RESEARCH ARTICLE

MAP3K1 function is essential for cytoarchitecture of the mouseorgan of Corti and survival of auditory hair cellsRizwan Yousaf1, Qinghang Meng2, Robert B. Hufnagel3, Ying Xia2, Chandrakala Puligilla4, Zubair M. Ahmed1 andSaima Riazuddin1,*

ABSTRACTMAP3K1 is a serine/threonine kinase that is activated by a diverse setof stimuli and exerts its effect through various downstream effectermolecules, including JNK, ERK1/2 and p38. In humans, mutantalleles of MAP3K1 are associated with 46,XY sex reversal. Untilrecently, the only phenotype observed inMap3k1tm1Yxia mutant micewas open eyelids at birth. Here, we report that homozygousMap3k1tm1Yxia mice have early-onset profound hearing lossaccompanied by the progressive degeneration of cochlear outerhair cells. In the mouse inner ear, MAP3K1 has punctate localizationat the apical surface of the supporting cells in close proximity to basalbodies. Although the cytoarchitecture, neuronal wiring and synapticjunctions in the organ of Corti are grossly preserved, Map3k1tm1Yxia

mutant mice have supernumerary functional outer hair cells (OHCs)and Deiters’ cells. Loss of MAP3K1 function resulted in thedownregulation of Fgfr3, Fgf8, Fgf10 and Atf3 expression in theinner ear. Fgfr3, Fgf8 and Fgf10 have a role in induction of the oticplacode or in otic epithelium development in mice, and their functionaldeficits cause defects in cochlear morphogenesis and hearing loss.Our studies suggest that MAP3K1 has an essential role in theregulation of these key cochlear morphogenesis genes. Collectively,our data highlight the crucial role of MAP3K1 in the development andfunction of the mouse inner ear and hearing.

KEY WORDS: Map3k1, Mekk1, Fgfr3, Fgf8, Fgf10, Hearing loss,Supernumerary outer hair cells, MAPK pathway, FGF signalingpathway

INTRODUCTIONMitogen-activated protein kinases (MAPKs) are responsible forregulating a wide array of cellular functions and processes. TheMAPK signaling cascade consists of three tiered phosphorylationsteps, starting with the phosphorylation of MAPK kinase kinases(MAP3Ks, MEK kinases or MKKKs) in response to a plethora ofstimuli, which in turn phosphorylate the MAPK kinases (MAP2Ks,MEK or MKKs) and then the MAPKs (Kyriakis and Avruch, 2001;Uhlik et al., 2004). MAP3K1, a member of the MAPK kinase

kinases family, plays a diverse cell-signaling function in variousbiological systems, including immune system development andfunction (Gallagher et al., 2007; Labuda et al., 2006), vasculatureremodeling (Li et al., 2005), tumor progression (Cuevas et al.,2006), cardiogenesis (Minamino et al., 2002), and injury repair(Deng et al., 2006).

MAP3K1 belongs to the serine/threonine kinase class that alsoparticipates in the regulation of the MAPK cascade (Cuevas et al.,2007; Hagemann and Blank, 2001; Uhlik et al., 2004). TheMAP3K1 protein contains an ubiquitin interaction motif (UIM), acaspase-3 cleavage site and a conserved kinase domain (Uhlik et al.,2004; Witowsky and Johnson, 2003). MAP3K1 is associated withthe plasma membrane and is tethered by α-actinin to actin stressfibers and by protein tyrosine kinase 2 (PTK2) to focal adhesions incells (Christerson et al., 2002; Cuevas et al., 2003). Following anapoptotic signal, caspase-3 cleaves MAP3K1 at phosphorylatedAsp874 (p.Asp874), resulting in the separation of the N-terminalUIM motif from the 91-kDa C-terminal kinase domain, releasing itfrom the cell membrane into the cytosol (Bonvin et al., 2002;Schlesinger et al., 2002). MAP3K1 is activated in response to anumber of different stimuli, such as cold, growth factors, mildhyperosmolarity, microtubule disruption, cell shape disturbance,pro-inflammatory cytokines and other physiological stresses(Sadoshima et al., 2002; Xia et al., 2000; Yujiri et al., 1998).Once activated, MAP3K1 exerts its effect through the JNK, ERK1/2and p38 MAPK pathways, as well as the transcription factors Junand NF-κB (Bonvin et al., 2002; Guan et al., 1998; Xia et al., 2000;Yujiri et al., 1998).

In humans, mutant alleles ofMAP3K1 are associated with 46,XYgonadal dysgenesis (Pearlman et al., 2010). These gain-of-functionalleles affect the downstream phosphorylation of p38 and ERK1/2,as well as binding of MAP3K1 with the cofactors RHOA andMAP3K4 (Loke and Ostrer, 2012). Additionally, in vitro studieshave suggested that these mutations alter the sex-determinationpathway by concomitantly upregulating β-catenin expression anddownregulating expression of the SOX9, SRY, FGF9 and FGFR2genes (Loke et al., 2014). However, mice carrying theMap3k1 loss-of-function allele display a minor testicular deficit in the developinggonad and have normal gross appearance besides the open-eyelidphenotype (Warr et al., 2011). MAP3K1-deficient mice display aneye open at birth (EOB) phenotype (Zhang et al., 2003) and haveimmune-system and wound-healing deficits, abnormal retinalvascularization, disintegration of retinal pigment epithelium, lossof photoreceptors, and retinal degeneration (Mongan et al., 2011).Additionally, cultured keratinocytes from these mutant mice displaya lack of actin stress fiber formation and deficient cell migration(Yujiri et al., 2000; Zhang et al., 2003).

Previous studies have demonstrated the role of MAPK-mediatedfibroblast growth factor (FGF) signaling in otic induction anddevelopment (Urness et al., 2010). Hearing depends on the preciseReceived 2 September 2015; Accepted 16 October 2015

1Department of Otorhinolaryngology Head & Neck Surgery, School of Medicine,University of Maryland, Baltimore, MD 21201, USA. 2Department of EnvironmentalHealth, University of Cincinnati, College of Medicine, Cincinnati, OH 45267, USA.3Divisions of Pediatric Ophthalmology and Human Genetics, Cincinnati Children’sHospital Medical Center, University of Cincinnati, Cincinnati, OH 45229, USA.4Department of Pathology & Laboratory Medicine, Medical University of SouthCarolina, Charleston, SC 29425, USA.

*Author for correspondence (sriazuddin@smail.umaryland.edu)

This is an Open Access article distributed under the terms of the Creative Commons AttributionLicense (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use,distribution and reproduction in any medium provided that the original work is properly attributed.

1543

© 2015. Published by The Company of Biologists Ltd | Disease Models & Mechanisms (2015) 8, 1543-1553 doi:10.1242/dmm.023077

Disea

seModels&Mechan

isms

organization of sensory hair cells and non-sensory supporting cellswithin the organ of Corti (OC). Any substantial alteration in thecellular number, alignment or patterning within the OC causeshearing loss, underlining the necessity of defined cytoarchitecture inthe OC for sound perception. Here, we report that mice lacking thekinase domain of MAP3K1 (Map3k1tm1Yxia) have an extra row ofOHCs and supporting cells in the OC and suffer from early-onsetprofound hearing loss. Our results demonstrate the indispensablerole of MAP3K1 in mouse inner-ear development and function.

