Adeno-Associated Virus Gene Replacement for Recessive Inner Ear Dysfunction: Progress and Challenges

Abstract

Approximately 3 in 1000 children in the US under 4 years of age are affected by hearing loss. Currently, cochlear implants represent the only line of treatment for patients with severe to profound hearing loss, and there are no targeted drug or biological based therapies available. Gene replacement is a promising therapeutic approach for hereditary hearing loss, where viral vectors are used to deliver functional cDNA to “replace” defective genes in dysfunctional cells in the inner ear. Proof-of-concept studies have successfully used this approach to improve auditory function in mouse models of hereditary hearing loss, and human clinical trials are on the immediate horizon. The success of this method is ultimately determined by the underlying biology of the defective gene and design of the treatment strategy, relying on intervention before degeneration of the sensory structures occurs. A challenge will be the delivery of a corrective gene to the proper target within the therapeutic window of opportunity, which may be unique for each specific defective gene. Although rescue of pre-lingual forms of recessive deafness have been explored in animal models thus far, future identification of genes with post-lingual onset that are amenable to gene replacement holds even greater promise for treatment, since the therapeutic window is likely open for a much longer period of time. This review summarizes the current state of adeno-associated virus (AAV) gene replacement therapy for recessive hereditary hearing loss and discusses potential challenges and opportunities for translating inner ear gene replacement therapy for patients with hereditary hearing loss.

1. Introduction:

Hearing loss is a common disorder affecting newborns and children. Clinically significant hearing loss is confirmed in nearly 3 out of 1000 children under the age of 4 in the United States each year^1-3. Approximately half of these affected children have a genetic cause of their hearing loss⁴. Currently, novel and rare variants in over 150 identified deafness genes have been reported⁵, and as a whole these individual genetic variants represent a prevalent problem⁶. Universal newborn hearing screening is mandatory in every US state to identify affected patients at birth, and genetic diagnosis typically takes place after a child has failed the initial screen and follow-up⁷. The majority (80%) of hereditary hearing loss involves autosomal recessive inheritance and causes sensorineural hearing loss, where the source of hearing loss is localized to specialized cells within the inner ear⁷. These cells are encased within the fluid-filled cochlea and include sensory hair cells, auditory neurons, and non-sensory cells involved in maintaining the unique structure and ionic environment of the inner ear⁸. Typically, those with autosomal recessive hearing loss are severely hearing impaired at birth and are born to hearing parents that are unknowing carriers of a loss-of-function mutation or deletion in the same gene⁷. A mutant allele of the gene inherited from each parent results in either limited or none of the protein product being produced by the cells, or the production of a non-functional protein.

The treatment options for patients with hereditary hearing loss are limited. While patients with mild-to-moderate hearing loss may benefit from hearing aids, those with severe to profound hearing loss usually require cochlear implantation to improve their auditory perception. While cochlear implants are revolutionary in their ability to improve speech perception for severely hearing-impaired patients, many aspects of hearing remain challenging for implant wearers, such as music perception, sound localization, and the ability to understand speech in a noisy environment⁹. Inner ear gene therapy has the potential of restoring normal cellular function in the cochlea, thereby improving auditory function. Over the past few years, several studies have successfully applied inner ear gene therapy to animal models of hereditary hearing loss^10-29. The most commonly used strategy for inner ear gene therapy is gene replacement, where normal copies of the cDNA of the target gene are delivered into the inner ear to compensate for the lack of functional gene product caused by the defective gene. Gene replacement is a useful strategy for treating loss-of-function mutations, and is the most straightforward approach in gene therapy. In contrast, dominant forms of hearing loss are often caused by mutations that lead to the production of an abnormal protein product which may interfere with normal cellular function³⁰. Therefore, variants associated with dominant forms of hearing loss are typically not good targets for gene replacement, but are good targets for strategies which inhibit the expression of the dominant mutant allele^19,31.

Gene replacement therapy has recently been approved by the FDA to treat a rare form of hereditary blindness called Leber’s congenital amaurosis (Luxturna, approved in 2017), as well as spinal muscular atrophy (Zolgensma, approved in 2019). The approval of these therapies represents a major milestone in the development of advanced precision therapeutics designed to treat inherited diseases. Both approved therapies are based on a gene delivery vector known as adeno-associated virus (AAV). AAV is a parvovirus that is dependent on adenovirus or herpesvirus co-infection for efficient replication^32,33. It is non-pathogenic in humans. Replication deficient AAV vectors are derived from the naturally occurring virus by replacing the viral DNA with an expression cassette containing the therapeutic transgene, which is then packaged into the viral capsid composed of AAV structural proteins³². This is accomplished by flanking the transgene cassette with the viral inverted terminal repeat (ITR) elements that act as DNA packaging signals for the AAV non-structural proteins. Recombinant vectors are generated and packaged in producer cells by providing a plasmid for the structural (Cap) and non-structural (Rep) viral proteins separate from the plasmid that contains the ITR flanked transgene cassette, as well as the adenovirus or herpes virus helper genes³². After target cell entry, AAV vector genomes typically persist in the nucleus in an extrachromosomal form, such as an episome or high molecular weight concatamers^34,35. An AAV vector-based approach for gene delivery has many advantages, such as high gene transfer efficiency, stable transgene expression, broad cellular tropism, low immunogenicity, and rare integration into the host DNA³². AAV vectors are produced from many different naturally occurring capsids, including serotypes 1-12 and over 100 primate isolates³⁶. The capsid proteins are semi-modular, with conserved regions that do not tolerate alteration and distinct variable regions amenable to protein engineering approaches that have been used to generate synthetic capsids with improved transduction efficiency and cell targeting³⁷. The main disadvantage of AAV is the fact that the packaging capacity for AAV is only approximately 4.2kb, assuming the minimal lengths for the ITR, promoter, and poly-A sequences³⁸. Despite this disadvantage, AAV remains one of the most commonly used viral vectors in human gene therapy clinical trials today due to its well characterized biosafety profile³⁹.

In this review, we will summarize the published studies on the application of AAV gene replacement therapy to various mouse models of autosomal recessive hereditary hearing loss. We will also discuss the opportunities and challenges that lie ahead for the translation of inner ear gene replacement therapy to patients with hereditary hearing loss.

2. Gene replacement studies for hereditary hearing loss (Figure 1, Table 1):

Figure 1: Variants in genes with diverse functions can cause hereditary hearing loss.

This figure illustrates the location of genes that have been successfully targeted for inner ear gene replacement therapy in the cochlea (figure modified from Faridi et al., 2018¹²³).

Table 1.

