SEARCH

SEARCH BY CITATION

Keywords:

  • Cecr2;
  • neurulation;
  • inner ear defects;
  • stereocilia bundles;
  • planar cell polarity

Abstract

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

The loss of Cecr2, which encodes a chromatin remodeling protein, has been associated with the neural tube defect (NTD) exencephaly and open eyelids in mice. Here, we show that two independent mutations of Cecr2 are also associated with specific inner ear defects. Homozygous mutant 18.5 days post coitus (dpc) fetuses exhibited smaller cochleae as well as rotational defects of sensory cells and extra cell rows in the inner ear reminiscent of planar cell polarity (PCP) mutants. Cecr2 was expressed in the neuroepithelium, head mesenchyme, and the cochlear floor. Although limited genetic interaction for NTDs was seen with Vangl2, a microarray analysis of PCP genes did not reveal a direct connection to this pathway. The mechanism of Cecr2 action in neurogenesis and inner ear development is likely complex. Developmental Dynamics 240:372–383, 2011. © 2011 Wiley-Liss, Inc.


INTRODUCTION

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

Neural tube defects (NTDs) are one of the most common congenital birth defects, affecting ∼1/1,000 human births worldwide with varying incidence between populations (Blom,2009). The process by which the neural tube closes involves many steps, from early neural induction to the elongation, elevation, bending, and fusion of the neural folds to create a hollow neural tube (Copp et al.,2003). Failure to properly complete any one of these processes can lead to a NTD. In humans, NTD susceptibility shows complex inheritance with both environmental and genetic contributions. One of the most notable environmental factors is perinatal folic acid intake (Smithells et al.,1976; Blom,2009). In mice, however, most NTDs are due to single gene defects (Harris and Juriloff,2007). Many genes have been implicated in NTDs, including pathways associated with Sonic Hedgehog (Shh), bone morphogenetic proteins (BMPs), and planar cell polarity (PCP; Copp et al.,2003; De Marco et al.,2006). Each pathway, when disrupted in mice, shows stereotypical phenotypes which include NTDs.

The PCP pathway establishes organization within an epithelial sheet. PCP was first demonstrated in Drosophila wings, eyes, abdomen, and notum (Gubb and Garcia-Bellido,1982; Fanto and McNeill,2004). These tissues have regular structures, such as the ommatidia of the eye or bristles on the abdomen and wing, which orient in regular patterns with respect to each other. A “core” set of PCP genes was established in Drosophila (Fanto and McNeill,2004). Homologues in vertebrates include Vangl1/2, Frizzled3 and 6, Disheveled1-3, Celsr1, Prickle, Scribbled, PTK7, and Ankrd6 (Schwarz-Romond et al.,2002; Carreira-Barbosa et al.,2003; Curtin et al.,2003; Montcouquiol et al.,2003; Lu et al.,2004; Wang et al.,2006a,b; Torban et al.,2007). Examples of PCP can be observed throughout vertebrate embryogenesis, most notably in the convergent extension (CE) movements required to elongate and shape the neural epithelium during early neurulation, as well as the highly organized epithelium within the cochlear sensory structures (Park and Moon,2002; Wang et al.,2005). PCP has also been implicated in the cell movements that form the eyelid (Murdoch et al.,2001; Wang and Nathans,2007). Defects in PCP genes constitute approximately 5% of all known murine NTD mutations and are usually associated with craniorachischisis (a lack of neural tube closure throughout its length), open eyelids, and rotational defects of the cochlear sensory cells (Kibar et al.,2001; Murdoch et al.,2001; Hamblet et al.,2002; Curtin et al.,2003; Montcouquiol et al.,2003; Lu et al.,2004; Wang et al.,2006b; Harris and Juriloff,2007; Torban et al.,2008).

The murine gene Cecr2 is associated with exencephaly, a perinatal-lethal cranial NTD, but the pathway in which Cecr2 acts is not clear. Two different Cecr2 mutations have been reported. A hypomorphic genetrap mutation (Cecr2Gt45Bic) develops exencephaly with 74% penetrance on a BALB/c genetic background (Banting et al.,2005). Open eyelids are present in most mutants. A deletion mutation Cecr2tm1.1Hemc was recently described (Fairbridge et al.,2010). Cecr2tm1.1Hemc mutants show 96% penetrance for exencephaly, usually accompanied by open eyelids. Neither mutation has ever been associated with spina bifida or craniorachischisis.

CECR2 is a chromatin remodeling protein that forms the CERF complex with SNF2L/SMARCA1 (Banting et al.,2005). The specific role of the CERF complex in neurulation is not known. Chromatin remodeling proteins typically are involved in the alteration of DNA-histone contacts and nucleosome repositioning and are of importance to cellular process such as transcription, replication, and repair (Aalfs and Kingston,2000). Other chromatin remodeling proteins have been associated with NTDs. For instance, heterozygotes for mutations in SWI/SNF chromatin remodeling components Brg1/Smarca4 and BAF155/Srg3/Smarcc1 are both associated with 14–20% exencephaly (Bultman et al.,2000; Kim et al.,2001).

We now report that both Cecr2 mutations causing exencephaly also result in inner ear defects in the form of shortened and wider cochlear ducts as well as misorientation of the sensory cells within the organ of Corti. Using the LacZ fusion protein of the Cecr2Gt45Bic mutation, we found that Cecr2 is expressed in the early otic placode and persists in the cochlear floor of the inner ear at later stages. The specific inner ear defect coupled with both a neural tube defect and open eyelids led us to investigate the PCP pathway. We demonstrated that during cranial neurulation Cecr2Gt45Bic homozygous mutants had a wider distance between the cranial neural folds. Cecr2Gt45Bic heterozygous embryos had a slight delay in closure, suggesting that Cecr2 shows a degree of dosage sensitivity. Genetic interaction for NTDs was seen when the Cecr2tm1.1Hemc mutation was crossed with Vangl2, but not for the inner ear defect. Microarray analysis and quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR) of whole embryos at the time of neurulation were performed to investigate a possible influence of Cecr2 on PCP pathway gene expression; however, no direct connection was found. We discuss the possible relationship between Cecr2 and PCP in neurulation and inner ear development.