RESULTSMAP3K1 is localized with basal bodies in supporting cellsTo comprehend the role of Map3k1 in inner-ear development andfunction, we used previously generated Map3k1tm1Yxia mice (Xiaet al., 2000; Zhang et al., 2003). In these mice, the exons encodingthe kinase domain of MAP3K1 have been replaced with thebacterial lacZ gene, resulting in the expression of a MAP3K1–β-galactosidase fusion protein (Fig. 1A). Immunolabeling of thewhole-mount preparation for the OC from these mice revealedpunctate expression of the MAP3K1–β-galactosidase fusion protein

at the apical surface of supporting cells (Fig. 1B). Interestingly, theMAP3K1–β-galactosidase fusion protein and basal-body markerpericentrin colocalize with a very exquisite pattern (Fig. 1C). TheMAP3K1–β-galactosidase fusion protein was expressed as twopuncta on either side of the pericentrin-labeled basal cell bodies(Fig. 1C, inset). Basal bodies are known to direct planar cell polarity(PCP) of sensory hair cells (Jones et al., 2008). Therefore, weinvestigated the orientation of the OHC bundles in Map3k1 mutantmice, and no significant deficit was observed (see Fig. S1).

Additionally, weak diffused cytoplasmic expression of Map3k1was also observed in the inner hair cells and outer hair cells(Fig. 1C), and in the marginal and intermediate cells of the striavascularis (see Fig. S2A). Moreover, β-galactosidase staining inadultMap3k1 heterozygous mice revealed expression in supportingcells of the lesser epithelial ridge, greater epithelial ridge, striavascularis, Reissner’s membrane and the spiral ganglion neurons(see Fig. S2B).

Map3k1tm1Yxia/tm1Yxia mice have supernumerary outer haircells and Deiters’ cellsUpon confocal imaging of the OC from Map3k1tm1Yxia

heterozygous and homozygous mice, we observed supernumeraryouter hair cells (OHCs) throughout development (Fig. 2). TheMap3k1tm1Yxia heterozygous mice have sparse one- to ten-cellstretches of an extra row of OHCs, whereas homozygousMap3k1tm1Yxia mice have a continuous extra row of OHCs in theapical, middle and basal cochlear turns (Fig. 2A). Thesesupernumerary OHCs also have correctly polarizedmechanosensitive hair bundles at their apical poles (Fig. 2B).Although no difference was observed in the number of inner haircells (IHCs) (Fig. 2C), a statistically significant increase in OHCswas observed throughout the cochlear duct in homozygous andheterozygous Map3k1tm1Yxia mice (Fig. 2C). However, incomparison with homozygous mutant mice, fewer extra OHCswere observed throughout the cochlear duct in heterozygousMap3k1tm1Yxia mice (Fig. 2C), suggesting a dose-dependent roleof MAP3K1-mediated signaling in the cytoarchitecture of themouse OC.

In higher vertebrates, each hair cell in the OC is enveloped bysupporting cells (Fig. 3A). In the mouse OC, the IHCs rest on theinner phalangeal cells, whereas each OHC rests upon a singleDeiters’ cell, and inner and outer pillar cells separate these two typesof sensory hair cells (Fig. 3A). To determine the effect of extra rowsof OHCs on the precise cytoarchitecture of the OC, weimmunolabeled the cochlear sections from postnal day 0 (P0)control and Map3k1tm1Yxia mutant mice with an anti-Prox1antibody, a marker for pillar and Deiters’ supporting cells. InMap3k1tm1Yxia mutant mice, the extra row of OHCs was alsosupported by an extra row of Deiters’ cells (Fig. 3B). However, theorganization of the inner and outer pillar cells, Claudius cells andthe tunnel of Corti was not affected in Map3k1tm1Yxia mutant mice(Fig. 3B,C). These findings suggest that MAP3K1 has a role in thedevelopment of controlled cytoarchitecture of mouse OC.

To assess the functional status of supernumerary hair cells, webriefly exposed (15 s) the OC explants from control andMap3k1tm1Yxia mutant mice to FM1-43, a styryl pyridinium dyethat enters the hair cells via partially open mechanotransductionchannels at rest (Gale et al., 2001; Meyers et al., 2003). Hair cellsfrom control and Map3k1tm1Yxia mice showed dye loading in allrows of sensory cells, including the extra row of OHCs (Fig. 3D).We also performed synaptophysin immunostaining to identifypostsynaptic endings of efferent neurons forming axodendritic

TRANSLATIONAL IMPACT

Clinical issueHearing loss, which can present during early life or as a late-onsetcondition, is one of the most common neurosensory disordersworldwide. Normal hearing depends on the precise organization ofsensory hair cells and non-sensory supporting cells within the organ ofCorti (OC) in the cochlea of the ear. Any significant alterations to cellnumber, alignment or patterning within the OC causes hearing loss,underlining the importance of a defined cytoarchitecture in the OC forsound perception. Thus, understanding the molecular signalingcascades that lead to inner-ear sensory-cell differentiation and functionis important for defining the molecular basis of hearing loss and devisingstrategies for hearing restoration. In this study, the putative role ofMAP3K1 (a serine/threonine kinase with a pivotal function in MAPKsignal transduction cascades) in inner-ear development and functionwas explored in mice.

ResultsTo determine whether MAP3K1 plays a part in hearing, the authorscharacterized homozygous Map3k1tm1Yxia kinase-null mutant mice. Inthe mouse inner ear, MAP3K1 is localized at the apical surface ofcochlear supporting cells, in close proximity to basal bodies. A hearingtest of themutant mice revealed early-onset profound hearing loss, alongwith progressive degeneration of outer hair cells (OHCs). The authorsshow that the cytoarchitecture, neuronal wiring and synaptic junctions inthe OC are unaffected; however, loss of MAP3K1 function results in anextra row of functional OHCs and Deiters’ cells (a type of cochlearsupporting cell). Loss of MAP3K1 function also results in downregulationof members of the fibroblast growth factor (FGF) signaling pathway:Fgfr3, Fgf8, Fgf10 and Atf3 expression in the inner ear. Previous studieshave shown that Fgfr3, Fgf8 and Fgf10 have a role in induction of theotic placode – from which the auditory system develops duringembryogenesis – or in otic epithelium development in mice.

Implications and future directionsFunctional deficits in the FGF signaling pathway are known to causedefects in cochlear morphogenesis and hearing loss in mice. This studyprovides evidence that MAP3K1 has an essential role in the regulationof the FGF signaling pathway during the development, function andsurvival of inner-ear hair cells. Homozygous Map3k1tm1Yxia micerepresent another model for the investigation of signaling pathwaysinvolved in hearing loss. In addition, elucidation of the central role of theMAP3K1 kinase protein could have implications for the controlledregeneration of inner-ear sensory cells for hearing restoration.

1544

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1543-1553 doi:10.1242/dmm.023077

Disea

seModels&Mechan

isms

synapses with the dendrites of spiral ganglion neurons in P12control and mutant mice (Fig. 3E). Interestingly, the synaptophysinlabeling was also observed at the base of an extra row of OHCs(Fig. 3E, arrowhead). Furthermore, no obvious difference in theneuronal wiring of the OC in the control andMap3k1tm1Yxia mutantmice was observed at P12 (Fig. 3F). However, the width of the areaof neurons around the OHCs was wider in Map3k1tm1Yxia mutantmice, likely due to the wiring of the extra row of OHCs (Fig. 3F).Collectively, we observed that Map3k1tmYxia mutant mice have acompletely developed, structurally supported, polarized andfunctional extra row of innervated OHCs and Deiters’ cells.

Map3k1tm1Yxia mutant mice have early-onset hearing lossNext, to assess the hearing function of Map3k1tm1Yxia mutant mice,we measured auditory brainstem responses (ABRs) and distortionproduct otoacoustic emissions (DPOAEs). Although the wild-typecontrol and Map3k1tm1Yxia heterozygous mice had comparablehearing thresholds across all the tested frequencies, significanthearing loss was observed in the homozygous Map3k1tm1Yxia miceat P16 and P30 (Fig. 4A,B). Similarly, the DPOAEs of wild-typeand Map3k1tm1Yxia heterozygous mice were comparable, whereasthe homozygous Map3k1tm1Yxia mice had no detectable DPOAEthresholds (Fig. 4C). Taken together with ABRs, these results

suggest that hearing loss in Map3k1tm1Yxia mutant mice is likely toresult from peripheral (cochlear) deficiencies. In contrast,Map3k1tm1Yxia mutant mice did not exhibit any observablevestibular dysfunction phenotypes, such as hyperactivity, head-tossing or circling behavior.