Summary of AAV-mediated inner ear gene replacemet thearpy for recessive inner ear dysfunction.

Study (authors and date published)	Therapeutic gene	Vector Used	Promoter	Mouse Strain	Age of Injection	Vector Copies per ear	Outcome
Akil et al., 2012	VGLUT3	AAVl	CBA	Vglut3 KO	P0-P3, P10	2.3e10 vg	50dB threshold improvement, best performers within 10dB of wild-type mice
yu, Wang et al., 2014	GJB2	AAV1	CB7	Gjb2 cKO	PD-P1	3e9 - 7.5e9 vg	No threshold improvement, improved cellular gap junction function
Crispino et al., 2017	GJB2	BAAV	CMV	Gjb6 KO	P4	1e9 vg	No threshld improvement
Chang, Wang et al., 2015	KCNQ1	AAVl	CB7	Kcnq1 KO	PO-P2	2.5e9 - 1.5e10 vg	50dB threshold improvement to within 20-30 dB of wild-type mice
Askew et al., 2015	TMC1, TMC2	AAVl	CBA	Tmc1 KO, TMC1 -Bth	P0-P2	1e10 vg	30dB threshold improvement at lower frequencies
Nist-Lund, Pan et al., 2019	TMC1	AAV1, Anc80	CMV	Tmc1 KO	P0-P2	1.4e11 vg	40dB threshold improvement, best performers within 10-20dB of wild-type mice. Improved breeding efficiency, litter survival and normal growth rate
Chien et al., 2015	WHRN	AAV8	CMV	Whrn wi/wi	P1-P5	5e10 vg	Increased stereocilia length. No auditory threshold improvement
Isgrig et al., 2017	WHRN	AAV8	CMV	Whrn wi/wi	P0-P5	1e10 vg	20dB threshold improvement, increased hair cell survival, improved vestibular function
Delmaghani et al., 2015	PJVK	AAVB	CAG, CB7	Pjvk KO	P3	2e10 vg	20-30dB threshold improvement in the low frequencies
Kim et al., 2016	MSRB3	AAV1	CMV	Msrb3 KO	E12.5	7e9 - 1e10 vg	70-80dB threshold improvement to within 10dB of wild type
Gyorgy, Sage et al., 2017	LHFPL5	AAV1, AAV9, exo-AAV	CBA	Lhfpl5 KO	P1-P2	2.7e9 vg	20-30dB threshold improvement in the mid frequencies
Pan, Askew, Galvin et al., 2017	USH1C	AAV1, Anc80	CMV	Ush1c c216G>A	P0-P1, P10-P12	1.6e9 - 2e9 vg	40-60dB threshold improvement, best performers close to wild-type at low-mid frequencies, reduced circling behavior
Emptoz et al., 2017	USH1G	AAVB	CB6	Ush1g KO	P0-P3	2e10 vg	20-30dB threshold improvement at low frequencies
Geng, Akil et al., 2017	USH3A	AAV2, AAV8	smCBA	Clrn1 cKO	P1-P3	2e10 vg	~40dB threshold improvement
Dulon et al., 2018	USH3A	AAVB	not reported	Clrn1 ex4−/−,cKO	P1-P3	1.2e10 vg	10-20dB threshold improvement in KO mice, 30-40dB improvement in cKO mice. Hearing loss progressed over time
Gyorgy, Meijer, Ivanchenko et al., 2019	USH3A	AAV-PHP.B	CBA	Clrn1 KO	P1	1.8e11 vg	~20dB threshold improvement at low frequencies, best performers had 50dB improvement
Kim et al., 2019	SLC26A4	AAV1	CMV	Slc26a4 KO, KI	E12.5	1.08e10 vg	60-70dB threshold improvement to within 10-20 dB of wild type, though hearing fluctuation and hearin loss occurs over time
Al-Moyed et al., 2019	OTOF	Dual AAV6	HBA/CMV	Otof KO	P6-P7	1.2-1.38e10 vg each vector	20-30dB threshold improvement in the low frequencies, best performers had 50dB improvement
Akil et al., 2019	OTOF	Dual AAV2 quad Y-F	CAG	Otof KO	P10, P17, P30	~1e10 vg each vector	50dB threshold improvement, best performers equal to wild-type mice

VGLUT3 (Akil et al., 2012)¹¹

The first study to successfully apply AAV-mediated gene replacement in a mouse model of hereditary hearing loss involved VGLUT3, which encodes a glutamate transporter necessary for inner hair cell synaptic transmission⁴⁰. Mutations in VGLUT3 are associated with the non-syndromic autosomal dominant hearing loss DFNA25⁴¹. In this study, AAV1-chick β-actin (CBA)-Vglut3 (2.3x10¹³ vg/mL) was injected into the cochlea of Vglut3 knockout mice. The authors used both cochleostomy and round window membrane (RWM) injection routes at either P0-P3 or P10 days of age. RWM injection at P0-P3 rescued hearing thresholds in all treated mice to within 10 dB of wild-type animals, and the hearing rescue persisted for over a year. Cochleostomy at P0-P3 also resulted in hearing rescue, but only in approximately 20% of injected mice. Mice injected at P10 demonstrated initial rescue at 7 weeks post injection, but their thresholds deteriorated with age, indicating that the rescue was not durable when gene delivery was performed at P10. Since the outer hair cells are not involved in the pathology in this animal model, and none of the hair cells degenerate over time, it was a good model to demonstrate the proof-of-principle for gene replacement therapy. However, VGLUT3-related deafness is very rare in humans, and when it is present it causes autosomal dominant hearing loss (DFNA25), which may not be amenable to gene replacement strategy, as discussed previously.

GJB2 and GJB6 (Yu and Wang et al., 2014, Crispino et al., 2017)^29,42

Connexins are a family of tetraspan proteins which form gap junctions⁴³. Gap junctions play an important role in the intercellular coupling between the non-sensory cells in the mammalian inner ear⁴⁴. GJB2 and GJB6 encode connexin 26 and connexin 30, respectively, and they are prominently expressed in the mammalian cochlea⁴⁵. Mutations in GJB2 and GJB6 are common causes of inherited non-syndromic hearing loss in the United States⁴⁶. In the study by Yu et al., Gjb2 gene replacement therapy was delivered to a conditional knockout mouse model of connexin 26 (cCx26KO)²⁹. The conditional knockout of Gjb2 was achieved using Foxg1-Cre, which limits the gene knockout to the inner ear (at E19). Since connexin 26 is critical for potassium homeostasis and maintenance of the endocochlear potential, the structural and physiological development of the cochlear sensory epithelium arrests during the first postnatal week, and cochlear hair cells begin to degenerate after P14 in these mutant mice²⁹. The authors delivered AAV1-CB7-Gjb2 (0.2-0.5μl of 1.5x10¹³ vg/mL) into the P0-P1 cCx26KO mouse cochlea via cochleostomy. They showed that over-expression of exogenous connexin 26 in wild-type cochlea is not detrimental to auditory function when using AAV1, even though AAV1 robustly infected hair cells (up to 44% transduced), which do not typically express connexin 26. Expression of exogenous GJB2 rescued propidium iodide diffusion between supporting cells, indicating restoration of the junctional complexes among these cells. Additionally, fluorescence recovery after photobleaching (FRAP) demonstrated that calcein AM dye diffuses from surrounding cells back into the bleached cell after AAV1-Gjb2 treatment, which does not occur in the knockout mouse. These assays demonstrated cellular level improvement of gap junction function in treated mutant mice. Unfortunately, no improvement in hearing thresholds were observed after treatment.

