Animals

Procedures involving animals were conducted in accordance with European Community (Council Directive 86/609) and French Agriculture Ministry directives (Council Directive 87–848). The project was approved by the “Cuvier” ethical committee of the French National Museum of National History (MNHN) and validated by the French Ministry of Agriculture (approval APAFIS#26087-2020061614327140).

All animals were housed in light, temperature (21°C) and humidity (50%−60%) controlled conditions. Food and water were available ad libitum. Mice were back-crossed and maintained with a B6D2F1/J background. All genotyping was made by polymerase chain reaction (PCR) on ear biopsies.

Males carrying loxP sequences flanking the exon 2-containing homeobox domain of Dlx5 and Dlx6 genes (Dlx5/6^flox/flox) [29] were crossed with Sox10^Cre/+ females [28] to inactivate Dlx5/6 (Sox10^Cre/+: Dlx5/6^flox/flox, Dlx5/6 cKO). For the genetic lineage tracing, Sox10^Cre/+: Dlx5/6^flox/+ females were crossed with males carrying Dlx5/6^flox/flox and the Rosa^{lox-stop-lox-lacZ} reporter [30] (Dlx5/6^flox/flox: Rosa^lacZ/lacZ). For embryo and foetus collection, females were euthanized by intraperitoneal injection of a lethal dose of pentobarbital, in accordance with ethical guidelines to minimize suffering. For all experiments, the controls of mutant specimens were from the same litter. We included as control genotypes the specimens heterozygous for the Cre and/or flox or homozygous for the flox only, which we had previously validated as showing a normal phenotype compared to non-genetically modified ‘wild-type’ specimens. For Dlx5 expression analysis, males carrying the lacZ reporter (Dlx5^lacZ/+) [18] were crossed with B6D2F1/J females.

Samples at E10.5, E11.5 and E17.5 stages were fixed in 4% paraformaldehyde (PFA) diluted in PBS overnight at 4°C and washed in PBS. Foetuses at stage E14.5 and E16.5 were fixed for 3 hours in 4% PFA diluted in PBS and 0.5% Triton (PBSTr) at 4°C and washed twice in Tween 0.1% diluted in PBS (PBST) and then again overnight at 4°C. For all samples, noon of the day of the vaginal plug was considered as E0.5.

X-Gal and Immunofluorescence staining

Whole-mount samples at E9.5, E10.5, E11.5 and E14.5 were collected and fixed in PFA 4% for 3 hours, washed and treated with X-gal to reveal β-galactosidase activity as previously described [7].

For whole-mount immunofluorescence staining, embryos at E11.5 were collected, fixed and dehydrated in a graded series of methanol (MeOH). They were then permeabilized in Dent’s solution (80% MeOH – 20% DMSO) overnight at 4°C and rehydrated in graded MeOH series and washed in PBS. The non-specific antigenic sites were blocked using a blocking solution (BS, PBS/ 20% goat serum/ 3% BSA/ 0.5% Triton) for 3 hours at room temperature (RT). The embryos were incubated with primary antibody diluted in BS for 2–3 days with gentle shaking at 4°C, washed several times over day in PBST and incubated with secondary antibodies also diluted in BS for 2–3 days with gentle shaking at 4°C in the dark. They were finally washed several times over day and cleared following the Cubic protocol as previously described [31].

For immunofluorescence staining on sections, foetuses were cryopreserved in 30% sucrose PBS and embedded in OCT for 20 μm sectioning using a Leica cryostat. Cryosections were dried for 30 minutes, washed in PBS and immersed in BS for 2 hours at RT. Primary antibodies were diluted in BS and sections were incubated overnight at 4°C. The sections were then washed and incubated with secondary antibodies diluted in BS for 2 hours at RT in the dark. Finally, the sections were washed in PBS and mounted with Fluoromount (Fisher Scientific, 15586276) for analysis (see S1 Table for antibody information).

Apoptosis analysis

Whole mount samples at E11.5 were collected, fixed, dehydrated in a graded series of MeOH and stored at −20°C. The apoptotic cells were revealed using the ApoptagⓇ Fluorescein in situ Apoptosis detection kit (Merck, S7110) following the kit guidelines.

Proliferation analysis

For the analysis of cell proliferation, control and mutant specimens at E12.5 were embedded in OCT, frozen, and 12 μm cryosections were prepared using a Leica cryostat. Immunostainings were then performed using anti-Sox9, anti-Desmin, and anti-PHH3 antibodies to identify chondrogenic, myogenic, and proliferating cells respectively. Confocal images were acquired using a Zeiss microscope on three consecutive sections from each region of interest. Manual counting of the total number of PHH3– and DAPI-positive cells was performed on Fiji software (ImageJ distribution, open-source) on each consecutive section. The PHH3/DAPI ratio was then determined to compare the mean number of proliferative cells in each region of interest in control and mutant conditions.

