3 Le-Cre;Jag1CKO/CKO eyes to be essentially devoid of Foxe3 expression (Figs. 3G,J, Supp Fig 1I, n=3 animals). At E18.5, only small patches of nonnuclear expression were observable in the absence of Jag1 (Supp Fig 1I). At high magnification of P3 lenses, the normal, orderly arrangement of AEL cells within a monolayer is apparent (Fig. 3H). But, in Le-Cre;Jag1CKO/+ lenses, Foxe3+ cells were irregularly arranged (Fig 3I). Colabeling with anti-Jag1 confirmed the loss of protein expression in P3 Jag1 lens mutants (compare Fig. 3G to Fig. 1J). We conclude that the AEL and transition zone compartments have broken down by birth in Le-Cre;Jag1CKO/CKO eyes, a more severe phenotype than that of Le-Cre;RbpjCKO/CKO lenses (compare Figure 3 to Figure 8 in Rowan et al., 2008).
The small number of β-Crystallin+ fibers and loss of the AEL and transition zone at birth suggests that without Jag1 function, lens development breaks down soon after primary fiber cells form. Therefore we assayed the expression of two lens progenitor cell markers: Foxe3 (green) and E-cadherin (Ecad, red) from E12.5 to E14.5, when the AEL compartment becomes established (Blixt et al., 2000; Medina-Martinez et al., 2005; Xu et al., 2002). Foxe3 specifically marks lens progenitor cells, while Ecad is expressed by both lens progenitor and corneal epithelial cells (arrowheads in Figs. 4D–F,J–L). Interestingly, both proteins become inappropriately downregulated in the AEL of Le-Cre;Jag1CKO/+ and Le-Cre;Jag1CKO/CKO embryos (n≥3/genotype). Already at E12.5 a significant loss of Foxe3 expression is apparent in Le-Cre;Jag1CKO/CKO lenses (Figs. 4C, and data not shown). Simultaneously, the Ecad domain is reduced in E12.5 Le-Cre;Jag1CKO/+, or completely missing from 50% of Le-Cre;Jag1CKO/CKO lenses (Figs 4E,F; n = 4 embryos per genotype). Although the Le-Cre driver removes Jag1 in both developing lens and corneal epithelium (Ashery-Padan et al., 2000), we find Ecad expression is only downregulated in the lens.
Loss of Foxe3 and Ecad expression persists at E14.5 (Figs 4G–L), where 50% of Le-Cre;Jag1CKO/CKO eyes examined have no Foxe3 expression (Fig. 4I), and the rest have sporadic, rare Foxe3+ nuclei (n = 4 embryos per genotype). Similarly, the Ecad lens domain is abnormal in both E14.5 Jag1 heterozygous and homozygous mutants (Figs 4K,L), with Ecad expression in the cornea unaffected (arrowheads). However, not all AEL characteristics are lost, since we observed Pax6 and Sox2 expression at E12-E14.5 are unaffected (n = 3/age and genotype; data not shown). We conclude that in the absence of Jag1, embryonic lens progenitor cells accumulate in the AEL, but are unable to maintain all of their epithelial features.
Dramatic reduction in Foxe3 and Ecad expression may reflect a loss of lens progenitor cells by either apoptotic cell death or premature differentiation. As is the case for Le-Cre;RbpjCKO/CKO mutants (Jia et al., 2007; Rowan et al., 2008), we assayed cleaved PARP expression from E10.5 to P3, and found no increase in apoptosis (not shown). Foxe3 is exclusively found in lens progenitor cells and when mutated, causes premature fiber cell differentiation (Blixt et al., 2000; Medina-Martinez et al., 2005; Valleix et al., 2006). Therefore, we determined the percentage of Foxe3-negative cells at E11, E12.5 and E14.5 (Figs. 5A,G–I), as a substitute for quantifying lens fibers directly. Interestingly, at E11 and E14.5, fiber cells are significantly increased. Because removal of Rbpj during this period of lens development produced the same outcome (Jia et al., 2007; Rowan et al., 2008), we conclude that Jag1-Notch signaling regulates aspects of primary and secondary fiber cell differentiation. However, neither β- or γ-Crystallin are expressed precociously or inappropriately in the AEL of E9.5 to E14.5 Jag1 mutant lenses (Suppl Figure 2 and data not shown).
