Figure 1.
Phenotype and hearing sensitivity of mtl and bsd adult mice. A
, Dorsal views show the different body size and tail lengths of wildtype and homozygotes. Ventral views show the presence of white belly patches of variable size in mtl and bsd homozygotes. B, Mean ABR thresholds (± standard deviation) plots for mtl/mtl compared to +/+ tested at 6–7 weeks of age (upper panel) and plots showing the mean ABR thresholds of mice with genotypes +/+, +/bsd and bsd/bsd around 6 weeks of age (lower panel). Clicks and tones were presented up to 95 dB SPL; all mutants showed no detectable ABR at even the highest sound level presented (these values are indicated by the arrows). The absence of discernible ABR at 95 dB SPL suggests a profound hearing loss.
Table 1.
Complementation tests drJ/mtl and bsd/mtl.
Figure 2.
Complementation test between mtl and bsd mice. A
, Phenotype of mtl and bsd adult mice from the complementation tests. Dorsal and ventral views are shown and genotypes indicated. We found mice which were compound mutants mtl/bsd showing the affected phenotype, shorter tails and white belly patches, whereas other mice were +/+, +/mtl or +/bsd and showed wildtype phenotype. B, Mean ABR thresholds (± standard deviation) are plotted for the complementation test at 16 weeks old. Clicks and tones were presented up to 95 dB SPL; all mice showed no detectable ABR at even the highest sound level presented (these values are indicated by the arrows), which suggests a profound hearing loss.
Figure 3.
A point mutation in mtl mutants leads to a splice defect. A
, Genomic sequence of the splice donor site at 3′ end of exon 4 of transcript Lmx1a-001. B, Base pair change c.173+1 G>A identified in mtl mutants. C, D, Partial traces and sequence of cDNA obtained from E10.5 whole embryos. C, Wildtype transcript cDNA pre and post splicing, exonic sequence in upper case and intronic sequence in lower case. The sequence trace shows cDNA of spliced transcript at the 3′ end of exon 4 and 5′ end of exon 5. D, cDNA from mtl mutant showing the irregular splicing between exons 4 and 5, causing an extension of 44 base pairs into intron 4 (sequence in lower case). This extension ends as it finds a cryptic splice acceptor site, GTAT (in red colour), within intron 4. E, Effect of the altered transcript is a continued open reading frame in mtl homozygotes of the exon 4 leading to a termination of protein translation at the TAA stop codon in the altered amino acid position 225* (arrow). F, PCR amplification of cDNA using primers located within exon 3 and exon 6. Bands of the proper size are amplified for wildtype and heterozygotes whereas in mtl mutants the amplified band is bigger in size due to the 44 base pair extension.
Figure 4.
Deletion of exon 3 of Lmx1a in bsd mutants. A
, PCR amplification of genomic DNA from +/+, +/bsd and bsd/bsd mice with primers specific to each exon of Lmx1a gene. One single band is amplified for each exon and sequenced for all bsd genotypes with no differences between mutants (M) and controls (wt, het) except for exon 3, where no band was amplified in bsd mutants (asterisk). B, Partial traces and sequence of Lmx1a intron 4–5 are shown. At 30 base pairs downstream of exon 4 we identified a single nucleotide polymorphism (SNP)-from A to G- that we used to genotype bsd mice. C, PCR amplification of bsd cDNA from +/+, +/bsd and bsd/bsd mice at E10.5 with primers specific to transcript Lmx1a-001 covering exons 2–8 (Table S2). Bands were amplified for controls (wt, het) but no band was detected in mutants (M). We designed primers for small fragments of the transcript (Table S2). Controls (+/+) showed bands for all combinations of primers (numbers on top indicate position of primers forward and reverse, respectively) whereas in bsd/bsd no bands were amplified by primers 13 and 35 (black asterisks). Interestingly, we found a small band of less than 100 bps with primers 24. This band is likely to correspond to the size of the amplicon (322 bp) minus the deleted exon 3 (233 bp) of bsd mutants. D, Diagram showing structure of Lmx1a gene in +/+, mtl/mtl and bsd/bsd. Initiation codon in exon 1 and termination codon in exon 8 are indicated. Point mutation in mtl homozygotes is indicated by a red asterisk and deletion in bsd mutants is shown by a red dotted line. Mtl mutation affects the homeodomain whereas in bsd the deletion of exon 3 involves the LIM2 domain.
Figure 5.
Morphology of inner ears and analysis of sensory patches in mutants.