Immunofluorescence acquisitions and analysis

Whole mount samples for lightsheet fluorescence imaging were prepared using CUBIC L-R(N) clearing protocol [31], after fixation and immunofluorescence staining. For light sheet fluorescence imaging, samples were included in agarose low melting temperature 2% (A9414, Sigma-Aldrich). We first incubated samples in 10 mL of 50% CUBIC-L reagent composed of 10% (w/w) N-butyldiethanolamine (471240, Sigma-Aldrich) and 10% (w/w) triton X-100 (Sigma-Aldrich) diluted in MilliQ water for 2 days at RT, for 2 days in 10 mL of 100% CUBIC-L at RT, and washed 3 times in PBS. They were then incubated in 10 mL of 50% CUBIC-R(N) reagent composed of 45% (w/w) Antipyrine (A5882, Sigma-Aldrich), 30% (w/w) Nicotinamide (A15970, Thermo Fisher) an 0,5% (w/w) N- butyldiethanolamine (471240, Sigma-Aldrich) diluted in MilliQ water for 2 days at RT and for 2 days in 10 mL of 100% CUBIC-L at RT. For lightsheet imaging, the samples were incubated in a fresh CUBIC-R(N) reagent in the tank and incubated for a few minutes for refractive index homogenization.

Epifluorescence imaging of embryos and foetuses in toto was performed with a Zeiss Macro-Apotome AxioZoom V16. Confocal imaging was made using a Zeiss LSM 980 with AiryScan 2 detector. Lightsheet imaging was performed with the Alpha3 system (PhaseView, France), equipped with a 10 × XL Plan N (XLPLN10XSVMP, Olympus, Japan) clearing objective with refractive index adaptive collar (RI 1.33–1.52), a sCMOs Orca Flash4 camera (Hamamatsu, Japan). 2.20 and the acquisition software QtSPIM. For the 561 nm excitation laser, the ET600/50m emission filter was used and for the 638 nm excitation laser, the emission filter ET670/50m (S2 Table).

Image reconstruction and analysis

Confocal and macroscope images were reconstructed and analyzed using Fiji ImageJ (Java version 1.8.0) and ZEN Blue 3.5 (Zeiss). Airyscan and lightsheet images were converted in arivis SIS Converter 3.1.1 and analyzed with arivis Vision4D 3.0.1 software (arivis AG, Germany).

X-ray computed microtomography (micro-CT) acquisitions and analysis

E17.5 foetuses were fixed overnight in PFA 4% and washed in several baths of PBST over day. They were dehydrated in upgrading EtOH and MeOH series and soaked in phosphotungstic acid (PTA) 1.5% diluted in MeOH 90% for 4 days. The PTA solution was changed every day with a fresh solution to ensure optimal penetration of the contrast agent. The micro-CT scanning was done by X-ray micro-CT system GE Phoenix v|tome|x L 240 (Waygate Technologies/ Baker Hughes Digital Solutions GmbH, Wunstorf) equipped with a 180 kV/15 W maximum power nanofocus X-ray tube and a high-contrast flat panel dynamic detector 41|100 with 4000 × 4000 pixels and a pixel size of 100 × 100 μm. The exposure time was 800 ms in 2000 positions over 360°. Three projections were captured in each position and an average of the signal was used to improve the signal-to-noise ratio. The microCT scan was carried out at 600 kV acceleration voltage and with 200 μA X-ray tube current. The beam was filtered by a 0.2 mm-thick aluminium filter to avoid beam hardening artefacts. The isotropic voxel size of obtained volumes was 9 μm for all four samples: 2 control and 2 Dlx5/6 cKO. The tomographic reconstruction was performed using GE phoenix datos|x 2.0 software (Waygate Technologies/ Baker Hughes Digital Solutions GmbH, Wunstorf, Germany). Reconstructed slices were imported to software Avizo 7.1 (Thermo Fisher Scientific, Waltham, MA, USA) for semi-automatic segmentation as previously described in our previous work [32]. The muscles and other anatomical structures of interest were outlined by the operator in every 3rd to 5th slice depending on the complexity of the structure and the rest was calculated by linear interpolation between manually outlined slices. The segmented structures were then transferred to polygonal mesh (STL format) and imported to VG Studio MAX 3.5 (Volume Graphics GmbH, Heidelberg, Germany) for further visualization and analysis.