RESULTS

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

During Neurulation Cecr2 Is Expressed in Both the Neural Epithelium and Head Mesenchyme

To better understand the role of Cecr2 during neurulation, we examined its expression during this time using the LacZ fusion protein of the Cecr2Gt45Bic mutation. Xgal staining at 13.5 days post coitus (dpc) previously showed that Cecr2 is expressed in the nervous system (brain, spinal cord, spinal ganglia, and forming retina), as well as the nasal epithelium, lens, limb, and intercostal mesenchyme (Banting et al.,2005). Whole-mount embryos at 9.5 dpc showed Xgal staining along the closed neural tube.

We observed sections of Cecr2Gt45Bic homozygous embryos between 5 and 16–17 somite pairs (cranial neural tube closure occurs between 10 and 14 somite pairs; Fig. 1). We documented strong expression across the entire neuroepithelium throughout this time window (Fig. 1A–C,E,F). Head mesenchyme stained strongly at five somite pairs, suggesting a possible role for Cecr2 in neural fold elevation through head mesenchyme proliferation (Fig. 1A,B). The foregut/midgut also strongly expressed Cecr2 at this early stage (Fig. 1C). Expression was not seen in the heart or the intraembryonic coelomic cavities (which form the future pericardio-peritoneal canal, Fig. 1B,C). At 16–17 somite pairs, Cecr2 expression persisted within the neural tube but appeared to fade caudally (Fig. 1E,F). Cecr2 was expressed in the developing hindgut and the pharyngeal region of the foregut as well as within the first branchial arches and the epithelial groove separating the branchial arch from the optic eminence (Fig. 1E). Representative controls showed no endogenous staining at these age points (Fig. 1D,G).

thumbnail image

Figure 1. Cecr2 is expressed in the neural tube and head mesenchyme during neurulation. A–C: Xgal-stained five somite pair Cecr2Gt45Bic homozygous mutant (rostral, middle, and caudal respectively). E,F: Xgal-stained 16–17 somite pair Cecr2Gt45Bic homozygous mutant (rostral and caudal respectively). D,G: Representative age-matched Xgal-stained wild-type control embryos. Arrowheads point to the neural folds, and an asterisk labels mesenchyme. Structures of interest have been labeled: the closing and closed neural tube (NT), optic vesicle (OV), developing gut (G), branchial arches (BA), coelomic cavity (CC) or the heart (H) and dorsal aorta (DA). A diagrammatic representation of each embryo stage (top right of each panel) indicates the level and plane of section. Scale bars = 50 μm.

Download figure to PowerPoint

Neural Folds in Cecr2Gt45Bic Homozygous Embryos Elevate but Are Abnormally Wide

We collected embryos at the time of neurulation to examine the extent of formation of the neural folds in mutants (Fig. 2). Cecr2Gt45Bic heterozygous embryos showed neural tube closure as expected, but the timing was delayed slightly. By 14 somite pairs, only 5.7% (3/53) of wild-type neural folds were still open, while 25.4% (16/63) of Cecr2Gt45Bic heterozygous neural folds remained open. By 16 somite pairs, all (24/24) wild-type neural folds observed had met at the dorsal midline, while 11.43% (4/35) of Cecr2Gt45Bic heterozygous neural folds remained open. At 18 somite pairs, one heterozygous Cecr2Gt45Bic sample out of 14 observed remained open (data not shown). Few of these late-closure embryos would be destined to develop exencephaly, since the frequency of exencephaly observed in the older Cecr2Gt45Bic heterozygous embryos collected in this study was 1.3% (1/73). By 19 somite pairs, all heterozygous Cecr2Gt45Bic embryos had closed neural tubes (n = 6, data not shown).

thumbnail image

Figure 2. Neural folds of Cecr2Gt45Bic homozygous embryos elevate but are wider than wild-type or Cecr2Gt45Bic heterozygous embryos. A–R: Whole-mount imaging of embryos ranging from 7–18 somite pairs of Cecr2Gt45Bic homozygous mutants, Cecr2Gt45Bic heterozygotes, and wild-type controls. Images were taken at various magnifications. G,I: Lines indicated where embryos were measured across the branchial arches as well at the medial hinge points to determine a head: neural fold width ratio. S: Graphical representation of neural fold to head width ratios of Cecr2Gt45Bic homozygotes, Cecr2Gt45Bic heterozygotes, and wild-type controls. The number within the bar gives the number of embryos observed in each category. P values are given in the text. T: Scanning electron micrograph of a Cecr2Gt45Bic homozygous embryo at 18 somite pairs. Scale bar = 200 μm.

Download figure to PowerPoint

Cecr2Gt45Bic homozygous embryos from 11–14 somite pairs were observed to form both medial and dorsolateral hinge points; however, after failure of the folds to close, the cranial neural folds then splayed open (Fig. 2R,T). Specifically, in approximately 71% of the mutant embryos, commensurate with the penetrance of exencephaly expected, the folds were likely too far apart to allow apposition and fusion. The separation between the neural folds was determined by measuring the distance between the widest portion of the cranial neural tube (the dorsal lateral hinge points) and branchial arches to create a ratio (Fig. 2G,I). The neural fold to head width ratio of Cecr2Gt45Bic homozygous embryos during neurulation was 25% larger than that of wild-type controls (Fig. 2S). The average Cecr2Gt45Bic homozygous neural fold distance to head width ratio was 0.755 compared with 0.533 for wild-type, which was significantly different by a two-tailed t-test (P value < 0.05). Cecr2Gt45Bic heterozygous embryos had a neural fold distance to head width ratio of 0.566, slightly wider but not significantly different from wild-type controls (P value = 0.117). The Cecr2Gt45Bic heterozygous neural fold distance to head width ratio was also significantly different from that of Cecr2Gt45Bichomozygotes (P value < 0.05). Approximately 21% of Cecr2Gt45Bic homozygous embryos observed (3/14) had a head to fold width ratio comparable to that of wild-type and Cecr2Gt45Bic heterozygous samples, suggesting that these embryos would have been nonpenetrant and would have closed their neural tubes.

