The RNA-binding protein TRIM71 is essential for hearing in humans and mice and times auditory sensory organ development

Abstract

The RNA-binding protein TRIM71 is essential for brain development, and recent genetic studies in humans have identified TRIM71 as a risk gene for congenital hydrocephaly (CH). Here, we show that mono-allelic missense mutations in TRIM71 are associated with hearing loss (HL) and inner ear aplasia in humans. Utilizing conditional Trim71 knockout mice carrying a CH and HL-associated mutation, we demonstrate that loss of TRIM71 function during early otic development (embryonic day 9–10) causes severe hearing loss. While inner ear morphogenesis occurs normally in Trim71 knockout mice, we find that early otic loss of TRIM71 function disrupts the highly stereotyped timing of cell cycle exit and differentiation within the inner ear auditory sensory organ (cochlea), resulting in the premature formation and innervation of mechano-sensory hair cells. Transcriptomic profiling of Trim71-deficient cochlear progenitor cells identifies Inhba and Tgfbr2 as targets of TRIM71 repression, and our analysis of Inhba-Tgfbr1 double knockout mice indicates that TRIM71 maintains hair cell progenitors in a proliferative and undifferentiated state by restricting TGF-β-type signaling. Characterization of hair cells and their associated neurons in adult Trim71 knockout mice revealed reduced presynaptic terminals and neuronal degeneration in the outer hair cell region, providing a basis for the observed hearing deficits in Trim71 knockout mice.

Introduction

The inner ear cochlea contains a highly specialized sensory organ dedicated to detecting sound. The core functional unit of this sensory organ is comprised of mechanoreceptors-termed hair cells, which are positioned atop supporting cells and are innervated by afferent spiral ganglion neurons (SGNs). Hair cells, supporting cells, and SGNs are generated in a highly stereotyped manner that requires strict spatial and temporal control of cell proliferation and differentiation (reviewed in (1)). Recent functional studies have revealed that the RNA-binding protein LIN28B and members of the let-7 (lethal) family of miRNAs play a central role in timing cell cycle exit and differentiation of hair cell and supporting cell progenitors (termed pro-sensory cells) in the murine (2) and avian auditory sensory organ (3). The let-7 miRNAs and their mutual antagonist LIN28B are part of an evolutionary highly conserved network of genes, first identified in C. elegans, which controls developmental timing and stemness (4, 5). While LIN28B promotes stemness, let-7 miRNAs promote cell cycle withdrawal and differentiation by targeting growth-related genes, including Lin28b itself (6). Another well-known let-7 target and evolutionarily conserved gene that promotes stemness is Trim71 (also referred to as Lin41) (5, 7). The expression of Trim71 peaks during early embryonic development (8, 9), and Trim71 continues to be expressed in various stem cell populations, such as germ cells, throughout later developmental stages (10). Highlighting TRIM71’s importance for stemness, TRIM71 has been found to enhance the reprogramming of human fibroblasts into induced pluripotent stem cells (iPSCs) when co-expressed along with pluripotency factors OCT4, SOX2, and KLF4 (11). TRIM71 belongs to the tripartite-motif (TRIM)-NHL protein family, characterized by an N-terminal RING domain with intrinsic ubiquitin E3 ligase activity and a C-terminal NHL domain responsible for RNA-binding (reviewed in (12)). Genetic and biochemical studies showed that TRIM71 cooperates with core elements of the miRNA-induced gene silencing complex to repress the translation of targeted mRNAs and facilitate their degradation (13–16). During early embryonic development, TRIM71 prevents the premature activation of neurogenic genes, and Trim71 gene trap (8, 9) and knockout mice (14) have severe neural tube closure defects due to decreased cell proliferation and premature onset of neural differentiation. In humans, missense mutations in the TRIM71 gene lead to a neurodevelopmental syndrome characterized by ventriculomegaly and hydrocephalus (17–20). Loss of function studies in mice revealed that brain abnormalities arise from the premature differentiation of neuroepithelial cells, adversely affecting cortical neurogenesis and subsequent cerebrospinal fluid biomechanics (19). We now show that CH-causing missense mutations in TRIM71 are linked to hearing loss (HL) in humans. Using conditional Trim71 knockout (KO) mice that carry the mouse equivalent of the human CH and HL-associated mutation, we demonstrate that loss of TRIM71 function during early otic development results in severe hearing loss. Transcriptomic and functional data demonstrate that TRIM71 times cell cycle exit and differentiation within the auditory sensory epithelium by repressing Inhba and Tgfbr2 expression, two key components of TGF-β-type signaling. Characterization of hair cells and their associated neurons in adult Trim71 KO mice revealed a reduction in presynaptic terminals and neuronal degeneration in the outer hair cell region. In sum, our research identifies TRIM71 as a syndromic HL gene and critical regulator of auditory-sensory development.

Results

Mono-allelic missense mutations in TRIM71 are associated with hearing loss in humans

Recent human genetics studies identified TRIM71 as a novel disease gene in congenital hydrocephaly (17–20). A subset of patients with de novo or transmitted mono-allelic TRIM71 mutations reported hearing problems and underwent further testing (Table 1). These patients have been reported in previous studies (19, 20). Three out of nine patients reported a hearing loss phenotype, and for two of these patients, details regarding clinical workup of hearing loss were available (Patients KCHYD673–1 and 18CY000656, also referred to as “patient 1” and “patient 2”, respectively, in the rest of the manuscript). Patient 1, a female who harbors a p.(Gln334Arg) mutation in TRIM71, which impairs the native subcellular localization of TRIM71 to P-bodies (20), presented with a congenital hypothyroidism with ectopic thyroid, a right mild sensorineural hearing loss, bilateral chorioretinal coloboma, bilateral renal cysts, axial hypotonia, brachydactyly clinodactyly of the 5th fingers, a square face, round external ears without lobule and bifid uvula at 7 months of age. At 5 years old, the hearing loss has progressed to a mixed severe right hearing loss (Figure S1A). Brain and internal auditory canals MRI showed bilateral lateral semicircular canals malformation with normal cochlea and olfactory bulbs (Figure S2A-C). She had motor delay with abnormal vestibular tests without other neurodevelopmental impairment. Patient 2, a female who harbors a missense mutation in TRIM71’s NHL domain [p.(Arg608His)], which impairs RNA binding (19), presented with a mixed severe left and a moderate conductive right hearing loss at 6 years old (Figure S1B). Temporal bone CT scans showed unilateral left semicircular canal malformation with normal cochlea, bilateral external ear canal hypoplasia, and ossicular malformations (Figure S2D-K). She also presented with speech and language delay, hydrocephalus, brachymesophalangia of the 5th left finger, ventricular septal defect, and only 4 lumbar vertebrae. These clinical findings suggest a potential link between TRIM71 dysfunction and human hearing loss.