Map3k1tm1Yxia mutant mice display rapid degeneration ofOHCs and spiral ganglion neuronsNext, we examined the morphology of the cochlear epithelium atvarious developmental stages to determine the underlying cause ofhearing loss observed in the Map3k1tm1Yxia mutant mice. Confocalimaging of myosin-VIIa-labeled OC of Map3k1tm1Yxia micerevealed normal development of IHCs and OHCs, which wereindistinguishable from those of wild-type and heterozygous miceuntil P12 (Fig. 5). By P14, Map3k1tm1Yxia homozygous micedisplayed obvious degeneration of OHCs in the basal turn, and byP16, no intact OHCs were observed in the basal turn (Fig. 5L).Furthermore, by P16, a varying degree of OHC degeneration wasobserved along the length of the cochlea (Fig. 5J-L). However, byP30, almost all of the OHCs in the basal and middle turn of thecochlea were degenerated, and many OHCs in the apical coil werealso degenerated (Fig. 5M-O). In contrast, the IHCs appearedlargely intact even at P75 (Fig. 5 and data not shown). The loss of

Fig. 1. Localization of MAP3K1 in the innerear. (A) MAP3K1 has two domains, a regulatorydomain and a kinase domain, whereas, inMap3k1 mutant mice, the kinase domain hasbeen replaced with a β-galactosidase reportercassette, resulting in aMAP3K1–β-galactosidasefusion protein. (B) Confocal imaging of the whole-mount preparation of organ of Corti (OC) fromP16 mice labeled with the anti-β-galactosidaseantibody (green) and phalloidin (red). Asanticipated, no β-galactosidase labeling wasobserved in control (Map3k1+/+) mice, whereas,in the Map3k1tm1Yxia mutant mice, a distinctexpression of the MAP3K1–β-galactosidasefusion protein was observed at the apical surfaceof supporting cells of the OC. Scale bar: 10 μm.(C) In sections of P0 mice, the MAP3K1–β-galactosidase fusion protein (green) colocalizeswith pericentrin (red), a marker for basal cellbodies. In addition, weak diffused cytoplasmicexpression of fusion protein was also observed ininner and outer hair cells. Scale bar: 10 μm.Insets show higher magnification ofβ-galactosidase and pericentrin localization.Scale bar: 1 μm.

1545

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1543-1553 doi:10.1242/dmm.023077

Disea

seModels&Mechan

isms

OHCs in Map3k1tm1Yxia mutant mice was followed by progressivedegeneration of the spiral ganglion, which was more pronounced inthe basal region at P90 (Fig. 6). Thus, the elevated hearingthresholds observed in MAP3K1-deficient mice are likely to becaused by the rapid degeneration of OHCs. These results suggestthat MAP3K1 function is essential for the maintenance of OHCs inthe mouse auditory system.

Genes of the FGF signaling pathway are downregulated inMap3k1 mutant miceWe reasoned that the extra row of OHCs and Deiters’ cells anddegeneration of hair cells in Map3k1 mutant mice is likely to becaused by impaired intracellular signaling during the developmentand maturation of mouse OC. Therefore, we examined theexpression of various genes associated with the Map3k1-mediated

Fig. 2. Map3k1tm1Yxia mutant mice havesupernumerary outer hair cells (OHCs).(A) Whole-mount preparation of the OC fromP12 mice labeled with myosin VIIa (green)and phalloidin (red). In contrast to three rowsof OHCs observed in wild-type mice,Map3k1tm1Yxia mutant mice have an extrarow of OHCs along the length of the cochlea,whereas sparse patches of a fourth row ofOHCs was observed in Map3k1tm1Yxia

heterozygous mice. Scale bar: 10 μm.(B) Scanning electron micrographs of P14mice revealed characteristic polarized ‘V’-shaped stereocilia bundles at the apicalsurface of supernumerary OHCs present inMap3k1tm1Yxia heterozygous andhomozygous mutant mice. Scale bar:10 μm. (C) Quantitation of inner hair cells(IHCs) and OHCs in control andMap3k1tm1Yxia mice at P12. Forquantification purposes, organ of Corti (OC)were isolated from four wild-type andMap3k1tm1Yxia mutant mice each and haircells were counted in the apical middle andbasal coil regions. No significant differencewas observed in the IHC number.Statistically significant (*P<0.05, **P<0.01)increases in the OHC number wereobserved in the Map3k1tm1Yxia homozygousmutant and heterozygous mice, with agradient from apex to base (mean±s.e.m.).

1546

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1543-1553 doi:10.1242/dmm.023077

Disea

seModels&Mechan

isms

Fig. 3. Cytoarchitecture of the organ of Corti (OC) is preserved inMap3k1tm1Yxiamutant mice. (A) Schematic representation of the developing mouse OC atP0 and P10. OHC, outer hair cell; IHC, inner hair cell; DC, Deiters’ cell; OPC, outer pillar cell; IPC, inner pillar cell; HC, Hensen’s cells; IPhC, inner phalangealsupporting cells. (B) Cross-section of Map3k1tm1Yxia mutant and wild-type control mice at P0, immunostained with Prox1 (green) and myosin VIIa (red). Thearrows point to the pillar cells, whereas Deiters’ cells are marked by the arrowheads.Map3k1tm1Yxia mutant mice have an extra row of OHCs accompanied by anextra row of Deiters’ cells. Scale bar: 10 μm. (C) Immunostaining with the anti-CD44 antibody, a marker for OPCs, including Claudius cells, revealed an intactgross cytoarchitecture of the OC in Map3k1tm1Yxia mutant mice at P0 and P10. Scale bar: 10 μm. (D) No apparent difference in the FM1-43 dye uptake wasobserved among control andMap3k1tm1Yxiamutant explants. FM1-43 dye was also taken up by the supernumerary OHCs present inMap3k1tm1Yxiamutant mice.Scale bar: 10 μm. (E) Supernumerary OHCs in Map3k1tm1Yxia mutant mice are innervated and have synaptic junctions (arrowheads), immunolabeled withsynaptophysin (green). Scale bar: 10 μm. (F)Map3k1tm1Yxiamutant mice have grossly intact neuronal wiring. Neurofilament (NF-200) protein immunostaining ofwild-type control and Map3k1tm1Yxia mutant mice revealed grossly intact neuronal wiring of an extra row of OHCs. Scale bar: 10 μm.

1547

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1543-1553 doi:10.1242/dmm.023077

Disea

seModels&Mechan

isms

signaling pathway (Fig. 7 and see Fig. S3) and transcriptionaltargets of FGF signaling, as well as the development-related genesin Map3k1tm1Yxia mutant mice at P10 (before the onset of hearingand OHC degeneration). As expected, the heterozygous mice hadreduced expression of Map3k1, whereas the homozygous mice hadno expression (Fig. 7). Among the 42 genes analyzed, we observedsignificant (P<0.001) downregulation of Atf3, Fgfr3, Fgf8 andFgf10 in the inner ear of Map3k1 mutant mice (Fig. 7). Previousstudies have shown that Fgfr3, Fgf8 and Fgf10 have a role in the oticplacode induction or in otic epithelium development in mice, andtheir functional deficits cause defects in cochlear morphogenesisand hearing loss (Hayashi et al., 2007; Mansour et al., 2009; Pannieret al., 2009). Our results suggest that MAP3K1 has an essential rolein the regulation of these key cochlear morphogenesis genes duringdevelopment.