Preparation of cDNA and RNA probes from embryos

RNA extractions were done using PureLink RNA Mini Kit with Trizol following the kit guidelines (ThermoFisher, 12183018A) and turned into cDNA libraries following the SuperScript III First-Strand Synthesis SuperMix for qRT-PCR (ThermoFisher, 11752-050) protocol. RNA probes were designed on primer BLAST (https://www.ncbi.nlm.nih.gov/tools/primer‐blast/) and synthesised by PCR, following the guidelines of Gel Extraction Kit (Qiagen, 28704), PCR Purification Kit (Qiagen, 28104) and using T7 to generate RNA probes for in situ hybridization.

In situ hybridization

Embryos at E10.5 were collected, fixed, dehydrated in a graded series of MeOH and stored at −20°C. They were rehydrated progressively in PBST, permeabilized by Proteinase K (10 µg/ml in PBST) and post fixed in PFA 4%. Samples were equilibrated in a solution of 50% Hybridization solution (Hb) – 50% PBST 10 minutes at room temperature under agitation and in 100% Hb for an hour at 70°C. Embryos were then incubated in Hb with 1ng/µL RNA probe overnight at 70°C.

After 4 successive washes of 30 minutes with Wash Buffer (WB) at 70°C, they were equilibrated in a graded series of TBST at room temperature. Embryos were blocked with Blocking Buffer (BB) and incubated overnight at 4°C in Antibody Solution (AbS) containing anti-DIG antibodies coupled with alkaline phosphatase (Merk, 11093274910). Samples were washed in TBST several times followed by 3 NTMT washes and revealed with BM-Purple (Merk, 11442074001) until they reached proper staining contrasts. Composition of solutions are presented in (S3 Table).

Dlx5/6 inactivation affects the entire vocal tract and auditory systems

To induce targeted inactivation of Dlx5/6 in CNCC and otic vesicle (Dlx5/6 cKO), we crossed Sox10^Cre/+ deleter mice with the previously-described Dlx5/6^flox/flox mouse line [9,29], in which both the homeodomain-encoding region (exon 2) of Dlx5 and Dlx6 are flanked by loxP sequences. Sox10 and Dlx5/6 genes are both expressed in post migrating CNCC populations, in the otic placodes and their derivatives (S1 Fig) [15,18,27,28].

The macroscopic phenotypic analysis before birth revealed that Dlx5/6 cKO foetuses presented craniofacial malformations similar to those observed in previous Dlx5/6 inactivated mouse models [9,15,18,19], including a transformed lower jaw exhibiting ectopic vibrissae and hypoplastic external ear (Fig 1A–A’).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Dlx5/6 genes are required for the proper formation of organs composing both the vocal and auditory systems.

(A–A’) Macroscopic view of control and Dlx5/6 cKO mutant phenotypes at E16.5. In mutants, the lower jaw is defective with ectopic vibrissae (vb*) and the external ear (pinna) shows severe hypoplasia. (B–D’) Immunofluorescence stainings for the cartilage marker Sox9 and muscle marker MyHC on coronal and frontal sections at the levels indicated in A. (B–B’) The Dlx5/6 mutant presents malformations of external, middle and inner ear components including reduced pinna, defective incus and semicircular canals, and enlarged cochlear duct. (C–D’) The Dlx5/6 mutant presents a transformed lower jaw characterized by an absence of Meckel cartilage and of associated masticatory musculature. The tongue is severely reduced with disorganized myofibers. The thyroid cartilage also seems reduced compared to control (n = 6 each condition, control genotype: Dlx5/6^flox/flox). (E–H’) 3D reconstructions of the different structures composing the vocal tract and auditory systems in control and Dlx5/6 mutant conditions at E17.5 (n = 2 each condition, control genotype: Dlx5/6^flox/flox and Dlx5/6^flox/+). (F–F’) The Meckel cartilage is missing (purple double asterisk in F’), the middle ear ossicles (incus and malleus) are defective and the vestibular and cochlear membranous labyrinths are swollen in mutants compared to controls. (G–H’) The lower jaw transformation in mutants is associated with a hypoplastic tongue and mispatterned extrinsic laryngeal (infrahyoid) muscles. The thyroid and hyoid cartilages also show defects of formation in the mutant context. All structures presenting a malformation in the Dlx5/6 mutant context are indicated with an asterisk. Absent or reduced structures are noted with a double asterisk. See also S1 and S2 Appendices for interactive PDFs of 3D reconstructions. Abbreviations: ac, arytenoid cartilage; cc, cricoid cartilage; co, cochlea; ee, external ear; hy, hyoid cartilage; ic, incisor; in, incus; lr, laryngeal muscles; ma, malleus; mc, Meckel cartilage; md, mandible; ml, molar; ms, masseter muscle; mx, maxilla; pt, petrous part of the temporal bone; ph, pharyngeal muscles; sc, semicircular canals; tc, thyroid cartilage; tg, tongue; vb, vibrissae. Scale bar in A’ for A-A’ 1000 µm, for B-D’ 200 µm.