Table 1.

Genetic and clinical phenotypes of patients with mutations in TRIM71.

Proband ID	Sex	Ethnicity	Inheritance	Position (GRCh 37)	AA Change	Hearing Loss?	Details about hearing
KCHYD154–1	F	European	De novo	3:32932519:G:A	p.R608H	-	N/A
KCHYD154–5	M	European	Transmitted from KCHYD154–1	3:32932519:G:A	p.R608H	-	N/A
KCHYD79–1	M	Mexican	De novo	3:32932519:G:A	p.R608H	-	N/A
18CY000656	F	European	De novo	3:32932519:G:A	p.R608H	+	Left: profound mixed deafness with malformation of external auditory canal.
KCHYD85–1	M	European	De novo	3:32933083:G:A	p.R796H	+	Left sensorineural hearing loss
KCHYD419–1	F	European	De novo	3:32933083:G:A	p.R796H	-	N/A
KCHYD425–1	F	European	De novo	3:32933083:G:A	p.R796H	-	N/A
KCHYD670–1	M	European	De novo	3:32932566:G:A	p.D624N	-	N/A
KCHYD673–1	F	European	De novo	c.1001G>A	p.Q334R	+	Unilateral sensorineural severe deafness

Trim71 is expressed in otic neural-sensory progenitor cells

To gain insights into how missense mutations in Trim71 may cause conductive and or sensorineural hearing loss, we utilized published single-cell RNA sequencing data to examine Trim71 expression in the developing mouse inner ear at stages E9.5, E11.5, and E13.5 (21) using Gene Expression Analysis Resource (gEAR) portal (22). The inner ear epithelium and its innervating afferent neurons derive from the otic placode located adjacent to the hindbrain. By E9.5, the otic placode-derived cells have formed an otocyst and by E11.5, many of the neuroblasts have delaminated and coalesced to form the otic ganglion, which later subdivides into cochlear (spiral ganglion) and vestibular ganglion, and by E13.5, both vestibular and auditory sensory epithelia are specified (reviewed in (23)).

At E9.5, Trim71, similar to Lin28b, was broadly expressed in both otic and peri-otic cells, with high expression in otic epithelial cells marked by Tbx2 and Lfng expression (Figure 1A). At E11.5 and E13.5, Trim71 expression became more refined, co-expressing with Lin28b in otic epithelial and ganglion cells at E11.5, and at E13.5 in inner ear sensory cells and cochleovestibular ganglion cells (Figure 1A). We independently confirmed Trim71’s early otic expression in stage E9.5 mouse tissue using anti-TRIM71 immuno-staining (Figure 1B) and RNA in situ hybridization (ISH) (Figure 1C). Trim71 transcript was highly expressed in the developing otocysts, particularly in the neural sensory competence domain marked by Lfng expression (24), and consistent with previous reports, in the ventricular zone of the adjacent hindbrain (19) (Figure 1C). Immunostaining with anti-TRIM71 antibody showed broad expression of TRIM71 protein in otic and peri-otic tissue and confirmed previous findings that TRIM71 protein resides in the cytoplasm (Figure 1B). Next, we used quantitative PCR (qPCR) to analyze changes in Trim71 expression prior (E13.5) and during cochlear hair cell differentiation (E14.5-E16.5). As a control, we examined the expression of pro-sensory genes Isl1 (24), Hmga2, and Lin28b (2), along with Atoh1. The transcription factor ATOH1 is highly expressed in nascent hair cells, and induction of Atoh1 expression within the cochlear sensory epithelium marks the onset of hair cell differentiation and cochlear differentiation in general (25). Like Isl1, Lin28b, and Hmga2, the expression of Trim71 was highest prior to differentiation (E13.5). However, while Isl1, which continues to be expressed in supporting cells, only modestly declined during differentiation, the expression of Trim71, Lin28b, and Hmga2 sharply declined at the onset of differentiation (E13.5-E14.5), and Trim71 was near undetectable in differentiating cochlear epithelia (E15.5- E16.5) (Figure 1D).

Figure 1. Trim71 is expressed in otic and neuro-sensory progenitor cells.

(A) Trim71 and Lin28b mRNA expression in otic and periotic cells stages E9.5, E11.5, and E13.5. Violin plots were generated from previously published scRNA-sequencing data sets. Tbx2, Lfng, and Sox2 served as marker genes. (B) Otic and peri-otic TRIM71 protein expression in stage E9.5 mouse embryo. (C) Otic and peri-otic Trim71 and Lfng mRNA expression in stage E9.5 mouse embryo. (D) RT-qPCR of Trim71, Lin28b, Hmga2, Isl1, and Atoh1 in cochlear epithelia isolated from stage E13.5, E14.5, E15.5, and E16.5 mouse embryos. Bar graph showing mean ± SD, n=3 biological replicates per group.

Trim71 knockout during early otic development impairs hearing in mice

We established Trim71 mutant mouse models to examine the role of TRIM71 in inner ear development and hearing. Non-conditional Trim71 homozygous mutant mice have severe neural tube closure defects and die mid-gestation (8, 14, 26). To circumvent embryonic lethality, we employed a conditional KO approach using Trim71 floxed (Trim71^f/f) mice. In these mice, exon 4, which codes for the NHL domain, is flanked by LoxP sites, which, upon Cre-mediated recombination, renders the Trim71 gene non-functional (14) (Figure S3A). To achieve broad yet stage-specific Trim71 deletion, we used a doxycycline (dox) inducible Cre strategy (R26^rtTA*M2 and TetO-Cre). Additionally, we generated Trim71 mutant mice (Trim7^R595H/Δ), where one floxed allele was replaced by an allele that carries the murine homolog (R595H) of the human CH and HL-associated missense mutation in TRIM71 (R608H) (19) (Figure S3B). We validated our loss-of-function strategy in pilot experiments where dox was administrated at E5.5 (Figure S3C).

As expected, dox administration at E5.5 (pre-gastrulation) resulted in severe neural tube closure defects in Trim71 KO (Trim7^Δ/Δ) or Trim71 mutant (Trim71^R595H/Δ) embryos (Figure 2D). A small fraction (<10%) of Trim71 heterozygous mutant embryos (Trim71^R595H/+) displayed neuro-developmental defects, as reported by (19). However, the majority of Trim71 heterozygous mutant embryos developed normally and were indistinguishable from their wild-type littermates (Figure S3D). Furthermore, dox administration at E8.5 or later allowed Trim71 KO and Trim71 mutant mice to survive into adulthood (Figure 2A). At one month, Trim71 KO (Trim71^Δ/Δ) and Trim71 mutant (Trim71^R595H/Δ) mice weighed significantly less than Trim71 heterozygous mutant (Trim71^R595H/+) or control (Trim71^f/f) mice (Figure 2B). Additionally, Trim71 KO and Trim71 mutant mice were smaller overall and had notably shorter tails than Trim71 heterozygous mutant or control mice (Figure 2A).