DISCUSSIONMAP3K1 is involved in the cellular response to a wide array ofstimuli, including growth factors, pro-inflammatory cytokines, cellshape disturbance, microtubule disruption and a variety of cell stresssignals (Cardone et al., 1997; Deak et al., 1998; Nakagami et al.,2001; Widmann et al., 1998; Xia et al., 2000; Yujiri et al., 1998),which translate into activation of the downstream JNK-, p38- andERK-mediated MAPK pathways (Schlesinger et al., 1998; Xu et al.,1996). The full-length MAP3K1 protein is tethered to insolublestructures such as membranes or the cytoskeleton by the actin-bundling protein α-actinin (Christerson et al., 1999), and has alsobeen found to be associated with actin fibers entering focaladhesions, so is implicated in controlling their turnover (Cuevaset al., 2003). MAP3K1 interacts directly with the signaling

molecule MEK1 (Karandikar et al., 2000), which, in the mouseoocyte, has been associated with the control of microtubuleorganization and colocalizes with the centrosomal proteinγ-tubulin (Yu et al., 2007). Interestingly, in Map3k1tm1Yxia mice,the MAP3K1–β-galactosidase fusion protein in the supporting cellsof the OC also colocalized with the centrosomal protein and,therefore, might also participate in microtubule organization in theinner ear non-sensory cells and in PCP (Jones et al., 2008). There aremany examples of PCP proteins, such as Vangl2, that are notexpressed in the inner ear sensory hair cells, but still severely affectthe stereocilia bundle orientation (Montcouquiol et al., 2006).However, in theMap3k1tm1Yxia mutant mice we did not observe anysignificant stereocilia bundle orientation deficit, which might reflectthat Map3k1 has either no direct role in PCP, or that there isfunctional redundancy with other family members.

In a parallel study, a splice-site mutation in the Map3k1 gene ingoya mutant mice was identified as part of an ENU mutagenesisscreening program (Parker et al., 2015). goya mice also exhibit aprogressive hearing-loss phenotype along with the supernumeraryOHCs. Interestingly, when maintained on the same geneticbackground (C3H), the goya and Map3k1tm1Yxia mutant mice werefound to have a similar progressive hearing-loss phenotype andwere profoundly deaf by 9 weeks (Parker et al., 2015). Modulationof the phenotype of a given allele by the genetic background of aninbred strain is a well-documented phenomenon (Doetschman,2009; Montagutelli, 2000). Intriguingly, in addition to the increasedhearing loss that we observed, a study involving Map3k1 mice onthe C57BL/6J background has also shown a drastic decrease in thenumber of animals surviving to maturity (Warr et al., 2011), whichfurther supports the notion of a Map3k1 modifier gene.

Fig. 4.Map3k1tm1Yxia mutant mice have elevatedhearing thresholds. Hearing thresholds of 3 wild-type, 7 heterozygous and 8 Map3k1tm1Yxia

homozygous mice were evaluated at P16, whereas8 wild-type, 11 heterozygous and 7 Map3k1tm1Yxia

homozygous micewere tested at P30. (A) Averagedauditory brainstem responses (ABR) thresholds ofwild-type, and heterozygous and homozygousMap3k1tm1Yxia mice at P16 and P30 in response toclick stimulus. The Map3k1tm1Yxia mutant miceshowed significantly (***P<0.001) elevatedthresholds compared with heterozygous and wild-type mice at both ages (mean±s.e.m.).(B) Averaged ABR thresholds of wild-type (whitediamonds), heterozygous (gray squares) andhomozygous Map3k1tm1Yxia mutant (blacktriangles) mice at P16 (solid lines) and P30 (dashedlines), in response to 8-kHz, 16-kHz and 32-kHztone-bursts. At both developmental stages,Map3k1tm1Yxia mutant mice showed significantly(***P<0.001) elevated thresholds compared withthe wild-type control and heterozygous mice at allfrequencies tested (mean±s.e.m.). (C) Distortionproduct otoacoustic emissions (DPOAEs) ofMap3k1tm1Yxia mutant (black triangles),heterozygous (gray squares) and wild-type control(white diamonds) mice at P30, represented as afunction of f2 stimulus frequencies. Map3k1tm1Yxia

mutant mice showed no responses, with valuesclose to the noise floor, indicating that the residualOHCs were non-functional. SPL, sound pressurelevel.

1548

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1543-1553 doi:10.1242/dmm.023077

Disea

seModels&Mechan

isms

Although the expression of MAP3K1 was observed in thesupporting cells of the OC, Map3k1tm1Yxia mutant mice exhibitedfour rows of OHCs. During development, extra OHCs can be foundin wild-type mice, but they are generally restricted in the apical turnsof the cochlea, and their frequency can vary depending on differentinbred mouse strains (Lim and Anniko, 1985), whereas, inMap3k1tm1Yxia mutant mice, the fourth row of OHCs was foundthroughout development and along the length of the cochlea. Onepossibility to explain such a phenotype includes the lack ofapoptosis of a population of cells in the prosensory domainthat would otherwise have undergone degeneration duringdevelopment. Normally, full-length MAP3K1 acts as an anti-apoptotic protein but, when cleaved by caspase 3, becomespro-apoptotic (Schlesinger et al., 2002). Relocalization of theMAP3K1 C-terminal 91-kDa fragment containing the kinasedomain is necessary for its pro-apoptotic function; however, in

the case ofMap3k1tm1Yxia mice, the kinase domain is replaced withthe lacZ domain, rendering it incapable of inducing apoptosis in thesupporting cells of the developing OC.

Alternatively, the supernumerary OHCs in Map3k1tm1Yxia couldstem from the deficit in the control of the prosensory domain sizeduring embryogenesis. The sensory epithelial cell differentiationinitiates with specification of the prosensory domain in the otocyst.Many signaling molecules, including the sonic hedgehog, Notchand Wnt pathway genes, act in concert to form a highly specializedpatterned sensory epithelium. Canonical Wnt (Wnt/β-catenin)signaling is also essential for the specification of the otic placodebecause either conditional deletion or activation of β-catenin inPax2-positive ectodermal cells or Foxg1-positive placodal cellsresults in a substantial reduction or expansion of the size of the oticplacode, respectively (Freyer and Morrow, 2010; Ohyama et al.,2006). MAP3K1, via direct interaction with Axin1, is known to

Fig. 5. Outer hair cells (OHCs) in Map3k1tm1Yxia mutantmice degenerate as early as P14. Maximum intensityprojections of confocal Z-stacks of whole-mount cochleaelabeled with the anti-myosin-VIIa antibody (green) andphalloidin (red) are shown. (A-C) Representative imagesfrom the apical, middle and basal turns of the organ of Corti(OC) of a wild-type control mouse at P30. (D-O) Images ofthe OC from the three turns of the cochlea of Map3k1tm1Yxia

mutant mice at P12 (D-F), P14 (G-I), P16 (J-L) and P30(M-O). The hair cells appear to have normal developmentand morphology at P12 in the apical and middle cochlearturns in Map3k1tm1Yxia mutant mice. Initial signs of OHCdegeneration are evident in the basal turn (F). At P14,obvious degeneration of OHCs was observed inMap3k1tm1Yxiamutant mice. (K,L) SevereOHC degenerationcan be observed by P16 in themiddle (K) and basal (L) turns.The OHC loss progresses rapidly and, by P30, severedegeneration is evident in all three cochlear turns. Incontrast, inner hair cells (IHCs) remained intact along thelength of the cochlea. Scale bar: 10 μm.