In the study by Crispino et al., they examined whether Gjb2 gene replacement therapy could improve the auditory function in the Gjb6 knockout mouse (Cx30^−/−)⁴². This is based on a previous study which showed that overexpression of connexin 26 using a modified bacterial artificial chromosome could rescue hearing in the Cx30^−/− knockout mice⁴⁷. When the authors delivered Gjb2 to the Cx30^−/− knockout mice at P4 using a bovine AAV, which has been shown to transduce inner ear hair cells and supporting (BAAV-CMV-Gjb2, ~1x10⁹ vg), they found that hearing was not rescued in these Cx30^−/− knockout mice. In summary, these two studies suggest that while gene replacement therapy may be able to restore gap junction function in the mouse inner ear, hearing recovery was not observed. This could be explained in part by the viral vectors used in these studies (AAV1 and BAAV), which may not be capable of transducing all the necessary cell types in the inner ear to restore auditory function. It is also possible that Gjb gene replacement therapy may need to be delivered at an even earlier time point in mouse auditory development (in utero) for hearing recovery to occur.

KCNQ1 (Chang and Wang et al., 2015)¹⁴

Variants of human KCNQ1 and KCNE1 are associated with Jervell and Lange-Nielsen syndrome, characterized by deafness and prolonged QT interval in heart muscle contractions. KCNQ1 is a voltage-gated potassium channel expressed in marginal cells (epithelial) of the stria vascularis. KCNE1 is a β-subunit of this channel and complexes with KCNQ1 to secrete potassium into the endolymph and participates in the generation of the endocochlear potential⁴⁸. The high potassium in the endolymph is critical for the maintenance of the endocochlear potential, which acts as an ionic battery that drives signaling in the hair cells. AAV1-CB7-Kcnq1 (0.5μl of 5x10¹² to 1.5x10¹³ vg/mL) was injected into the inner ears of Kcnq1 knockout P0-P2 pups. Cochleostomy injection of AAV1-CB7-Kcnq1 in the Kcnq1 knockout mice resulted in transduction of 60-80% of marginal cells along the whole cochlea, and ectopic expression could be seen in neurons, hair cells, and fibrocytes. Gene replacement therapy improved auditory thresholds to within 20-30 dB of the wild-type animals. Endocochlear potential was also restored to wild-type level (~85mV). However, hearing began to deteriorate 4 months after injection. It is unclear if the endocochlear potential also degenerated along with the elevation in hearing thresholds, since it was not measured at later stages. This study was the first to demonstrate functional hearing rescue in a mouse model of lateral wall dysfunction.

TMC1 (Askew et al., 2015; Nist-Lund and Pan et al 2019)^13,26

TMC1 is part of the mechanotransduction channel at the tips of shorter-row stereocilia on the apical surface of hair cells and is required for the transduction of mechanical movement of the stereocilia bundle into electrical signals^49-51. Loss of TMC1 results in loss of mechanotransduction, hair cell degeneration, and deafness^49,52. Mutations in TMC1 are associated with the non-syndromic autosomal recessive hearing loss DFNB7/11, as well as the non-syndromic autosomal dominant hearing loss DFNA36⁵³ In a study by Askew et al. (2015), AAV1-CBA-Tmc1 (1.0μl of 1x10¹³ vg/mL) was injected into P0-P2 Tmc1 mutant mice (Tmc1^Δ/Δ) via RWM. The authors showed that AAV1-CBA-Tmc1 treatment in Tmc1^Δ/Δ mutant mice restored hair cell mechanotransduction to near wild-type levels. However, since viral transduction was mostly confined to the inner hair cells and not the outer hair cells, the auditory improvement was approximately 20-30 dB SPL at low frequencies. Interestingly, inner ear delivery of AAV1-CBA-Tmc2 (a paralog of Tmc1) improved inner hair cell survival in the dominant gain-of-function Tmc1 mutant (Bth), demonstrating proof-of-principal for using gene substitution strategy as a treatment for autosomal dominant hearing loss.

In a subsequent study by the same group (Nist-Lund et al., 2019), the synthetic AAV AAV2/Anc80L65 was used to deliver Tmc1 cDNA into the inner ears of Tmc1^Δ/Δ mutant mice. The AAV2/Anc80L65 was generated in the lab using modeling of possible ancestral sequences of the AAV capsid gene which could be common to most identified serotypes⁵⁴. It has been shown to transduce both inner and outer hair cells in the cochlea, as well as vestibular hair cells with high efficiency⁵⁴. The authors showed that the increased outer hair cell transduction by AAV2/Anc80L65 significantly improved the auditory function in treated Tmc1^Δ/Δ mutant mice. ABR thresholds improved approximately 40dB SPL, with the best performers within 10-20dB of wild-type mice. In addition, the authors reported that AAV2/Anc80L65-CMV-Tmc1 delivery also restored vestibular function, breeding efficiency, litter survival and normal growth rate in the Tmc1^Δ/Δ mutant mice²⁶. These studies represent the first attempts at using a gene replacement strategy to treat genetic mutations affecting hair cell mechanotransduction. These studies also demonstrate that increasing viral transduction efficiency in targeted cell types is critical for the success of inner ear gene replacement therapy.

WHRN (Chien et al., 2015, Isgrig et al., 2017)¹⁵,²³

Whirlin (WHRN) is a scaffold protein that is important for stereocilia elongation⁵⁵. Mutations in the gene encoding whirlin are associated with Usher syndrome type 2D⁵⁶, as well as the non-syndromic autosomal recessive hearing loss DFNB31⁵⁵. The whirler mouse (Whrn^wi/wi) has a non-sense mutation in Whrn which leads to the absence of functional whirlin protein production⁵⁵. As a result, the Whrn^wi/wi mutant mice have short and dysfunctional stereocilia, which results in deafness and vestibular dysfunction⁵⁵. In a study by Chien et al. (2015), AAV8-CMV-Whrn (~0.5 μl at 1x10¹³ vg/ml) was injected into neonatal Whrn^wi/wi mutant mice (P1-P5) via RWM. While stereocilia length was improved in transduced inner hair cells, auditory function was not improved. This was attributed to the low levels of inner hair cell transduction (~10-15%)¹⁵.