https://doi.org/10.1371/journal.pone.0337426.g001

First, we performed immunofluorescence staining for MyHC and Sox9 on sections to investigate respectively the muscular and skeletal phenotypes of control and mutant foetuses at E16.5. Regarding the auditory system, Sox9 immunostainings revealed an hypoplasia of the middle ear ossicles of CNCC origin [13] (Fig 1B–B’). The inner ear showed defects of the membranous and bony labyrinths of the semicircular canals and cochlea (Fig 1B–B’). At the level of the vocal tract, we observed that the transformed lower jaw was associated with tongue hypoplasia and masticatory masseter muscle aplasia (Fig 1C–C’), as previously described after Dlx5/6 inactivation [9,15]. The hyoid and the laryngeal thyroid cartilages derived from CNCC [7,12] seemed also impacted but the extent of malformation was difficult to access on sections.

To push forward the phenotypic analysis, we performed micro-CT scanning of mutant and control foetuses at E17.5 to access the complex layout of the craniocervical structures and to reconstruct in 3D the different components of the vocal tract and auditory system (Fig 1E–H’, S1 and S2 Appendices). This approach clearly revealed the extent of inner ear malformations including swollen membranous labyrinths with indiscernible semicircular canals and wider cochlear duct (Fig 1F–F’). Moreover, we observed the presence of two small mispatterned skeletal structures at the site of the incus and malleus ossicles associated with an absence of Meckel cartilage in mutants (Fig 1F–F’). The 3D reconstruction of the jaw musculoskeletal elements highlighted the severe malformations of the lower jaw (mandible) and of the tongue musculature (Fig 1G–G’). The hyoid and thyroid cartilages were severely affected and extrinsic laryngeal (infrahyoid) muscles showed patterning defects (Fig 1H–H’). Pharyngeal and intrinsic laryngeal muscles presented morphology differences compared to controls adapted to the defective hyoid and thyroid cartilages to which they connect. In contrast, the cricoid and arytenoid cartilages of mesodermal origin [7,12] appeared unaffected by Dlx5/6 mutation (Fig 1D–D’, S1 and S2 Appendices).

Dlx5/6 genes regulate early CNCC and otic vesicle patterning

We then used genetic lineage tracing approaches to analyze the morphogenetic events at the basis of the musculoskeletal malformations in Dlx5/6 mutants. We first examined the fate of CNCC by Sox10 genetic lineage tracing using the Rosa^{lox-stop-lox-lacZ} (Rosa^lacZ/+) [30] reporter analysis at E11.5 in control and mutant conditions (Fig 2A, B). β-galactosidase (β-gal) positive cells corresponding to placode derivatives, were detected in the tympanic ring and the endolymphatic sac of the otic vesicle in controls (Fig 2A–A’). We also observed β-gal-positive cells corresponding to CNCC derivatives within the mesenchyme of the maxillary and mandibular parts of the PA1, the hyoid arch (PA2) and more posterior arches (PA3–6) (Fig 2A”), which later give rise to the skeletal components of the jaws, the ossicles and the hyoid and thyroid cartilages. In Dlx5/6 cKO mutants, both compartments showed organization defects of β-gal-positive cells. In the developing ear, the endolymphatic sac was missing and the tympanic ring was mis-patterned, whereas β-gal-positive CNCC were scattered and disorganized within the PA1–6 (Fig 2B–B”). We did not detect any obvious differences between controls and mutants in cell apoptosis within the pharyngeal arches at E11.5, nor in cell proliferation within otic and vocal components at E12.5 (S3 Fig). Remarkably, no additional defects were observed along the body axis, indicating that our targeted Dlx5/6 mutation specifically affects vocal tract and auditory precursors.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Dlx5/6 regulate the patterning of CNCC in the PAs and of the otic vesicle without affecting their connective tissue fate.