In addition to a NTD, Cecr2Gt45Bic mutant embryos also showed 81% penetrance of open eyelids in exencephalic mutants (13/16 plus one with only one eye open). Open eyelids were not observed in Cecr2Gt45Bic homozygotes nonpenetrant for exencephaly or heterozygotes (data not shown).

Cecr2 Is Expressed Throughout Inner Ear Development

The combination of a NTD associated with a wider neural fold separation, as well as a high frequency of open eyelids, was reminiscent of a PCP phenotype and led us to examine the expression of Cecr2 in the forming inner ear. Stages examined ranged from the otic placode at 10 somite pairs (Fig. 3A) to 18.5 dpc showing a nearly mature cochlear duct (Fig. 3E).

thumbnail image

Figure 3. Cecr2 is expressed during inner ear development. A–E: Xgal staining of Cecr2Gt45Bic homozygous embryos are shown at 8.0 days post coitus (dpc), 9.5 dpc, 11.5 dpc, 13.5 dpc, and the inner ear at 14.5 dpc, respectively. Arrowheads point to the developing neural tube and brain (A, B, and C, respectively). Various stages of the developing ear have been labeled: otic placode (P), otic vesicle (OV), and optic vesicle (asterisks). Other structures: semicircular canals (SCC), statoacoustic ganglion (SAG), cochlear floor (CF). F: A representative Xgal stain control at 14.5 dpc. G,J: Whole-mount inner ears of both wild-type and Cecr2Gt45Bic homozygotes respectively, indicating the cochlea (C, bracket) and vestibule (V, bracket). H,K: Xgal-stained 18.5 dpc cochlear ducts of both wild-type and Cecr2Gt45Bic homozygotes, respectively. I,L: Xgal-stained 18.5 dpc dissected cochlear ducts in paraffin section for both wild-type and Cecr2Gt45Bic homozygotes respectively; the cochlear floor (CF) and location of the sensory hair cells (brackets) have been labeled.

Download figure to PowerPoint

Xgal staining of Cecr2Gt45Bic homozygous embryos revealed very early expression in the otic placode at 10 somite pairs (Fig. 3A). Cecr2 expression persisted through the formation of the otocyst (Fig. 3B), and during the stages of early inner ear morphogenesis between 11.5 dpc and 13.5 dpc (Fig. 3C,D). The cochlear duct was extracted at 14.5 dpc and 18.5 dpc and individually stained with Xgal. Cecr2 was expressed in the developing cochlear floor at both 14.5 dpc and at 18.5 dpc (Fig. 3E,L). Cecr2 was expressed in the developing statoacoustic ganglion (SAG) at 11.5 dpc and 14.5 dpc (Fig. 3C,E). Cecr2 wild-type controls showed no detectable endogenous staining in any of the cochlear structures or the SAG (Fig. 3F,I).

Xgal staining of Cecr2Gt45Bic homozygous whole-mount cochlear ducts at 18.5 dpc revealed that Cecr2 was expressed in a band along the longitudinal axis of the duct, with the most staining occurring at the apical end, where the most immature hair cells were located (Fig. 3K). This staining pattern faded out basally, where the most mature cells were located. Paraffin sectioning of these samples revealed that Cecr2 localized to the cochlear floor medial to the sensory cells (Fig. 3L).

Cecr2 Mutants Show a Smaller, Wider Cochlea, and Stereocilia Disorganization

Inner ears were dissected from 18.5 dpc embryos and imaged for gross morphological differences. Whole inner ears from exencephalic homozygous Cecr2Gt45Bic embryos were found to be approximately 1/3 smaller when compared with wild-type or heterozygous Cecr2Gt45Bic litter mates (Fig. 3G,J). The inner ears from homozygous Cecr2Gt45Bic samples that were nonpenetrant for exencephaly were comparable in size to that of wild-type and Cecr2Gt45Bic heterozygous littermates (data not shown). The dissected cochlear ducts from homozygous Cecr2Gt45Bic exencephalic embryos were found to be shorter and wider than wild-type control embryos (Fig. 3H,K).

We analyzed for inner ear defects on both the hypomorphic allele Cecr2Gt45Bic, and the deletion allele Cecr2tm1.1Hemc. Rotational defects of sensory cells were analyzed by dissecting the organ of Corti, followed by fluorescent staining and imaging by confocal microscopy at the basal (5%), middle (50%), and apical (75%) positions along the organ of Corti epithelial length (examples of the middle position data are presented in Fig. 4A). Cecr2Gt45Bic and Cecr2tm1.1Hemc homozygous samples showed an increase in the number of misaligned cells (represented in red, which fall outside the ± 30 degree rotation from the medial lateral axis). Cecr2Gt45Bic heterozygous samples were also affected but to a lesser extent. Cecr2tm1.1Hemc heterozygotes showed a similar intermediate pattern (data not shown).

thumbnail image

Figure 4. Loss of Cecr2 results in PCP-like inner ear defects. A: Confocal images of fluorescently stained organ of Corti with phalloidin (actin in green) and acetylated anti-tubulin (kinocilium in magenta) of wild-type, Cecr2Gt45Bic heterozygotes, Cecr2Gt45Bic homozygotes, and Cecr2tm1.1Hemc homozygotes. Sensory cell rows have been labeled as IHC (inner hair cells) and OHC 1–3 (outer hair cells), and supportive cells as PSC. Below each confocal micrograph, are diagrammatic representations of the sensory cells indicating misaligned cells (red) and correctly aligned cells (black). B: Histograms indicate orientation of cells within the apical OHC3 layer for each genotype. C: Graphical representations for each cell layer indicating the percent of cells rotated more than ± 30 degrees over the basal, middle, and apical regions of the organ of Corti. Significant differences are indicated by brackets for Cecr2Gt45Bic and an asterisk for Cecr2tm1.1Hemc to wild-type comparisons.