In a follow up study by the same group, Isgrig et al. (2017) showed that when AAV8-CMV-Whrn (1 μl at 1x10¹³ vg/ml) was delivered through the posterior semicircular canal approach to Whrn^wi/wi mutant mice (P0-P5), hair cell transduction was greatly improved not only in the vestibular organs, but also in the cochlea²³. Partial hearing recovery was observed, with ABR thresholds at 60-70dB SPL in some treated mutant mice. In addition, vestibular function was also improved, as evidenced by a reduction in circling behavior and improvement in rotarod and swim tests. Improvement in circling behavior lasted for at least 4 months, although the partial hearing recovery deteriorated during that time. Hair cell survival in treated Whrn^wi/wi mutant mice was improved compared to untreated mutant mice. These two studies delivered the same viral gene replacement therapy (AAV8-CMV-Whrn) to neonatal Whrn^wi/wi mice using two different surgical approaches (RWM vs. posterior semicircular canal), and showed that the posterior semicircular canal approach resulted in better auditory and vestibular recovery compared to the RWM gene delivery, likely due to the higher hair cell transduction efficiency obtained using the posterior semicircular canal approach. The results from these two studies demonstrate the importance of surgical approach selection for the success of inner ear gene replacement therapy.

PJVK (Delmaghani et al., 2015)¹⁶

Pejvakin (PJVK) is a peroxisome-associated protein that is involved with oxidative stress-induced peroxisome proliferation¹⁶. Mutations in PJVK are associated with the non-syndromic autosomal recessive hearing loss DFNB59⁵⁷. Pjvk knockout mice have elevated hearing thresholds and are hyper-vulnerable to sound exposure¹⁶. When AAV8-Pjvk was delivered to the inner ears of the Pjvk knockout mice via the RWM (2μl, 1x10¹³ vp/ml), ABR threshold improvement of 20-30 dB SPL was achieved at low frequencies. In addition, the hair cell peroxisome proliferation response to noise exposure was also restored in treated Pjvk mutant mice¹⁶. This was the first study to apply gene replacement therapy to treat genetic mutations affecting the peroxisome system in the mammalian inner ear.

MsrB3 (Kim et al., 2016)²⁴

Methionine sulfoxide reductase B3 (MsrB3) plays an important role in hair cell homeostasis⁵⁸. Mutations in MSRB3 are associated with the non-syndromic autosomal recessive hearing loss DFNB74⁵⁹, and the MsrB3 knockout mouse is profoundly deaf, similar to DFNB74 patients. In this study, AAV2/1-CMV-MsrB3-GFP (1.31.10¹³ vg/ml) was injected into the otocyst of MsrB3 knockout mice at E12.5, using a previously described in utero gene delivery protocol⁶⁰. Significant improvement in auditory thresholds was observed in treated MsrB3 knockout mice compared to untreated mutant mice (70-80 dB SPL improvement). Unfortunately, the hearing improvement was not permanent, with hearing deterioration being observed starting at approximately 7 weeks postnatally²⁴ The authors also found that the treated MsrB3 knockout mice have restored stereocilia morphology compared to untreated mutant mice, but the restoration of stereocilia morphology was also short-lived, with stereocilia degeneration occurring starting at 5 weeks postnatally²⁴. This was the first attempt to apply an in utero gene replacement strategy in a mouse model of hereditary hearing loss.

LHFPL5 (Gyorgy and Sage et al., 2017)²²

LHFPL5 (also known as TMHS) is present at the stereocilia tips of both inner and outer cochlear hair cells as well as vestibular hair cells⁶¹. It plays an important role in hair cell mechanotransduction by regulating tip-link assembly and channel conductance⁶¹. Loss of Lhfpl5/Tmhs causes both auditory and vestibular dysfunction and hair cell degeneration⁶¹. Mutations in LHFPL5 are associated with the non-syndromic autosomal recessive hearing loss DFNB67⁶². This study is unique in that they tested whether exosome-associated AAV (exo-AAV) improves viral transduction in the inner ear for AAV1 and AAV9. Additionally, they used self-complementary AAV vector genomes instead of single-stranded genomes used in prior studies. They found that exosome-associated AAV (1x10¹¹ vg) increased hair cell transduction efficiency in vitro for AAV1 (20% AAV1 vs. 50% exo-AAV1) and AAV9 (70% AAV9 vs 100% exo-AAV9). While in vivo transduction of inner hair cells reaches up to 100%, outer hair cell transduction with AAV was approximately 15% with AAV alone and approximately 20-30% with exo-AAV. When exo-AAV1-CBA-Lhfpl5 (2.7x10⁹ vg) was injected into the inner ears of P1-P2 Lhfpl5 knockout mice, partial hearing recovery was observed (up to 30 dB SPL improvement at low frequencies). The vestibular function in the Lhfpl5 knockout mice was also improved after exo-AAV1-CBA-Lhfpl5 treatment, as evidenced by a reduction in head tossing and circling behavior. This was the first study to utilize exosome-associated AAV in inner ear gene replacement.

USH1C (Pan, Askew, Galvin et al., 2017)²⁷

The USH1C gene encodes harmonin, a protein that is present at the tips of stereocilia and functions as a scaffold protein that anchors the upper tip-link⁶³. Loss of harmonin results in reduced mechanotransduction currents and eventually hair cell death^64,65. Mutations in USH1C cause a form of Usher syndrome type 1, which is a common cause of deafness-blindness^66,67. The knock-in Ush1c mutant mouse which carries an analogous human mutation (Ush1c c.216G>A) is deaf and develops retinal degeneration⁶⁸. In addition, these mutant mice also exhibit significant vestibular dysfunction, which is also seen in patients with type 1 Usher syndrome⁶⁹. In a study by Pan et al. (2017), AAV2/Anc80L65 was used to deliver Ush1c cDNA to the knock-in Ush1c mutant mice²⁷. AAV2/Anc80L65-CMV-Ush1c (titer 2x10¹² vg/mL, 0.8-1.0 μl) was injected via RWM at age P0-P1. This resulted in average rescue of auditory thresholds to within 20dB of wild-type for low- to mid-frequency hearing, while high-frequency hearing thresholds remained elevated. Balance behavior was also rescued, as evidenced by the reduction in circling behavior, and improvement on the rotarod test²⁷. Cochlear hair cell survival was significantly increased, resulting in nearly 100% of hair cells surviving in the low- to mid-frequency locations, while 80-90% of hair cells in the high-frequency location survived. However, injection at P10-P12 did not rescue any auditory function, which highlights the importance of determining the time interval when there is a therapeutic window for gene delivery to achieve hearing recovery.

