Gata3 is Required in Late Proneurosensory Development for Proper Sensory Cell Formation and Organization

It has been previously shown that zinc-finger transcription factor Gata3 has dynamic expression within the inner ear throughout embryonic development and is essential for cochlear neurosensory development. However, the temporal window to which Gata3 is required for the formation of the cochlear neurosensory epithelia remains unclear. To investigate the role of Gata3 on cochlear neurosensory development in the late prosensory stages, we used the Sox2-creERT2 mouse line to target and conditionally delete Gata3 at E11.5 before the cells have fully committed to a neurosensory fate. While the inner ears of Sox2-creERT2: Gata3 f/f mice appear morphologically normal, the sensory cells in the organ of Corti are partially lost and disorganized in a basal to apical gradient with the apex demonstrating the more severe phenotype. Additionally, spiral ganglion neurons display aberrant peripheral projections, such as increased distances between radial bundles and disorganization upon reaching the organ of Corti. Furthermore, heterozygous Sox2-creERT2: Gata3 f/+ mice show a reduced phenotype in comparison to the homozygous mutant, supporting the concept that Gata3 is not only required for proper formation at the later proneurosensory stage, but also that a specific level of Gata3 is required. Therefore, our studies confirm that Gata3 plays a time-sensitive and dose-dependent role in the development of sensory cells in the late proneurosensory stages.


Introduction
The mammalian inner ear is comprised of six unique sensory organs, but only one of these organs is responsible for the sense of hearing: the cochlea. The cochlea contains the organ of Corti (OC) which is comprised of mechanosensory hair cells (HCs) and their corresponding supporting cells (SCs). HCs transduce sound energy into electrical impulses via innervation by spiral ganglion neurons (SGNs) that project into the hindbrain for further auditory processing. The development of these three sensory cell types has been extensively studied, but there are still gaps in knowledge regarding the transcription factors and gene networks that control the spatial and temporal aspects of this process at later proneurosensory stages.
The inner ear is derived from the otic placode, which will invaginate to form the otic cup before developing into the otocyst around embryonic day 9 (E9) 1,2 . While several transcription factors are important for neurosensory development in this time frame, the zinc-nger transcription factor Gata3 is particularly interesting due to its dynamic expression throughout development. While Gata3 is initially expressed as early as E8.5 throughout the otocyst, by E10.5 its expression is restricted to the proneurosensory regions [3][4][5][6][7][8][9] . Gata3 continues to be expressed in SGNs until postnatal day 14 (P14) and remains highly expressed in SCs, with lower levels in HCs, throughout adulthood [10][11][12][13][14] .Therefore, it has been postulated that Gata3 plays an important and dynamic role in inner ear development and neurosensory cell formation during this embryonic temporal window.
Previous studies have shown that loss of Gata3 in the early proneurosensory region around E8.5 leads to loss of all cochlear neurosensory cells 5,15 , while loss of Gata3 one day later around E9.5 leads to a patchy loss of HCs and SCs, and disorganization and patchy loss of SGNs. Other studies have investigated the role of Gata3 postnatally in the maintenance of HCs and SCs 11,12 . These studies found that Gata3 is necessary later on to maintain OHCs and to functionally develop IHCs, while loss of Gata3 from postnatal SCs results in an increase in some types of SCs through downregulation of other genes. However, there exists a gap in knowledge about the role of Gata3 later in embryonic development as the proneurosensory cells start to differentiate into HCs, SCs, and SGNs. Speci cally, it remains to be seen how long expression is required for proper embryonic development of neurosensory cells before Gata3 switches to its postnatal maintenance role. Additionally, while we know the presence of Gata3 is necessary for proper neurosensory development, the precise level of Gata3 expression is also important for maintenance and function. For example, both Gata3 haploinsu ciency and Gata3 over-expression, as a result of gene duplication, cause human hypoparathyroidism, sensorineural deafness, and renal dysplasia (HDR) syndrome [16][17][18][19][20][21] . While the triad of symptoms of HDR syndrome range in severity, nearly all patients exhibit deafness 16,19,22,23 . Uniquely, deafness is the only symptom of HDR syndrome which can present singularly 16,19,21,23 . This suggests that not only is continued expression of Gata3 required for proper inner ear development, but speci c levels of Gata3 are also required. Continued investigation of the dose-dependent requirements of Gata3 would also contribute to the elds' overall understanding of inner ear gene regulatory networks.
Therefore, in this study, we explored the window of developmental plasticity which is governed by Gata3 as a follow up to previous studies showing its loss is detrimental to cochlear neurosensory epithelia 5-7, 10,21,24,25 . Using the Sox2-cre ERT2 mouse line 26 , we conditionally deleted Gata3 from proneurosensory cells via tamoxifen (TMX) injection at E11.5. Our results show that deletion of Gata3 causes severe loss and disorganization of HCs, SCs, and SGNs in a basal to apical gradient, with a more severe phenotype presenting in the apex. Interestingly, the mutant ears were morphologically normal in that it presented with a full-length cochlea unlike previous Gata3 deletion studies. Overall we show that, while Gata3 is not necessary at E11.5 for overall morphological development and elongation of the cochlea, however, Gata3 is required in the late proneurosensory stages for cochlear sensory epithelia and SGNs both to form and organize properly.