Download figure to PowerPoint

The variance and spread of rotation defects was visualized by plotting histograms of individual cell variance from the medial lateral axis (Fig. 4B). The majority of wild-type cells showed negligible rotation variance with an upper limit of 50 degrees. Cecr2Gt45Bic heterozygotes and both Cecr2Gt45Bic and Cecr2tm1.1Hemc homozygotes showed an increased spread of rotations throughout the full range of 0 to 180 degrees, and the average bulk of cells tended to drift farther away from medial lateral axis (Fig. 4B).

Each sensory cell row (one row of inner hair cells [IHC] and three rows of outer hair cells [OHC1, OHC2, and OHC3]) was independently assessed and compared between genotypes for basal, medial, and apical positions (Fig. 4C). The general trend was that the Cecr2Gt45Bic homozygotes were the most severe over all layers and Cecr2Gt45Bic heterozygotes were intermediately affected when compared with wild-type controls (Fig. 4C). The mature sensory cells at the basal position showed the least misalignment for all genotypes and cell layers, whereas misalignment increased as imaging moved more apically to more immature sensory cells. The IHC was the least affected cell layer in all genotypes, with a progressive increase in misalignment severity through the OHC1, OHC2, and OHC3 layers, with the OHC3 being the most disorganized (Fig. 4C). These observations agree with different waves of maturation occurring within the organ of Corti at this time (Kelly and Chen,2007). The Cecr2tm1.1Hemc homozygous mutants appeared less severe than the Cecr2Gt45Bic line at the IHC, OHC1, and OHC2 layers (Fig. 4C).

To compare the differences between each genotype at each position, a chi-square test for independence was used to analyze the data set. Statistical significance was determined to be a P value < 0.016 using a Bonferroni downward adjustment for comparisons made within the Cecr2Gt45Bic data set. Within the IHC layer of all positions imaged, no significant differences were found (Fig. 4C). Within the OHC1 cell layer, we found no significant differences in the basal position; however, toward the middle position of the organ of Corti, Cecr2Gt45Bic homozygous mutants were different and more severe from wild-type or Cecr2Gt45Bic heterozygotes. OHC1 in the apical region showed differences between wild-type and both Cecr2Gt45Bic heterozygotes and homozygotes. At the OHC2 layer, the Cecr2Gt45Bic homozygous samples were different from the other genotypes at all positions. Wild-type and Cecr2Gt45Bic heterozygotes were indistinguishable until the apical comparison, where they then showed significant difference. Finally, the OHC3 layer had the greatest difference between genotypes. Only two comparisons did not show significance, those being wild-type to Cecr2Gt45Bic heterozygotes basally and Cecr2Gt45Bic heterozygotes to homozygotes apically.

Cecr2Gt45Bic homozygous embryos that were nonpenetrant for the exencephaly phenotype were also assessed for misalignment of the sensory cells. Generally, this category was found to be within the Cecr2Gt45Bic heterozygous and the homozygous exencephalic misalignment ranges; however, the small sample size was limiting (data not shown).

The second Cecr2tm1.1Hemc line was compared independently with wild-type samples using a chi-square test of independence. Statistical differences were determined using a P value <0.05. Cecr2tm1.1Hemc homozygous samples were found to be statically different basally only in the OHC3 layer. Within the middle position, Cecr2tm1.1Hemc homozygous samples showed greater misalignment than wild-type in all layers except OHC1. Cecr2tm1.1Hemc homozygous mutant to wild-type comparisons in the apical position were only significant in OHC3, although this class of comparisons contained smaller sample sizes than other categories (Fig. 4C).

The overall appearance of the rows of hair cells was also affected in both Cecr2 mutations in the apical region (Fig. 5). In 3/10 Cecr2Gt45Bic homozygous samples there were regions where an abnormal fourth row of OHC was present (Fig. 5B). Another 3/10 Cecr2Gt45Bic homozygous samples observed showed extra OHC cells present; however, not enough for a complete fourth row. Extra cells protruding from the IHC row were also seen 3/10 Cecr2Gt45Bic homozygous samples. Cecr2tm1.1Hemc homozygous mutants were affected in a comparable manner (Fig. 5C). Cecr2Gt45Bic heterozygotes and wild-type rarely showed a small number of extra cells in the OHC layer, with a frequency of 2/17 and 1/13 respectively.

thumbnail image

Figure 5. Cecr2 mutants show increased cell rows at the apical region. Imaging was optimized to show cell outlines rather than stereocilia structures. A–C: Confocal images of fluorescently stained organ of Corti with phalloidin (actin in green) of wild-type, Cecr2Gt45Bic, and Cecr2tm1.1Hemc homozygous mutants. Arrowheads show the number of cell layers. Extra cells are indicated by an asterisks.

Download figure to PowerPoint

Cecr2 Does Not Alter PCP Transcript Expression During Cranial Neural Closure

Since the inner ear defects resembled those seen in PCP defects, we analyzed a previous 9.5 dpc whole embryo microarray data set (Fairbridge et al.,2010) to examine how PCP and PCP-related transcripts responded to the disruption of Cecr2Gt45Bic during neural tube closure (the Cecr2tm1.1Hemc line was not available when the microarray was performed). A total of 94 genes (with multiple transcripts), both core PCP genes, and genes related to this pathway, were present on the Affymetrix Mouse Genome 430 2.0 microarray. All PCP transcripts with significant changes (P value < 0.05) are shown in Table 1. A subset of these transcripts was further validated by qRT-PCR along with several candidates that were either not present on the microarray or whose P value was only suggestive (Table 2). The qRT-PCR was performed on Cecr2tm1.1Hemc homozygous embryonic head samples collected around the time of neural closure, to validate on a second, independent Cecr2 mutation.

Table 1. PCP Genes or Homologues of PCP Genes, and Genes Related to PCP Were Identified in a Microarray Analysis of Embryos During Neurulationa
SymbolNameFold changeP valueProbe ID
  • a

    Genes with at least one probe showing a P value <0.05 are listed. Bolded genes indicate probe sequences used for quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR) primer design. Five genes with P values <0.1 are also listed as they were also tested by qRT-PCR.

