Inner Ear Gene Delivery: Vectors and Routes

Abstract

Objectives:

Current treatments for hearing loss offer some functional improvements in hearing, but do not restore normal hearing. The aim of this review is to highlight recent advances in viral and non-viral vectors for gene therapy and to discuss approaches for overcoming barriers inherent to inner ear delivery of gene products.

Data Sources:

The databases used were Medline, EMBASE, Web of Science, and Google Scholar. Search terms were [(“cochlea*” or “inner ear” or “transtympanic” or “intratympanic” or “intracochlear” or “hair cells” or “spiral ganglia” or “Organ of Corti”) and (“gene therapy” or “gene delivery”)]. The references section of resulting articles was also used to identify relevant studies.

Results:

Both viral and non-viral vectors play important roles in advancing gene delivery to the inner ear. The round window membrane is one significant barrier to gene delivery that intratympanic delivery methods attempt to overcome through diffusion and intracochlear delivery methods bypass completely.

Conclusions:

Gene therapy for hearing loss is a promising treatment for restoring hearing function by addressing innate defects. Recent technological advances in inner ear drug delivery techniques pose exciting opportunities for progress in gene therapy.

INTRODUCTION

Hearing loss affects 466 million people worldwide and is expected to double to just under one billion people affected in the next thirty years according to recent World Health Organization estimates. [1] The primary contributor to worldwide hearing loss is sensorineural hearing loss (SNHL), for which current treatments are limited to sound amplification with hearing aids and electrical stimulation of nerves via cochlear implants. However, while these treatments offer a functional improvement in hearing, they are not curative. Gene therapy offers a potential alternative solution to SNHL that addresses core defects in the cochlea. Successful gene therapy to restore hearing involves implementing safe and effective vectors to carry genetic material to specific tissue while overcoming the barriers to and challenges of local delivery to the inner ear. This review summarizes recent advances in gene therapy for hearing loss with a focus on vectors that target the cochlea and approaches to overcome the anatomical barriers inherent to inner ear drug delivery.

VECTORS

More than half of cases of congenital hearing loss are genetic in origin, and up to 93% of these cases are autosomal recessive. [2] Autosomal recessive forms of hearing loss result from a missing gene product, and restoration of this gene product may be curative. Delivery of genes into the inner ear requires packaging of the genes into a vector that can enter cells. Broadly, these vectors can be categorized into viral and non-viral vectors (Table 1). Compared to non-viral vectors, viral vectors have increased efficiency of gene transfer; however, non-viral vectors exhibit reduced inflammation and toxicity.

Table 1:

Vectors

Vector Type	Vector	Advantage	Limitation
Viral	Adenovirus	High transduction efficiency Can accommodate large gene inserts	Transient expression of transgene
	Adeno-associated virus	Non-pathogenic Can transduce pre- and post-mitotic cells Can achieve long-term transgene expression	Can only accommodate small gene inserts <5 kb; however, can undergo homologous recombination, making dual-AAV approach possible
	Lentivirus	Can achieve long-term transgene expression Can accommodate large gene inserts	May be immunogenic and ototoxic Random integration into genome possible
	Herpes simplex virus	Neurotropic vector	May elicit strong host immune responses
	Sendai virus	Low pathogenicity High transduction efficiency Located exclusively within cytoplasm	Less thoroughly investigated
Non-Viral	Polymeric-based nanoparticles	Small size Can protect against enzymatic breakdown Low toxicity	Low efficacy Low biodegradability
	Lipid-based nanoparticles	Low immunogenicity	Potential toxicity Decreased efficacy
	Peptide-based nanoparticles	Can be combined with cationic lipoplexes or polyplexes for greater functionality	Potential toxicity