1549

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1543-1553 doi:10.1242/dmm.023077

Disea

seModels&Mechan

isms

modulate Wnt/β-catenin signaling pathway activity (Sue Ng et al.,2010). Previously, fibroblasts from Map3k1tm1Yxia mutant micedisplayed a tenfold increase in transcription factor Lef/Tcf reporteractivity in response to Wnt3a expression compared to an onlythreefold increase observed in wild-type cells (Jin et al., 2013).Because Wnt expression regulates proliferation in the earlyprosensory domain and hair cell differentiation in the later stage,the level of Wnt/β-catenin signaling can affect hair cell formation ina dose-dependent manner (Jacques et al., 2012). It is plausible that,during development, Wnt/β-catenin signaling might have resultedin relative overactivation of downstream transcription factors,leading to the expansion of the prosensory domain and excessivedifferentiation of OHCs inMap3k1tm1Yxiamutant mice. Future workusing cell-specific markers and cochlear tissue from variousembryonic developmental stages will provide knowledgeregarding the role of MAP3K1 in sensory-cell specification anddifferentiation.Map3k1tm1Yxia mutant mice exhibit downregulation of at least

four genes, which are known for their role in inner-ear developmentand function. ATF3, a transcription factor that is induced inresponse to a number of stress stimuli (Hai et al., 1999; Kyriakiset al., 1994; Liang et al., 1996), is significantly downregulated inMap3k1tm1Yxiamutant mice. ATF3 is also implicated in the survival,

repair and neurite outgrowth in association with heat shock protein27, Akt and Jun (Nakagomi et al., 2003; Pearson et al., 2003). Also,Fgfr3 mutant cochlea show neuronal wiring pattern disruption(Puligilla et al., 2007). However, inMap3k1tm1Yxia mutant mice, wedid not observe any gross deficit in the neuronal wiring.Furthermore, the DPOAE data are suggestive of a functionaldeficit of hair cells in Map3k1tm1Yxia mutant mice. These resultssuggest that downregulation of Atf3 is unlikely to be the reason forthe supernumerary OHCs and Deiters’ cells, and hearing deficits,observed in Map3k1tm1Yxia mutant mice.

Before the onset of hair-cell degeneration at P10, we foundsignificant downregulation of Fgfr3, Fgf8 and Fgf10 expression intheMap3k1tm1Yxia mutant mouse OC cells. Dysfunction of Fgfr3 inmice, either due to loss (Hayashi et al., 2007) or gain (Mansouret al., 2009; Pannier et al., 2009) of function, results in hearing loss.Interestingly, both Fgfr3 mutant alleles have supernumerary OHCs(Hayashi et al., 2007; Mansour et al., 2009), as is observed inMap3k1tm1Yxia mutant mice. This implies that, in inner-eardevelopment, Map3k1 and Fgfr3 might participate in the samesignaling cascade to control the precise cytoarchitecture of the OC.However, in contrast to Fgfr3mutant mice, which have either loss ofpillar or Deiters’ supporting cells in the OC (Hayashi et al., 2007;Mansour et al., 2009; Pannier et al., 2009), Map3k1tm1Yxia mutant

Fig. 6. Map3k1tm1Yxia mice at P90displayed the degeneration of spiralganglions in the basal turn. Cochlearsections from control andMap3k1tm1Yxia

mutant mice were stained withhematoxylin and eosin. Scale bar:20 μm.

Fig. 7. Fgfr3, Fgf8, Fgf10 and Atf3 genes are downregulated in the organ of Corti (OC) of Map3k1tm1Yxia mutant mice. Semi-quantification expressionanalysis of various developmental and MAP3K1-mediated signaling pathway genes normalized against Gapdh and actin (endogenous controls) at P10. Asanticipated, a dose-dependent, statistically significant reduction in the expression of Map3k1 was observed in heterozygous and homozygous mutant mice.Furthermore, the expression of Fgfr3, Fgf8, Fgf10 and Atf3 was also significantly abolished (Student’s t-test, ***P<0.001), suggesting a role of Map3k1 inregulating their expression in the inner ear (mean±s.e.m.). **P<0.01; *P<0.05.

1550

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1543-1553 doi:10.1242/dmm.023077

Disea

seModels&Mechan

isms

mice do not exhibit any obvious deficit in the differentiation of pillarcells; rather, they have an extra row of Deiters’ cells. Thephenotypes of overproduction of OHCs and loss of pillar cellsreported for the gain-of-function allele of Fgfr3 phenotype isremarkably similar to Sprouty2, an antagonist of FGF signaling,mutant mice (Shim et al., 2005). Intriguingly, the hearing-lossphenotype in Sprouty2 mutant mice can be partially rescued bygenetically reducing the Fgf8 expression level (Shim et al., 2005).Furthermore, reduction of Fgf10 expression reverted the fate-switched supporting cells back and restored hearing in Fgfr3mutantmice (Mansour et al., 2013). Besides Fgfr3, Fgf8 and Fgf10 are alsodownregulated inMap3k1tm1Yxia mutant mice, which could accountfor the grossly intact supporting cells in these mutant mice ascompared to Fgfr3 mutant mice.Our studies further highlight the complexity of the signaling

pathway(s) required for formation of the precise cytoarchitecture ofmouse OC and the maintenance of OHCs. We also found thatMAP3K1 function was necessary for regulation of the FGF-mediated pathway in the auditory system and hearing in mice.

MATERIALS AND METHODSMap3k1 mutant miceMap3k1tm1Yxia mice were generated as described previously (Xia et al.,2000; Zhang et al., 2003). In summary, the targeting vector replaced theentire kinase domain in the Map3k1 locus, resulting in the formation of aMAP3K1–β-galactosidase fusion protein. These targeted embryonic stem(ES) cells (Map3k1+/tm1Yxia) were injected into mouse blastocysts, and theresulting chimeras were crossed with C57BL/6 mice to obtain mice withgermline transmission of the Map3k1tm1Yxia allele. All experiments wereapproved by the Animal Care and Use Committees at the University ofMaryland, School of Medicine in accordance with the National Institutes ofHealth (NIH) Guide for the Care and Use of Laboratory Animals.

ABR and DPOAE measurementsHearing function was evaluated by ABR analyses at two different timepoints (P16 and P30) in mice of all the three genotypes, i.e. wild-type,heterozygous and Map3k1tm1Yxia homozygous mice. Mice wereanesthetized with intraperitoneal injections of Avertin (0.4-0.75 mg/gbody weight, Sigma-Aldrich, St Louis, MO). All recordings wereperformed in a sound-attenuated chamber using an auditory-evokedpotential diagnostic system RZ6 (Tucker-Davis Technologies Inc.,Alachua, FL) as previously described (Nayak et al., 2013). Experimentsrepresent the mean and standard error of mean (s.e.m.) of three or moreanimals. Significance was analyzed using Student’s t-test.

DPOAE recordings were performed at P30 with an acoustic probe (ER-10C, Etymotic Research, Elk Grove Village, IL) using a DP2000 DPOAEmeasurement system version 3.0 (Starkey Laboratory, Eden Prairie, MN).Two primary tones, with a frequency ratio of f2/f1=1.2, where f1 representsthe first tone and f2 represents the second, were presented at intensity levelsL1=65 dB sound pressure level (SPL) and L2=55 dB SPL. Also, f2 wasvaried in one-eighth octave steps from 8 to 16 kHz. DP grams comprised2f1–f2 DPOAE amplitudes as a function of f2.

Confocal imagingThe temporal bones from the control and mutant mice were isolated andfixed for 1 h at room temperature or overnight at 4°C in 4%paraformaldehyde (Electron Microscopy Sciences, Hatfield, PA),followed by three washes with phosphate buffered saline (PBS). Finelydissected cochlear coils were permeabilized and blocked in 5% BSA, 2%normal goat serum and 0.1% Triton X-100 in PBS for 1 h. The tissuesamples were washed and probed with primary antibody overnight at 4°C.We used the following primary antibodies in our studies: β-galactosidase(1:1000 dilution; MP Biomedicals, Solon, OH), myosin VIIa (1:200;Proteus BioSciences, Ramona, CA), synaptophysin (1:200; Abcam,Cambridge, MA), NF-200 (1:200; Sigma-Aldrich), Prox1 (1:100;Millipore, Billerica, MA) and Pericentrin (1:200; Millipore, Billerica,

MA). After three washes with PBS, samples were probed with afluorescently labeled Alexa-Fluor-488 or -546 secondary antibody (1:500;Life Technologies, Grand Island, NY) for 1 h at room temperature.Rhodamine-phalloidin or Alexa-Fluor-647-conjugated phalloidin (1:250;Life Technologies, Grand Island, NY) was used to label actin. Samples weremounted using ProLongGold (Life Technologies, Grand Island, NY) andviewed under an LSM 700 confocal microscope (Zeiss Microimaging Inc.,Thornwood, NY) using a ×63, 1.4 N.A. oil-immersion lens.