Figure 2. Early embryonic loss of Trim71 impairs hearing in mice.

Doxycycline (dox) inducible Cre strategy (R26rtTA*M2; TetO-Cre) was used in mice to conditionally delete Trim71 during early otic development (E8.5-E10.5). (A) P30 Trim71^Δ/Δ and Trim71^R595H/Δ mice that received dox starting at E8.5 are reduced in size compared to Trim71^f/f (WT, control) and Trim71^R595H/+ mice. (B) Quantification of the body weight in (A) (n=4 animals, two independent experiments). (C-H) Auditory Brainstem Response (ABR) thresholds were measured in 1-month old mice. (C) ABR thresholds were not elevated (normal hearing) in Trim71^R595H/+ mice compared to Trim71^+/+ wild type littermates. (D) ABR thresholds were not elevated in Trim71^+/Δ mice compared to Trim71^f/f (control) littermates. Dox was administered starting at E8.5 (n=3 animals per group). (E) ABR thresholds were elevated in Trim71^Δ/Δ mice across all frequencies compared to Trim71^f/f control littermates after dox was administered starting at E8.5 (n=9 animals in control group and n=8 animals in Trim71^Δ/Δ group, three independent experiments). (F) ABR thresholds were elevated at high frequency (32 kHz) in Trim71^Δ/Δ mice compared to control Trim71^f/f littermates after dox was administered at E9.5 (n=5 animals in control group, n=4 animals in Trim71^Δ/Δ group). (G) ABR thresholds were normal in Trim71^Δ/Δ mice compared to Trim71^f/f control littermates after dox was administered at E10.5 (n=3 animals in control group, n=3 animals in Trim71^Δ/Δ group). (H) ABR threshold was significantly elevated at high frequency in Trim71^Δ/R595H mice compared to Trim71^f/f control littermates after the dox was administered at E10.5 (n=4 animals per group). One-way ANOVA with Tukey’s correction was used to calculate P values in B. Two-tailed, unpaired Student t test was used to calculate P values in C-H.*P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. All data are collected from at least two independent experiments.

Next, we used auditory brainstem response (ABR) threshold measures to characterize the hearing of Trim71 homozygous and heterozygous KO and mutant mice, as well as their control littermates (Trim71^f/f and Trim71^+/+). Our analysis revealed that Trim7^{R595H /+} mice had similar ABR thresholds compared to their wild-type littermates, indicating normal hearing (Figure 2C). Similarly, Trim71 heterozygous KO mice, with one floxed allele (Trim71^+/Δ), exhibited normal hearing after receiving dox at E8.5 (Figure 2D). In contrast, Trim71 homozygous KO mice (Trim7^Δ/Δ) had elevated ABR thresholds at low (4, 8 kHz), mid (16, 24 kHz), and high frequencies (32 kHz) compared to control littermates (Trim71^f/f) after receiving dox at E8.5 (Figure 2E). Conditional deletion of Trim71 after dox administration at E9.5 still impaired hearing at high frequency (Figure 2F, 32 kHz), but homozygous deletion of Trim71 floxed alleles after stage E10.5 had no adverse effect on hearing (Figure 2G). Trim71 mutant mice that carried the R595 mutant allele and a floxed allele (Trim71^△/R595H) displayed high-frequency hearing loss after receiving dox at E10.5, indicating that the R9595 missense mutation impairs hearing (Figure 2H). In summary, our findings show that TRIM71 function during early otic development (E8.5-E10) is essential for proper hearing.

Trim71 knockout during early otic development results in premature pro-sensory cells cell cycle exit and differentiation

To gain insights into the genesis of the observed hearing deficits, we analyzed whether loss of TRIM71 during early otic development (dox starting between E8.5-E10.5) alters the behavior of auditory pro-sensory cells. Cochlear (auditory) hair cells and supporting cells derive from a pool of SOX2-expressing pro-sensory cells within the developing cochlear epithelial duct (27, 28). Cochlear pro-sensory cells exit the cell cycle in an apical-to-basal gradient that in mice starts at around E12.5 and peaks at E13.5 (29). Following terminal mitosis at around E14.0, pro-sensory cells located at the cochlear mid-base start to differentiate, initiating a basal-to-apical wave of differentiation that reaches the cochlear apex at around E18.5 (25, 30). To capture the terminal mitosis of pro-sensory cells, we administered a single pulse of EdU at E12.5 (apex final mitosis) or E13.5 (mid-base final mitosis) and analyzed EdU incorporation in hair cells and supporting cells (Deiter’s cells and pillar cells) five days later.

We found that an EdU pulse at E12.5 labeled 2-fold fewer apical hair cells (Figure 4A, B) and apical supporting cells (Figure 3A, C) in E17.5 Trim71 KO (Trim71^Δ/Δ) mice than control littermates, indicating that in the absence of TRIM71 pro-sensory cells exit the cell cycle prematurely. Furthermore, we found that the length of the cochlear sensory epithelium in Trim71 KO mice was about 20% shorter compared to control littermates (Figure 3D). A similar phenotype was observed with Trim71 mutant (Trim71^R595H/Δ) mice. We found that an EdU pulse at E13.5 labeled 2-fold fewer hair cells in the cochlear base (Figure 3E, F) and more than 2-fold fewer supporting cells in the cochlear mid and apical region in Trim71 mutant mice than control littermates (Figure 3E, G). However, loss of TRIM71 function did not disrupt the apex-to-base gradient of pro-sensory cell cycle exit, and in both control and Trim71 mutant mice, EdU incorporation in hair cells and supporting cells was highest in the cochlear base and lowest in the apex (Figure 3E-G). These results are consistent with our recent findings that cochlear epithelial cells isolated from stage E13.5 Trim71 KO (Trim7 ^Δ/Δ) or Trim71 mutant (Trim71^△/R595H ) mice (dox E5.5) proliferate at a significantly lower rate than cochlear epithelial cells isolated from control littermates when cultured as organoids in vitro (31).

Figure 4. Loss of Trim71 increases the expression of genes involved in sensory organ differentiation and extracellular matrix organization.