Viral Vectors

Viral-mediated gene delivery into the inner ear has been applied to animal models using various viral vectors including adenovirus [3–6], adeno-associated virus [7–10], lentivirus [11–13], herpes simplex virus [14–16], vaccinia virus [16], and Sendai virus [17, 18]. These viral vectors have previously been reviewed in detail. [19] While adenovirus (AdV) has high transfection efficiency and can accommodate large gene inserts, AdV-mediated gene therapy results in only transient expression of the transgene. [20] Lentivirus, a retrovirus, can achieve long term gene expression and can also accommodate large gene inserts, but may be immunogenic and ototoxic and carries the danger of random integration into the genome. [12] Herpes simplex virus (HSV), a neurotropic vector, has been used to transduce neurotropin-3 within mice to prevent cisplatin-induced damage. [14] Both HSV and vaccinia virus may elicit strong immune responses within the host. Sendai virus, of the paramyxoviridae family, is a promising vector because of its low pathogenicity, high transduction efficiency, and location exclusively within the cytoplasm. [17, 18] However, this vector has been less thoroughly investigated for inner ear gene delivery compared to other viral vectors.

Adeno-associated virus (AAV) was the viral vector used when Lalwani et al. demonstrated in 1996 the first in vivo expression of a gene transduced within the mammalian inner ear. [7] Since then, AAV, a single-stranded DNA parvovirus, has become the most investigated viral vector for gene delivery to the inner ear due to a number of attributes: it is non-pathogenic, able to transduce both pre- and post-mitotic cells, which includes the neurosensory epithelia of the inner ear, and able to achieve long-term expression of the transgene. [8] In one recent study, AAV was successfully used to overexpress neurotrophin-3 for protection against noise exposure. [21]

A disadvantage of AAV is that it can only accommodate small gene inserts of less than 5 kb. However, because of the ability of AAV to undergo homologous recombination, a dual-AAV approach can be employed in which the larger insert can be split and carried by two recombinant AAV vectors, allowing the full-length protein to be expressed upon recombination. [22, 23] Using this dual-AAV approach, two groups have successfully restored hearing to otoferlin knock-out mice by delivering the 6 kb otoferlin cDNA. Concurrently, recent work has focused on identifying serotypes AAV with high specificity and efficiency for hair cells or neurons within the inner ear, including synthetic serotypes such as Anc80L65 and PHP.B, a capsid variant of AAV9. [24–27] Bovine AAV demonstrated high transduction efficiency compared to AAV2, AAV5, and AAV4 and was then used to restore connexin expression in connexin-deficient mice. [28, 29] Until recently, studies investigating the usage of viral vectors for inner ear gene therapy have predominantly used rodents. In 2019, Gyorgy et al. conducted the first viral transduction of the non-human primate inner ear using AAV9-PHP.B; both inner and outer hair cells were shown to be transduced with great efficiency. [30] AAV-PHP.B was also successfully used to restore hearing in a mouse model of Usher Syndrome 3A.

Promoters for Inner Ear Gene Expression

In addition to the choice of viral vector, transduction efficiency is also dependent on the promoter chosen to drive transgene expression. Moreover, the choice of promoter impacts the expression of reporter genes in different cell types. For instance, while the cytomegalovirus promoter (Cmv), commonly used with AAV, has a high transmission efficiency for a variety of cell types [31], the myosin 7A promoter (Myo) is recognized as the promoter most specific to hair cells. [32] This section focuses on promoters that have been used to drive expression of transgenes carried by AAV (Table 2).

Table 2:

Properties of Select Promoters

Promoter	Cell types
Cytomegalovirus IE enhancer and chicken β-actin promoter	Inner hair cells, spiral ganglion cells, spiral ligament cells, Reissner’s membrane, inner sulcus cells, Hensen’s cells, mesenchymal cells, with high efficiency [36]
Cytomegalovirus promoter	Inner and outer hair cells [27], spiral ganglion cells, spiral ligament cells, Reissner’s membrane, inner sulcus cells, Hensen’s cells, mesenchymal cells, with intermediate to high efficiency [36]
Glial fibrillary acidic protein promoter	Support cells of the spiral limbus, pillar cells in Organ of Corti [27]
Elongation factor 1α promoter	Spiral ganglion cells, spiral ligament cells, mesenchymal cells, with marginal efficiency [36]
Myosin 7A promoter	Inner hair cells [32, 36]