Since diminished Foxe3 and Ecad AEL expression is followed by loss of this lens compartment, we wished to understand better when and to what extent cell proliferation is affected. First, we quantified the percentage of S-phase progenitor cells in BrdU pulsed-labeled embryos and found significant loss of BrdU+ cells during primary and early secondary fiber cell genesis (Figs. 5B, J–L). This could be generally correlated with significant reduction in the percentage of CyclinD1+ cells in E10.5 and E14.5 Le-Cre;Jag1CKO/CKO eyes (Figs. 5C, M–O). But, Cyclin D1-expressing cells rebounded beyond wild type numbers only at E12.5 (Fig. 5C). Interestingly, in Le-Cre;RbpjCKO/CKO lenses, a similar but less robust phenomenon was observed for Cyclin D2 expression (Fig. 6G of Rowan et al., 2008). Paradoxically, in Jag1 conditional lens mutants, the number and pattern of Cyclin D2+ cells was unaffected (Fig. 5D), suggesting that Jag1-Notch signaling specifically regulates an aspect of Cyclin D1 expression, while Cyclin D2 is regulated by the Notch pathway independent of Jag1. Interestingly, the expression of the CKI p27Kip1 was profoundly reduced at E12.5 in Le-Cre;Jag1CKO/CKO lenses (Fig. 5E). This is similar to the reduced numbers of these cells in Le-Cre;RbpjCKO/CKO eyes (Fig. 6K of Rowan et al., 2008), but the temporal kinetics differ between the two lens mutants. Finally, p57Kip2 expressing cells were quantified, and found to be significantly increased only at E14.5 (Fig. 5F). In E14.5 Le-Cre;Jag1CKO/CKO eyes, Cyclin D2 and p57Kip2 expression were properly localized to the lens equator (not shown), suggesting that in the absence of Jag1, the transition zone, like the AEL, initiates formation but cannot be sustained.
Although the mutant phenotypes of Jag1 and Rbpj conditional lens mutants are grossly identical, the loss of Jag1 is more severe and onsets at an earlier stage of lens development. For example, Rbpj conditional lens mutants do not exhibit the same dramatic changes in Foxe3 or Ecad expression at E12.5-E14.5 (Fig. 6 of Rowan et al, 2008). Moreover, different shifts in markers of cell proliferation versus cell cycle exit were found between these two mutants. Thus, we wished to explore the molecular epistatic relationships among several Notch pathway components to determine if there is feedback upon Jag1 expression and how early Jag1 is deleted during lens development. For the latter question, anti-Jag1 immunolabeling demonstrated efficient elimination of the targeted protein in the mouse inner ear (Brooker et al., 2006), so we also used this approach here. In E10.5 sections of embryonic eyes from Le-Cre;Jag1CKO/+intercrosses double labeled with anti-Jag1 and anti-Hes1, Jag1 protein was completely gone from the lens pit of Le-Cre;Jag1CKO/CKO embryos (Fig. 6C), and reduced in those of Le-Cre;Jag1CKO/+embryos (Fig. 6B; n = 3 embryos per genotype). The adjacent optic cup domain was unaffected (asterisks in Figs. 6A–C). Although Hes1 expression appears normal in E10.5 Jag1 lens mutants (arrow in Fig. 6F), by E12.5, it was completely missing from Le-Cre;Jag1CKO/CKO lens vesicles. Thus, the loss of Hes1 protein in Jag1 lens mutants occurs two days earlier than in Rbpj lens mutants.
Next, we reciprocally examined Jag1 expression in Le-Cre;RbpjCKO/+and Le-Cre;RbpjCKO/CKO mutants. From E10.5 to E15.5 no obvious change in Jag1 expression was found (Figs 6J,6K and data not shown; n = 3/3 mutants). To independently confirm this outcome, we also scrutinized Jag1 expression in E10.5 or E13.5 Hes1 germline mutants, but could detect no appreciable difference in the Jag1 expression pattern relative to wild type controls (n = 3/3 mutants; data not shown). At older ages (P0-P3), Jag1 expression is progressively diminished with increasing removal of Rbpj function (Jia et al., 2007; Rowan et al., 2008). But because multiple transition zone and AEL markers and lens size and morphology are all affected in the postnatal lens, these experiments all suggest that Jag1 expression is not regulated by Notch-Rbpj feedback in the embryonic lens.
Here we demonstrate that elimination of Jag1 from early stages of lens development has a moderate effect on primary fiber cell formation, but it is catastrophic for lens cell growth and secondary fiber cell genesis. In Jag1 conditional mutants, AEL progenitor cells and newly postmitotic equatorial transition zone cells cease developing prenatally. This leads to loss of the AEL and transition zone around birth, and in the adult eye, anterior chamber defects that include aphakia.