Scanning electron microscopy (SEM)For SEM studies, the inner ears were isolated at P14 and fixed for 1.5 h in afixative containing 2.5% glutaraldehyde and 2 mM CaCl2 in 0.1 M sodiumcacodylate buffer, and later washed three times with 0.1 M sodiumcacodylate buffer. The inner ears were post-fixed in 1% osmium tetroxidein 0.1 M sodium cacodylate buffer for 1 h at room temperature. Afterwashing three times with PBS buffer, inner ears were decalcified byincubating in 0.25 M EDTA for 1-2 days at 4°C. The samples were thenfinely dissected to expose the sensory epithelium and to remove the tectorialmembrane and the stria vascularis. The cochlear tissues were thendehydrated in gradient alcohol changes, critical point dried, sputter coatedwith platinum, and imaged on a field-emission SEM.

Cochlear explants and FM1-43 dye uptakeOC explants from P3 wild-type andMap3k1tm1Yxia mutant mice were finelydissected and cultured on collagen-coated glass-bottom Petri dishes(MatTek Corporation, Ashland, MA) in DMEM medium supplementedwith 10% fetal bovine serum (Life Technologies, Grand Island, NY) and10 mg/ml ampicillin (Millipore, Billerica, MA) at 37°C and 5% CO2. Theexplants were kept in vitro for 2 days to allow for complete adhesion to thedish. To test the FM1-43 dye {N-(3-triethylammoniumpropyl)-4-[4-(dibutylamino)-styryl]pyridinium dibromide} (Life Technologies, GrandIsland, NY) uptake by the hair cells, the culture was exposed to 3.0 µMFM1-43 in Hank’s Balanced Salt solution (HBSS) for 15 s and then quicklywashed three times with HBSS. The culture was mounted with Fluoro-Gel(Electron Microscopy Sciences, Crofton, MD) and immediately imagedunder an LSM 700 confocal microscope (Zeiss Microimaging Inc.).

CryosectionsHemisected heads of P0 mice or inner ears of P10 and older mice werecollected and fixed in 4% paraformaldehyde (PFA) at 4°C overnight. Forolder mice, the temporal bones were decalcified in 0.25 M EDTA for 1 to2 days at 4°C. The inner ears were then equilibrated with 30% sucrose inPBS overnight at 4°C, embedded in OCT, and immediately frozen byplacing the block on an ethanol/dry-ice mix. The frozen tissue blocks weresectioned with a cryostat at 14 μm thickness.

Reverse transcriptase PCR (RT-PCR) and real-time PCRTotal RNAwas isolated from fine dissected OC isolated from P10 inner eartissue of wild-type, heterozygous and Map3k1tm1Yxia homozygous mice(five mice each) using the RiboPure RNA isolation kit (Life Technologies,Grand Island, NY) and cDNA was prepared using an oligo-dT primer andSMARTScribe Reverse Transcriptase enzymes (Clontech, Mountain View,CA). To determine the differential expression of various genes, SYBR-Green-based real-time primers (available upon request) were designed usingIntegrated DNA Technologies online PrimeTime qPCR assay design tool(http://www.idtdna.com/Scitools/Applications/RealTimePCR/). The real-time PCR assays were performed in triplicate using an ABI StepOnePlusReal-Time thermalcycler (Life Technologies, Grand Island, NY). CT valueswere normalized using Gapdh and actin as an endogenous control, and foldchanges of different genes were calculated using SABiosciences onlinesoftware (http://pcrdataanalysis.sabiosciences.com/pcr/arrayanalysis.php).The genes with a twofold change and with a P-value less than 0.05 basedon a Student’s t-test analysis were considered significant.

X-gal stainingThe inner ears were isolated from wild-type and Map3k1tm1Yxia

heterozygous mice at P30 and were fixed for 10 min at room temperature

1551

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1543-1553 doi:10.1242/dmm.023077

Disea

seModels&Mechan

isms

in LacZ fixative (1% formaldehyde, 0.2% glutaraldehyde and 0.02% NP-40in PBS). Then, the samples were washed twice with PBS containing 0.02%NP-40 and 2 mM MgCl2, 10 min each, followed by incubation at 37°Covernight in a staining solution [1 mg/ml X-Gal, 5 mM K3Fe(CN)6, 5 mMK4Fe(CN)6 and 0.1 MMgCl2]. Following staining, samples were washed inPBS, decalcified and cryosectioned as described above. The sections werecounterstained with hematoxylin and eosin and imaged using a ×40 oilimmersion lens on a Zeiss Axioplan Apotome-equipped microscope.

AcknowledgementsWe thank Drs A. P. Giese and E. M. Richard for their technical assistance,discussion of the results and review of the manuscript.

Competing interestsThe authors declare no competing or financial interests.

Author contributionsR.Y., R.B.H., Z.M.A. and S.R. designed the research. R.Y., Y.X., Z.M.A. and S.R.designed and/or contributed new reagents and performed research. R.Y., Q.M.,R.B.H., C.P., Y.X., Z.M.A. and S.R. analyzed data. All authors drafted or revised themanuscript, and all authors approved the final version.

FundingThis study was supported by National Institute on Deafness and OtherCommunication Disorders (NIDCD/NIH) research grants R01DC011803 andR01DC011748 to S.R. and R01DC012564 to Z.M.A. This work was also supportedby the Action on Hearing Loss grant to S.R.

Supplementary informationSupplementary information available online athttp://dmm.biologists.org/lookup/suppl/doi:10.1242/dmm.023077/-/DC1

ReferencesBonvin, C., Guillon, A., van Bemmelen, M. X., Gerwins, P., Johnson, G. L. andWidmann, C. (2002). Role of the amino-terminal domains of MEKKs in theactivation of NF kappa B and MAPK pathways and in the regulation of cellproliferation and apoptosis. Cell. Signal. 14, 123-131.

Cardone, M. H., Salvesen, G. S., Widmann, C., Johnson, G. and Frisch, S. M.(1997). The regulation of anoikis: MEKK-1 activation requires cleavage bycaspases. Cell 90, 315-323.

Christerson, L. B., Vanderbilt, C. A. andCobb,M. H. (1999). MEKK1 interacts withalpha-actinin and localizes to stress fibers and focal adhesions. Cell Motil.Cytoskeleton 43, 186-198.

Christerson, L. B., Gallagher, E., Vanderbilt, C. A., Whitehurst, A. W., Wells, C.,Kazempour, R., Sternweis, P. C. and Cobb, M. H. (2002). p115 Rho GTPaseactivating protein interacts with MEKK1. J. Cell. Physiol. 192, 200-208.

Cuevas, B. D., Abell, A. N., Witowsky, J. A., Yujiri, T., Johnson, N. L., Kesavan,K., Ware, M., Jones, P. L., Weed, S. A., DeBiasi, R. L. et al. (2003). MEKK1regulates calpain-dependent proteolysis of focal adhesion proteins for rear-enddetachment of migrating fibroblasts. EMBO J. 22, 3346-3355.

Cuevas, B. D., Winter-Vann, A. M., Johnson, N. L. and Johnson, G. L. (2006).MEKK1 controls matrix degradation and tumor cell dissemination duringmetastasis of polyoma middle-T driven mammary cancer. Oncogene 25,4998-5010.

Cuevas, B. D., Abell, A. N. and Johnson, G. L. (2007). Role of mitogen-activatedprotein kinase kinase kinases in signal integration. Oncogene 26, 3159-3171.