Fzd1frizzled homolog 11.14>0.051422985_at
Fzd1frizzled homolog 11.210.051437284_at
Fzd10frizzled homolog 101.00>0.051440182_at
Fzd10frizzled homolog 101.180.051455689_at
Fzd2frizzled homolog 2−1.03>0.051418534_at
Fzd2frizzled homolog 21.210.051418533_s_at
Fzd2frizzled homolog 21.320.051418532_at
Fzd7frizzled homolog 71.210.051450044_at
Fzd7frizzled homolog 71.26>0.051450043_at
Frzbfrizzled-related protein−1.500.011448424_at
Frzbfrizzled-related protein−1.370.031416658_at
Lix1limb expression 1−1.553.33E-041421180_at
Nlknemo like kinase−1.330.041456467_s_at
Nlknemo like kinase−1.120.011435970_at
Nlknemo like kinase−1.07>0.051419112_at
Pard3par-3 (partitioning defective 3)1.01>0.051436764_at
Pard3par-3 (partitioning defective 3)1.150.051436765_at
Rac1RAS-related C3 botulinum substrate 1−1.086.01E-041423734_at
Rac1RAS-related C3 botulinum substrate 11.00>0.051437674_at
Ror1receptor tyrosine kinase-like orphan receptor 1−1.120.041429313_at
Ror1receptor tyrosine kinase-like orphan receptor 11.02>0.051451014_at
Ror1receptor tyrosine kinase-like orphan receptor 11.14>0.051429312_s_at
Ror2receptor tyrosine kinase-like orphan receptor 2−1.199.12E-041423428_at
Vangl2vang-like 2−1.05>0.051436118_at
Vangl2vang-like 21.11>0.051419218_at
Vangl2vang-like 21.270.051455592_at/
Wnt5bwingless-related MMTV integration site 5B−1.594.16E-031439373_x_at
Wnt5bwingless-related MMTV integration site 5B−1.03>0.051422602_a_at
Table 2. A Subset of PCP Genes, Identified in a Microarray Analysis of Embryos During Neurulation With at Least One Probe Being Suggestive, Showing a P Value < 0.1a
SymbolNameFold changeP valueProbe ID
  • a

    Bolded genes indicate probe sequence used for qRT-PCR primer design. PCP, planar cell polarity.

Celsr2cadherin, EGF LAG seven-pass G-type receptor 21.010.871435336_at
Celsr2cadherin, EGF LAG seven-pass G-type receptor 21.260.061422073_a_at
Gsk3bglycogen synthase kinase 3 beta−1.180.061454958_at
Gsk3bglycogen synthase kinase 3 beta−1.080.301434439_at
Gsk3bGlycogen synthase kinase 3 beta1.060.761439949_at
Gsk3bglycogen synthase kinase 3 beta1.160.211439931_at
Gsk3bglycogen synthase kinase 3 beta1.230.061451020_at
Gsk3bglycogen synthase kinase 3 beta1.390.141437001_at
Prickle1prickle like 1−1.210.071452249_at
Prickle1prickle like 11.010.341444759_at
Wnt7bwingless-related MMTV integration site 7B1.050.601420891_at
Wnt7bwingless-related MMTV integration site 7B1.410.091420892_at

A total of eight PCP genes had transcripts that indicated significant expression change in the microarray data set, and a further four from the suggestive list were chosen (Tables 1, 2). Despite the significant P values, many of these 12 genes demonstrated low fold changes. The highest fold changes obtained from the RNA microarray of whole embryos were Frzb (−1.50), Lix1 (−1.55), and Wnt5b (−1.59; Table 1). When the qRT-PCR analysis focused only on the Cecr2tm1.1Hemc heads at the same time point, no transcript showed significant fold changes (Table 3). Further qRT-PCR analysis of post-closure 16–18 somite pairs Cecr2tm1.1Hemc embryos confirmed a significant decrease in Lix1 expression, and qRT-PCR of embryonic bodies did confirm Frzb expression changes (data not shown). Thus, there remains the possibility of Cecr2/PCP interactions in other tissues and processes during development, but we detected no expression changes at the time of neural tube closure in the head. Thus, the Cecr2 exencephaly phenotype is unlikely to be directly influenced by misregulation of PCP genes, although this does not rule out an effect on localization of PCP proteins.

Table 3. qRT-PCR Analysis of a Subset of PCP Genes Chosen From a Microarray Analysis or That Were Not Represented on the Microarraya
SymbolFold changeP value
  • a

    Gene not represented on microarray. qRT-PCR, quantitative reverse transcriptase-polymerase chain reaction; PCP, planar cell polarity.

Frzb−1.1> 0.05
Fzd1−1>0.05
Fzd10−1.21>0.05
Fzd21.1>0.05
Fzd5−1.07>0.05
Fzd 71.21>0.05
Gsk3b1.08>0.05
Lix1−1.21>0.05
Prickle 11.02>0.05
Vangl1a1.07>0.05
Vangl2−1.18>0.05
Wnt5b1.04>0.05
Wnt7b−1.06>0.05

Cecr2tm1.1Hemc/tm1.1Hemc; Vangl2Lp/+ Embryos Show Increases in NTD Occurrence

A role in PCP has been suggested for several genes by looking for genetic interactions with the Vangl2Lp (Looptail) mutation. We, therefore, harem bred N6 BALB/c Cecr2tm1.1Hemc heterozygous females, which carry the more severe Cecr2 allele (Fairbridge et al.,2010), with a C3H/C57Bl/6J hybrid Vangl2Lp/+ male (obtained from Dr. Philippe Gros). Although Vangl2Lp mutants show 100% craniorachischisis, only 5% of heterozygotes show spina bifida (Merte et al.,2010). A similar penetrance was seen for both Vangl2Lp heterozygotes (8.2%, 4/49) and, Cecr2tm1.1Hemc; Vangl2Lp double heterozygotes (5.6%, 3/54; Table 4). The Cecr2tm1.1Hemc/tm1.1Hemc;Vangl2Lp/+ embryos generated showed a genetic interaction, displaying 58.3% spina bifida (4/12 with spina bifida only and 3/12 with both spina bifida and exencephaly, Table 4). This result is significantly different from the Vangl2Lp/+ spina bifida frequency using chi-square test for independence (P value < 0.0006; Table 4). Of interest, of the 12 Cecr2tm1.1Hemc homozygous embryos observed, only 7 showed exencephaly (Table 4). This is less than the expected 96% (Fairbridge et al.,2010), which probably reflects the now mixed genetic background of this experiment (BALB/c X LPT/LeJ, a C3H/C57Bl/6J hybrid).