(A-B) Sox10 genetic lineage tracing of CNCC and otic placode derivatives using the Rosa-lacZ reporter in control and Dlx5/6 mutant embryos at E11.5. In mutants, the otic vesicle shows a lack of endolymphatic sac (grey double asterisk in B’) and defect of the tympanic ring compared to controls (A’–B’). β-gal-positive CNCC compose the mesenchymal population of the PAs in controls that show aberrant patterning in mutants (A”–B”). (C–F’) Immunofluorescence stainings on sections for β-galactosidase labelling CCNC derivatives, and for Tcf4, MyHC and Sox9 that mark connective tissue fibroblasts, muscles and cartilages respectively in control and mutant foetuses at E14.5. Tcf4-positive fibroblasts of CNCC origin are detected between masticatory myofibers associated with the lower jaw in controls (C–C’). β-gal/Tcf4-positive fibroblasts are maintained in the masticatory region of mutants despite the absence of differentiated musculature (D–D’). At the level of the glottis and hyoid cartilage, we observe differences in the distribution of CNCC derivatives that still keep their cartilaginous identity (E–F’). All structures presenting a malformation are marked with an asterisk. Absent or reduced structures are noted with a double asterisk (n = 3 each condition, control genotype: Sox10^Cre/+; Dlx5/6^flox/+; Rosa^lacZ/+). Abbreviations: ed, endolymphatic sac; gl, glottis; hy, hyoid cartilage; mr, masticatory region; ms, masseter muscle; ov, otic vesicle; PA, pharyngeal arch; tp, tympanic ring. Scale bar in B for A–B 400 µm, for A’–B’ 150 µm, for A”-B” 200 µm, in F for C–F 200 µm, for C’–F’ 20 µm.

https://doi.org/10.1371/journal.pone.0337426.g002

We then followed the fate of β-gal derivatives during musculoskeletal differentiation at E14.5. We performed immunofluorescence stainings on sections for anti-β-gal and for markers of cartilaginous, fibroblastic or tendinous derivatives (Sox9, Tcf4 and Tnc respectively) of placodal and CNCC populations. In controls, CNCC-derived β-gal cells corresponded to Tcf4-positive fibroblasts forming the muscle connective tissue along the myofibers of the masticatory masseter and tongue muscles (Fig 2C–C’, S2A Fig) [7]. In mutants, the β-gal/Tcf4-positive muscle connective tissue fibroblasts are present, even in the absence of masseter musculature (Fig 2D–D’, S2B Fig). The β-gal/Tcf4-positive populations of the external ear were also preserved in the hypoplastic pinna of mutants that however showed defects in organization compared to controls (S2C, S2D Fig). The β-gal-positive CNCC-derived population showed aberrant patterning at the level of the glottis, the hyoid and thyroid cartilages but kept their cartilaginous and tendinous identity as observed by Sox9, Tnc and β-gal colocalization in the structures of both controls and mutants (Fig 2E–F’, S2E–S2H Fig).

We then investigated whether such alterations in Dlx5/6 expression could also affect the innervation of vocal tract and auditory components. We analyzed the neuromuscular system of control and mutant embryos by in toto immunofluorescence stainings at E11.5 targeting the neurofilament (NF) and Desmin proteins that mark respectively the developing peripheral nervous and muscular systems (Fig 3). In controls, the maxillary and mandibular branches of the trigeminal ganglion (cranial nerve CN V) project into the upper and lower jaw buds to innervate notably the sensory vibrissae and the masticatory muscle precursor (Fig 3A–A’). We also observed the facial nerve (CN VII) that innervates the facial muscles and its chorda tympani branch that later join the lingual nerve (CN V) in the mandibular arch [33]. More dorsally, the vestibulocochlear nerve (CN VIII) innervates the inner ear, and posteriorly the glossopharyngeal, the vagus and accessory nerves (CN IX, X, XI) project to innervate the tongue, throat and laryngeal components (Fig 3A–A’, S1 Movie) [34]. In Dlx5/6 cKO, the mandibular branch of the trigeminal ganglion shows increased distal arborization characteristic of the maxillary branch to innervate the ectopic vibrissae observed on the transformed lower jaw (Figs 1A’, E’, 3B, S4 Fig). The mandibular neuronal projection toward the masticatory muscle precursor is also reduced in E12.5 mutants (S4 Fig). The vestibulocochlear nerve (CN VIII) shows defect of projection on the otic vesicle in the mutants compared to controls (Fig 3B–B’). Moreover, one branch of the chorda tympani that innervates the distal part of tongue later during development was missing. However, we did not detect any obvious differences in the hypoglossal nerve (CN XII) while the hypoglossal cord forming the tongue was already reduced in E11.5 mutants. In contrast, the masticatory muscle precursor was present at this stage as previously described in constitutive Dlx5/6 mutants, the defect of masticatory muscle differentiation occurring during late embryonic stages in mutants (Fig 2, S4 Fig) [15]. We did not notice differences in the configuration of the glossopharyngeal, the vagus and accessory nerves (CN IX, X, XI) between controls and mutants (Fig 3B’–D, S2 Movie).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Dlx5/6 inactivation induces innervation defects of some vocal and auditory components.