(A) Experimental design: Cochlear epithelia were isolated from E13.5 Trim71^Δ/Δ embryos and Trim71^f/f (control) littermates following dox administration at E8.5. (B) RT-qPCR was used to validate the conditional ablation of Trim71 (exon 4) and its associated transcripts in Trim71^Δ/Δ cochlear epithelia (n=3 animals per group, two independent experiments). (C) RT-qPCR reveals elevated Atoh1 transcript levels in Trim71^Δ/Δ cochlear epithelia compared to control (n=3 animals per group, two independent experiments). (D) Volcano plot of RNA-seq data. Plotted is the beta-value (x-axis) versus −log10 q-value (y-axis). Transcripts that are significantly upregulated in Trim71^Δ/Δ cochlear epithelia are marked in red dots, and transcripts that are significantly downregulated are marked in blue dots. (E) Intersectional analysis to identify hearing loss genes that are altered by Trim71 knockout. (F) Gene ontology enrichment analysis. Biological processes and pathways associated with genes that are upregulated in E13.5 Trim71^Δ/Δ cochlear epithelia are ranked by adjusted P value. (G) Immunoblots for p-SMAD2/3 and β-actin (loading control) using protein lysates of acutely isolated cochlear sensory epithelia from E13.5 Trim71^Δ/Δ embryos and control littermates (n=3, two independent experiments). Student t test was used to calculate the P values in (B) and (C). *P < 0.05, **P < 0.01, P >0.05 is deemed not significant (n.s.).

Figure 3. Loss of Trim71 leads to premature cell cycle withdrawal and differentiation of auditory pro-sensory cells.

(A-D) The timed-mated pregnant dam was fed doxycycline (dox)-containing feed starting at E8.5, a single EdU pulse was given at E12.5 and EdU (green) incorporation in Trim71 knockout (Trim71^Δ/Δ) mice and control (Trim71^f/f) littermates was analyzed at E17.5. Cochlear length of Trim71^Δ/Δ mice and control (Trim71^f/f) littermates was analyzed at P0. (A) Representative confocal images of EdU incorporation (red) in hair cells (MYO7a, magenta) and supporting cells (SOX2, blue) at the cochlear base, mid and apex in E17.5 Trim71^Δ/Δ mice and control littermates are shown. (B) Quantification of hair cell-specific EdU incorporation in (A) (n=3 animals per group, two independent experiments). (C) Quantification of supporting cell-specific EdU incorporation in (A). Deiters cells (DC) and pillar cells (PC) are quantified (n=3 animals per group, two independent experiments). (D) Length of cochlear sensory epithelia in Trim71 knockout (Trim71^Δ/Δ) mice is shorter than in control mice (n=5 animals per group, two independent experiments). (E-G) The timed-mated pregnant dam was fed dox-containing feed starting at E10.5, EdU was injected at E13.5, and cochlear tissue of Trim71^Δ/R595 mice and Trim71^f/f (control) littermates was collected and analyzed at E18.5. (E) Representative confocal images of EdU incorporation (red) in hair cells (MYO7a, magenta) and supporting cells (SOX2, blue) at the cochlear base, mid, and apex in E18.5 Trim71^Δ/R595 mice and control littermates are shown. (F) Quantification of hair cell-specific EdU incorporation in (E) (n=3 animals per group, three independent experiments). (G) Quantification of supporting cell-specific EdU incorporation in (E) (n =3 animals per group, three independent experiments). (H) Representative confocal images of neuronal innervation pattern (NFH, red) at the cochlear base, mid and apex of Trim71^Δ/R595 mice and control littermates are shown. (I) Representative confocal images of actin-rich hair cell stereocilia (phalloidin, grey) at the cochlear base, mid, and apex of Trim71^Δ/R595 mice and control littermates. (J) Quantification of hair cells in (I) (n=3 animals per group, three independent experiments). Student t test was used to calculate P values. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

How far the differentiation of hair cells and supporting cells has advanced can be judged by the morphology of their actin-rich apical protrusions (termed stereocilia) and by the pattern of SGN innervation, respectively. Two distinct subtypes of SGNs innervate inner and outer hair cells (reviewed in (32)). A single inner hair cell is innervated by multiple type I SGNs, whereas a single type II SGN, after a short turn towards the cochlear base, forms synapses with multiple outer hair cells. The type II SGN axon turning is guided by planar cell polarity signaling cues established by supporting cells (33). By E17.5/ E18.5, hair cells located at the base of the cochlea have formed stereocilia, and outer hair cells are innervated by type II SGNs. In contrast, hair cells at the apex have not yet formed stereocilia, and type II SGNs have not yet oriented toward the base or reached their targets.

Our analysis of hair cell and type II SGN morphology in stage E18.5 Trim71 mutant (Trim71^R595H/Δ) and control littermates revealed that the innervation of outer hair cells at the base, mid, and apex was more advanced in Trim71 mutant cochleae compared to control cochleae (Figure 3H). Furthermore, the development of hair cell stereocilia in the basal, mid and apical portion of Trim71 mutant cochleae was more advanced than that of corresponding hair cell stereocilia in control cochleae (Figure 3I). Likely due to the shift in the timing of cell cycle exit and differentiation, we found that the hair cell density at the cochlear base and mid-turn was mildly reduced in E18.5 Trim71 mutant mice compared to control littermates (Figure 3J). No changes in the timing of pro-sensory cell cycle exit and differentiation were observed in Trim71 heterozygous mutant (Trim71^R595H/+) mice (Figure S4), which have normal hearing. To determine whether deletion of Trim71 after E10.5 would lead to precocious cell cycle exit and differentiation, we used Emx2^Cre/+ to delete Trim71 specifically in cochlear epithelial cells, including pro-sensory cells, around E11.5-E12.0 (34). We found that Emx2^Cre/+ mediated deletion of Trim71 did not affect the timing of pro-sensory cell cycle exit (Figure S5A, B), nor did it impact the timing of outer hair cell innervation or the timing of stereocilia formation (Figure S5C, D). Consistent with our data showing that deletion of Trim71 after E10.5 (Figure 2 G) does not affect hearing, we found that Emx2^Cre/+ induced Trim71 KO mice (Emx2^Cre/+Trim71^Δ/Δ) did not experience significant hearing loss compared to control littermates (Trim71^+/+, Trim71^f/+ or Trim71^f/f), Trim71 heterozygous KO mice (Emx2^Cre/+Trim71^Δ/+), or mice that carried the Emx2^Cre/+ transgene (Figure S5E).

