Learn more: PMC Disclaimer | PMC Copyright Notice
The Nuclear Receptor Nor-1 Is Essential for Proliferation of the Semicircular Canals of the Mouse Inner Ear
Abstract
Nor-1 belongs to the nur subfamily of nuclear receptor transcription factors. The precise role of Nor-1 in mammalian development has not been established. However, recent studies indicate a function for this transcription factor in oncogenesis and apoptosis. To examine the spatiotemporal expression pattern of Nor-1 and the developmental and physiological consequences of Nor-1 ablation, Nor-1-null mice were generated by insertion of the lacZ gene into the Nor-1 genomic locus. Disruption of the Nor-1 gene results in inner ear defects and partial bidirectional circling behavior. During early otic development, Nor-1 is expressed exclusively in the semicircular canal forming fusion plates. After formation of the membranous labyrinth, Nor-1 expression in the vestibule is limited to nonsensory epithelial cells localized at the inner edge of the semicircular canals and to the ampullary and utricular walls. In the absence of Nor-1, the vestibular walls fuse together as normal; however, the endolymphatic fluid space in the semicircular canals is diminished and the roof of the ampulla appears flattened due to defective continual proliferative growth of the semicircular canals.
The mouse inner ear develops during the second half of gestation from an otic epithelial vesicle that undergoes a complex series of shape changes to give rise to specific functional compartments including the cochlea, which comprises the auditory apparatus, and the vestibule, which is responsible for sensing motion and gravity. The endolymphatic fluid-filled membranous labyrinth is surrounded by perilymphatic fluid and is contained within an osseous structure (tympanic bone). The vestibular part of the inner ear contains two principal sets of sensory structures, the maculae (saccule and utricle), sensors of linear acceleration and gravity, and the cristae ampullaris, sensors of angular acceleration that are contained within three ampullae. Stimuli related to vestibular functions are produced by movement of the endolymphatic fluid and are recorded by sensory hair cells of the maculae and cristae.
Development of the murine inner ear is initiated at E8.5 by the formation of an otic epithelial placode, which arises as a thickening of the head ectoderm between rhombomeres 5 and 6 in the hindbrain, followed by invagination and separation from the ectoderm to become a closed vesicle embedded in the head mesenchyme at E9.5. Structural perturbances begin to occur in the vesicle at E11 to 11.5 with the outgrowth of the ventral wall of the otic vesicle to form a hollow tube that gives rise to the cochlea and a dorsal extension that forms the endolymphatic duct. Cells localized in the dorsolateral walls give rise to the semicircular canals, while the medial and medioventral regions give rise to the utricle and saccule (38). Restricted and asymmetric expression patterns in the otic vesicle have been described for several genes (14, 38) at this developmental stage. These genes then participate in the regional morphogenesis of the inner ear.
The semicircular canals originate from bilayered outpocketings of the dorsolateral otic vesicle. The lateral wall of each epithelial outpocketing delaminates from the underlying mesenchyme and grows toward the corresponding medial wall, forming a fusion plate (27). Subsequently, the fusion plate disappears in order to generate the closed tubular form of the canal by E13.5. At this stage, all of the major structural changes associated with semicircular canal formation are complete and the shape resembles that of an adult inner ear. Several genes, such as Otx1, Nkx5.1, and Netrin1, have been shown to be required in the proper formation and fusion of the semicircular canals (reviewed in reference 15).
Inner ear differentiation and formation are followed by a period of continual growth (27). While a pathway involving Msx1 and bone morphogenic protein 4 (BMP4) has been implicated in the proliferative continual growth of the semicircular canals in the chicken (4), the mechanism leading to vestibular continual growth in mice has not been established.
In this paper, we identify nuclear receptor Nor-1 as an essential mediator of semicircular canal continual growth. Nor-1 is a member of the nuclear receptor family of transcription factors (7). This family comprises a large group of structurally related transcription factors that program developmental, physiological, and behavioral responses to a variety of chemical signals, most of which are nuclear receptor ligands (13, 24, 25, 39). Together with two additional highly conserved proteins within this family (Nur77 and Nurr1), Nor-1 represents a subclass of nuclear receptors that interact with overlapping cis-acting DNA elements and can function as constitutively active transcription factors without an apparent requirement for ligand binding (18, 29, 43). Although little information is available on the spatiotemporal expression, physiological functions, or target genes regulated by Nor-1, the protein was shown to be induced in primary cultures of forebrain neurons undergoing apoptosis (30) and plays a key role together with Nur77 in mediating negative selection of self-reactive T-lymphocyte cells via apoptosis (3, 5, 22, 44, 46). Nor-1 has also been identified as a fusion partner of the Ewing’s sarcoma gene in the recurrent t(9;22) chromosomal translocation found in extraskeletal myxoid chondrosarcomas (2, 6). The fusion is thought to result in inappropriate activation of Nor-1-dependent target genes, leading to cartilaginous neoplasia.
In order to establish the essential physiological functions of Nor-1, we have generated Nor-1-null mice by gene targeting. Here we demonstrate that inactivation of the Nor-1 gene in mice results in abnormal hyperactive and partial bidirectional circling behavior. Further, we show that the circling behavior is associated with a defect in proliferative continual growth of the semicircular canals of the vestibule after their formation and an associated reduction in the endolymphatic fluid space within the canals and their corresponding ampullae.