USH1G (Emptoz et al., 2017)¹⁸

Another gene that is involved with type 1 Usher syndrome is USH1G, which encodes for the protein SANS. SANS is localized at the stereocilia tip and is important for proper hair cell function⁷⁰. The Ush1g knockout mouse exhibits deafness as well as significant vestibular dysfunction, as evidenced by circling behavior¹⁸. The stereocilia bundles in cochlear and vestibular hair cells are abnormal and lack functional tip-links, which are critical for hair cell mechanotransduction. In this study, AAV8-CB6-Ush1g (1x10¹³ vg/mL, 2μl) was injected into the Ush1g knockout mice (P0-P3) via RWM. The authors found that AAV8-CB6-Ush1g treatment was able to rescue stereocilia bundle morphology in the Ush1g knockout mice Partial hearing improvement was also observed in treated mutant mice (ABR thresholds of ~80-100 dB SPL), though the hearing improvement appears to be short-lived, with progressive hearing loss by approximately 12 weeks postnatally.

USH3A (Geng and Akil, et al., 2017; Dulon et al., 2018; Gyorgy et al., 2019)^17,20,21

Clarin-1 is a transmembrane protein that is critical for hair cell stereocilia bundle morphogenesis^71,72. Mutations in clarin-1 (CLRN1) are associated with Usher syndrome type 3A⁷³. Clarin-1 is an attractive target for inner ear gene therapy due to the fact that unlike most other genes that cause hearing loss due to loss-of-function mutations, mutations associated with CLRN1 usually cause delayed-onset hearing loss⁶⁷. This potentially allows for a longer therapeutic window for inner ear gene therapy to act. The first study on Clrn1 gene replacement therapy was performed by Geng et al. (2017)²⁰, where a mouse model of Ush3a was created which develops progressive hearing loss after birth (KO-TgAC1). This was achieved by creating a Clrn1 knockout mouse that carries a transgene containing the Clrn1 cDNA and its 3’ UTR under the control of the Atoh1 3’ enhancer/β-globin basal, which causes significant downregulation of Clrn1 soon after birth⁷⁴. The authors used both AAV2 and AAV8 to deliver Clrn1 cDNA to neonatal (P1-P3) KO-TgAC1 mutant mice (8.60x10¹² and 3.41x10¹³ vg/mL, respectively). Gene delivery was done through the RWM. They found that while delivery of AAV-smCBA-Clrn1 did not result in hearing improvement, delivery of AAV-smCBA-Clrnl-UTR did result in the delay of the progression of hearing loss, with average hearing improvement of 38 dB SPL up to 150 days postnatally²⁰.

In a second study of Clrnl gene replacement, Dulon et al. (2018) generated two mutant mouse models of Ush3a, one with Clrn1 exon 4 deletion (Clrn1^ex4−/−), and the other with postnatal Clrn1 exon 4 deletion restricted to the hair cells (Clrn1^ex4fl/fl Myo15-Cre⁺⁻ The Clrn1^ex4−/− mutant mouse has profound hearing loss, whereas the Clrn1^ex4fl/fl Myo15-Cre⁺⁻ mutant mouse develops progressive hearing loss from P15-P60. When AAV2/8-Clrn1 was delivered to P1-P3 Clrn1^ex4−/− mutant mice through the RWM, small amount of hearing recovery (~10-15 dB) was observed. Delivery of AAV2/8-Clrn1 into P1-P3 Clrn1^ex4fl/fl Myo15-Cre⁺⁻ mutant mice resulted in greater hearing improvement, with hearing thresholds of ~60-70 dB SPL at P60. Unfortunately hearing loss progresses from P60-P120, likely reflective of continued outer hair cell degeneration. AAV2/8-Clrn1 gene delivery also resulted in improved stereocilia structural preservation compared to untreated Clrn1^ex4fl/fl Myo15-Cre⁺⁻ mutant mice¹⁷.

One thing in common for the two above studies is the use of conventional AAVs (AAV2 and AAV8) for inner ear gene delivery, which transduces the inner hair cells at higher rate than outer hair cells. In a third study on Clrn1 gene replacement, a synthetic AAV, AAV9-PHP.B, was used to deliver Clrn1 cDNA to the Clrn1 knockout (Clrn1^−/−) mice through the RWM²¹. Unlike patients with Usher syndrome type 3A, where hearing loss is progressive in nature, the Clrn1 knockout mouse has profound hearing loss throughout life. When AAV9-PHP.B-CBA-Clrn1 (1.8x10¹¹ vg) was injected through the RWM of P1 Clrn1 knockout mice, high level of inner and outer hair cell transduction was achieved. Unfortunately, only partial hearing recovery was observed at low frequencies (4-8 kHz), likely due to a narrow window of therapeutic opportunity in this mouse model. These three studies of Clrnl gene replacement show that the hearing outcomes of inner ear gene therapy can vary greatly depending on the therapeutic window, animal models, or viral vectors used for gene delivery.

SLC26A4 (Kim et al., 2019)⁷⁵

SLC26A4 encodes pendrin, which is a CI/HCO₃ anion exchanger that is critical for fluid homeostasis in the inner ear^76-78. Mutations in SLC26A4 cause Pendred syndrome (sensorineural hearing loss and thyroid goiter), which is a common cause of hereditary hearing loss⁷⁹. In this study, the authors used two mouse models to study Slc26a4 gene replacement therapy: a Slc26a4 knockout mouse (Slc26a4^Δ/Δ80 and a Slc26a4 knock-in mouse (Slc26a4^{m1Dontuh/tm1Dontuh})⁸¹. Since Slc26a4 is expressed as early as E11.5 in the mouse endolymphatic sac⁷⁶, and the expression of Slc26a4 is critical during E16.5-P2 for hearing recovery⁸², the authors delivered AAV2/1-CMV-Slc26a4 (1.08x10¹⁰ vg) to the otocyst of the Slc26a4^Δ/Δ and Slc26a4^{m1Dontuh/tm1Dontuh} mutant mice in utero at E12.5. They found that Slc26a4 gene replacement therapy restored pendrin expression in the endolymphatic sac, increased outer hair cell survival, and prevented cochlear hydrops. Slc26a4 gene replacement therapy also improved the ABR thresholds to within 10-20 dB of wild-type mice, and improved endocochlear potential. However, hearing recovery was unstable in the injected mutant mice, with many animals developing fluctuating and progressive hearing loss between 3-11 weeks of age. In addition, vestibular function was not rescued by Slc26a4 gene replacement, as measured by absence of improvement in the rotarod test. The lack of stable hearing recovery and improvement in vestibular function may be due to the fact that pendrin expression was only transient and was only found in the endolymphatic sac, but not in the cochlear and vestibular organs. This could be a result of the viral vector used (AAV2/1), the gene delivery approach (in utero), and/or the biology of Slc26a4 in the mouse inner ear.