Green fluorescent protein (GFP) was established early on as an effective reporter of gene transfer [33], and several studies have utilized GFP to compare the ability of various promoters to drive transgene expression. Lee et al. evaluated three promoters, synapsin, glial fibrillary acidic protein (Gfap), and Cmv, and found that a combination of Cmv and the vector AAV-PHP.B was most efficient at transducing both inner and outer hair cells. [27] One study compared Cmv, chicken-β-actin (Cba), synapsin 1, and mouse phosphoglycerate kinase 1. [34] While both Cmv and Cba were found to efficiently drive expression of enhanced GFP in hair cells, Cba was chosen for further study and successfully drove the expression of transmembrane channel-like 1 (Tmc1) in inner hair cells of deaf mice with deletion or dominant point mutation of Tmc1.

The combination of the cytomegalovirus immediate early enhancer and Cba promoter (CAG) is promising and was shown to drive high levels of transgene expression in cells of the cochlea. [35] Liu et al. compared six different promoters, including CAG, Cmv, neuron-specific enolase promoter (NSE), myosin 7A promoter (Myo), elongation factor 1α promoter (EF-1α), and Rous sarcoma virus promoter (RSV). [36] The highest level of expression of enhanced GFP was achieved using the CAG promoter, while the Myo promoter was the most specific, achieving reporter expression only in the inner hair cells of the cochlear. CMV and NSE achieved an intermediate level of expression; EF-1α achieved only marginal expression; and the RSV promoter failed to drive expression. The CMV immediate early enhancer has also been combined with the Cmv promoter. This CMV-beta-globin promoter was found to be efficient in hair cells. [37] Ultimately, the choice of promoter can influence the transduction efficiency and specificity and is an important consideration in vector design.

Non-Viral Vectors

There are several non-viral vectors that offer a potential alternative to the more commonly used viral vectors for gene delivery. Non-viral vectors allow for more specificity in the targeting of therapeutics to certain cell types, do not have a DNA size limit, are relatively easy to produce, and exhibit a diminished immune response in comparison to viral vectors. [38] The three primary categories of non-viral vectors are inorganic particles such as gold and silica; synthetic or biodegradable particles such as peptide-based, lipid-based, and polymeric-based nanoparticles; and physical delivery methods such as electroporation. [39] Most studies involving non-viral vectors for the delivery of gene therapy to the inner ear implement synthetic or biodegradable nanoparticles such as lipoplexes, dendrimers, or polymers.

Cationic liposomes are structures composed of a positively charged lipid bilayer and act as a shield for nucleic acids from breakdown by extracellular enzymes and neutralization by antibodies. Their bilayer structure also enables endocytosis by target cells [40]. Wareing et al. (1999) demonstrated the first successful cationic liposome-mediated gene transfer. [41] Okano et al. demonstrated another early successful example of cell–gene delivery of therapeutic molecules to the inner ear without the use of viral vectors. [42] Using a liposome-mediated delivery method, NIH3T3 cells were transfected with recombinant brain-derived neurotrophic factor (BDNF). The cells were then transplanted to the cochlea, which led to a significant increase in production of BDNF in the inner ear. More recently, Gao et al. used cationic lipid nanoparticles to deliver Cas9-guide RNA-lipid complexes in vivo that targeted the Tmc1 allele via scala media injection. The results indicated enhanced hair cell survival and reduced hearing loss in a mouse model of human genetic deafness. [43]

Cationic polymers, such as polyethylenimine, represent another category of nanoparticles that have demonstrated success as vectors for inner ear gene delivery in vitro. Chen et al. used polyethylenimine-polyethylene glycol (PEI-PEG) to transfect spiral ganglion cells using a liposome (lipofectamine™ 2000) as a control. The PEI-PEG group demonstrated a higher transfection rate with lower toxicity to spiral ganglion cells compared to the liposome group [44]. Although polymers and other nanoparticles offer promising alternatives to viral vector gene delivery, they are not without concern and have not been extensively studied for treating inner ear disorders. [45] Zhou et al. indicated that cationic polymer PEI nanoparticles could cause significant ototoxicity of the cochlear structures, therefore offsetting much of the potential beneficial effects of non-viral nanoparticle transfection. [46] For future clinical use, more studies concerning potential ototoxicity should be explored to elucidate the utility of nanoparticles as non-viral vectors for gene delivery.