MATERIALS AND METHODS
Gene targeting.
The targeting vector was constructed in the plasmid pKos (Stratagene) and contains a 5.2-kb cassette containing, in the 5′- to -3′ direction, an internal ribosomal entry site, a β-galactosidase gene, and a neomycin resistance cassette flanked on its 5′ side by the Mc-1 promoter. The reporter and selection cassette are flanked on the 5′ side by 8.5 kb of the Nor-1 genomic sequence extending 5′ of exon 2 and flanked on its 3′ side by a 1.5-kb Nor-1 genomic fragment containing the 3′ end of exon 2 and downstream sequences. The targeting vector inserted into exon 2 of the Nor-1 gene replaced the region encoding amino acids 212 to 231 located upstream of the DNA binding domain of the Nor-1 protein. Two SfiI sites at nucleotide positions encoding amino acids 212 and 231 were introduced into the genomic clone by yeast recombination (42), and the selection cassette was inserted as an SfiI fragment.
The targeting vector was linearized at its 3′ end and used to electroporate 129SvEv embryonic stem (ES) cells. After selection of neomycin-resistant clones, ES cell DNA was screened for homologous recombination. The mice were generated at Lexicon Genetics Inc., The Woodlands, Tex.
To identify the Nor-1 mutation, Southern blot analysis was performed on genomic DNA isolated from the ES cell clones. DNA samples were digested with EcoRI, resolved by electrophoresis, and transferred onto nylon membranes for hybridization with a radiolabeled 360-bp genomic DNA fragment located outside of but immediately adjacent to the disrupted Nor-1 gene. This 3′ probe was obtained as a PCR product using oligonucleotides ACAGGGGGTCAACTCATAAT and GAACTCCAGAGTCAAGAGAA as primers.
Three oligonucleotides were used in a single PCR for genotyping. They consisted of two 5′ primers, one (GGCCGCAGCTGCACTCAGTC) located in a 5′ portion of exon 2 that was deleted in the targeting vector to allow selective detection of the wild-type allele as a 960-bp product and the other (GTGGCGGACCGCTATCAGGAC) located in the neo gene to generate a 1,200-bp PCR product representing the mutant allele. The 3′ primer (GTTCTGCCACCACAGAGCATCTTG) was located in the 3′ end of exon 2. Two independent targeted ES cell clones were used to generate chimeric mice. Both clones yielded chimeric mice that contributed to the germ line.
β-Galactosidase staining.
Whole embryos (the afternoon of the plug day was assigned as gestational day E0.5) and P1 heads were fixed for 30 min in 2% formaldehyde-0.2% glutaraldehyde-0.02% NP-40 in phosphate-buffered saline (PBS) (E10.5 to E12.5) or in 2% paraformaldehyde for 2 h (E13.5 to P1) and washed in PBS. The tissues were then incubated overnight in a 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside (X-Gal) staining solution (1.3 mM MgCl2, 15 mM NaCl, 44 mM HEPES [pH 7.9], 3 mM potassium ferro- and ferricyanides, 0.5 mg of X-Gal per ml, 0.05% NP-40 in H2O). After staining, samples were washed with PBS, refixed in 10% formalin overnight, dehydrated in graded ethanols, cleared in histoclear, and embedded in paraffin. Sections (6 μm) were cut and counterstained with nuclear fast red.
Balance tests.
Balance tests were performed as previously described (21).
Immunohistochemistry.
The heads of embryos and P1 pups were fixed in Bouin’s fixative overnight, washed in 70% ethanol; processed for paraffin sectioning, and cut to 6 μm. For E-cadherin staining, heads were fixed in 4% paraformaldehyde for 2 h, cryoprotected in 30% sucrose overnight, embedded in Tissue-Tec O.C.T compound, and then cut to 7 μm. For phalloidin detection, heads were fixed in 4% paraformaldehyde; the inner ears were dissected and processed as whole mounts. The specimens were stained in accordance with standard avidin-biotin immunohistochemical procedures (Vector). Primary antibodies included those against proliferating cell nuclear antigen (PCNA; 1:1,000; Santa Cruz), E-cadherin (1:1,200; Calbiochem), and phalloidin (2 μg/ml; Sigma).
Terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labeling (TUNEL) assay.
The heads of embryos were fixed in 10% formalin overnight, processed to paraffin, and cut at 6 μm. For staining, the sections were rehydrated and digested with proteinase K (40 ng/ml) for 7 min. After washes with PBS, the sections were blocked with TdT buffer (30 mM Tris [pH 7.4], 140 mM Na cacodylate, 1 mM CoCl2) and incubated with TdT/dUTP mix (8 μl of UTP and 6 μl of TdT [Roche] in 1,200 μl TdT buffer) for 1 h at 37°C. The reaction was stopped with 2.5 M NaCl-250 mM Na citrate. The signal was detected by using Vectastain Elite and peroxidase substrate kits (Vector). Methyl green was used as a counterstain.
Cell counts and statistical analysis.