Table 4. NTD Phenotype Analysis of the Cecr2tm1.1Hemc Mutation Crossed With the Vangl2Lp/+ Mutationa
GenotypeTotal nNo NTDSBexenSB and Exen% SB% exen
  • a

    Cecr2tm1.1Hemc/tm1.1Hemc:Vangl2Lp/+ embryos showed an increase in spina bifida compared to Vangl2Lp/+ embryos (P value < 0.0005). SB, spina bifida; exen, exencephaly; NTD, neural tube defect.

+/+50500000%0%
Cecr2tm1.1Hemc/+36360000%0%
Cecr2tm1.1Hemc/tm1.1Hemc1250700%58%
Vangl2Lp/+49454008.2%0%
Cecr2tm1.1Hemc/+: Vangl2Lp/+54513005.6%0%
Cecr2tm1.1Hemc/tm1.1Hemc: Vangl2Lp/+12243358.3%50%

The inner ears of the Cecr2tm1.1Hemc; Vangl2Lp double heterozygotes were also examined (Fig 6.D,E). Both single heterozygotes are known to show a moderate degree of sensory cell misalignment. However, the Cecr2tm1.1Hemc; Vangl2Lp double heterozygote showed no increase in misalignment (P value >0.05), indicating a lack of genetic interaction for this phenotype with this genotype.

thumbnail image

Figure 6. Organ systems analyzed when the Cecr2tm1.1Hemc mutation was crossed with the Vangl2Lp mutation. A–C: Whole-mount images of embryonic lower spine in wild-type (left), Vangl2Lp/+ (middle, showing a curled tail) and Cecr2tm1.1Hemc/tm1.1Hemc; Vangl2Lp/+ (right, showing a bent tail). A black arrow highlights spina bifida in the Cecr2tm1.1Hemc/tm1.1Hemc; Vangl2Lp/+ embryo. D,E: Confocal images of fluorescently stained organ of Corti with phalloidin (actin in green) and acetylated anti-tubulin (kinocilium in magenta) of Vangl2Lp/+ and Cecr2tm1.1Hemc; Vangl2Lp double heterozygotes. Examples of misaligned cells are indicated by an asterisk.

Download figure to PowerPoint

DISCUSSION

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

This is the first study to show that Cecr2 mutants have inner ear polarity defects. Multiple signaling pathways have been implicated in the development of the inner ear, including PCP, notch, fibroblast growth factor (FGF), BMP/Noggin, and hedgehog signaling (De Marco et al.,2006). Both Cecr2Gt45Bic and Cecr2tm1.1Hemc homozygous mutants show defects in the orientation of the stereocilia bundles and the organization of the hair cell rows. In addition, we have previously shown that Cecr2 mutants have NTDs (exencephaly) and open eyelids (Banting et al.,2005). This combination of defects lead us to examine whether the PCP pathway was being affected by the loss of Cecr2.

Most PCP genes cause the more severe NTD craniorachischisis (Wang and Nathans,2007). However, there are also several PCP mutants that do show exencephaly. The Fz3 mutation on a 129/SVJ x C57BL6 background shows a small percentage of exencephaly, which appears to be dependent on modifiers (Wang et al.,2002). Embryos deficient in Dvl2 show a low percentage of exencephaly and spina bifida, whereas most Dvl1−/−; Dvl2−/− double mutants show craniorachischisis on a 129SvEv inbred strain and a mixed 129SvEv x NIH Black Swiss strain (Hamblet et al.,2002). Other combined mutations have also shown a percentage of exencephaly, including Dvl3±; Vangl2Lp+ (5 cranio, 2 exen /22), whereas Dvl3−/−; Vangl2Lp/+ showed only craniorachischisis (6/16) on a mixed genetic background with 129S6 (Etheridge et al.,2008). Smurf ubiquitin ligases1 and 2, involved in degradation of the PCP core protein Prickle1, show exencephaly and/or spina bifida as well as inner ear defects in 25% of embryos which lose three of four Smurf alleles (Narimatsu et al.,2009). The crooked tail allele of Lrp6 allele on a mixed DBA/2J x A/J genetic background shows ∼20–30% exencephalic embryos, and not craniorachischisis or spina bifida (Carter et al.,2005). Although associated with canonical Wnt signaling, Lrp6 is necessary for Xenopus convergent extension (Tahinci et al.,2007), yet the exencephalic mice are of normal size, similar to Cecr2. Finally, mutation of Wnt5a results in inner ear defects and neural tube defects: of 34 Wnt5a null animals, 1 had craniorachischisis and 3 had exencephaly (Qian et al.,2007). Therefore, the presence of exencephaly rather than craniorachischisis is within the spectrum of PCP phenotype, although it may be indicative of a secondary or uncharacterized aspect in the pathway.