Results
Gata3 is deleted from HCs, SCs, and SGNs at E11.5 Previous studies have characterized Sox2-cre ERT2 expression at the placode stage (E8.5), otocyst stage (E10.5), and the late otocyst stage (E12.5) 27-32 . At E10.5, Sox2 is present in both the nonsensory cochlear oor and roof 31 . By E12.5, Sox2 is exclusively expressed in OC sensory cells 31 . In order to con rm complete knockout of Gata3 from HCs, SCs, and SGNs, in situ hybridization was performed using a Gata3 riboprobe. While the control shows high expression of Gata3 in all cell types from base to apex, the homozygous mutant shows no expression in the HCs and SCs and greatly reduced expression in the SGN cell bodies (Fig. 1A-D'), demonstrating that our model is indeed reducing levels of Gata3 in the cell types of interest.

Gata3 is required for sustained formation and organization of HCs
We rst analyzed the effect of deletion of Gata3 on HCs, as previous Gata3 CKOs show either no HC development or only patches of HCs [5][6][7]10 . In order to assess the phenotype of the deletion of Gata3, two different controls were used: Gata3 f/f ( Fig. 2A-A") and Sox2-cre ERT2 (Fig. 2B-B"). Other studies have previously demonstrated that the knock-in Sox2-cre ERT2 line shows inner hair cell (IHC) duplets, which was con rmed in our study (Fig. 2B-B''; white circles). It was important to investigate the IHC duplets in the heterozygous mutant compared to the Sox2-cre ERT2 control to ensure that the phenotype seen is not the result of using this Cre line ( Fig. 2C-C"). While the base, middle and apex of the heterozygous mutant all contain IHC duplets similar to the Sox2-cre ERT2 control, it should be noted that the third row of outer hair cells (OHCs) is lost in the middle region of the OC into the apical region ( Fig. 2C-C"). We found that the heterozygous genotype shows a continuous formation of HCs from base to apex, while the homozygous mutant shows some disturbances in the apical region, similar to the previous Gata3 CKO study that showed the presence of patches 6 . The homozygous mutant also shows a worsening phenotype compared to the heterozygous mutant. While the base contains all three rows of OHCs, progressive rows of OHCs are lost ( Fig. 2D-D'). Just two rows of OHCs are present in the middle region and almost no rows of OHCs are present in the apex (Fig. 2E-F"). Additionally, the entire cochlea contains Myosin VIIa positive cells in the GER with the highest number appearing in the apex, which is similar to a postnatal Gata3 CKO from SCs using this same Cre line 11 . Furthermore, these cells in the GER are associated with SGN endings. Ectopic HCs have been seen in the GER in both CKO and over-expressor models previously [33][34][35][36][37][38] . While ectopic HCs generally are not seen in combination with missing rows of OHCs, previous studies have shown that loss of Gata3 results in missing OHCs postnatally 12,21 . The phenotype of both ectopic HCs and missing rows of OHCs as a result of embryonic loss of Gata3 is unique and further supports a role for Gata3 in this speci c temporal window in this speci c cell type.