To measure the proliferative index and the amount of cell death from the anterior semicircular canal cross-sections, all PCNA- and TUNEL-positive cells and all of the cells from each semicircular canal were counted. Data collected from each experimental group are expressed with the standard error of the mean (SEM). Student’s t test was used for statistical analysis. For all stages and genotypes, sections from six animals were counted, except for Nor-1+/+ mice at E17.5, where n = 4, and P1, where n = 5 (PCNA).
In situ hybridization.
Sections for radioactive and dioxigenin in situ hybridization were fixed in Bouin’s fixative and 2% paraformaldehyde, respectively, and treated as previously described (28, 32). The Nor-1 probe was prepared by PCR amplifying nucleotides 1089 to 1518 in accordance with the rat Nor-1 cDNA (GenBank accession no. D38530). The primers used were 5′GCCCAGCACCTCCATGTACTTC and 3′CAGCAGGCTGGACGCGGTAGGG. The resulting 430-bp fragment was cloned into vector pCRII (Invitrogen). The following probes have been previously described: Netrin1 (33) and BMP4 (20).
RESULTS
Targeted disruption of the Nor-1 gene in mice.
The Nor-1 gene was disrupted in mouse ES cells by using the targeting vector shown in Fig. Fig.1A.1A. The vector contained a 10-kb fragment of the mouse Nor-1 gene interrupted within exon 2 by the introduction of a β-galactosidase reporter gene (lacZ) and a neomycin selection cassette downstream of the ATG translation initiation codon and upstream of the DNA binding domain of Nor-1, disrupting the protein at amino acids 212 to 231 (GenBank accession no. D38530). Targeted integration of this vector into the Nor-1 genomic locus was detected by Southern blot analysis of EcoRI-digested ES cell DNA with a 32P-labeled 360-bp probe located outside of but immediately 3′ to the disrupted Nor-1 gene. A 13-kb band represents the wild-type Nor-1 gene, whereas the mutated Nor-1 allele results in a shorter 2.5-kb band (Fig. (Fig.1B).1B). Two independent targeted ES cell clones were used to generate chimeric mice. Both clones yielded chimeras that contributed to the germ line, and the offspring were genotyped by PCR. In these analyses (see Materials and Methods), the wild-type Nor-1 gene is represented by a 960-bp PCR product while the mutant allele is amplified as a 1,200-bp product (Fig. (Fig.1C).1C). Genotype analysis indicated that Nor-1-null mice were born in expected Mendelian ratios (data not shown).
Targeted inactivation of the Nor-1 gene in mouse ES cells and generation of null mice. (A) Schematic diagram of the strategy used to target the Nor-1 gene. (Top) The 10-kb targeting vector used for electroporation. Numbered boxes represent exons. (Middle) Genomic structure of the Nor-1 gene using homology to the human Nor-1 gene (GenBank accession no. D85243). (Bottom) Representation of the structure of the inactivated Nor-1 gene. (B) Southern blot analysis of control and targeted DNAs from ES cells. DNA was digested with EcoRI and hybridized to a 360-bp genomic probe located outside of but immediately 3′ to the disrupted Nor-1 gene. This probe hybridized to 13- and 2.5-kb fragments from wild-type and mutant alleles, respectively. Lanes: C, control; 1, ES cell clone 1; 2, ES cell clone 2. (C) PCR analysis of mouse tail DNA. Two PCR products are shown at 960 and 1,200 bp, corresponding to the wild-type and mutant alleles, respectively. (D) In situ hybridization of a coronal section of a P0 wild-type mouse brain at the level of the hippocampus using an antisense Nor-1 probe (top) and β-galactosidase staining in the same region of a Nor-1+/− mouse brain demonstrate that β-galactosidase faithfully reports the correct spatial expression of Nor-1 observed for the endogenous gene.
To determine whether the β-galactosidase-encoding gene introduced into the Nor-1 locus reports the correct spatiotemporal expression of Nor-1 in heterozygote mice, we compared the expression pattern of Nor-1 mRNA observed by in situ hybridization in P0 coronal brain sections in the region of the hippocampus (Fig. (Fig.1D,1D, top) with that obtained after β-galactosidase staining of heterozygous Nor-1lacZ mice (Fig. (Fig.1D,1D, bottom). Staining for Nor-1 was observed in the cortex, hippocampus, and habenular nucleus, confirming that β-galactosidase is expressed in a pattern similar to that observed for endogenous Nor-1 by in situ hybridization.
Nor-1 heterozygotes were indistinguishable from wild-type animals, and matings between Nor-1 heterozygous mice resulted in viable Nor-1-null animals. The Nor-1 homozygous mice were normal in appearance; however, abnormal hyperactive and partial bidirectional circling behavior was observed by 3 weeks of age in 15% of these mice (C57BL × 129/SV). The circling behavior was interspersed with noncircling periods of feeding, grooming, and sleep. This circling phenotype was never observed in the heterozygous or wild-type mice. The hyperactivity and bidirectional circling behavior are characteristics of mouse mutants with functional defects in the vestibular apparatus of the inner ear (11, 23).
Nor-1 is expressed in the nonsensory epithelium during otogenesis.