Cecr2Gt45Bic heterozygotes show exencephaly at a penetrance ranging from 0–4%, depending on the strain background (Banting et al.,2005). The collection of Cecr2Gt45Bic embryos for inner ear analysis produced 1/73 exencephalic heterozygotes (1.3%). In this study, we have shown that the fusion of the neural folds is slightly delayed in Cecr2Gt45Bic heterozygotes, which may explain the elevated susceptibility to exencephaly. This dosage sensitivity suggests the defect is near a threshold and, therefore, may be sensitive to changes in genes with which it genetically interacts. Cecr2tm1.1Hemc; Vangl2Lp double heterozygotes did not show genetic interaction; however, Cecr2tm1.1Hemc/tm1.1Hemc; Vangl2Lp/+ embryos showed a significant increase in spina bifida, an NTD not seen in Cecr2tm1.1Hemc homozygous embryos (inner ear defects were not assessed due to the small number of embryos available, but may also show an interaction). This demonstrates that Cecr2 can act as a dominant enhancer of this Vangl2 phenotype. Thus the two genes are in the same developmental pathway, although perhaps not the same molecular pathway. Furthermore, microarray data did not conclusively reveal direct Cecr2 regulation of any PCP gene levels. One promising result is the expression level changes of Lix1, which is expressed in the early neural tube (Moeller et al.,2002). In Drosophila, the Lix1 homologue Lowfat (Lft) is known to modulate the Fat signaling pathway, which in turn regulates PCP (Mao et al.,2009). Cytoplasmic Lft interacts with membrane proteins Fat and Dachsous and directly affects their levels. Although our microarray analysis detected a decrease in Lix1 levels in Cecr2Gt45Bic mutant embryos, this was not confirmed using qRT-PCR in 10–14 somite pairs Cecr2tm1.1Hemc homozygous mutants (Fairbridge et al.,2010) before neural closure. However, significant Lix1 expression changes were confirmed in the post-closure 16–18 somite pairs (Fairbridge et al.,2010) and a closer look at the timing of these changes in the cranial neural tube specifically is warranted.

Cecr2Gt45Bic homozygous mutants also have smaller inner ears and misaligned sensory cells (Figs. 3G,J, 4). Inner ear defects in the Cecr2tm1.1Hemc homozygous mutants are slightly milder, and may reflect the fact that this mutation was only backcrossed three generations onto the BALB/c background for this experiment. For both Cecr2 mutations the defects are milder than those seen in Vangl2 mutants, which shows greater proportions of cells rotated; from 70% misalignment in the IHC to 15%, 60–70%, and 90% misalignment in the OHC1, OHC2, and OHC3, respectively (Torban et al.,2008). In contrast, in Cecr2Gt45Bic homozygous mutants the misalignment is most severe in the OHC3 layer with a relatively unaffected IHC layer. At the 50% level, OHC3 has 61% misalignment, while OHC2 and 1 have 36% and 12%, respectively, and the IHC layer has only 11%. This pattern is more similar to that seen in PCP mutants such as circletail (Scrib1), and PTK7 (Montcouquiol et al.,2003; Lu et al.,2004), where the OHC3 is predominantly affected and the IHC is relatively unaffected. Cecr2 and Ptk7 mutants show other similarities. In the apical region, both can have an extra fourth row of OHC cells as well as extra IHC cells offset below the row. Both also show a general gradient of severity from basal to apical. The phenotypic similarities between these PCP mutants and Cecr2 suggest that Cecr2 influences the establishment of polarity within the developing inner ear.

Cecr2 is expressed in the cochlear floor and statoacoustic ganglion in both 14.5 dpc and 18.5 dpc inner ears. Vangl2 mRNA shows similar expression in both locations at 14.5 dpc and 16.5 dpc (Montcouquiol et al.,2003). Celsr1, Dvl1, and Dvl2 are found in the statoacoustic ganglion (Wang et al.,2005). Unlike Cecr2, all the core PCP genes tested have also been found expressed in the sensory hair cells themselves (Wang et al.,2005; Jones and Chen,2007; Seifert and Mlodzik,2007). However, Wnt5a expression also appears in the cochlear floor and not the hair cells (Qian et al.,2007). The latter suggests that Cecr2 may play a role in cells controlling signals that act externally on the hair cells to provide orientation.

A PCP-related pathway in which Cecr2 could influence polarity is ciliogenesis. Cilia defects, especially those with increased activity of the Shh pathway, are often associated with exencephaly (Murdoch and Copp,2010). However, Cecr2 mutants do not have the other typical features of ciliogenesis mutants. Nearly all such mutants are associated with polydactyly, which has never been seen in Cecr2 mutants. Other typical features of ciliogenesis mutants include left/right asymmetry, limb defects, heart looping, and early embryonic death, none of which are seen in Cecr2 mutants. Furthermore, the kinocilia in Cecr2 mutants are correctly positioned, while the kinocilium of Bbs6/Mkks mutants are separated from the stereocilia, which show subtle flattening of the V-pattern, but not rotational defects (Ross et al.,2005). Thus, the Cecr2 phenotype spectrum does not support a role for Cecr2 in ciliogenesis.

Inner ear polarity changes can also result from non-PCP defects in the development of the inner ear precursor cells. The homeobox gene Emx2 is thought to regulate proliferation and differentiation of hair cell precursors, and Emx2 mutants show hair cell polarity defects (Holley et al.,2010) However, Emx2 mutants also show significant loss of OHCs and support cells, with remaining cells forming indistinct rows. The effect is similar throughout the cochlea, and not more disorganized apically as with Cecr2. Disruption of pathways such as notch, FGF, and BMP/Noggin signaling can show inner ear defects, but they are usually much more severe, with defects different from those observed in Cecr2 mutants. Defects within the notch pathway result in an increase in the number of sensory cell rows of the organ of Corti. In particular, Hes1 mutations cause increases in the number of IHC, whereas Hes5 mutations lead to increased rows of OHCs (Zine et al.,2001). Notch1 mutations lead to increases in both OHC and IHC layers (Hayashi et al.,2008). Fgfr3−/− mice show a deficiency in support cell differentiation and a slight increase in the number of OHC rows, with normal stereocilia bundle orientations (Hayashi et al.,2007). Exogenous Bmp4, a known target of Fgf3, produced an increase in the number of OHC rows from three to seven–nine rows when applied to cochlear explants by Bmp4-soaked beads (Puligilla et al.,2007). The increased cellular rows and extra cells seen in Cecr2 mutants are observed at the apical turn but not nearly as excessive as those mentioned above.

In summary, we have shown that two independent mutations in Cecr2 cause polarity defects of the inner ear. Although these defects resemble those seen in PCP gene mutants, and combining the Cecr2tm1.1Hemc/+ and Vangl2Lp/+ mutations shows limited genetic interaction involving NTDs, it is very unlikely that Cecr2 acts directly on the PCP pathway. As a chromatin remodeling gene, Cecr2 may regulate a broad array of independent pathways, and at least one likely intersects with PCP. We have previously identified several transcription factors that are down regulated in neurulating Cecr2 mutants and, therefore, form a starting point for examination of these Cecr2-associated pathways (Fairbridge et al.,2010). The role of Cecr2 in neural tube and inner ear formation is likely complex, and its study may reveal novel genes and pathways that could interconnect the chromatin state with PCP.