Gata3 is required for corresponding SC formation and organization
Previous studies that have deleted Gata3 have shown either no SC development or only limited patches of SC formation localized to the HC patches 5,6,10 . However, we still observe SCs in our model throughout the majority of the length of the cochlea. Similar to the HC phenotype in this model, we found that the homozygous mutant shows almost continuous formation of SCs, except for some patches in the apex (Fig. 3). The apex also shows disorganization of the SC rows. The homozygous mutant shows a worsening phenotype compared to the heterozygous mutant, similar to that seen in the HCs. The base and middle show disorganized SCs and complete loss of some outer SC rows in the middle region. The apex contains the most severe phenotype in which SC appear to cluster together which is very similar to the SC phenotype seen in other Gata3 CKO studies 10 . Ultimately, the phenotype in the HCs and SCs are consistent in their appearance of progressive loss of OHCs from base to apex, mirroring the loss of SCs from base to apex in the homozygous mutant. While this SC disorganization in our model is also similar to the phenotype seen in other Gata3 CKO studies, it is important to note that the previous study did not also observe ectopic HCs in the GER 10 . Further studies are needed in order to tease apart the speci c requirement for Gata3 within this speci c time window to determine if Gata3 deletion in one cell population can in uence another cell population.

Gata3 Is Required For Organization Of Sgn Peripheral Projections
Our observation of Myosin VIIa positive cells associated with SGN endings in the homozygous mutant led us to investigate the peripheral projections of SGNs to con rm that they were developing properly. Previous studies examining the effect of Gata3 deletion from the proneurosensory region of the developing otocyst observed a severe reduction in the number of SGNs present in mutant samples, while the SGNs that did form displayed aberrant projection patterns towards the developing OC 6,10 . Another study in which Gata3 deletion was restricted to SGNs saw proper formation of SGNs with disorganized peripheral projections 14,24 . We rst examined peripheral projections in a Sox2-cre ERT2 mutant sample to establish whether the Cre knock-in displays a SGN phenotype. When compared with control samples ( The area between the radial bundles in the apex of control and homozygous mutant samples was measured and quanti ed (Fig. 4J). The method for radial bundle distance quanti cation can be found as Supplementary Fig. S1 online. Analysis of the data showed a statistically signi cant increase in the distance between radial bundles in the homozygous mutant relative to the control (p < 0.0001).
Additionally, the values for the area in mutant samples was highly variable, further supporting that loss of Gata3 results in disorganization of peripheral projections of SGNs.
We also examined the peripheral projections where the neurites reach the OC (Fig. 4G-I"). The basal region of the heterozygous mutant is comparable to the control (Fig. 4G-H), but peripheral projections are progressively fewer and become disorganized in the heterozygous mutant with progression to the middle and apex (Fig. 4G'-G", H'-H"). Peripheral projections in the base of the homozygous mutant appear slightly disorganized upon reaching the OC. Additionally, the density of neurites in the homozygous mutant appears to be less when compared to the base of the control (Fig. 4I). The disorganization of the neurites and decreased density is even more pronounced in the middle and apex of the homozygous mutant ( Fig. 4I'-I"). Fewer neurites project into the OHC region of the OC in the middle and few-to-no neurites project to the OHC region in the apex. In these regions, not all neurites that are present within the OHC region properly turn towards the base but rather, turn towards the apex.
Based upon our results, Gata3 expression is important for the formation of radial bundles with regards to appropriate density and distance between bundles, as well as for proper branching patterns and overall organization. Additionally, Gata3 is needed for peripheral neurites to reach the OC, particularly the OHC region, and to form proper connections with HCs. Importantly, the loss of Gata3 has a phenotype that progressively worsens along the length of the cochlea, with the greatest phenotype observed in the apex.