Decreasing Gene Expression

Autosomal dominant forms of hearing loss whose mechanisms are gain-of-function or dominant negative typically require gene editing or gene silencing technologies to decrease gene expression. Approaches to these disorders often implement tools such as the CRISPR/Cas endonuclease system to invalidate specific mutated alleles or RNA interference (RNAi) to prevent translation by degradation of mRNA. Yukihide et al. were the first to show prevention of hearing loss using gene silencing techniques of a dominant negative disorder. The team used siRNA to post-transcriptionally target GJB2 expression in vitro and in vivo and successfully reduced the expression of the mutant allele in mice by greater than 70% of control GJB2 levels. [47] Yoshimura et al. recently demonstrated that RNAi-mediated gene silencing can slow progression of hearing loss and improve inner hair cell survival in a mouse model of human TMC1 deafness. This study demonstrated this effect in adult mice, whereas previous gene therapy studies have focused on treating neonatal mice. [48] Achieving successful RNAi-mediated gene silencing in a mature organ of Corti is an important step in revealing whether gene therapy can potentially treat human hearing loss by decreasing gene expression.

CRISPR/Cas9 endonucleases work at the transcriptional level to offer another promising approach to decrease gene expression by disrupting the gene of interest. Using a PAM variant of Staphylococcus aureus Cas9 endonuclease carried via viral vector, Gyorgy et al. showed they could prevent deafness in Beethoven mice without toxicity for up to one year by disrupting a mutant allele. The disruption did not affect the wild-type Tmc1/TMC1 allele and did not exhibit off-target cleavage. [49] Kang et al. recently developed a genome editing strategy using AAV8-Cre injection producing Cas9 protein activation that allows for more efficient gene editing in neonatal and adult mice without damaging hearing. Gene knockout in the inner ear typically requires breeding the offspring of FLOX mice with Cre mice, but Kang’s approach offers a more efficient method because it screens greater than two sgRNAs in one AAV. [50] CRISPR/Cas9 is a promising method for addressing autosomal dominant hearing disorders, but its application for hearing loss so far has only been investigated in rodents.

ROUTES FOR GENE DELIVERY

A major challenge to the delivery of gene therapy is the anatomically isolated nature of the cochlea, a small complex structure encased within the densest portion of the temporal bone. Anatomic barriers that may limit the delivery of gene therapy into the inner ear include the blood-labyrinth barrier (BLB), the round window membrane (RWM), the oval window, and the Eustachian tube. [51] The BLB is a network of vascular endothelial cells coupled together by tight junctions that divides the vasculature of the inner ear from the inner ear fluid spaces. [52] Generally, the BLB poses a challenge to systemic delivery, and ear-specific activation following systemic delivery has not yet been developed. The RWM is a semipermeable three-layered membrane at the base of the cochlea that serves as a communication between the scala tympani and the middle ear. Properties of the RWM itself, such as thickness, and properties of the diffusing molecule, such as size and electrical charge, can affect RWM permeability, leading to inconsistent diffusion that can limit intratympanic injection. [53] The oval window is the communication between the scala vestibuli and the middle ear space and is covered by the stapes footplate. [54] Finally, the eustachian tube, which connects the middle ear space to the nasopharynx, poses a challenge to intratympanic injection as drugs can pass through the nasopharynx into the nasopharynx, reducing drug diffusion into the inner ear. [55] Recent technological advances in inner ear drug delivery techniques have focused on bypassing these anatomic barriers and pose exciting opportunities for progress in gene therapy. This section of the review discusses the various routes through which gene therapy can be delivered into the inner ear (Table 3).