Removal of the Notch pathway effector Rbpj during early embryonic lens development results in abnormal progenitor cell growth and differentiation, postnatal lens degeneration and adult microphthalmic eyes, with dysgenic lenses, anterior chamber deformities (Jia et al., 2007; Rowan et al., 2008). Perhaps not surprisingly these phenotypes are largely identical to those of Jag1 lens mutants, except that nearly all Jag1 phenotypes are more severe. Because Rbpj, in Notch receiving cells, integrates all canonical pathway input, logically the loss of Rbpj should be just as severe, if not more so, than that of a single ligand or receptor. But, signal pathway effectors like the Rbpj orthologue Su(H) can have multiple, opposing functions (Koelzer and Klein, 2003; Koelzer and Klein, 2006; Morel and Schweisguth, 2000). Both vertebrate Rbpj and Su(H) proteins act as transcriptional repressors, unless they are in a complex with NotchIC and MAML/mastermind where they behave as transcriptional activators (reviewed in Lai, 2002). Thus, deletion of Rbpj simultaneously removes both repressor and activator activities, thereby modulating the Notch-dependent functions. Furthermore, Rbpj can complex with the bHLH factor Ptf1a, independent of Notch signaling (Hori et al., 2008; Masui et al., 2007), meaning that not all Rbpj functions in the lens may be Notch-dependent.
Although we strongly favor the idea that opposing Rbpj functions are sufficient to dampen its mutant phenotypes during lens growth and differentiation, there are other potential explanations for the more severe defects in Jag1 mutants lenses. First, it is plausible that the Le-Cre transgene may delete the Jag1 floxed allele more efficiently than it does the Rbpj floxed allele. Second, the Rbpj protein may perdure longer than Jag1 protein after Cre-mediated deletion. Our data are consistent with this particular idea, since Jag1 protein is completely removed within 24 hours of Le-Cre deletion, resulting in total loss of the Rbpj target gene Hes1 two days earlier than in Rbpj conditional mutants (Figure 6 and Rowan et al., 2008). Finally, it is plausible that Jag1 may regulate some aspect of lens development independent of canonical Notch signaling (Six et al., 2004). In support of this possibility, we observed that Ecadherin and Foxe3 expression are much more severely affected in Jag1 conditional mutants. Further experiments are needed to distinguish among these possibilities. It will be critical to determine the number of Notch ligands, receptors and Rbpj downstream target genes that are present in the embryonic lens, and whether each one acts in the transition zone, fiber cell-AEL boundary, or both.
Activation of Notch signaling can either prevent (lateral inhibition) or promote (lateral induction) ligand production (reviewed in Eddison et al., 2000; Lewis, 1998). In the first situation, a ligand-producing cell successfully signals its neighbor to reduce ligand expression, which reinforces the ability of the first cell to maintain or enhance its own ligand production. Therefore, cells with differing amounts of ligand adopt discrete developmental fates. In tissues where lateral inhibition is active, ligand expression is predicted to be mosaic, in either an on-off or high-low configuration. Two well-known examples of lateral inhibition occur during C. elegans vulval and Drosophila sensory bristle formation (Simpson, 1990; Sternberg, 1988). Alternatively, lateral inductive signaling occurs when a ligand-expressing cell stimulates those nearby to turn up ligand expression, promoting coordinated cell fate specification among a group of cells. Here, ligand expression is predicted to be patchy with precise boundaries. Lateral inductive signaling has been well studied in the Drosophila wing margin, vertebrate limb bud and inner ear (reviewed in Irvine and Vogt, 1997; Lewis, 1998).
Across embryonic lens development, the Jag1 expression pattern changes several times, but is never mosaic, as would be expected for lateral inhibition. In the lens placode and vesicle, Jag1 protein and mRNA expression are essentially ubiquitous, with progressive restriction to the posterior vesicle during primary fiber cell formation. Here, posterior vesicle cells also display strong, uniform Jag1 expression. Another hallmark of lateral inhibition is temporal acceleration of differentiation, which we did not observe in Jag1 lens mutants for primary or secondary fiber genesis. Although the proportions of primary fiber cell differentiation and proliferation shift, neither are accelerated, or abolished, by the loss of Jag1. Therefore, a different Notch ligand probably regulates lateral inhibition for at least the first wave of fiber differentiation.