OTOF (Al-Moyed et al., 2019; Akil et al., 2019)^10,12

A major limitation of using AAV for gene delivery is the fact that packaging capacity for AAV is only approximately 4.2kb, assuming minimal lengths for the ITR, promoter, and poly-A sequences³⁸. This limits the number of candidate genes that can be targeted by AAV-mediated gene replacement. One way to overcome this limitation is by employing a dual-AAV approach, in which a large cDNA is broken into two smaller parts, and the smaller segments of the cDNA are delivered using two separate AAVs⁸³. Two studies applied the dual-AAV approach to deliver otoferlin into the inner ears of otoferlin knockout mice^10,12. Otoferlin (OTOF) is expressed primarily in the inner hair cells, and it is involved with synaptic vesicle reformation and exocytosis⁸⁴. The otoferlin knockout mouse (Otof^−/− has profound hearing loss throughout life⁸⁴. In a study by Al-Moyed et al. (2019), dual-AAV2/6 was used to deliver the two segments of Otof cDNA into the inner ears of otoferlin knockout mice through the RWM. The human β-actin promoter/CMV enhancer was used to drive transgene expression and viral titer was 1.2-1.38 x 10¹⁰ vg/μl. Gene delivery was done at P6-P7. They found that inner ear gene delivery resulted in restoration of OTOF expression in approximately 50% of the inner hair cells. Auditory function was partially improved, with ABR thresholds approximately 60-80 dB SPL¹².

In another study on otoferlin gene replacement by Akil et al. (2019), Otof segments were delivered using a modified AAV2 where four surface tyrosine residues have been changed to phenylalanine residues (AAV2 quadY-F). A chimeric CMV-CBA promoter was used to drive gene expression and viral titers were 4.5-6.3 x 10¹² vg/ml (2μl was delivered in each animal). Gene delivery was performed through the RWM at P10, P17, and P30. The authors found that Otof gene delivery not only resulted in restoration of OTOF expression in the inner hair cells in the Otof knockout mice, it also significantly improved the auditory functions of the treated mutant mice at all three gene delivery time points. The click ABR threshold improved to the level of wild-type animals and lasted up to 20 weeks after gene delivery. These studies demonstrate that large genes may be amenable to gene replacement therapy using the dual-AAV approach. The study by Akil et al. is also the first study to show hearing improvement when inner ear gene delivery is performed after the onset of hearing (P10-P12) in the mouse¹⁰.

3. Opportunities for gene replacement therapy as a treatment for hereditary hearing loss:

Since 2001, universal newborn hearing screening has been mandatory in all 50 states, and it has resulted in testing of 95% of children born in the US each year (CDC report Williams et al., 2015, https://www.cdc.gov/mmwr/preview/mmwrhtml/mm6413a4.htm). Newborn hearing screens are justified both because methods of treatment (e.g. hearing aids, cochlear implants) are available, and it is critical to identify children with hearing loss as early as possible in order to maximize their chances of learning and developing spoken language. Although cochlear implants are imperfect as a replacement for natural hearing, they are highly effective for enabling speech perception and facilitating verbal communication. Average cost of cochlear implantation currently runs around $50,000-$100,000 per ear. If gene therapy can improve upon the sound perception over that of cochlear implants, the biotech/pharma sector can rationalize pricing structures that should incentivize the development of gene therapy for hearing loss. The FDA recently approved Luxturna as the first AAV gene therapy in the US, for which there are an estimated 1,000-2,000 individuals in the US and 6,000 world-wide with an RPE65 gene deficiency that are candidate patients. Congenital blindness has a prevalence of 4 in 10,000 people in developed nations, which is roughly 1/8^th the number of individuals with congenital deafness⁸⁵. Although it is challenging to estimate the number of patients by single genetic target, surely many forms of inherited hearing loss will be attractive targets for companies interested in developing gene therapy treatments. Most (73%) of the identified deafness genes are expressed in the sensory hair cells of the inner ear. Furthermore, many identified deafness genes are under 4kb in length, the size that is currently feasible for AAV delivery in a single vector (Figure 2). Recently developed dual vector systems are now able to split larger genes between two AAV vectors for recombination once delivered to the same cell. Routes of administration are actively being investigated for large animal models and humans⁸⁶. Compared to the eye, which contains over 100 million photoreceptors, there are just approximately 16,000 hair cells to target in the human cochlea and thus fewer cells to correct for sensory restoration.

Figure 2: Most genes that cause hereditary hearing loss affect mechanosensory hair cells in the cochlea.

This figure lists the genes that are associated with hereditary hearing loss by cDNA size. Most of the genes (73%) that cause hereditary hearing loss affect proper mechanosensory hair cell function (shown in black). The size limitations for single AAV and dual AAV gene delivery are also indicated.

4. Challenges to the translation of gene replacement therapy for hereditary hearing loss:

Genetic heterogeneity and window for therapeutic intervention

There are 300-400 human genetic loci predicted to be associated with hearing loss, where just over 150 deafness genes have been identified (www.hereditaryhearingloss.org). These genes are involved with a wide variety of cellular functions, and variants in most of these genes lead to early-onset hearing loss. It is a daunting task to explore a treatment for each gene, but the design of successful gene replacement therapy will require in-depth understanding of the unique functions of each individual protein and the optimal timing for gene replacement to take place. In many cases, we still do not understand enough about the functions of individual proteins, where they are expressed, and at what stage in development they are necessary. This provides a long list of opportunities to researchers eager to tease apart basic mechanisms. Furthermore, the most clinically treatable cases where gene replacement can be successfully applied will likely involve patients with progressive or delayed-onset hearing loss. This is because in patients with profound congenital hearing loss, the sensory structures of the inner ear may already be damaged at the time of treatment for gene replacement to be effective. Currently, the field has already begun to identify genes causing mild or progressive hearing loss^87-90, which provides hope that there are still many genes left to be discovered that are amenable to postnatal treatment.