TGFβ-type signaling is upregulated in Trim71 deficient cochlear progenitor cells

To investigate the molecular mechanisms leading to the premature differentiation of Trim71-deficient auditory pro-sensory cells, we administered dox at E8.5 and harvested cochlear epithelia from Trim71 KO (Trim71^Δ/Δ) mice and their control littermates (Trim71^f/f) at E13.5. We then isolated RNA for RT-qPCR and bulk RNA-sequencing (seq) experiments (Figure 4A). First, we validated Trim71 KO using RT-qPCR. As expected, cochlear epithelia from E13.5 Trim71^Δ/Δ mice lacked Trim71 transcripts containing the critical exon4 but produced truncated, non-functional transcripts containing upstream exon2 and 3 (Figure 4B). Furthermore, consistent with early differentiation of hair cells, we found that the Atoh1 transcript, which is expressed in nascent hair cells, was significantly upregulated in cochlear epithelia from E13.5 Trim71^Δ/Δ mice compared to their control littermates (Figure 4C). Analysis of RNA-seq data identified 398 differentially expressed genes (DEG) (p-value <0.01) when comparing the transcriptome of Trim71^Δ/Δ cochlear epithelial cells to those of the controls (Figure 4D) (Dataset S1). Over three-quarters (340) of DEGs were upregulated in Trim71 KO cochlear epithelial cells, which aligns with TRIM71's known inhibitory function in gene expression.

We next investigated whether the expression of genes associated with hearing loss (HL) was affected (up or downregulated) by Trim71 deletion. We curated a list of 190 HL genes using the Walls WD, Azaiez H and Smith RJH Hereditary Hearing Loss Homepage (https://hereditaryhearingloss.org/ ) (Dataset S2). In total, we found that 15 HL genes were differentially expressed in Trim71^Δ/Δ cochlear epithelial cells compared to control, with three being downregulated and twelve being upregulated (Figure 4E).

To gain insights into the biological processes/ pathways that may be affected by Tim71 KO, we conducted gene ontology enrichment analysis for up and downregulated genes (35). ‘Trans-synaptic signaling,’ ‘embryonic morphogenesis’, and ‘sensory perception of sound’ were among the biological processes that were significantly overrepresented in the list of downregulated genes (Dataset S3). Among the top 10 pathways/ biological processes that were significantly overrepresented in the list of upregulated genes were developmental processes such as ‘sensory organ’ and ‘ear development’ as well as processes associated with TGFβ-type signaling, including ‘extracellular matrix organization’, ‘connective tissue development’ and ‘regulation of cellular response to growth factor stimuli’ (Figure 4F) (Dataset S4). Notably, Inhba, which encodes for Activin A, a key ligand for Activin receptor signaling, and Tgfbr2, which encodes for TGFβ type II receptor (TGFBR2) were both upregulated. Activation of TGFβ type I and II receptors or Activin I and II receptors leads to the phosphorylation of effector proteins SMAD2 and SMAD3. After phosphorylation, these proteins translocate into the nucleus, where they bind to specific DNA- binding elements and activate/repress transcription (reviewed in (36)). To determine whether loss of Trim71 leads to an over-activation of TGFβ type signaling, we analyzed phospho-SMAD2/3 levels in E13.5 cochlear epithelial protein lysates obtained from control and Trim71^Δ/Δ embryos (n=3) (Figure 4G). We found that phospho-SMAD2/3 levels were higher in Trim71 deficient cochlear epithelia than in control, suggesting that TRIM71 acts to repress premature activation of TGFβ-type signaling in cochlear epithelial cells.

Disruption of Activin/TGFβ signaling delays cell cycle exit and differentiation of cochlear pro-sensory cells

We have recently shown that TGFβ2/TGFβ-RI signaling restricts cell cycle reentry and proliferation of early postnatal cochlear supporting cells in organoid culture (37). However, whether TGF-β signaling limits pro-sensory cell proliferation during cochlear development is currently unknown. To characterize the role of TGFβ signaling in the developing cochlea, we generated Tgfbr1 conditional KO mice using Pax2-Cre, which is transiently expressed in otic progenitors (38). To determine the timing of terminal mitosis of pro-sensory cells, we administered a single pulse of EdU at E13.5 and analyzed EdU incorporation in hair cells and supporting cells (Deiter’s cells and pillar cells) at E18.5 in control Tgfbr1 KO mice. We found that in Tgfbr1 KO mice, the percentage of EdU-labeled hair cells (Figure S6A, B) and Deiter’s cells (supporting cell sub-type) (Figure S6A, D) in the cochlear apex was significantly higher than in control littermates. However, this mild delay in pro-sensory cell cycle exit did not increase the density of cochlear hair cells in the Tgfbr1 deficient mice compared to control littermates (Figure S6C).

In previous research, we showed that the conditional knockout of Inhba using Pax2-Cre results in a mild delay in hair cell differentiation without affecting the timing of cell cycle withdrawal (39). To determine whether combined loss of Tgfbr1 and Inhba exacerbates the defects in cell cycle withdrawal and differentiation observed in Tgfbr1 and Inhba single KO mice, we generated Inhba and Tgfbr1 double KO (DKO) mice using Pax2-Cre. EdU pulse-chase experiments, which involved a single EdU pulse at E13.5, revealed a significant increase in the number of EdU⁺ hair cells (Figure 5A, B) in the cochlear base and apex of E18.5 Tgfbr1-Inhba DKO mice compared to control. The mild delay in cell cycle exit did not significantly increase hair cell density in Tgfbr1- Inhba DKO mice compared to control littermates (Figure 5C). Furthermore, analysis of the outer hair cell innervation pattern (Figure 5D) and morphology of stereocilia (Figure 5E) revealed less mature phenotypes in Tgfbr1-Inhba DKO mice compared to control littermates. In sum, these findings suggest that Trim71 maintains prosensory cells in a proliferative and undifferentiated state, at least in part by repressing TGFβ-type signaling.

Figure 5. Loss of Tgfbr1 and Inhba delays cochlear pro-sensory cell cycle exit and differentiation.

(A-E) Timed pregnant dams received EdU at E13.5, and cochlear tissue from Tgfbr1-Inhba double knockout mice (DKO) (Pax2-Cre; Tgfbr1^f/f; Inhba^f/f) and control littermates (Tgfbr1^f/f; Inhba^f/f) were collected and analyzed at E18.5. (A) Representative confocal images of EdU (green) incorporation in hair cells (MYO7A, magenta) and supporting cells (SOX2, blue) at cochlear apex, mid, and base in Tgfbr1-Inhba DKO mice and control littermates. (B) Quantification of EdU incorporation in hair cells in (A) (n=3 animals per group, three independent experiments). (C) Quantification of hair cell density in (A) (n=3 animals per group, three independent experiments). (D) Representative confocal images of innervating neurons (NFH, magenta) at the cochlear base, mid, and apex of Tgfbr1-Inhba DKO mice and control littermates. (E) Representative confocal images of hair cell stereocilia (phalloidin, grey) at the cochlear base, mid, and apex of Tgfbr1-Inhba DKO mice and control littermates. Student t test was used to calculate the P values in B and C.