Gata3 Is Required For Proper Central Path nding Of Sgns
Given that homozygous mutants display aberrant peripheral projections of SGNs, with the phenotype progressively getting more severe in a basal to apical manner (Fig. 4), we next investigated whether central projection of SGNs to the cochlear nucleus (CN) was also affected. Previous studies examining the role of Gata3 in spiral ganglion neuron central path nding have shown varied results depending on the location and timing of Gata3 deletion 6,24 . Early deletion of Gata3 throughout the entire inner ear at E9.5 results in central SGN bers bifurcating at several branch points, with terminal bers projecting nonspeci cally throughout the CN 6 . However, deletion of Gata3 within delaminated SGNs at E9.5 results in normal projection of SGNs within the CN with tonotopy maintained 24 . Taken together these two studies suggest that Gata3 may be affecting SGN neuron central path nding in a cell non-autonomous and timedependent manner. In order to investigate this further, lipophilic dyes were applied to the base (red) and apex (green) of Sox2-cre ERT2 control, as well as heterozygous and homozygous mutant cochlea (Fig. 5A) to visualize the projections of SGNs into the CN. Sox2-cre ERT2 control SGNs entered the hindbrain and bifurcated sending ascending and descending process towards the anteroventral cochlear nucleus (AVCN) and dorsal cochlear nucleus (DCN)/posteroventral cochlear nucleus (PVCN) respectively (Fig. 5B).
Sox2-cre ERT2 control SGNs remained segregated with basal bers extending more dorsally and apical bers more ventrally (Fig. 5B). This stereotyped central wiring was also maintained in heterozygous (Fig. 5C) mice. In contrast SGNs in homozygous mice display less segregation between apical and basal bers. Apical bers often project more dorsally into spaces occupied by basal bers. Additionally, some apical bers upon reaching the hindbrain project outside of cranial nerve VIII into areas outside of the CN (Fig. 5D). These results provide further evidence for the idea that Gata3 expression plays an important role in the development and wiring of SGNs centrally. Our data along with previous studies 6, 24 suggest that Gata3 is acting in a cell non-autonomous manner at or before E11.5 to promote proper central wiring of SGNs. Further investigations are needed to elucidate what cell populations require early Gata3 expression in order to promote proper central path nding of SGNs.
Gata3 deletion at E11.5 results in full morphologic development of the cochlear duct and vestibular system, but shows progressive neurosensory epithelial loss and disorganization Previous Gata3 deletion studies have shown a variety of phenotypes that include morphologic and cochlear neurosensory epithelia defects 5-7, 10,21,24,25 . Gata3 null mice display a severely truncated cochlear and vestibular system which were devoid of sensory epithelia except for a small patch of HCs and SGNs in a portion of the saccule 5,7 . Gata3 deletion at E8.5 using the Foxg1-cre mouse line resulted in a truncated cochlea which contained no HCs and abnormal morphologic development of the vestibular system 6 . Gata3 deletion at E9.5 using the Pax2-cre mouse line resulted in similar morphologic defects including a truncated cochlea and abnormal vestibular system. However, deletion at E9.5 resulted in patchy sensory cell development of HCs, SCs, and SGNs 6, 10 . In studies that have conditionally deleted Gata3 from only SGNs, HCS and SCs form properly 24,25 . We contribute results for Gata3 deletion at E11.5, a time in development in which proneurosensory cell differentiation is occurring. Our ndings show that Gata3 deletion at E11.5 results in a morphologically sound structure with a full length cochlea and well developed vestibular system (data not shown). Within the cochlea, the sensory cells in the OC are mostly present and have a varying phenotype depending on the cochlear region. In the homozygous mutant cochlear base, HCs and SCs are present with only mild disorganization (Fig. 6), while the homozygous mutant basal radial bundles have larger spacing than normal but the neurons are well organized. This contrasts the phenotype seen in the apex since the peripheral projection density of the mutant apex is decreased and those projections which are present appear disorganized (Fig. 6). Additionally, the tonotopy of SGN central projections is maintained within the CN in both heterozygous and homozygous mutants (Fig. 5). In comparison, the mutant apical HCs are severely reduced to patchy clusters with some ectopic HCs that appear in the GER, while the apical SCs are not organized in rows and instead cluster together (Fig. 6). Our data demonstrates a role for Gata3 in all neurosensory cells after their initial speci cation.