It is important to remember that while the mouse auditory system has many similarities to human ears in terms of anatomy, function, and genetics, there are some important differences. The most critical difference between the two species is in the timing of inner ear development: the human auditory system is developed by gestational week 19⁹¹, whereas the mouse auditory system is not fully developed until ~12 days postnatally⁹². Since most of the genes that cause recessive hearing loss are necessary early in the development of the inner ear⁹³, it is possible that even though the window of therapeutic opportunity extends post-birth for some genes in mice, the window of opportunity for gene therapy may occur in utero in humans. The greatest potential for human translation of inner ear gene replacement begins with genes where critical inner ear cell types are preserved (e.g. OTOF, VGLUT3), or genes causing post-lingual or progressive hearing loss (e.g. CLRN1) that could be addressed by gene replacement after birth.

Effectiveness of gene delivery

Many of the genes that are involved with autosomal recessive hearing loss encode proteins that are non-secreted and must function within individual cells of the inner ear (e.g. GJB2, TMC1, WHRN). This presents the challenge that the vast majority of an entire cell population must be targeted and corrected with high efficiency by inner ear gene therapy. Therefore, finding the right viral vector with high transduction efficiency is critical for the success of inner ear gene therapy. Different AAV vectors have been shown to transduce various cell types in the inner ear, including hair cells, supporting cells, fibrocytes, mesenchyme, stria vascularis, spiral limbus, and spiral ganglion neurons⁹⁴. However, most studies indicate that conventional AAVs tend to transduce the inner hair cells in the cochlea with the highest efficiency, and other cells types at much lower efficiency^{11,13,15,54,95}. More recently, several synthetic AAVs have been reported to transduce various inner ear cells with high efficiency^21,54,96-98. The higher transduction efficiency of synthetic AAVs to various cell types in the inner ear will likely improve the efficacy of inner ear gene therapy, as illustrated by Tmc1 gene therapy with AAV2/Anc80L65 vs. AAV1^13,26. Continued research to improve and refine viral transduction in the inner ear will be critical for the future success of inner ear gene therapy.

Surgical approaches for inner ear gene delivery

In most inner ear gene replacement studies that have successfully improved auditory function in mouse models of hereditary hearing loss, the round window or canalostomy approach was used for gene delivery^{10-13,17,18,20-23,26-28,31}. AAV gene therapy can be introduced safely into the perilymphatic fluid of the mammalian inner ear by round window membrane or canalostomy injection routes without disrupting the high potassium endolymphatic space^99,100. This avoids perturbation of the polarized ionic environment that creates the endocochlear potential which serves as the battery to drive mechanosensory signaling and is necessary for proper auditory hair cell function^95,101,102. In neonatal mice, the otic capsule is cartilaginous. Therefore, a glass micropipette can penetrate directly through the posterior canal wall for gene delivery. Similarly, the bulla in neonatal mice is much easier to open to expose the round window membrane. In contrast, the adult mouse otic capsule is completely ossified. Therefore, the bony covering of the posterior semicircular canal needs to be carefully opened for gene delivery. Similarly, the bulla in adult mouse needs to be carefully removed in order to expose the round window membrane. This makes adult inner ear gene delivery more challenging. The inner ear fluid volume is ~2μl in mice⁸⁶. Generally, a 1μl volume is sufficient to perfuse reporter vectors throughout the mouse inner ear, and several prior studies have established that a minimum of 1x10⁹ viral genomes (vg) of AAV is necessary for vector transduction, while doses in the range of 1x10⁹ to 1x10¹⁰ are typical^10-29. The inner fluid volume in rhesus macaque is roughly 60μl in volume, and delivery of 10μl-30μl containing 1x10¹¹ to 1x10¹² viral genomes of AAV vector have been shown to effectively transduce inner ear hair cells^21,86. In addition, delivery of up to 30μl of fluid volume in the rhesus macaque inner ear, via either the round window or the stapedotomy, has been shown to be safe, with no adverse effect on the auditory and vestibular functions, as measured by auditory brainstem responses and vestibulo-ocular reflexes, respectively⁸⁶.

Several surgical approaches are currently in use to access the human inner ear. The round window approach is commonly used for cochlear implantation¹⁰³. Surgical procedures to access all three semicircular canals have been described: the superior semicircular canal is accessed for superior semicircular canal dehiscence repair¹⁰⁴; the posterior semicircular canal is accessed for plugging in patients with intractable benign positional vertigo¹⁰⁵; and the horizontal semicircular canal is accessed during fenestration procedure for otosclerosis (which has been supplanted by stapedotomy)¹⁰⁶ Stapedotomy, commonly used to treat otosclerosis, offers access to the vestibule¹⁰⁷. In addition, the endolymphatic sac is commonly accessed for endolymphatic shunt placement in patients with Meniere disease who suffer from intractable vertigo (though the efficacy of this surgical procedure for controlling vertigo is highly controversial)¹⁰⁸. The human inner ear fluid volume is roughly ~190 μl⁸⁶. In the “Safety, Tolerability, and Efficacy for CGF166 in Patients with Unilateral or Bilateral Severe-to-Profound Hearing Loss” clinical trial (NCT02132130, ClinicalTrials.gov), where the human Atonal transcription factor HATH1 cDNA is delivered to the inner ears of patients with severe-to-profound hearing loss using an adenovirus (AdV5) vector, 60 μl is injected through the stapedotomy approach. While the results of this study are unavailable at the time of this writing, no significant adverse side-effects have been reported in study patients according to the principal investigator (Dr. Hinrich Staecker). It remains to be seen which surgical approach would be the best for AAV gene delivery in human inner ears.