(A–D) In toto immunofluorescence staining of control and mutant embryos at E11.5 for the neurofilament (NF) and Desmin that label the peripheral nervous system and muscle precursors respectively. Lateral (A–B’) and ventral (C–D) views with magnification at the level of vocal tract and auditory precursors (A’–B’). In mutants, the mandibular branch of the trigeminal nerve (V) presents characteristics of the maxillary branch with increased distal arborization to connect ectopic vibrissae. The vestibulocochlear nerve (VIII) shows defective projection toward the otic vesicle. The chorda tympani branch of the facial nerve (VII) (blue arrowheads in controls A’, C) misses the neuronal projection that later innervates the tongue (blue double asterisk) in mutants B’, D). Note that the masticatory muscle precursor is present at E11.5 while the hypoglossal cord giving rise to tongue muscles is already reduced in mutants. All structures presenting a mutant phenotype are noted with an asterisk. Absent or reduced structures are noted with a double asterisk (n = 3 each condition, control genotype: Sox10^Cre/+; Dlx5/6^flox/+). See also S1 and S2 Movies for interactive view of the neuromuscular system in control and mutant conditions. Abbreviations: ct, chorda tympani; h, heart; hc, hypoglossal cord; md, mandibular branch of the trigeminal ganglion; mm, masticatory muscle precursor; mx, maxillary branch of the trigeminal ganglion. Scale bar in A for A-B, C-D 300 µm, for A’-B’ 200 µm.

https://doi.org/10.1371/journal.pone.0337426.g003

Altogether, our results show that Dlx5/6 expression concurrently regulates the proper patterning and the innervation of both PA and otic placode derivatives, without altering their connective tissue fate during the development of the vocal tract and auditory components.

Dlx5/6 modulate BMP signaling in both pharyngeal arches and otic vesicle

It was previously reported that several actors of the BMP signaling pathway are downstream targets of Dlx5 in the otic vesicle [35]. Some BMP signaling genes were also shown to be expressed during pharyngeal arch patterning and affected by Dlx5/6 constitutive inactivation [36]. We thus wondered if Dlx5/6 transcription factors may act as regulators of the BMP signaling pathway in both vocal tract and auditory precursors.

We selected key genes, Bmper, Msx1 and Hand2, and analyzed their expression by in situ hybridization during early development at E10.5. Bmper (BMP-binding endothelial cell precursor-derived regulator) is a secreted protein present in migrating CNCC and the mesodermal core of PAs [36–39]. It was reported that Bmper-inactivated mice showed malformations of the thyroid cartilage [38]. Msx1 and Hand2 are downstream components of the BMP signaling cascade, they are expressed in CNCC and are regulated by Dlx5/6 in pharyngeal arches [36,40,41]. Msx1 and Hand2 deficient mice show malformations of the middle ear, the hyoid and thyroid cartilages [42,43].

All the genes analyzed were expressed in both otic vesicle and pharyngeal arches at E10.5 in control embryos (Fig 4). Bmper was expressed in the tympanic ring of the otic vesicle and in the mesodermal core of the PA1–2 (Fig 4A). Expression was also noticed in the distal part of PA1–2 and in PA3–6. In Dlx5/6 mutant embryos, expression in the distal PA1–2 and otic vesicles was reduced and undetectable in the PA1–2 mesodermal cores and posterior arches (PA3–6) (Fig 4B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Dlx5/6 transcription factors regulate BMP signaling in both pharyngeal arches and otic vesicle.