EXPERIMENTAL PROCEDURES

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

Cecr2 Mutant Mice and Genotyping

Mice were housed under a 14 hr light/10 hour dark cycle at 22 ± 2°C. Stock mice were fed Laboratory Rodent Diet 5001 (LabDiet) while breeding mice were fed Mouse Diet 9F 5020 (LabDiet). Cecr2Gt45Bic mutant mice were backcrossed for ≥10 generations onto a BALB/c background. Also used in this analysis was new mouse line carrying a more severe allele of Cecr2 (termed Cecr2tm1.1Hemc). The Cecr2tm1.1Hemc line was backcrossed onto a BALB/c genetic background three generations before heterozygous crosses were analyzed, and six generations for the cross to Vangl2. This mutant has been described in (Fairbridge et al.,2010). To obtain timed embryos, females were harem bred and checked for a post copulatory plug (considered 0.5 days post coitus [dpc]). Pregnant dams were then euthanized by CO2 asphyxiation at specific time points and embryos were dissected in phosphate buffered saline. Extraembryonic membranes or a tail biopsy were collected for DNA extraction. All experiments were approved by the Animal Care and Use Committee of the University of Alberta. Genomic DNA was isolated from extraembryonic membranes or tail biopsies. Genotyping for the Cecr2Gt45Bic mutation and Cecr2tm1.1Hemc mutation was done using PCR as previously described (Fairbridge et al.,2010; Banting et al.,2005, respectively).

Xgal Staining, Immunostaining, Scanning Electron Microscopy, and Histology

Xgal staining was performed on whole-mount embryos and inner ears as previously described (Banting et al.,2005) with few modifications. Whole-mount embryos were stained in Xgal stain for at least 16 hr and inner ears samples for 24–48 hr. Samples were embedded in paraffin wax using standard protocols with slight modifications. Processing was done from 100% methanol rather than standard 100% ethanol. Samples were sectioned 5-μm-thick and counterstained with acidified eosin in 95% ethanol, then imaged using a brightfield light microscope (LM- Leica DMRXA with Optronics Macrofire LM CCD digital camera) and Picture Frame 2.3 software.

To assess stereocilia bundle disorganization, cochleae from Cecr2 wild-type, Cecr2Gt45Bic heterozygous, Cecr2Gt45Bic homozygous, and Cecr2tm1.1Hemc homozygous 18.5 dpc embryos on a BALB/c background were collected. From each embryo, the sensory epithelium was extracted and fluorescently labeled with an acetylated tubulin antibody and phalloidin to visualize filamentous actin (Torban et al.,2008).

For scanning electron microscopy, Cecr2Gt45Bic homozygous mutant embryos were processed as previously described (Davidson et al.,2007).

Stereocilia Bundle Analysis

Stereocilia bundle analysis was modeled after (Montcouquiol et al.,2003) with few deviations. The sensory epithelium was imaged using confocal microscopy at the basal (5%), middle (50%), and apical (75%) positions. For each position, a minimum of 10–20 cells were assessed for orientation for each hair cell row; one row of inner hair cells (IHC) and three rows of OHCs (OHC1, OHC2, and OHC3). The stereocilia bundle orientation was determined by comparing the location of each stereocilia bundle with respect to a line perpendicular to the medial lateral axis. All data were analyzed independently by two individuals. Bundles that were rotated more than 30 degrees away from the medial lateral axis were considered misaligned. A multiple comparisons chi-square test of independence using a Bonferroni adjustment was used to downward adjust the desired probability of type 1 error (EER) to an adjusted probability (alpha) of comparison-wise type 1 error (CER).

Generation of Cecr2tm1.1Hemc:Vangl2lp Double Heterozygous and Cecr2tm1.1Hemc/tm1.1Hemc: Vangl2lp/+ Embryos

A Vangl2lp/+ male on C3H/C57Bl/6J hybrid background was a gift from Philippe Gros (McGill University, Montreal, Quebec). Cecr2tm1.1Hemc heterozygous females were harem bred to the Vangl2lp/+ male to generate a F1 generation on a mixed Balb/c × C3H/C57Bl/6J hybrid background. F1 females were harem bred to F1 males (both of various genotype combinations) and killed by CO2 asphyxiation when visibly pregnant. Embryos ranging from 12.5 to 18.5 dpc were scored for the presence of NTDs or any visual gross abnormalities. The presence of a looped tail was used to track the Vangl2 genotype.

Microarray and qRT-PCR

The microarray analysis compared BALB/c 11–14 somite pair embryos (just before cranial neural tube closure), with 4 biological replicates for each of wild-type and Cecr2Gt45Bic homozygous mutants. The Affymetrix Genechip Mouse Genome 430 2.0 Arrays were used (described in Fairbridge et al.,2010). The qRT-PCR was done using 4 biological replicates (10–14 somite pairs, heads collected above the second brachial arch) of wild-type and Cecr2tm1.1Hemc homozygous mutants described in Fairbridge et al.,2010) which shows an enhanced phenotype (∼96% penetrance of exencephaly). Reactions were set up with probe sets and primers from the Roche ProbeFinder Version: 2.45, then analyzed using the Fluidigm 48.48 Access Array Integrated Fluidic Circuit, according to manufacturer's protocols as described in Fairbridge et al. (2010).

Acknowledgements

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

We thank Dr. Andrew Waskiewicz and Dr. Kirst King-Jones for advice as well as the University of Alberta Advanced Microscopy Facility for technical assistance with this study. C.D., M.K., and N.F. were supported by scholarships from the Natural Sciences and Engineering Research Council of Canada (NSERC). N.F. was also supported by a scholarship from the Alberta Heritage Foundation for Medical Research.

REFERENCES

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