To examine whether Nor-1 is expressed during development of the vestibular system, we used β-galactosidase as a reporter of Nor-1 gene expression in Nor-1lacZ heterozygote mice. For these analyses, Nor-1lacZ embryos from stage E8.5 to E17.5 and P1 pups were collected, stained for β-galactosidase expression as whole mounts, and then serially sectioned. The results in Fig. Fig.22 demonstrate that Nor-1 is not expressed during the earliest stages of vesicle formation (E10.5; Fig. Fig.2A).2A). However, expression of Nor-1 begins coincidentally with the initiation of regional shape changes in the otic vesicle at E11.5 and is restricted to the dorsolateral region that is destined to form the three semicircular canals of the vestibule (Fig. (Fig.2B).2B). This onset of expression was confirmed by Nor-1 in situ hybridization (Fig. (Fig.2C2C).
Nor-1lacZ expression during otic development. No staining for Nor-1 was detected at E10.5 (A) or earlier. Expression of Nor-1 was localized in the future fusion plate-forming cells of the otic vesicle at E11.5 (B and C, arrows). (C) Nor-1 in situ hybridization corresponding to the β-galactosidase-stained E11.5 Nor-1+/− specimen shown in panel B. Nor-1 and BMP4 expression compared at E11.5 (D and E). Nor-1 mRNA was localized to the canal region that is destined to form the fusion plate (D, arrows). An adjacent ventral section shows BMP4 expression in the presumptive anterior and posterior cristae (E), which do not overlap the Nor-1-positive areas. (F) Nor-1lacZ expression in the fusion plates of the semicircular canals at E12.5. After closure of the semicircular canals, Nor-1 is localized in the inner side of each canal, as shown at E13.5 (G) and P1 (H and I) at the level of the anterior semicircular canal. (I) Nor-1 in situ hybridization corresponding to the β-galactosidase-stained Nor-1+/− specimen shown in panel H. (J) P1 whole-mount β-galactosidase-stained Nor-1+/− left ear vestibule. At P1, Nor-1 expression was also detected in the ampullary (K) and utricular (L) wall. Scale bars are 50 μm (C, D, E, G, H, and I) and 100 μm (A, B, F, J, K, and L). Abbreviations: AC, anterior crista; ASC, anterior semicircular canal; ED, endolymphatic duct; LSC, lateral semicircular canal; PC, posterior crista; PSC, posterior semicircular canal; S, saccule; U, utricle; D, dorsal; M, medial; P, posterior; A, anterior.
To find out whether Nor-1 is expressed in the early sensory areas before they can be distinguished morphologically, we performed in situ hybridization for Nor-1 and BMP4, a marker specific for developing cristae in the otic vesicle (28) at E11.5. Nor-1 signal (Fig. (Fig.2D)2D) is excluded from the BMP4-positive areas (Fig. (Fig.2E)2E) in adjacent sections of the otic vesicle, confirming that expression of Nor-1 is limited to the nonsensory epithelium.
The expression at E12.5 (Fig. (Fig.2F)2F) is localized in the canal forming fusion plates in a pattern that overlaps that previously observed for Netrin1 (31).
At E13.5, all of the major structural changes that give rise to the general shape of the membranous labyrinth have been completed and the shape resembles that of the adult inner ear. At this stage, the expression of Nor-1 is limited to the inner edge of the semicircular canal epithelium, flanking the site of fusion (Fig. (Fig.2G),2G), and this restricted expression continues at P1 (Fig. (Fig.2H).2H). The expression in the inner side of the semicircular duct at P1 was confirmed by Nor-1 in situ hybridization (Fig. (Fig.2I)2I) and is also shown as a β-galactosidase-stained whole-mount picture of the vestibule (Fig. (Fig.2J2J).
Expression of Nor-1 was also seen in the nonsensory epithelial cells that form the roof of the ampulla (Fig. (Fig.2K)2K) and the utricle (Fig. (Fig.2L)2L) at P1 but is absent from the sensory epithelium of the cristae and maculae (Fig. 2K and L).
Finally, Nor-1 expression is not observed in the auditory and vestibular nuclei in the hindbrain or the auditory and vestibular ganglia that innervate the cochlea and vestibule of the inner ear, respectively (data not shown).
The nonoverlapping expression patterns of Nor-1 (Fig. (Fig.2D)2D) and BMP4 (Fig. (Fig.2E)2E) in adjacent sections analyzed at E11.5 and the lack of staining observed in the sensory areas at later stages support our conclusion that Nor-1 is specific to the nonsensory epithelium of the developing inner ear. Thus, the restricted spatiotemporal expression of this factor during the formation and growth of the semicircular canals suggests that Nor-1 plays an important role in morphogenesis of the semicircular canals of the vestibule.
Balance tests.
In order to identify other signs, in addition to the circling behavior, indicative of inner ear defects, we performed behavioral tests to assess vestibular function. The tests included elevated-platform, contact-righting, reaching response, and swimming tests (21). They were performed at three different ages, 2.5, 4.5, and 8 months, with a total of 18 Nor1−/− and control animals (Nor-1+/+ and Nor-1+/− mice were used as one control group [n = 18]).