Whole mount in situ hybridization for Bmper, Msx1 and Hand2 in E10.5 control and mutant embryos (A–F). Bmper expression is observed in the mesodermal core and the distal part of PA1-2 and in the tympanic ring of controls (A). In mutants, Bmper expression is lost in the mesodermal cores and expression is reduced in the distal PAs and otic vesicle (B). In controls, Msx1 and Hand2 are expressed in all PAs and otic vesicles (arrowheads in C, E). In mutants, expression is undetectable in the otic vesicle and reduced or undetectable in PAs (D, F). All structures presenting undetectable or reduced gene expression in the mutant are marked with an asterisk (n = 4 each condition, control genotypes: Dlx5/6^flox/+ and Dlx5/6^flox/flox). Abbreviation: e, eye; nt, neural tube; ov, otic vesicle; PA, pharyngeal arch. Scale bar in F for A-F 200 µm.

https://doi.org/10.1371/journal.pone.0337426.g004

In the control, Msx1 was expressed in the distal cap of PA1–2, in posterior PAs, as well as the otic vesicle (Fig 4C). In our mutants, Msx1 expression was lost in PA2–6 and otic vesicles, and reduced in PA1 (Fig 4D). Hand2 was strongly expressed in all pharyngeal arches and expression was detected in the otic vesicle in controls (Fig 4E). The Dlx5/6 mutant embryos showed a drastic loss of expression in both compartments (Fig 4F). Our analysis indicates that Dlx5/6 genes directly regulate the BMP signaling pathway through Bmper early during the development of both vocal tract and auditory precursors.

A common Dlx5/6-BMP axis acts in Sox10 derivatives during auditory and vocal tract development

In this study, we show that Dlx5/6 transcription factors control the early patterning and innervation of pharyngeal arch and otic derivatives by regulating the BMP signaling pathway in both compartments.

The BMP signaling pathway is implicated in many developmental processes including cellular differentiation, growth and apoptosis [44]. It has been previously shown that Dlx genes and BMP signaling interact for proper craniofacial development [22,45]. In our study, we selected and analyzed actors of the BMP signaling pathway that were shown to be direct downstream targets of Dlx5 in the otic placode [35] and that we noticed being also expressed in both PAs and otic vesicles. We observed that the gene coding for the secreted protein Bmper (Cv2) was expressed in distal CNCC and within the mesodermal cores of PA1–2 in controls. In our mutants, Bmper expression was reduced in CNCC and not detectable in the mesodermal cores in which Dlx5/6 is not expressed, indicating indirect regulation of Bmper expression by Dlx5/6 in the latter compartment. We previously demonstrated that Dlx5/6 expression in CNCC is necessary to instruct the adjacent mesoderm for the proper differentiation of masticatory muscles [9,15], but the molecular actors involved in CNCC-mesoderm interaction remained elusive. We hypothesize that the secreted Bmper, expressed in both CNCC and mesodermal populations and regulated by Dlx5/6, may handle such a role. Bmper-inactivated mice show defects of laryngeal cartilages and hypoplastic skull vault, however the head muscular phenotype had not been investigated [38]. Further functional analyses in the mouse embryos are still required to test if Bmper could be a mediator between Dlx5/6 in CNCC and the adjacent mesoderm to regulate head muscle differentiation.

We have also shown that inactivation of Dlx5/6 severely impacts the expression of Hand2 and Msx1 in both otic vesicles and PAs. These genes are known to act downstream of the BMP signaling and it has been proposed that Dlx5/6 regulate their expression to direct lower jaw patterning [36,41]. Moreover, Chip-seq data have identified Msx1 as a direct target of Dlx5 in the inner ear precursor [35]. The data indicate that Hand2 and Msx1 are regulated directly or indirectly by Dlx5/6 during PA and otic vesicle development.

It has been shown that the Hedgehog (HH) signalling pathway, including its key transducer Gli3, controls larynx formation and function [12,46]. Notably, Shh and Gli3 mutants show patterning defects of CNC derivatives forming the laryngeal connective tissue, correlated with changes in the acoustic structure of vocalizations [12]. A study demonstrated that Gli3 and Hand2 transcription factors act in a synergistic manner in cephalic neural crest derivatives for proper skeletal and glossal development [47]. HH signaling and the BMP pathway regulated by Dlx5/6 may work together to control laryngeal development.

In contrast to Wnt1 that is an early and specific marker of neural crest derivatives, the Sox10 lineage also includes the otic vesicle [28]. In chick, a common enhancer element controls Sox10 expression in both the otic placode and CNCC [27], indicating that the two compartments are co-regulated by Sox10 activation during early development. Altogether, the data thus suggest that a common Dlx5/6-BMP genetic program acts within the Sox10 lineage during the morphogenesis of the vocal and auditory complex (Fig 5).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. A common Dlx5/6-BMP axis acts in Sox10 derivatives during the formation of both auditory and vocal tract systems.