Deak, J. C., Cross, J. V., Lewis, M., Qian, Y., Parrott, L. A., Distelhorst, C. W. andTempleton, D. J. (1998). Fas-induced proteolytic activation and intracellularredistribution of the stress-signaling kinase MEKK1. Proc. Natl. Acad. Sci. USA95, 5595-5600.

Deng, M., Chen, W.-L., Takatori, A., Peng, Z., Zhang, L., Mongan, M.,Parthasarathy, R., Sartor, M., Miller, M., Yang, J. et al. (2006). A role for themitogen-activated protein kinase kinase kinase 1 in epithelial wound healing.Mol.Biol. Cell 17, 3446-3455.

Doetschman, T. (2009). Influence of genetic background on genetically engineeredmouse phenotypes. Methods Mol. Biol. 530, 423-433.

Freyer, L. and Morrow, B. E. (2010). Canonical Wnt signaling modulates Tbx1,Eya1, and Six1 expression, restricting neurogenesis in the otic vesicle. Dev. Dyn.239, 1708-1722.

Gale, J. E., Marcotti, W., Kennedy, H. J., Kros, C. J. and Richardson, G. P.(2001). FM1-43 dye behaves as a permeant blocker of the hair-cellmechanotransducer channel. J. Neurosci. 21, 7013-7025.

Gallagher, E., Enzler, T., Matsuzawa, A., Anzelon-Mills, A., Otero, D., Holzer, R.,Janssen, E., Gao, M. and Karin, M. (2007). Kinase MEKK1 is required for CD40-dependent activation of the kinases Jnk and p38, germinal center formation, B cellproliferation and antibody production. Nat. Immunol. 8, 57-63.

Guan, Z., Buckman, S. Y., Pentland, A. P., Templeton, D. J. and Morrison, A. R.(1998). Induction of cyclooxygenase-2 by the activated MEKK1→SEK1/MKK4→p38 mitogen-activated protein kinase pathway. J. Biol. Chem. 273,12901-12908.

Hagemann, C. and Blank, J. L. (2001). The ups and downs of MEK kinaseinteractions. Cell. Signal. 13, 863-875.

Hai, T., Wolfgang, C. D., Marsee, D. K., Allen, A. E. and Sivaprasad, U. (1999).ATF3 and stress responses. Gene Expr. 7, 321-335.

Hayashi, T., Cunningham, D. and Bermingham-McDonogh, O. (2007). Loss ofFgfr3 leads to excess hair cell development in themouse organ of Corti.Dev. Dyn.236, 525-533.

Jacques, B. E., Puligilla, C., Weichert, R. M., Ferrer-Vaquer, A., Hadjantonakis,A.-K., Kelley, M. W. and Dabdoub, A. (2012). A dual function for canonical Wnt/beta-catenin signaling in the developing mammalian cochlea. Development 139,4395-4404.

Jin, C., Chen, J., Meng, Q., Carreira, V., Tam, N. N. C., Geh, E., Karyala, S., Ho,S.-M., Zhou, X., Medvedovic, M. et al. (2013). Deciphering gene expressionprogram of MAP3K1 in mouse eyelid morphogenesis. Dev. Biol. 374, 96-107.

Jones, C., Roper, V. C., Foucher, I., Qian, D., Banizs, B., Petit, C., Yoder, B. K.and Chen, P. (2008). Ciliary proteins link basal body polarization to planar cellpolarity regulation. Nat. Genet. 40, 69-77.

Karandikar, M., Xu, S. and Cobb, M. H. (2000). MEKK1 binds raf-1 and the ERK2cascade components. J. Biol. Chem. 275, 40120-40127.

Kyriakis, J. M. andAvruch, J. (2001). Mammalianmitogen-activated protein kinasesignal transduction pathways activated by stress and inflammation. Physiol. Rev.81, 807-869.

Kyriakis, J. M., Banerjee, P., Nikolakaki, E., Dai, T., Rubie, E. A., Ahmad, M. F.,Avruch, J. and Woodgett, J. R. (1994). The stress-activated protein kinasesubfamily of c-Jun kinases. Nature 369, 156-160.

Labuda, T., Christensen, J. P., Rasmussen, S., Bonnesen, B., Karin, M.,Thomsen, A. R. and Ødum, N. (2006). MEK kinase 1 is a negative regulator ofvirus-specific CD8(+) T cells. Eur. J. Immunol. 36, 2076-2084.

Li, Y., Minamino, T., Tsukamoto, O., Yujiri, T., Shintani, Y., Okada, K.-I.,Nagamachi, Y., Fujita, M., Hirata, A., Sanada, S. et al. (2005). Ablation of MEKkinase 1 suppresses intimal hyperplasia by impairing smooth muscle cellmigration and urokinase plasminogen activator expression in a mouse blood-flow cessation model. Circulation 111, 1672-1678.

Liang, G., Wolfgang, C. D., Chen, B. P. C., Chen, T.-H. and Hai, T. (1996). ATF3gene: genomic organization, promoter, and regulation. J. Biol. Chem. 271,1695-1701.

Lim, D. J. and Anniko, M. (1985). Developmental morphology of the mouse innerear: a scanning electron microscopic observation. Acta Otolaryngol. Suppl. 422,5-69.

Loke, J. and Ostrer, H. (2012). Rapidly screening variants of uncertain significancein the MAP3K1 gene for phenotypic effects. Clin. Genet. 81, 272-277.

Loke, J., Pearlman, A., Radi, O., Zuffardi, O., Giussani, U., Pallotta, R.,Camerino, G. and Ostrer, H. (2014). Mutations in MAP3K1 tilt the balance fromSOX9/FGF9 to WNT/beta-catenin signaling. Hum. Mol. Genet. 23, 1073-1083.

Mansour, S. L., Twigg, S. R. F., Freeland, R. M., Wall, S. A., Li, C. and Wilkie,A. O. M. (2009). Hearing loss in a mouse model of Muenke syndrome. Hum. Mol.Genet. 18, 43-50.

Mansour, S. L., Li, C. and Urness, L. D. (2013). Genetic rescue of Muenkesyndrome model hearing loss reveals prolonged FGF-dependent plasticity incochlear supporting cell fates. Genes Dev. 27, 2320-2331.

Meyers, J. R., MacDonald, R. B., Duggan, A., Lenzi, D., Standaert, D. G., Corwin,J. T. and Corey, D. P. (2003). Lighting up the senses: FM1-43 loading of sensorycells through nonselective ion channels. J. Neurosci. 23, 4054-4065.

Minamino, T., Yujiri, T., Terada, N., Taffet, G. E., Michael, L. H., Johnson, G. L.and Schneider, M. D. (2002). MEKK1 is essential for cardiac hypertrophy anddysfunction induced by Gq. Proc. Natl. Acad. Sci. USA 99, 3866-3871.

Mongan, M., Wang, J., Liu, H., Fan, Y., Jin, C., Kao, W. Y.-W. and Xia, Y. (2011).Loss of MAP3K1 enhances proliferation and apoptosis during retinaldevelopment. Development 138, 4001-4012.

Montagutelli, X. (2000). Effect of the genetic background on the phenotype ofmouse mutations. J. Am. Soc. Nephrol. 11 Suppl. 16, S101-S105.

Montcouquiol, M., Sans, N., Huss, D., Kach, J., Dickman, J. D., Forge, A.,Rachel, R. A., Copeland, N. G., Jenkins, N. A., Bogani, D. et al. (2006).Asymmetric localization of Vangl2 and Fz3 indicate novel mechanisms for planarcell polarity in mammals. J. Neurosci. 26, 5265-5275.

Nakagami, H., Morishita, R., Yamamoto, K., Yoshimura, S.-I., Taniyama, Y.,Aoki, M., Matsubara, H., Kim, S., Kaneda, Y. and Ogihara, T. (2001).Phosphorylation of p38 mitogen-activated protein kinase downstream of bax-caspase-3 pathway leads to cell death induced by high D-glucose in humanendothelial cells. Diabetes 50, 1472-1481.