In the elevated-platform and reaching response tests, the Nor-1−/− mice performed as did the Nor-1+/+ and Nor-1+/− controls. Similarly, the majority of the Nor-1−/− mice showed a normal ability to stay afloat during the swimming tests, with only 2 out of 18 Nor-1−/− animals failing the test. In contrast, 6 out of 18 Nor-1−/− animals failed the contact-righting test, in which the mouse must determine its orientation without tactile cues, compared to 0 out of 18 animals in the control group. No additional behaviors indicative of progressive deterioration of inner ear function were observed as the mice aged.
Taken together, our test results show that Nor-1-null mice display only slight balance defects in addition to circling behavior.
Disruption of the Nor-1 gene leads to abnormal vestibular development.
To determine whether Nor-1 plays an essential role in embryonic morphogenesis and/or outgrowth of the semicircular canals, the membranous labyrinths of inner ears of stages E13.5 to P1 were paint filled to monitor the morphological development of the vestibular apparatus.
By E13.5, the three semicircular canals of the vestibule have acquired the mature pattern but their size continues to increase. As shown in Fig. Fig.3B,3B, the semicircular canals were formed in the Nor-1−/− inner ears by E13.5. However, the diameter of the canals was reduced as the development continued through E16.5 (Fig. (Fig.3D)3D) and at P1 (Fig. (Fig.3F)3F) compared to that of the control specimens (Fig. 3C and E). In addition, the ampullas, the sensory organs for the semicircular canals, appear flattened in the Nor-1−/− specimens at these stages (Fig. 3D and F).
Morphological defects in Nor-1−/− mouse inner ears. Paint-filled right inner ears of Nor-1+/− (A, C, and E) and Nor-1−/− (B, D, and F) mice from E13.5 to P1. Panels A to D are lateral views, and panels E to F are medial views. The Nor-1−/− semicircular canals are formed at E13.5 (B); however, all Nor-1−/− inner ears at E16.5 (D) and P1 (F) show thinned canals (asterisk) and flattened ampullae (arrows). Scale bars are 50 μm (A and B) and 300 μm (C to F). Abbreviations: AA, anterior ampulla; ASC, anterior semicircular canal; CC, common crus; ED, endolymphatic duct; ES, endolymphatic sac; LA, lateral ampulla; LSC, lateral semicircular canal; PA, posterior ampulla; PSC, posterior semicircular canal; D, dorsal; A, anterior; P, posterior.
In contrast, comparison of the dissected and paint-filled labyrinths did not reveal any overt differences in the formation of the cochlear duct in the Nor-1−/− animals (data not shown).
The vestibular sensory areas appear normal in Nor-1−/− animals.
To determine whether the circling behavior and structural abnormalities in the ampullary wall are associated with additional defects in vestibular sensory areas, the adult inner ear stereocilia of the sensory epithelia were examined by phalloidin staining as whole mounts. The stereocilia on the apical surface of the hair cells are key elements necessary for detection of endolymphatic fluid movement and conversion of this mechanical stimulus into electrical impulses. No apparent degeneration or disorganization of the stereocilia in the crista ampullaris (Fig. (Fig.4B)4B) or the macula utriculi (Fig. (Fig.4D)4D) was detected in circling Nor-1-null mice compared to their noncircling wild-type counterparts (Fig. 4A and C). Thus, the sensory epithelium of the vestibule appears to be unaffected in circling Nor-1−/− mice.
Phalloidin labeling of stereocilia of sensory hair cells from posterior cristae and macula utriculi from 2.5-month-old Nor-1+/+ (A and C) and circling Nor-1−/− (B and D) mice. Panels A and B show surface views of posterior cristae, and panels C and D show surface views of macula utriculi. Stereocilia appear normal in the circling Nor-1−/− mouse cristae (B) and macula (D). Scale bars are 20 μm.
The amount of cell death is not increased in Nor-1−/− mouse semicircular canals.
To determine whether the reduced semicircular canal size is associated with changes in apoptosis during continual canal growth, inner ears from stages E13.5 to E17.5 were analyzed by TUNEL staining.
At E13.5 TUNEL-positive cells were observed frequently (Fig. 5A and B); however, at E15.5 (Fig. 5E and F) and E16.5 (Fig. 5G and H), positive cells were sparse in both the Nor-1+/− and Nor-1−/− mouse canals analyzed. The TUNEL-positive cells in the sections were counted, and the results are summarized in Fig. Fig.5I.5I. No significant difference was detected in the amount of cell death in the Nor-1−/− animals compared to the control specimens. Thus, defective semicircular canal outgrowth does not appear to be associated with increased cell death.
TUNEL assay of Nor-1+/− and Nor-1−/− anterior semicircular canals. TUNEL-stained cross-sections of anterior semicircular canals of Nor-1+/− (A, C, E, and G) and Nor-1−/− (B, D, F, and H) mice. In both the Nor-1+/− and Nor-1−/− mouse sections at E13.5, TUNEL-positive cells were detected around the semicircular canal epithelium (A and B). By E14.5, the amount of cell death was greatly reduced in the canals (C and D), and at E15.5 (E and F) and E16.5 (G and H), only occasional TUNEL-positive cells were observed in mice of both genotypes. (I) Comparison of the amounts of cell death in Nor-1+/− and Nor-1−/− mouse anterior semicircular canals. The histograms present percentages of TUNEL-positive cells versus the total number of cells. The error bars show the SEM. Scale bars are 12.5 μm (A to F) and 25 μm (G and H). He, heterozygous; Ho; homozygous.