Loss of Trim71 results in hair cell stereocilia and synaptic defects

HL-patients with mono-allelic missense mutations in TRIM71 have, in addition to various middle and inner ear abnormalities, brain malformations (hydrocephaly). To determine whether brain malformations occur in our Trim71 deficient mouse models, we imaged adult brains (P30) Trim71 KO (Trim71^△/△) and Trim71 mutant (Trim71^R595H/△) mice and their control littermates using magnetic resonance imaging (MRI). Consistent with previous reports, we found that deletion of Trim71 at E8.5 or thereafter does not cause hydrocephaly (19) (Figure S7A). To characterize inner ear/ middle ear morphology in adult (P30) Trim71 KO (Trim71^△/△) and Trim71 mutant (Trim71^R595H/△) mice and their control littermates, we used Computed Tomography (CT) imaging. Our analysis revealed that Trim71 KO and Trim71 mutant mice, like their control littermates (Trim71^+/+ or Trim71^R595H/+), had no obvious inner ear/ middle ear malformations (Figure S7B). To further address the underlying mechanism of hearing loss, we analyzed whether hair cells degenerate in Trim71 KO (Trim71^△/△) and Trim71 mutant (Trim71^R595H/△) mice or their stereocilia are lost or malformed. Our analysis revealed there was no reduction in the number of outer hair cells or inner hair cells in Trim71 KO and Trim71 mutant mice compared to their control littermates (Figure S8A, C, D).

We next analyzed whether loss of Trim71 alters the synaptic connections of hair cells with innervating type I and type II SGNs. To visualize hair cell-specific pre-synaptic complex, we used pre-synaptic marker protein CtBP2/RIBEYE (40). We found that the number of CtBP2 puncta per outer hair cells (Figure 6A, B) and inner hair cells (Figure 6A, C) was significantly reduced in the mid and basal portion of the cochlea but unchanged in cochlear apex in Trim71^△/△ and Trim71^R595H/△ mice compared to Trim71^f/f or Trim71^R595H/+ control littermates. Neurofilament H staining revealed a severe reduction in neuronal innervation of the outer hair cell region in Trim71^△/△ and Trim71^R595H/△ mice compared with Trim71^f/f or Trim71^R595H/+ control littermates (Figure S8A, B). Together, these data suggest that the loss of hearing observed in Trim71^△/△ and Trim71^R595H/△ mice may result from defects in synaptic transmission and or loss of neuronal modulation of outer hair cell function.

Figure 6. Loss of Trim71 results in fewer hair cell synapses.

Dox was administered starting at E8.5, and Trim71^Δ/Δ and Trim71^R595H/Δ mice and their control littermates Trim71^f/f (Ctrl), Trim71^R595H/+ were analyzed at P30. (A) Representative confocal images of pre-synaptic puncta (CTBP2, green) in outer hair cells (OHC) and inner hair cells (IHC). Images were taken at the cochlear apex, mid, and base. (B) Pre-synaptic puncta in outer hair cells are reduced in Trim71^Δ/Δ mice. Quantification of CTBP2 puncta per outer hair cell (OHC) (n=3–4 for Ctrl, Trim71^R595H/+, Trim71^Δ/Δ, n=2 for Trim71^R595H/Δ). (C) Pre-synaptic puncta in inner hair cells are reduced in Trim71^Δ/Δ and Trim71^R595H/Δ mice. Quantification of CTBP2 puncta per inner hair cell (n=3–4 for Ctrl, Trim71^R595H/+, Trim71^Δ/Δ, n=2 for Trim71^R595H/Δ). One-way ANOVA was used to calculate the P values. P < 0.05, **P < 0.01 and ***P < 0.001.

Discussion

Congenital hearing loss is the most common sensory disability, with an estimated three-quarters of cases having a genetic origin (41). Our study identifies TRIM71 as a novel gene linked to hearing loss (HL). TRIM71 missense mutations were first associated with a group of patients suffering from congenital hydrocephaly (CH) (17). Upon clinical reevaluation, it was determined that pathogenic variants in TRIM71 are responsible for a syndromic phenotype that includes various features such as developmental delay, hearing loss, limb defects, craniofacial abnormalities, dysmorphism, and congenital cardiovascular defects with or without ventriculomegaly. Hereditary HL exhibits significant clinical and genetic heterogeneity (reviewed in (42)). Hundreds of syndromic associations have been documented in which hearing loss coexists with abnormalities or malformations in other organs (43).

Most of the genes that are implicated in either isolated or syndromic HL code for proteins that regulate the development and or function of hair cells. In the case of syndromic HL, these proteins may also play roles in the development of semicircular canals, the middle ear, and the temporal bone (e.g. EYA1, SIX1, POU3F4, CHD7) (44). Some of these syndromes can present similar clinical features, and only genetic analysis can provide a definitive diagnosis. A clinical diagnosis of CHARGE syndrome was initially given to patient 1 due to the association of hearing loss with malformations of the external ear, middle ear, and lateral semicircular canals (45). Identifying a heterozygous pathogenic variant in TRIM71 has allowed for the invalidation of this clinical diagnosis, leading to improved prognostic and genetic counseling for the patient.

Generating Trim71 KO (Trim71^Δ/Δ) and Trim71 mutant mice that harbor the murine homolog of the human HL-associated missense mutation (Trim71^R595H/Δ), we demonstrate a causative link between TRIM71 dysfunction and hearing loss. We find that loss of TRIM71 function around~ E9-E10, which correlates with the peak of murine Trim71 expression in otic neural-sensory progenitors, impairs hearing, without causing structural inner ear or middle defects. Trim71 homozygous mutant mice, which carry the R595H missense mutation and a conditional knockout allele, exhibited similar developmental defects as Trim71 homozygous KO mice. However, subtle differences in hearing sensitivity were observed. Specifically, Trim71^R595H/Δ mice that received dox at E10.5 demonstrated high-frequency hearing loss, while the same timing of deletion in Trim71^{Δ/ Δ} mice did not affect hearing. The R595H missense mutations in TRIM71 (R608H in humans) reside within TRIM71’s NHL domain, which previous studies have shown disrupts TRIM71’s RNA-binding ability without altering TRIM71’s ubiquitin ligase activity (19). Thus, the observed subtle differences in hearing sensitivity between Trim71^R595H/Δ and Trim71^{Δ/ Δ} mice may be due to the preservation of TRIM71’s ubiquitin ligase activity in Trim71^R595H/Δ mice. Only a small subset (<10%) of Trim71 heterozygous mutant (Trim71^R595H/+) mice manifest neuro-developmental defects (19). The only sporadic occurrence of neuro-developmental defects may explain why none of the Trim71 heterozygous mutant (Trim71^R595H/+, n=6) or heterozygous knockout mice (Trim71^Δ/+, n=3) tested for hearing showed hearing deficits and suggests the existence of modifier gene(s).