Discussion
Gata3 was previously shown to be necessary for both proper morphology and cochlear neurosensory epithelia early in development when its expression is high throughout the entire otocyst 5-7, 10,21,39 . However, the role of Gata3 in HC, SC, and SGN formation after its restriction to the proneurosensory region was unknown. Our study reveals novel ndings that Gata3 plays both a dose-dependent and necessary role in the formation and organization of neurosensory cell types, but does not have an impact on the overall morphology of the inner ear at this speci c developmental time point.
This project contributes new knowledge about the role of Gata3 on proneurosensory epithelia formation in a temporal window that lls a gap between previous studies that investigated Gata3 deletion. From our results we nd that deletion of Gata3 from the proneurosensory domain at E11.5 results in a fully formed cochlear duct (data not shown). Regardless of the single or dual loss of Gata3 alleles, both ears formed morphologically normal cochleas. Therefore, Gata3 is not required for morphologic development at E11.5. Given that previous Gata3 deletion studies did not see normal morphology of the cochlea 5,6 , it is intriguing that deletion of Gata3 about two days later results in a morphologically sound inner ear with a fully formed cochlea. While HCs, SCs and SGNs do form, they are highly disorganized and this phenotype progressively worsens from base to apex. In Gata3 heterozygous null mice, studies have found that OHC loss occurs without Gata3 12,21,40 . This phenotype is mirrored in our study, despite the difference in timing of which Gata3 is deleted. It is also noteworthy that the heterozygous mutant had a subtler phenotype compared to the homozygous mutant, suggesting that precise levels of GATA3 are needed for proper formation and organization of the proneurosensory epithelia. If precise levels of Gata3 are truly necessary, then increased levels of Gata3 should also have a phenotype in our model. Several other overexpressor studies have been previously studied that demonstrated ectopic HCs in the GER [33][34][35][36]38 . Previous studies have even used the Gata3 over-expressor model in combination with upregulation of other sensory genes in order to increase the e ciency of ectopic HC formation 35,36 . Investigating the overexpression of Gata3 in this model would be useful in determining the detrimental effects, if any, of higher levels of Gata3 in the cochlea. While this would further elucidate the speci c role of Gata3 in the cochlea, the investigation of Gata3 over-expression is especially pertinent since extra alleles of Gata3 have also been known to cause HDR syndrome 20 .
Finally, it should also be noted that we deleted Gata3 from three different cell types: HCs, SCs, and SGNs. While these cell types work together, it is unclear if loss of Gata3 in just one of the cell types is enhancing the overall phenotype we see in our model. Despite the fact that HCs in our model are innervated by SGNs throughout the entire cochlea, we are unable to determine if the HC disorganization is causing the improper SGN peripheral projections when using the Sox2-cre ERT2 model, or vice versa. Likewise, since the SCs and HCs are connected via tight junctions in the normal OC, our model is unable to determine if a phenotype in one of these cell types is exacerbating the phenotype overall. Therefore, in order to tease apart the role of Gata3 at this speci c time point, future studies could use other more cell-speci c Cre lines to delete Gata3. Comparison of our phenotype in this model to Gata3 CKO in HC-speci c, SCspeci c, or SGN-speci c lines could elucidate the exact role of Gata3 in the proneurosensory stage of development.
In conclusion, our work demonstrates that Gata3 is essential for proper cochlear neurosensory epithelia development and organization in the late proneurosensory stage at E11.5. Because our study demonstrates a phenotype in the heterozygous mutant in addition to a more severe phenotype in the homozygous mutant, we can con rm that correct levels of Gata3 are also required for proper development. Furthermore, our study performs the latest embryonic Gata3 deletion known in the eld and contributes to the understanding that Gata3 is required for proper formation and organization of the cochlea sensory epithelia at E11.5, but not for overall cochlea morphology.

Mouse model and genotyping
All animal care and procedures were approved by Western Michigan University Institutional Animal Care and Use Committee (IACUC) following the guidelines for use of laboratory animals (IACUC #20-11-01). All experiments were carried out in accordance with the ARRIVE guidelines, and all methods were carried out in compliance with all relevant regulations. The following mouse strains were used: Sox2-cre ERT2 (Jackson Labs) 26 , tdTomato Ai9 (Jackson Labs) 41 , and Gata3 Flox were provided by Dr. Maxime Bouchard 42 . Sox2-cre ERT2 males were bred with Gata3 f/f females to produce males that were Sox2cre ERT2 : Gata3 f/+, who were viable. Sox2-cre ERT2 : Gata3 f/f mice were produced by breeding Sox2cre ERT2 : Gata3f/+ or Sox2-cre ERT2 : Gata3 f/f males with Gata3 f/f or Gata3 f/+ females. Genotyping was performed using the following primers:

Whole-mount Immunohistochemistry
Whole mount immunohistochemistry was performed on previously xed tissue 43 . Ears were washed in phosphate buffered saline (PBS), then washed ve times ve minutes in PBS/0.05% Tween20 followed by blocking for one hour in 5% normal donkey serum, 1% bovine serum albumin, and 0.5% TritonX-100 in PBS. The tissue was incubated in primary antibodies, diluted in blocking buffer, at 4°C for three nights.