Nor-1 is required for proliferation of the nonsensory epithelium of the semicircular canals.
To determine whether the observed morphological defects were due to abnormal proliferative growth of the semicircular canals, cross-sections through the canals and cristae ampullaris were prepared from E13.5 to P1. The proliferative index of the nonsensory epithelial cells was determined by immunostaining with antibodies directed against PCNA. These sections were then counterstained with hematoxylin and eosin. Histological comparison of corresponding sections through the anterior semicircular canals of Nor-1+/− (Fig. 6A to E) and Nor-1−/− (Fig. 6F to J) mice confirmed a marked reduction in the diameter of Nor-1−/− canals from E15.5 onward.
Decreased proliferation in the Nor-1−/− mouse vestibule. PCNA-immunostained cross-sections of anterior semicircular canals of Nor-1+/− (A to E) and Nor-1−/− (F to J) mice are shown. In the Nor-1+/− mouse sections at E13.5 (A) and E15.5 (B), proliferating cells were detected around the semicircular canal epithelium, excluding the Nor-1-positive cells in the inner side of the canals at E15.5 (B). At E16.5 (C) and E17.5 (D), the PCNA positive-cells were concentrated on the lateral sides of the canals, and by P1, proliferation in the Nor-1+/− mouse canal was greatly reduced (E). A reduction in PCNA-positive cells was observed in Nor-1−/− mouse semicircular canals at E15.5 (G), E16.5 (H), E17.5 (I), and P1 (J) compared to those in the Nor-1+/− specimens. PCNA-immunostained sections of anterior ampullas of Nor-1+/− specimens at E16.5 (K and M) and E17.5 (L and N) and corresponding Nor-1−/− specimens at E16.5 (O and Q) and E17.5 (P and R) show a decrease in proliferation also at the level of the ampullary roof of Nor-1−/− mice. Panels M, N, Q, and R are magnifications of the roof of the ampulla shown in panels K, L, O, and P. (S) Comparison of proliferation in anterior semicircular canals of wild-type (WT), heterozygous (He), and homozygous (Ho) Nor-1 mice. The histograms present percentages of PCNA-positive cells versus the total number of cells. The error bars show the SEM. *, P < 0.05 (Student t test). Scale bars are 25 μm (A to J) and 50 μm (K, L, O, and P).
At E13.5 (Fig. (Fig.6A)6A) and E15.5, the level of proliferation in the Nor-1+/− mouse semicircular canals was high throughout the epithelium, with the exception of the Nor-1-expressing cells at E15.5 (compare Fig. Fig.6B6B to Fig. Fig.7A).7A). By E16.5 (Fig. (Fig.6C)6C) and E17.5 (Fig. (Fig.6D),6D), the PCNA-positive area became more restricted to the lateral walls, flanking the Nor-1-positive cells. After birth, at P1, the amount of proliferation was reduced, with a few PCNA-positive cells observed (Fig. (Fig.6E).6E). Comparison of the PCNA-positive cells in the epithelium of Nor-1−/− mice at these stages indicates that there is a significant reduction in the number of proliferating cells that is most obvious at E16.5 and E17.5 (Fig. 6F to J).
Reduced β-galactosidase staining in Nor-1−/− mouse semicircular canals (A to D). Cross-sections through anterior semicircular canals of Nor-1+/− (A and C) and Nor-1−/− (B and D) mice show comparable β-galactosidase staining in mice of both genotypes at E15.5 (A and B). However, by P1 (C and D), the Nor-1lacZ staining is greatly decreased in the Nor-1−/− mouse semicircular ducts. Altered cell morphology of the Nor-1-expressing cells in the Nor-1−/− mouse semicircular canals (E to H). Hematoxylin-eosin-stained cross-sections of anterior semicircular canals of Nor-1+/− (E and G) and Nor-1−/− (F and H) animals. The Nor-1-expressing cells at the inner edge of the semicircular ducts of Nor-1−/− mice have the normal columnar shape at E15.5 (F); however, at P1, these cells have lost their specialized cell shape and appear flattened (H) compared to corresponding Nor-1+/− mouse sections at E15.5 (E) and P1 (G). E-cadherin staining of cross-sections of the anterior semicircular canals shows similar staining at E15.5 in Nor-1+/− (I) and Nor-1−/− (J) mouse ducts. Together with loss of the columnar cell shape in the Nor-1−/− mouse ducts, there is a decrease in E-cadherin staining at P1 (L) while staining for E-cadherin is seen in the Nor-1+/− mouse cells (K). (M to P) Comparison of Nor-1 (M) and Netrin1 (N) expression in Nor-1+/− mouse anterior semicircular canals shows an overlapping pattern on the inner side of the semicircular duct; however, Nor-1 is more limited to the small group of columnar cells. Netrin1 expression is retained in the Nor-1−/− mouse ducts (O). Panel P is a bright-field image of Panel O. Scale bars are 25 μm (A to D), 12.5 μm (E to L), and 50 μm (M to P).