A–F, Paint-filled inner ears from mtl and bsd mutants at E16.5. Inner ears (front and back views) from wildtype (A, B), mtl/mtl (C, D) and bsd/bsd (E, F) mice are shown. Both homozygotes show under-developed cochlear ducts (red arrowheads) compared to their wildtype littermates. Vestibular system in homozygotes appears like a cyst without semicircular canals (white arrowheads). G, H, Confocal imaging and 3D reconstruction of P0 inner ears of one mtl mutant (G) and one wildtype (H). We used Myo7a (green) for sensory hair cells, rhodamine-phalloidin (red) for actin and DAPI (blue) for cell nuclei. The expression of Myo7a shows the main sensory patches in the inner ear (anterior, posterior and lateral cristae, utricle and saccule in vestibular system and organ of Corti in the cochlea) (H). In mtl mutants cristae can be identified by their relative position in the otocyst and utricle and saccule domains look abnormal compared to those in wildtypes (G, H). In mtl mutants there are two continuous Myo7a-positive hair cell populations descending to the truncated cochlea (insets). One of these populations show a quite uniform organization of the hair cells within the organ of Corti but without a tectorial membrane (I, area boxed in G), whereas the other population extending to a more distal location within the cochlea (K, L, area boxed in G) displays properties of a developing organ of Corti containing hair cells and a tectorial membrane (L). (K and L are reconstructions from the same confocal image stack. K, Organ of Corti surface, image sub-stack underneath tectorial membrane; L is a rotated and tilted 3D reconstruction of the same image stack). Asc, anterior semicircular canal; cd, cochlear duct; ed, endolymphatic duct; lsc, lateral semicircular canal; psc, posterior semicircular canal; sac, saccule, OHC, outer hair cells; IHC, inner hair cells; TM, tectorial membrane. Scale bars: A–F, 200 µm; G–H, 500 µm.
Figure 6.
Analysis of sensory patch formation in cochlear and vestibular sections of mtl mutants at E16.5.
Expression of markers for sensory patch development, Jag1 (A, A’, B–D, M–O, R, R’, S), Myo7a (E, E’, F–H, P, T, T’, U) and Prox1 (I, I’, J–L, Q, V, V’, W) are shown (A’, E’,I’, R’, T’ and W’ are high magnifications of A, E, I, R, T and W, respectively). (A–L, A’, E’, I’) are wildtypes at E16.5. (M–W, R’, T’, V’) are mtl/mtl at E16.5. (D, H, L) Posterior crista (PC) is positive for these markers at E16.5. In mtl mutants, vestibular labyrinth formation fails and, instead, a large cystic otocyst with a wide cochlear duct can be observed (M, N), with enlarged endolymphatic space (R–W). Sensory patches are positive for Jag1 and the expression is detected in mtl mutants although in abnormally extended pattern. The mutant cristae (O–Q) show expression of the three markers and a similar degree of organization (multilayered, bulged epithelium, layered expression of Myo7a and Prox1) as in wildtypes. The utrico-saccular sensory patches in mtl mutants are disorganized but express the sensory patch markers. In the base of the mutant cochlear duct (R, T, V) all three markers are expressed, but the pattern is aberrant (R’, T’, W’) labelling two sensory patches compared to controls where the single organ of Corti is labelled. The apex of the truncated mutant cochlear duct (S, U, W) shows Jag1 expression, faint Prox1 expression and no Myo7a expression, reflecting the baso-apical differences in hair cell development of the littermate controls. Asterisks show the endolymphatic compartment, which is extremely enlarged in mtl mutants (R, T, V) compared to wildtypes (A, E, I). Scale bars: A, E, I, B, F, J, C, G, K, M, N, R–W, 200 µm; A’, E’, I’, D, H, L, R’, T’, V’, 50 µm.
Figure 7.
Expression analysis of Lmx1a, FOXI1 and FGF9 in the developing inner ear. A–F
, In situ hybridization of Lmx1a on wildtype inner ear sections at E10.5 (A, C, E) and E12.5 (B, D, F). Lmx1a mRNA expression is shown in blue and nuclear counterstain in pink. In wildtype inner ears Lmx1a is expressed in the otic vesicle and in the endolymphatic sac (black arrow in A, B and D), whereas delaminating neuroblasts are Lmx1a-negative (A, C, E, white arrows). At E12.5 Lmx1a expression is also found in the region of the fusion plates which will later form the semicircular canals (asterisk in B). Lmx1a mRNA expression appears to be restricted to prospective non-sensory inner ear tissue. G–H, In wildtypes FOXI1 is specifically expressed in a subpopulation of cells within the endolymphatic sac (G) whereas in mtl mutants the typical evagination of the endolymphatic sac/duct fails and no FOXI1-positive cells are detected in or adjacent to the otic vesicle/otocyst (H). I–S, At E11.5 FGF9 expression highlights areas of epithelial thickening on the luminal as well as the mesenchymal side of the otic vesicle epithelium, such as the outpocketings of the developing semicircular canals (arrows in I). The endolymphatic duct is negative for FGF9 (asterisk in I). J–S, Series of sections covering the complete otocyst arranged from lateral to medial, of one mtl homozygote showing only a rudimentary otocyst epithelium rearrangement. Sections also show complete absence of an endolymphatic duct/sac structure. FGF9 is expressed in the mutant epithelium and marks epithelial thickening, but unlike in wildtypes there is no semicircular canal formation. The luminal and mesenchyme sides of the thickened mutant epithelium looked underdeveloped (arrows in I and P). SG, spiral ganglion. Scale bars: 100 µm.