(A) Schematic views of Sox10 and Dlx5/6 co-expression in the otic vesicle (ov) and pharyngeal arches in E10.5 mouse embryos (in blue), and role of Dlx5/6 in regulating BMP signaling in placode and CNCC (in blue) and in PA mesodermal core derivatives (in pink). (B) Summary of the phenotype resulting from Dlx5/6 conditional mutation affecting both acoustic receptor and effector organs of placode/CNCC (in blue) or mesodermal origins (in pink).

https://doi.org/10.1371/journal.pone.0337426.g005

Dlx5/6 pleiotropic role may have supported acoustic communication diversification

The Dlx gene family has an ancient origin in bilaterians and is present across all vertebrate clades [16]. Expression analyses in chick, xenopus and fish embryos revealed that Dlx5/6 expression in the pharyngeal arches and otic vesicle is conserved across vertebrate species [48–52]. Our data show that Dlx5/6 genes simultaneously regulate the formation of all structures composing the vocal tract and auditory systems (Fig 5), receptor and effector organs of acoustic communication. This fully corresponds to the concept of pleiotropy since a single locus affects distinct phenotypic traits [24,25].

Research has demonstrated that acoustic communication originated once in the land vertebrates [1,2] and that the larynx is the main site of vocal production within the vertebrate family [4]. The mammalian clade comprises diverse ecological niches, a diversity that is explained by the rapid adaptive radiation during “the age of mammals” and terrestrial changes like continental rearrangements [4,53]. Mammals show a wide variety of acoustic transmission varying from infra to ultrasounds, possibly originating from the adaptation to new environments and for predator avoidance [54]. The differences in tone and frequency can be explained by the morphological diversity of the larynx in mammals as well as by the adaptation of the receptor organ capable of hearing infra or ultrasounds [3,4]. It has been suggested that mammalian vocal characteristics and hearing sensitivity co-evolved in the forest mammals, following the sensory drive hypothesis [55]. One of the characteristics of the mammalian clade is the presence of ossicles of CNCC origin. Their morphology is dependent on environment and behavior making the auditory apparatus competent to pick up a wider variety of frequencies [54]. Beyond sound emission and reception, a study revealed that targeted inactivation of Dlx5/6 in GABAergic neurons results in increased vocalization and socialization behaviors in adult mice [56]. We thus propose that the pleiotropic role of Dlx5/6 genes might have been involved in the co-adaptation of vocal and auditory acoustic organs, but also in the regulation of cognitive communication capacities in mammals.

Our Dlx5/6 mutants display malformations that also affect traits of domestication [57]. It has been proposed that the “Neural Crest Domestication Syndrome” (NCDS) results from a reduction of CNCC-derived tissues of behavioral relevance and inducing morphological changes of the jaw, ear or larynx [57–59]. In domesticated mammalian species such as cats and dogs, pets interact with humans with acoustic emissions that are not recorded in the wild. The pleiotropic role of Dlx5/6 in development and cognition may have supported the morphological changes predicted by the NCDS hypothesis by adapting both the auditory and vocal systems and by inducing a domesticated behavior [56].

The origin of speech in humans has been approached through anatomical analyses of the brain, the face and the vocal tract and, more recently, through the comparison of genes involved in language acquisition [60,61]. It has been shown that genes associated with face and vocal tract anatomy went through particularly extensive methylation changes in modern humans after the split from Neanderthals and Denisovans [61]. It has been proposed that DLX5 had played a role in the evolution of the human’s linguistic capacity and skull globularization, and that DLX5/6 had contributed to human “self-domestication” [56,60]. Interestingly, the DLX5/6 locus shows methylation motifs that changes linearly with age in mouse and human tissues including the brain [56]. The concomitant modifications of craniofacial morphology, cognition, and language abilities may have been controlled by a common genetic network including DLX5/6 genes during human evolution.

Incidence for understanding the etiology of human syndromic forms

In humans, the Split Hand/Split Foot malformation 1 (SHFM1, OMIM #183600) is characterized by ectrodactyly caused by pathogenic variants affecting the DLX5/6 locus [62]. Dlx5/6 mutant mice have been investigated to understand the etiology of the limb phenotype in SHFM1 [63]. In patients, the limb malformations can be associated to other defects including cranial malformations, hearing loss and intellectual disabilities (OMIM #220600). It has been proposed that a minimal SHFM1 chromosomal region containing DLX5/6 enhancers would be compromised and related to associated phenotypes [62]. Moreover, the DLX5/6 intergenic regulatory sequence includes SNPs associated with either hyper-(Williams–Beuren syndrome) or hypo- (ASDs) sociability and vocalization syndromes [56]. The pleiotropic role of Dlx5/6 in vocal tract and otic development, and in socialization and vocalization abilities is thus relevant to better understand the hearing loss, craniofacial features and cognitive traits associated with the limb phenotype in SHFM1 patients.