Nakagomi, S., Suzuki, Y., Namikawa, K., Kiryu-Seo, S. and Kiyama, H. (2003).Expression of the activating transcription factor 3 prevents c-Jun N-terminalkinase-induced neuronal death by promoting heat shock protein 27 expressionand Akt activation. J. Neurosci. 23, 5187-5196.

Nayak, G., Lee, S. I., Yousaf, R., Edelmann, S. E., Trincot, C., Van Itallie, C. M.,Sinha, G. P., Rafeeq,M., Jones, S. M., Belyantseva, I. A. et al. (2013). Tricellulin

1552

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1543-1553 doi:10.1242/dmm.023077

Disea

seModels&Mechan

isms

deficiency affects tight junction architecture and cochlear hair cells. J. Clin. Invest.123, 4036-4049.

Ohyama, T., Mohamed, O. A., Taketo, M. M., Dufort, D. and Groves, A. K. (2006).Wnt signals mediate a fate decision between otic placode and epidermis.Development 133, 865-875.

Pannier, S., Couloigner, V., Messaddeq, N., Elmaleh-Berges̀, M., Munnich, A.,Romand, R. and Legeai-Mallet, L. (2009). Activating Fgfr3 Y367C mutationcauses hearing loss and inner ear defect in a mouse model of chondrodysplasia.Biochim. Biophys. Acta 1792, 140-147.

Parker, A., Cross, S. H., Jackson, I. J., Hardisty-Hughes, R., Morse, S.,Nicholson, G., Coghill, E., Bowl, M. R. and Brown, S. D. M. (2015). The goyamouse mutant reveals distinct newly identified roles for MAP3K1 in thedevelopment and survival of cochlear sensory hair cells. Dis. Model. Mech. 8,1555-1568.

Pearlman, A., Loke, J., Le Caignec, C., White, S., Chin, L., Friedman, A., Warr,N., Willan, J., Brauer, D., Farmer, C. et al. (2010). Mutations in MAP3K1 cause46,XY disorders of sex development and implicate a common signal transductionpathway in human testis determination. Am. J. Hum. Genet. 87, 898-904.

Pearson, A. G., Gray, C. W., Pearson, J. F., Greenwood, J. M., During, M. J. andDragunow, M. (2003). ATF3 enhances c-Jun-mediated neurite sprouting. BrainRes. Mol. Brain Res. 120, 38-45.

Puligilla, C., Feng, F., Ishikawa, K., Bertuzzi, S., Dabdoub, A., Griffith, A. J.,Fritzsch, B. and Kelley, M. W. (2007). Disruption of fibroblast growth factorreceptor 3 signaling results in defects in cellular differentiation, neuronalpatterning, and hearing impairment. Dev. Dyn. 236, 1905-1917.

Sadoshima, J., Montagne, O., Wang, Q., Yang, G., Warden, J., Liu, J., Takagi,G., Karoor, V., Hong, C., Johnson, G. L. et al. (2002). The MEKK1-JNK pathwayplays a protective role in pressure overload but does not mediate cardiachypertrophy. J. Clin. Invest. 110, 271-279.

Schlesinger, T. K., Fanger, G. R., Yujiri, T. and Johnson, G. L. (1998). The TAO ofMEKK. Front. Biosci. 3, D1181-D1186.

Schlesinger, T. K., Bonvin, C., Jarpe, M. B., Fanger, G. R., Cardinaux, J.-R.,Johnson, G. L. and Widmann, C. (2002). Apoptosis stimulated by the 91-kDacaspase cleavage MEKK1 fragment requires translocation to soluble cellularcompartments. J. Biol. Chem. 277, 10283-10291.

Shim, K., Minowada, G., Coling, D. E. and Martin, G. R. (2005). Sprouty2, amouse deafness gene, regulates cell fate decisions in the auditory sensoryepithelium by antagonizing FGF signaling. Dev. Cell 8, 553-564.

Sue Ng, S., Mahmoudi, T., Li, V. S. W., Hatzis, P., Boersema, P. J., Mohammed,S., Heck, A. J. and Clevers, H. (2010). MAP3K1 functionally interacts with Axin1in the canonical Wnt signalling pathway. Biol. Chem. 391, 171-180.

Uhlik, M. T., Abell, A. N., Cuevas, B. D., Nakamura, K. and Johnson, G. L. (2004).Wiring diagrams of MAPK regulation by MEKK1, 2, and 3. Biochem. Cell Biol. 82,658-663.

Urness, L. D., Paxton, C. N., Wang, X., Schoenwolf, G. C. and Mansour, S. L.(2010). FGF signaling regulates otic placode induction and refinement bycontrolling both ectodermal target genes and hindbrain Wnt8a. Dev. Biol. 340,595-604.

Warr, N., Bogani, D., Siggers, P., Brixey, R., Tateossian, H., Dopplapudi, A.,Wells, S., Cheeseman,M., Xia, Y., Ostrer, H. et al. (2011). Minor abnormalities oftestis development in mice lacking the gene encoding the MAPK signallingcomponent, MAP3K1. PLoS ONE 6, e19572.

Widmann, C., Gerwins, P., Johnson, N. L., Jarpe, M. B. and Johnson, G. L.(1998). MEK kinase 1, a substrate for DEVD-directed caspases, is involved ingenotoxin-induced apoptosis. Mol. Cell. Biol. 18, 2416-2429.

Witowsky, J. A. and Johnson, G. L. (2003). Ubiquitylation of MEKK1 inhibits itsphosphorylation of MKK1 and MKK4 and activation of the ERK1/2 and JNKpathways. J. Biol. Chem. 278, 1403-1406.

Xia, Y., Makris, C., Su, B., Li, E., Yang, J., Nemerow, G. R. and Karin, M. (2000).MEK kinase 1 is critically required for c-Jun N-terminal kinase activation byproinflammatory stimuli and growth factor-induced cell migration.Proc. Natl. Acad.Sci. USA 97, 5243-5248.

Xu, S., Robbins, D. J., Christerson, L. B., English, J. M., Vanderbilt, C. A. andCobb, M. H. (1996). Cloning of rat MEK kinase 1 cDNA reveals an endogenousmembrane-associated 195-kDa protein with a large regulatory domain.Proc. Natl.Acad. Sci. USA 93, 5291-5295.

Yu, L.-Z., Xiong, B., Gao, W.-X., Wang, C.-M., Zhong, Z.-S., Huo, L.-J., Wang, Q.,Hou, Y., Liu, K., Liu, J. et al. (2007). MEK1/2 regulates microtubule organization,spindle pole tethering and asymmetric division during mouse oocyte meioticmaturation. Cell Cycle 6, 330-338.

Yujiri, T., Sather, S., Fanger, G. R. and Johnson, G. L. (1998). Role of MEKK1 incell survival and activation of JNK and ERK pathways defined by targeted genedisruption. Science 282, 1911-1914.

Yujiri, T., Ware, M., Widmann, C., Oyer, R., Russell, D., Chan, E., Zaitsu, Y.,Clarke, P., Tyler, K., Oka, Y. et al. (2000). MEK kinase 1 gene disruption alterscell migration and c-Jun NH2-terminal kinase regulation but does not cause ameasurable defect in NF-kappa B activation. Proc. Natl. Acad. Sci. USA 97,7272-7277.

Zhang, L., Wang, W., Hayashi, Y., Jester, J. V., Birk, D. E., Gao, M., Liu, C.-Y.,Kao, W.W.-Y., Karin, M. and Xia, Y. (2003). A role for MEK kinase 1 in TGF-beta/activin-induced epithelium movement and embryonic eyelid closure. EMBO J. 22,4443-4454.

1553

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1543-1553 doi:10.1242/dmm.023077

Disea

seModels&Mechan

isms