Figure 8.
Expression analysis of Lmx1a, Hmx3, Netrin1 and Pax2 mRNA levels in bsd mutants.
Quantitative real-time PCR on cDNA generated from normalized total RNA from bsd E10.5 whole embryos (only bsd results are shown). A, Lmx1a is significantly down-regulated in bsd homozygous mice at E10.5 (n = 5+/+, n = 5 bsd/bsd; *P<0.05, t test; mtl data not shown). B, Hmx3, Netrin1 and Pax2 are also significantly down-regulated in bsd mutants (n = 5+/+, n = 5 bsd/bsd; *P<0.05, t test; mtl data not shown). Error bars represent standard deviations. Quantities normalized to Hprt levels. Hprt was used as control for the quantity of material.
Figure 9.
Expression analysis of Hmx3, Netrin1 and Pax2 by immunohistochemistry.
Expression of Hmx3 protein was performed on wildtype (A, A’, B, B’) and mtl mutant sections (C, C’, D, D’) at E10.5 (A, A’, C, C’) and E12.5 (B, B’, D, D’). At E10.5 Hmx3 protein expression in the otocyst looks diffused in wildtypes and mutants (A’, C’) and expression in CNS is also strong (A’, C’, arrowheads). At E12.5 there is strong expression of Hmx3 in the ED/ES (B’) and in the outpocketings of the developing semicircular canals (B, B’, arrowhead). In mtl mutants no semicircular canals or ED/ES are found and a simple cyst is observed instead (D, D’). Netrin1 protein expression was analysed in wildtype (E, E’, E’’, G, G’, G’’,H, I) and mtl mutants (F, F’,F’’,K,K’,J) at E10.5 (E, E’, E’’, F, F’, F’’) and E12.5 (G,G’,G’’,H,I,J, K,K’). Netrin1 protein was detected in the dorsolateral margin of the wildtype otocyst (E’, E’’) whereas mutants did not show any expression (F, F’’). In wildtypes at E12.5 strong expression of Netrin1 protein is found in the outpocketings of the developing semicircular canals (G’, magnified view in G’’) but no expression was found in the ED/ES (H). Similar expression was detected in the neural tube of wildtypes and mtl mutants (I, J, arrowheads). Pax2 protein expression was analysed in wildtype (L, L’, M, M’) and mtl mutants (N, N’,O, O’) at E10.5 (L, L’, N, N’) and E12.5 (M, M’, O, O’). At E10.5 Pax2 protein expression is observed in the ventral region of the otocyst in wildtypes and mutants (L’, N’). In wildtypes at E12.5 strong signal is found in ED/ES, in outpocketings of the semicircular canals and in a ventral margin of the otic vesicle (M’, arrowheads). In mtl mutants the ED/ES and semicircular canals are missing and no expression can be found except for the marginal region of the presumptive cochlear duct, which still shows Pax2 expression (O’, arrowheads). Scale bars: (A–G, I–O), 200 µm; (A’–G’, H, K’, L’–O’) 50 µm; (E’’–G’’) 10 µm.
Figure 10.
Schematic diagram showing potential interactions between Lmx1a and other markers expressed in the otocyst at around E10.5.
The diagram illustrates the effects of Lmx1a on expression of eight downstream proteins: Ntn1 (this work), Sox2 [13], and Bmp4, Dlx5, Wnt2b, Fgf3, Tlx3 and NeuroD [14]. The effects were revealed by expansion or contraction of the expression domain in an Lmx1a mutant otocyst. Observations from early stages only were included because interpretation of expression in later stages would be confounded by the abnormality morphology of inner ear components. Interactions amongst the downstream proteins were based upon previously published data, not necessarily from the ear, identified using Ingenuity (www.ingenuity.com) [52], [53], [54], [55], [56]. All interactions may be indirect. Dotted line represents weak upregulation.