Safety and toxicity

The inner ear is encased in bone and relatively isolated from the rest of the body, making it both an attractive target and a challenge for gene therapy. While direct delivery of vectors to the inner ear fluid can result in distribution along the cochlear spiral, in animal models the vector can also be communicated into the cerebrospinal fluid (CSF), transducing the brain as well as transducing the opposite ear via a patent cochlear aqueduct^27,54. Off-target administration may be less of an issue in humans since the human cochlear aqueduct is usually not completely patent¹⁰⁹. Many hearing loss genes are unique to hair cells and require a complex protein environment to function, which means off-target functioning should be limited in non-hair cells. However, ectopic expression of ion channel or gap junction proteins could be detrimental if expressed by the wrong cell types. Theoretically, hair cells could be electrically silenced by expressing gap junction or channel proteins that ionically couple them to nearby supporting cells or the extracellular fluid. Additionally, the same issue has the potential to affect generation and maintenance of the endocochlear potential, as well as neuronal action potential generation and propagation. Therefore, when working with a specific subset of genes it may be necessary to ensure that the viral vector is restricted to expressing the transgene only in targeted cell types in order to avoid off-target transgene expression. One approach is to limit off-target expression or functioning of therapeutic genes using cell-specific microRNA binding elements that are incorporated into the vector cassette and can silence gene expression in cells expressing that specific microRNA¹¹⁰. This approach is facilitated by recently developed single-cell transcriptomic databases specific to the inner ear (gEAR portal, https://umgear.org/; and the SHIELD portal, https://shield.hms.harvard.edu/) to search for and identify novel cell-specific microRNAs. Another method to regulate gene expression is through the use of a target cell-specific promoter or enhancer. Currently, most vectors make use of ubiquitous promoters like the CMV or CBA promoter with gene expression activity that is not restricted to specific cell types. Limited inner ear cell-specific promoters have been identified, especially ones that are small enough to fit within an AAV transgene cassette while leaving enough room for the therapeutic transgene¹¹¹. This is due to the challenge of identifying tissue specific enhancers and promoters, which can be found hundreds of kilobases upstream or downstream of the gene of interest. However, newly developed bioinformatics tools and experimental techniques have enabled powerful reductionist approaches to interrogate up to a megabase of DNA for enhancer/promoter function^112-114. Finally, from a dosage perspective of gene expression, recent studies have demonstrated that strong constitutive promoters in the eye can have detrimental consequences on cell survival in the retina presumably due to protein over expression¹¹⁵. This effect can be mitigated by decreasing the overall vector dose but will likely come at the expense of decreased numbers of corrected cells. Therefore, control over protein expression using gene appropriate promoters or enhancers is the most attractive method to maximize the number of corrected cells while minimizing toxicity from both over-expression and off-target expression. Continued research in these areas is necessary to allow investigators to target specific cell types in the mammalian inner ear without off-target transgene expression, in order to improve the safety of inner ear gene therapy.

Longevity of therapeutic effects

Even though gene replacement therapy has been successfully applied to several mouse models of hereditary hearing loss, the longevity of the therapeutic effect varies greatly, with many studies showing hearing deterioration after a couple of months and others showing sustained hearing improvement up to 12 months of age. The longevity of various inner ear gene replacement therapy is likely dependent on the pathophysiology of hearing loss associated with specific genes, as well as the animal models used in these studies. Whether the longevity of therapeutic effectiveness from mouse studies can be extrapolated to humans with hearing loss remains to be seen. Since many of the cell types in the mammalian inner ear are post-mitotic and lack the ability to regenerate¹¹⁶, this should theoretically allow the transduced target cells to have persistent transgene expression throughout the cells’ entire lifespan without the need for re-administration. However, should the need for gene therapy readministration arise, its efficacy may be reduced due the production of neutralizing antibodies against the viral vector and/or transgene¹¹⁷.These considerations should be kept in mind when designing future inner ear gene therapy clinical trials.

Diagnosis and clinical translation

One of the major barriers to the translation of inner ear gene therapy to patients with hereditary hearing loss is the fact that genetic testing for hearing loss is presently not commonly performed. Currently, successful identification of genetic causes of hearing loss is achieved in only 40%-50% of the tested patients^118-120, leaving patients without answers and those researching potential treatments without a solid estimate of the size of the patient pool. Early identification of the underlying genetic cause of hearing loss will allow clinicians to collect data on the natural history of hearing loss caused by specific mutations. This information is critical for the determination of the optimal therapeutic window for inner ear gene delivery. As the inner ear gene replacement animal studies reviewed above illustrate, identification of the appropriate therapeutic window for gene delivery is a key factor for successful hearing recovery. Another major barrier for translating inner ear gene therapy to patients is the fact that there is currently no reliable tool to monitor the level of cell survival in the inner ear in real time.

In order for inner ear gene therapy to be successful, the targeted cell type in the inner ear must be viable at the time of gene delivery. The existing imaging modalities do not have the required resolution to yield this critical information in real time. However, improvements are being made in this area^121,122, and electrophysiological measures may also be helpful to monitor the status of the inner ear (e.g. distortion product otoacoustic emissions, electrocochleography).

5. Conclusions:

Over the past few years, several studies have demonstrated that inner ear gene replacement therapy can be used to improve auditory function in mouse models of hereditary hearing loss. These studies targeted genes that mediate a wide range of inner ear cellular functions (Figure 1), including synaptic transmission (e.g. Vglut3, Otof), ion channels (e.g. Tmc1, Kcnq1, Slc26a4), gap junctions (e.g. Gjb2, Gjb6), peroxisome proliferation (e.g. Pjvk), stereocilia function and maintenance (e.g. Whrn, Lhfpl5, Ush1c, Ush1g, Clrn1), and defense against oxidative stress (MsrB3). The fact that inner ear gene replacement therapy has been successfully applied to these genes with diverse roles in the mammalian inner ear illustrates the great potential for inner ear gene replacement therapy as a treatment for hearing loss caused by different etiologies. However, many challenges still lie ahead for the successful translation of gene replacement therapy to patients with hereditary hearing loss. A deep understanding of the gene target is required for assessing the therapeutic window for effective inner ear gene delivery. Viral capsid engineering will hopefully continue to provide the field with viral vectors that have high transduction efficiency and specificity for the targeted cell types. The use of cell-specific promoters and microRNAs can control the safety and dosage of gene expression while further limiting unwanted off-target transgene expression. In addition, better imaging and electrophysiological tests will eventually allow investigators to select the appropriate patients for inner ear gene replacement therapy administration and monitor its efficacy in real time. Given the recent FDA approval of gene therapy treatments for Leber’s congenital amaurosis and spinal muscular atrophy, it is our hope that there will be new gene therapy treatments available for patients with hereditary hearing loss in the foreseeable future.

Highlights.

1. Over the past few years, several studies have demonstrated that inner ear gene replacement therapy can be used to improve auditory function in mouse models of hereditary hearing loss.

2. Because of its strong biosafety profile, adeno-associated virus (AAV) was the viral vector of choice for the majority of these studies.

3. Inner ear gene replacement therapy has been successfully applied to genes with diverse roles in the mammalian inner ear.

4. A deep understanding of the gene target is required for assessing the therapeutic window for effective inner ear gene delivery.

5. Viral capsid engineering can develop viral vectors that have high transduction efficiency and specificity for the targeted cell types.

6. The use of cell-specific promoters and microRNAs can limit unwanted off-target transgene expression.

7. Better imaging and electrophysiologic tests will allow investigators to select the appropriate patients for inner ear gene replacement therapy administration and monitor its efficacy in real time.