Similarly, comparison of PCNA-positive cells in the ampullary roof of Nor-1+/− and Nor-1−/− mice at E16.5 and E17.5 showed almost complete absence of proliferation in this region in Nor-1−/− mice (Fig. 6O to R) relative to Nor-1+/− mice (Fig. 6K to N).
The proliferative index was measured by quantitation of PCNA-positive cells versus the total number of cells from transverse sections through the anterior semicircular canals of Nor-1+/+, Nor-1+/−, and Nor-1−/− mice. A significant decrease in proliferation in Nor-1−/− mice was confirmed at E15.5 and throughout the continual canal growth phase until P1 (Fig. (Fig.6S6S).
Taken together, the results show that a reduction in the proliferation of nonsensory epithelial cells seems to be the primary cause of the canal and ampullary wall defects observed in Nor-1−/− inner ears. Due to the restricted expression of Nor-1 in the inner side of the semicircular canals and its segregation from proliferating cells, we conclude that Nor-1 regulates continual growth of the semicircular canals by a paracrine mechanism.
Degeneration of nonsensory epithelial cells of the semicircular duct in the absence of Nor-1.
The restricted expression of Nor-1 in the inner side of the semicircular ducts of the canals identified a previously uncharacterized regional specification of the nonsensory epithelium, the functional significance of which is unknown. Comparison of Nor-1 expression in this region between Nor-1+/− and Nor-1−/− mice using β-galactosidase indicated strong expression of Nor-1 in both genotypic groups at E15.5 (Fig. 7A and B) that was markedly reduced selectively in Nor-1−/− mice by P1 (Fig. 7C and D). Histological analysis of the Nor-1-positive cells at this stage indicated that loss of Nor-1 expression is associated with degeneration of epithelial cell structure rather than cell death. Comparison of Nor-1+/− and Nor-1−/− mouse sections at E15.5 showed a normal columnar cell shape of the Nor-1-positive cells in Nor-1−/− mouse semicircular canals (Fig. 7A and B and E and F). However, corresponding sections at P1 showed a distinct change in cellular morphology from the normal tall columnar shape (Fig. 7C and G) to a markedly flattened shape of the Nor-1−/− animals (Fig. 7D and H). The tall columnar shape of the Nor-1-positive cells implies that they could be polarized, with the apical surface facing to the endolymphatic fluid. The zonula adherens, an actin belt encircling the cell just below its apical face, is one of the crucial components for maintenance of both the polarized shape of individual epithelial cells and the integrity of cell sheets. These two functions can be attributed to the activity of cell adhesion protein E-cadherin. Mutations or loss of any of the components involved in the establishment or maintenance of the zonula adherens result in loss of cell polarity and adhesion and can lead to the disintegration of the entire epithelium (37, 45). Immunohistochemical analysis of the expression of E-cadherin at E15.5 indicated that the protein localized to the apical side of Nor-1-positive cells in Nor-1+/− (Fig. (Fig.7I)7I) and Nor-1−/− (Fig. (Fig.7J)7J) animals. In the Nor-1−/− mouse inner ear, however, loss of the polarized cell shape correlated with loss of E-cadherin staining at P1 (Fig. (Fig.7L)7L) relative to the Nor-1+/− mouse epithelium (Fig. (Fig.7K7K).
Despite the significant alteration in cellular morphology, the expression of Netrin1 (Fig. (Fig.7N),7N), a marker for the semicircular duct inner epithelium whose expression partially overlaps that of Nor-1 (Fig. (Fig.7M),7M), was retained in Nor-1−/− mice (Fig. 7O and P).
Together, these data indicate that Nor-1 is necessary for maintenance of a regionally specified polarized epithelial cell group within the nonsensory epithelium. The morphology of Nor-1-expressing cells is similar to that of secretory or absorptive epithelial cells and suggests that these cells directly contribute to endolymphatic fluid composition, in addition to their role in regulation of proliferation of neighboring epithelial cells.
DISCUSSION
In this paper, we have identified a novel essential role for nuclear receptor Nor-1 in the development of the semicircular canals of the inner ear vestibule. Ablation of Nor-1 expression in mice results in partially penetrant hyperactive and bidirectional circling behaviors that are indicative of inner ear dysfunction (11). We have demonstrated that Nor-1 is expressed in the fusion plates of the otic vesicle and in the nonsensory epithelium of the semicircular canals of the vestibule during the critical periods of canal morphogenesis and during continual canal growth.
The expression of Nor-1 is regionally restricted in the semicircular canals and identifies a previously uncharacterized group of polarized epithelial cells within the nonsensory epithelium of the semicircular duct that require Nor-1 for their maintenance. We have shown that restricted expression of Nor-1 within the epithelium is required for proliferation of the nonsensory epithelium of the semicircular canals and their corresponding ampullas to maintain normal lumen volume and canal growth.