When secondary fiber cell production initiates, Jag1 expression is further restricted to the transition zone. This particular domain has a sharp anterior boundary with the AEL, but paradoxically is graded peripheral to central. Jag1+ cells passing out of the transition zone appear to cooperatively adopt a secondary fiber cell fate, and without Jag1 this larger, second wave of fiber cell formation ceases by birth. Intriguingly, the transition zone domain is reminiscent of Jag1 expression in prosensory patches of the inner ear (Brooker et al., 2006; Kiernan et al., 2005; Kiernan et al., 2006). Based on both its expression pattern and genetic requirements in the lens, we propose that Jag1 transduces a lateral inductive signal during primary fiber cell genesis that is relatively weak, perhaps because a separate lateral inhibitory signal is the primary mode of Notch regulation for this cell type. Then during secondary fiber cell formation, Jag1-dependent inductive signaling becomes concentrated in the transition zone. Because loss of Jag1 causes increased number of fiber cell differentiating at the same time that progenitor cell proliferation decreases, it remains unclear from our analyses whether at the equator Jag1 strictly regulates cell cycle exit.
Finally, Jag1 is expressed along the anterior edge of fiber cells, where they border the AEL growth zone. AEL progenitor cells are devoid of Jag1 ligand, but express at least two Notch receptors, plus the downstream effectors Rbpj and Hes1 (Bao and Cepko, 1997; Jia et al., 2007; Rowan et al., 2008; Weinmaster et al., 1992). Thus, a lateral inhibitory signal between fiber cells and the AEL might be predicted here, in which activated Notch in AEL cells suppresses Jag1 expression. However, loss of Rbpj or Hes1 did not result in derepression of Jag1 in the AEL; nor did misexpression of activated Notch downregulate Jag1 (Rowan et al., 2008, this paper). This strongly implies that if a Notch lateral inhibition signal does traverse this tissue boundary, Jag1 neither communicates it, nor is regulated by such a signal. Instead, our findings are consistent with the idea that Jag1+ fiber cells act concertedly to keep Notch activity high in the AEL, which blocks premature progenitor cell differentiation. Interestingly, Notch signaling at AEL/fiber cell boundary is quite similar to the dorsal-ventral tissue boundaries of the fly wing and vertebrate limb bud, where Serrate-expressing cells abut the wing margin or limb AER, which each undergo outgrowth (Irvine, 1999; Irvine and Vogt, 1997). In both types of appendages, the Serrate-mediated signal is modulated to be unidirectional through the activity of fringe (Irvine and Wieschaus, 1994). It is tantalizing to speculate whether a mammalian fringe-like gene is present the developing lens. In the future it will be important to test for expression and function of other Notch pathway ligands at the AEL-fiber cell boundary, and remove Jag1 function specifically in fiber cells. The latter experiment should indicate if Jag1 transduces a lateral inductive signal to anterior proliferating lens progenitor cells, and help assign particular Jag1 phenotypes to fiber cell or transition zone expression domains.
A–C) The β-Crystallin expression domain (red) is progressively lost and mislocalized to the anterior in Le-Cre; Jag1CKO/+(B) and Le-Cre; Jag1CKO/CKO eyes (C). D–F) E-cad (green) and p57Kip2 (red) double labeling shows analogous loss of the AEL and transition zone compartments. G-I) Foxe3+ cells are patchy in Le-Cre; Jag1CKO/+lenses (H) and essentially missing from Le-Cre; Jag1CKO/CKO eyes. Remnant staining seen in mutants at this age is nonnuclear. Bar in A,B,D–H = 5 microns; in C,I = 10 microns; n ≥ 3 animals per genotype.
A–F) Anti β-Crystallin (white) and DAPI (blue) colabeling at E12.5 (A–C) and E14.5 (D–F). At both ages, Crystallin expression is properly localized in the posterior compartment. G–I) γ-Crystallin (white) and DAPI (blue) at E14.5 also shows proper localization in the posterior compartment. Although mutants in F and I appear to potentially have reduced Crystallin expression or fewer Crystallin+ lens fiber cells, this was not reproducible in other mutants. Bar in A = 20 microns, distal is up in all panels.
The authors thank Julian Lewis for Jag1CKO mutant mice; Tasuku Honjo for RbpjCKO mice; Ruth Ashery-Padan and Richard Lang for Le-Cre transgenic mice; Peter Carlsson, Marten Smidt, Richard Lang, Grant Mastick and Guillermo Oliver for antibody reagents; Sheldon Rowan, Senthil Saravanamuthu, Peggy Zelenka and Brian Gebelein for valuable discussion and critical reading of this manuscript. This work was supported by NIH grants EY18097 and CHRF research funds to NLB.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.