The proper timing of auditory sensory development is disrupted by early (~E9-E10) Trim71 deletion in Trim71^R595H/Δ and Trim71^{Δ/ Δ} mice. Cell cycle withdrawal and differentiation of auditory pro-sensory cells occur at precise times and follow opposing longitudinal gradients. Experiments conducted in murine and avian auditory organs have revealed that the RNA-binding protein LIN28B and members of let-7 miRNA play a significant role in timing the exit from the cell cycle and differentiation of pro-sensory cells (2, 3). However, the precise mechanism remained elusive. In this study, we demonstrate that TRIM71, a let-7 target and member of the Lin28-let-7 pathway, is essential for preventing precocious pro-sensory cell cycle exit and differentiation. Our transcriptomic analysis revealed that cochlear progenitors in Trim71^{Δ/ Δ} mice expressed Activin A coding gene Inhba and TGFBRII coding gene Tgfbr2 at significantly higher levels than their control counterparts. Our previous research has shown that Activin A acts as a pro-differentiation signal in the developing cochlea (39). However, while overexpression of Activin antagonist follistatin led to severe delay in pro-sensory cell cycle exit and hair cell differentiation, the differentiation defect in Inhba knockout mice was relatively mild. Also, we did not observe any defects in the timing of pro-sensory cell cycle exit in Inhba knockout mice, suggesting that other members of the TGFβ-family are compensating for the loss of Inhba. Indeed, our data from Tgfbr1 single KO and Tgfbr1-Inhba DKO experiments indicate that TGF-β signaling cooperates with Activin signaling in driving pro-sensory cells out of the cell cycle and in promoting their differentiation into hair cells and supporting cells. In addition to Inhba and Tgfbr2, several other known ‘pro-differentiation’ genes were upregulated in Trim71 deficient cochlear progenitor cells. Notable among them are Nfix and Ebf3, which encode transcription factors that promote neurogenesis and gliogenesis (46–48). Future studies are warranted to address their function in cochlear sensory differentiation.

Besides the ear-related developmental defects, we found that Trim71 knockout (Trim71^Δ/Δ) mice and Trim71 mutant (Trim71^R595H/Δ) mice weighed less and had smaller bodies than Trim71 heterozygous mutant/knockout or wild-type littermates. Most strikingly, both the Trim71 KO and Trim71 mutant mice had unusually short tails. The small stature and the shortened tails resemble a phenotype reported for let-7g overexpressing mice, which lack caudal vertebrae due to a shortened developmental time window for axial elongation and somitogenesis (49).

How does early embryonic Trim71 deficiency lead to hearing loss in mice? In humans, Trim71 missense mutations lead to both sensorineural and conductive hearing loss, which is associated with malformations of the middle ear and temporal bone. However, MRI and CT imaging of the inner ear temporal bone and middle ear bones of adult Trim71^R595H/Δ or Trim71^Δ/Δ mice (dox E8.5) showed no defects. This suggests that the observed hearing loss in Trim71 homozygous mutant/ KO mice is due to the dysfunction of hair cells and/or neurons. While the precise mechanism remains unresolved, our transcriptomic and morphological analysis have identified several candidates. We found that in Trim71 deficient cochlear epithelial progenitors the expression of mesenchymal transcription factors (e.g., Pou3f4, Snai2, and Zeb2) is upregulated. The vital role of otic mesenchyme for cochlear development and hearing is best exemplified by the mesenchymal-specific transcription factor POU3F4. POU3F4 regulates temporal bone development (50), but also spiral ganglion fasciculation (51) and mutations in POU3F4 (DFNX2) are the leading cause of non-syndromic X-linked hearing loss in humans (52, 53).

Furthermore, we found that two well-known HL genes, Espn (DFNB36) (54) and Atp2b2 (DFNA82) (55), were downregulated in Trim71 deficient cochlear progenitor cells. Both ATP2B2 (also referred to as PMCA2) and ESPN proteins localize to hair cell stereocilia, where they regulate Ca2+ homeostasis (56) and stereocilia morphology (57) respectively. Further studies are needed to determine whether hair cell stereocilia function is altered by the loss or dysfunction of TRIM71. Synaptic transmission and signal amplification defects are other possible causes for the loss of hearing observed in Trim71 homozygous mutant/ KO mice. Specifically, we found that inner and outer hair cells in Trim71 homozygous mutant/ KO mice had fewer pre-synaptic puncta than their control littermates. Additionally, in a subset of Trim71 homozygous mutant/ KO mice, neuronal innervation of the outer hair cell region- which is essential for signal amplification and modulation- was nearly absent.

It is noteworthy to consider that the changes in the timing of terminal mitosis and differentiation within the auditory sensory epithelium may have led to the hearing loss observed in Trim71 homozygous mutant/ KO mice. Alternatively, TRIM71 might serve additional functions during early otic development that are not related to its effect on developmental timing. Future studies are needed to investigate TRIM71's role during the early stages of otic development.

Materials and Methods

Patient data

Whole-exome sequencing was performed on three members of patient 1’ family, using the method previously described (58). Patient 2’s genetic testing has been described (59). Patient 1 and patient 2 were clinically examined by Ear Nose and Throat physicians and clinical geneticists. They had air and bone conduction pure tone audiograms, and hearing impairment was defined according to the recommendations of the International Bureau for Audiophonology (BIAP) https://www.biap.org/en/recommandations/recommendations/tc-02-classification. MRI and CT scans of the internal auditory canal, labyrinth, and middle ear were performed and analyzed using methods previously described (60, 61). This study was approved by the Institutional Review Board at Massachusetts General Hospital and Yale University. All genetic testing was obtained with written informed consent prior to collection from patients.

Mouse breeding and genotyping

All experiments and procedures were approved by the Johns Hopkins University Institutional Animal Care and Use Committees protocol, and all experiments and procedures adhered to National Institutes of Health-approved standards. Trim71^R595H mutant mice (RRID: MGI:7661051) were provided by Kristopher Kahle, Harvard Medical School. Trim71 floxed mice (RRID: MGI:7661049) were provided by Waldemar Kolanus, Universität Bonn. Emx2^Cre/+ mice were obtained from Shinichi Aizawa, RIKEN (RRID: MGI: 3579416). Pax2-Cre transgenic mice (RRID: MGI:3046196) were obtained from Andrew Groves, Baylor College. Inhba floxed mice (RRID: MGI:3758877) were obtained from Martin Matzuk, Baylor College. Tgfbr1 floxed (RRID: IMSR_JAX:028701), TetO-Cre (RRID: IMSR_JAX:006234) and R26^rtTA*M2 (RRID: IMSR_JAX:006965) mice were purchased from Jackson Laboratories (Bar Harbor, ME). To induce Cre expression, doxycycline (dox) was delivered to time-mated females via ad libitum access to feed containing 2g/kg dox. Mice were genotyped by PCR as previously published. Genotyping primers are listed in Table S1. Mice of both sexes were used in this study. Embryonic development was considered as E0.5 on the day a mating plug was observed.