Spiral Ganglion Neuron Quanti cation
For radial bundle quanti cation, shown in Fig. 4 and Supplementary Fig. S1 online, cochlea were imaged at the same magni cation in the apex for each genotype. Using FIJI imaging software (Version 1.8.0_66), eight spaces between radial bundles were outlined. All area results were recorded in Graph Pad Prism (Version 9.1.2) and a TTEST analysis was performed. Data point plot graphs were constructed, and signi cance was set at P < 0.05.

In situ hybridization
Gata3 mRNA labeling was achieved using a previously described in situ hybridization protocol 43 . Mice were xed in 4% PFA and inner ears were dissected in 0.4% PFA. Control ears and experimental ears were run together throughout the experiment to ensure both ears received the same experimental conditions. Ears were dehydrated overnight in 100% methanol and rehydrated through a graded methanol series. Ears were digested with Proteinase K in PBS (Ambion, Austin, TX, USA). Samples were hybridized overnight at 60°C to the Gata3 riboprobe in hybridization solution consisting of 50% (v/v) formamide, 50% 2X saline sodium citrate (SSC), and 6% (w/v) dextran sulphate. Unbound probe was removed by performing washes with 2X SSC. Samples were then incubated with anti-digoxigenin antibody conjugated with alkaline phosphatase (Roche Diagnostics GmbH, Mannheim, Germany) overnight at room temperature. Ears were extensively washed with 1X washing buffer throughout the day, then left overnight in 1X washing buffer at room temperature. Samples were then incubated at room temperature in detection buffer (Roche) before being thoroughly saturated with nitroblue phosphate/5-bromo, 4-chloro, 3-indolil phosphate (BM purple substrate, Roche). Control and mutant samples were developed in BM purple for the same length of time. Ears were mounted in glycerol on a slide and imaged with a Nikon Eclipse E600 microscope and Canon EOS Rebel T7i camera. Images were edited in Corel Draw (version 19.0; 2017).

Lipophilic Dye Tracing
Neuronal tracing of spiral ganglion neurons was conducted as previously described 43 . Brie y, the lateral half of the inner ear was exposed, and pieces of lipophilic dye-soaked paper was inserted into the base (NeuroVue® Red) and apex (NeuroVue® Maroon) of the cochlea. Heads were then placed into glass vials lled with 4% paraformaldehyde and incubated at 37°C for 3 days to allow for proper dye diffusion. Following incubation, the brains were removed, and the brain stem was at mounted with the lateral side facing up in glycerol on a slide and imaged within 1 hour. All imaging was performed using a Leica Stellaris 5 confocal microscope with LAS X software and images were compiled in ImageJ and edited in CorelPhoto Paint (Version 19.0; 2017). Declarations imaged at the cochlear base and apex. Gata3 expression appears in the HCs, SCs, and SGNs of the control and is absent in the HCs and SCs and decreased in the SGNs of the homozygote mutant. Scale bar: 100 µm Figure 2 Deletion of Gata3 results in loss of HCs in a basal to apical gradient (A-F') Representative images from the basal, middle, and apical regions of the cochlea for HCs indicated by MYOSIN7A+ staining. Two different controls were used, Gata3 f/f and Sox2-cre ERT2 , in order to account for the haploinsu cent phenotype of the Cre line used. Both the heterozygous and homozygous mutant show IHC duplets (white circles) and missing rows of OHCs (white brackets), while the homozygous mutant also shows ectopic HCs in the GER. Scale bar: 50 µm    Loss of Gata3 follows in the timeline of previous studies showing progressive basal to apical loss of sensory cells. Schematic of Gata3mutant ear at E18.5, indicating the cochlea (red), vestibular system (green), and endolymphatic duct (purple). Images are representative of the phenotype observed along the length of the cochlea and are taken from gures 2,3, and 4. Our study deleted Gata3 later than previous studies at E11.5 and found that the overall morphology of the inner ear is intact unlike previous studies. In addition, there was a basal to apical loss of sensory cells indicating Gata3 is still required for their formation and organization.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.