Table 3:

Inner Ear Delivery

	Advantage	Disadvantage
Intratympanic	Minimally invasive procedure with a good safety profile Moderate efficacy	Unpredictable diffusion into the cochlea Loss of therapeutic via the eustachian tube Potential trauma to middle ear structures
Intracochlear	Circumvents the inconsistency of diffusion through the RWM Can achieve a high peak concentration of the therapeutic	Invasive procedure with potential for hearing loss and trauma, but RWM microinjections have shown ability to heal

Intratympanic Delivery

Injecting therapeutics through the tympanic membrane into the middle ear space is a safe and commonly used approach in drug delivery. The ultimate goal of this approach is to have the therapeutic reach the inner ear space by diffusing across the RWM or the oval window. The inconsistency of diffusion across the RWM, inadvertent clearance via the eustachian tube and difficulty with cellular uptake are the principal difficulties with intratympanic injection. These impediments typically lead to a decreased amount of therapeutic delivered to the inner ear than desired. To overcome the difficulties with RWM permeability, researchers have applied chemical agents and endotoxins [56], created microperforations in the RWM [57–59], and increased contact time of therapeutic using hydrogels and gelatin sponges. [60] To overcome leakage through the eustachian tube, approaches include seeking to replete the cleared drug by repeated dosing or continuous infusion and decreasing the drainage through the use of gels. [61] These methods have recently been reviewed by Szeto et al. [51]

There have been multiple examples of successful gene therapy delivery through an intratympanic route. Yoon and Lim performed intratympanic injection of an adenoviral vector to the middle ear space without creating an immune reaction. [62] Furthermore, their gene delivery product led to an overexpression of β-defensin 2 which led to suppressed NTHi adhesion onto the HMEEC cells. Yoon et al. successfully delivered oligoarginine-conjugated nanoparticles via intratympanic injection demonstrating successful in vivo gene delivery to inner ear hair cells in mice. [63] Importantly, the oligoarginine-conjugated nanoparticles successfully traversed the RWM, which is one of the important barriers when delivering through the tympanic membrane. Another intratympanic approach for delivery involves surgically drilling the bulla to expose the middle ear space, bypassing the tympanic membrane, and placing the therapeutic directly on the RWM. Wu et al. used hydroxyapatite nanoparticles to carry genetic material across the intact RWM of chinchillas with this method and showed a high transfection efficiency for the targeted cells. [64]

Intracochlear Delivery

Direct intracochlear delivery of vector can overcome many of the limitations of transtympanic delivery by bypassing RWM variability, reducing leakage into the eustachian tubes resulting in higher peak concentration within the cochlea, and reducing the basal-apical gradient of delivered agent. [65] In early studies of inner ear gene delivery using animal models, cochleostomy was used to introduce genes by direct injection or infusion using osmotic minipumps. [7, 8, 11, 41] Others have introduced genes into the inner ear of rodents via a canalostomy. [66, 67] The introduction of fenestration in the posterior semicircular canal has also been used to increase the efficiency of trans-RWM injection. [68] However, these methods are traumatic and unlikely to be the method of choice for inner ear gene therapy in humans.

The more promising methods of inner ear gene delivery include direct injection into the RWM or via stapedectomy, and these methods have been investigated in non-human primates. [69] The first transduction of the inner ear in non-human primates was achieved by injection of vector via the RWM using a 29-gauge needle. [30] Examination of the cochlea within the injected ear demonstrated mild hemorrhage as a result of this method. Microneedle technology designed for round window membrane perforation may increase the safety of gene delivery by trans-RWM injection and is a promising avenue for investigation. [57, 59, 70]

CONCLUSION

Gene therapy for hearing loss is the most promising treatment for restoring hearing function by addressing inherent defects of the inner ear. By restoring missing gene products through the delivery of viral and non-viral vectors carrying a transgene, gene therapy can offer a curative solution for autosomal recessive genetic forms of hearing loss. For autosomal dominant forms of hearing loss, techniques such as RNAi and the CRISPR/Cas9 endonucleases can reduce gene expression through silencing dominant negative alleles. While barriers to accessing the inner ear have posed a major challenge to the delivery of gene therapy, recent technological advances in inner ear drug delivery techniques pose exciting opportunities for progress in gene therapy.