RNA extraction and RT-qPCR

Organoids were harvested using Cell Recovery Solution (Corning, no.354253). Total RNA from organoids/tissue was extracted using the miRNeasy Micro Kit (QIAGEN, no. 217084). mRNA was reverse transcribed into cDNA using the iScript cDNA synthesis kit (Bio-Rad, no. 1708890). Q-PCR was performed on a CFX-Connect Real Time PCR Detection System using SYBR Green Master Mix reagent (Thermo Fisher Scientific, no. 4385612). Gene-specific primers used are listed in Table S2. Rpl19 was used as the endogenous reference transcript. Relative gene expression was calculated using ^ΔΔCT method (62).

RNA sequencing and data analysis

To induce Trim71 deletion, a timed pregnant female received dox-containing feed starting at E8.5. Offspring were collected at E13.5, and cochlear epithelial ducts from individual Trim71 KO (Trim71 ^Δ/Δ) and control (Trim71^f/f) animals were isolated. After RNA extraction, samples were processed using Illumina's TruSeq stranded Total RNA kit, according to the manufacturer's recommendations. The samples were sequenced on the NovaSeq 6000, paired-end, 2×50 base pair reads. Kallisto (v0.46.1) (63) was used to pseudo-align reads to the reference mouse transcriptome and to quantify transcript abundance. The transcriptome index was built using the Ensembl Mus Musculus v96 transcriptome. The companion analysis tool Sleuth was used to identify differentially expressed genes (DEGs) (64). These lists were then represented graphically using sleuth, pheatmap and ggplot2 packages in R v1.3.1093. Gene identifier conversion, gene annotation, and enrichment analysis were conducted using Metascape (35).

Immunohistochemistry

Cochleae were fixed with 4% (vol/vol) paraformaldehyde in PBS (Electron Microscopy Sciences, no. 15713) overnight at 4 degrees. The P30 cochlea tissue was decalcified in 5% EDTA for 2 days following fixation. To permeabilize cells and block unspecific antibody binding, organoids/explants were incubated with 0.5% (vol/vol) TritonX-100/10% (vol/vol) FBS in 1× PBS for 30 min. Antibody labeling was performed according to manufacturer’s recommendations. Antibody information is listed in Table S3.

Cell Proliferation

EdU (25mg/kg, E10187, Thermo Fisher Scientific) was intraperitoneal injected on embryonic day E13.5 and cochleae were harvested at E18.5. EdU incorporation was detected using Click-iT Edu Alexa Fluor 555 imaging Kit (Thermo Fisher Scientific, no. C10338) following the manufacturer’s specifications.

Immunoblotting

Cochlear epithelia cells from E13.5 were lysed with RIPA buffer (Sigma-Aldrich, no. R0278) supplemented with protease inhibitor (Sigma-Aldrich, no.11697498001), phosphatase Inhibitor mixture 2 (Sigma-Aldrich, no. P5726) and phosphatase inhibitor mixture 3 (Sigma-Aldrich, no. P0044). Western blots were generated as described previously (65). Antibody information is listed in Table S3.

Auditory brainstem response (ABR) measurements

ABR measurements were performed following published procedures (66). Briefly, 4–5 week–old mice were anesthetized with Ketamine (100 mg/kg) and Xylazine (20 mg/kg) and placed on a heating pad inside a sound-attenuating chamber. Subdermal platinum needle electrodes (E2, GrassTechnologies, West Warwick, RI) were placed on the left pinna (inverting), vertex (non-inverting), and the leg muscle (ground). ABR stimuli were delivered through a free-field speaker (FD28D, Fostex, Tokyo, Japan) placed 10 cm away from the animal’s head. ABR stimuli generation and signal acquisition were controlled by a BiosigRz software interfacing TDT WS4 high-performance computer workstation. ABRs, pure tone bursts at 4, 8, 16, 24, and 32 kHz were presented at a rate of 21/s, with a 5 ms duration. Stimuli were presented in descending sound levels from the maximum speaker output level in 10dB increments. Responses were collected using a low impedance head stage (RA4L1; TDT), pre-amplified and digitized (RA4PA preamp; TDT), and sent to an RZ6 processing module. The signal was filtered (300–3000 Hz) and averaged over 512 presentations. The sound threshold was defined as the sound level at which the peak-to-peak ABR signal magnitude in wave I was two standard deviations above the average background noise level.

Magnetic resonance imaging (MRI) and computer tomography (CT)

Mouse heads were fixed with 4% (vol/vol) paraformaldehyde in PBS overnight and mouse structural CT imaging data were collected with TriFoil Small Animal Tomography Systems in the core facility of Johns Hopkins University. Mouse structural MR imaging data were collected with Bruker 11.7T in the core facility of Johns Hopkins University.

Statistical Analysis

All results were confirmed by at least two independent experiments. The sample size (n) represents the number of animals analyzed per group. Animals (biological replicates) were allocated into control or experimental groups based on genotype and/or type of treatment. To avoid bias, masking was used during data analysis. Data was analyzed using GraphPad Prism 8.0. Relevant information for each experiment, including sample size, statistical tests, and reported p-values, are found in the legend corresponding to each figure. In all cases, p-values ≤0.05 were considered significant and error bars represent standard deviation (SD).

Significances.

Timing is an essential feature of animal development, as defects in the temporal pattern of proliferation and differentiation can result in tissue malformations or dysfunction. Despite their significance, the mechanisms that regulate developmental timing remain poorly understood. Our study identifies the RNA-binding protein TRIM71 as a key component of the timing mechanism that controls the temporal pattern of auditory sensory development through antagonizing TGFβ-type signaling. Furthermore, we find that missense mutations in TRIM71 predispose humans to hearing loss. Generating Trim71 knockout mice that carry the HL-associated mutation, we show that TRIM71 function during early otic development is essential for hearing and its loss leads to reduced pre-synaptic terminals and impaired neuronal survival, identifying TRIM71 as a syndromic HL gene.

Data and materials availability

RNA sequencing data has been deposited in the Gene Expression Omnibus data repository under accession GSE281437