Approximately 15% of the Nor-1−/− mice studied exhibited hyperactivity and bidirectional circling behavior similar to that previously observed in shaker/waltzer mutants associated with inner ear vestibular dysfunction (11, 23). Additional balance testing of randomly selected Nor-1−/− mice revealed a selective loss of contact-righting ability in over 30% of these animals, demonstrating an inability to determine orientation without tactile cues, a defect indicative of compromised function of the sensory structures of the vestibule (35). The shaker/waltzer phenotype can result from a variety of different structural and functional abnormalities in the vestibular system. Mutations in genes required for innervation of the inner ear by the vestibular ganglion (10, 12) and morphogenesis of the semicircular canals (1, 9, 17, 19, 31, 41) or sensory structures (12, 34), as well as those involved in regulating endolymphatic fluid production and resorption by the nonsensory dark cells adjacent to the sensory structures and endolymphatic sac, respectively (8, 26, 40), all show significant defects in perception of balance.
Despite the partial penetrance of the observed balance defects in Nor-1−/− mice, all of the null mice studied showed overt structural defects in development of the semicircular canals. While the sensory hair cells appear normal, the endolymphatic fluid-filled membranous ducts of the semicircular canals are severely reduced in all homozygotes whether circling or not and the roof of the sensory ampullas are compressed. Although the circling phenotype was only partially penetrant, no differences were observed in the severity of the defects in the inner ear between circling and noncircling Nor-1−/− mice. The significant constriction of lumen space would be expected to alter the amount and pressure of endolymphatic fluid passing through the canals and sensory structures, resulting in abnormal stimulation of the vestibular mechanosensory hair cells and a circling behavioral response that may vary in severity between animals. In this regard, the phenotype observed in Nor-1−/− mice most closely resembles, although is less severe than, that previously observed in EphB2 receptor tyrosine kinase-null mice (8), which exhibited partially penetrant circling behavior associated with severely reduced lumen volume within the canals. However, in the case of EphB2-null mice, the reduced endolymphatic fluid space was reduced due to decreased endolymph production from the vestibular dark cells. The phenotype observed in Nor-1-null mice might be confounded by disruption of fluid homeostasis in the endolymph. However, unlike that of EphB2, Nor-1 expression is not associated with regions of the inner ear that are known to be active in maintaining fluid homeostasis.
The different stages of development of the semicircular canals have been categorized into four steps: 1, outgrowth; 2, patterning and specification; 3, fusion and resorption; 4, continual growth (27). Recent studies have identified a number of factors that play an important role in this multistep process. The Prx1 and Prx2 genes have been shown to be required for a normal epithelial outgrowth to form the bilayered outpocketings of the otic vesicle (36). Nkx5.1 is required to determine the fusion plate-forming area, and Netrin1, along with Nkx5.1, is necessary for the final fusion of the opposing walls of the outpocketings (17, 31, 41). After formation of the semicircular canals, the mature inner ear is formed through continual growth.
Our data indicate that inactivation of Nor-1 does not affect initial structural differentiation of the semicircular canals but is required for continual proliferative growth of the semicircular canal nonsensory epithelium once the canals are formed. The very restricted expression of Nor-1 in the epithelium and its segregation from proliferating cells support the conclusion that Nor-1 regulates the proliferation of neighboring epithelial cells in a paracrine manner. In addition, the fact that the decrease in epithelial cell proliferation in Nor-1−/− mice precedes the loss of the differentiated-cell phenotype of Nor-1-expressing cells implies that this function is specific to Nor-1 and not due to the loss of cell differentiation.
While the murine pathway for continual growth of the semicircular canals has not been established, roles for BMP4 and its downstream mediator Msx1 have recently been implicated in the regulation of this phase of proliferative canal growth in the chicken (4, 16). In these studies, the implantation of noggin, an inhibitor of BMP2 and BMP4 signaling, in the chick inner ear after formation of the semicircular canals inhibited proliferation of the canal epithelium and was associated with a decrease in expression of Msx1 in epithelial cells (4). The regulation of proliferation by Nor-1 in the mouse suggested that this transcription factor might control canal growth by influencing BMP signaling pathways. In contrast to previous observations in the chick system, however, we were unable to detect expression of Msx1 in the nonsensory epithelium of the mouse semicircular canals. However, in situ hybridization did identify expression of Msx2 in a regionally restricted pattern that coincides with that previously observed for Msx1 in the chicken and is localized opposite to that expressing Nor-1 (data not shown). Our analysis of the expression of three members of the BMP family, BMP2, BMP4, and BMP7, has established that BMP2 and BMP7 are expressed in the nonsensory epithelium of the semicircular canals (data not shown). However, because of the very low to undetectable levels of expression of some of these RNAs in the canal, we cannot make a conclusive statement about their involvement in continual semicircular canal growth in mice. Thus, the precise molecular pathway by which Nor-1 influences semicircular canal proliferation remains to be established.
Taken together, our results identify Nor-1 as a key mediator of the proliferative growth phase of semicircular canal development. Future studies to identify target genes that are regulated by Nor-1 should help elucidate the molecular pathways mediating canal growth, as well as provide important new insights into the molecular mechanisms underlying regional specification of the nonsensory epithelium of the vestibular apparatus.
Acknowledgments
We thank Silvia Briones for mouse breeding and genotyping. We also thank M. Tessier-Lavigne, B. Hogan, D. Sassoon, T. Lufkin, and R. Maas for providing plasmids used to generate probes for in situ hybridization.
This work was supported by National Institutes of Health grants DK52429 and DK57743 to O.M.C. and by the NIDCD intramural program (D.K.W.).