This article is mentioned in:

1. Introduction

The cochlea is a highly differentiated and anatomically isolated sensory organ located within the temporal bones of mammals, including humans. The hearing function of the cochlea is mainly dependent on the performance of outer and inner hair cells in the organ of Corti and spiral ganglion neurons (SGNs) in the Rosenthal's canal. These hair cells convert sound signals into electrical signals, which are delivered to the auditory pathway of the brain by SGNs along the auditory nerve.

However, the hair cells of the organ of Corti as well as SGNs of the Rosenthal's canal are terminally differentiated cells, and their failure to recover or regenerate following exposure to ototoxic insults, including chemicals and noise, as well as aging, results in irreversible hearing loss in mammals (1–4). Despite advances in the development of neuroprotective drugs, no clinically effective means of repairing or preventing acoustic dysfunction are available as yet (5–7).

The cochlea is a relatively isolated structure anatomically. Although it is difficult for systemically delivered drugs to cross the blood-labyrinth barrier and enter the inner ear, locally injected drugs can be concentrated in the cochlear lymph fluid (8,9). There has been marked progress in target gene identification as well as vector preparation technologies, thus raising the potential for local gene transfection into the cochlea.

More than 90 genes affecting inner ear development have been identified to date, of which 19 are involved in non-syndromic deafness. Viral and non-viral vectors have been experimentally applied to the cochlea using different approaches (10–13). Careful characterization of the mechanisms underlying dysfunctions, such as presbycusis and genetic deafness, is likely to expand the horizons of this novel therapeutic modality. Thus, the number and types of cases to which gene transfection can be applied for the treatment of chronic degenerative disorders and congenital diseases of the ear using gene modification approaches is likely to increase.

2. Principles of gene transfection in the cochlea

Local gene transfection involves the insertion, alteration or removal of genes in target tissues using a viral or non-viral vector to drive or inhibit the expression of a functional protein (14–16).

The most evident advantage of local gene transfection into the cochlea is the steady and long-term expression and effect on the target area, which shows marked contrast to the effects of cytokine treatment (17). Exogenous expression of the X-linked inhibitor of apoptosis protein (XIAP) gene delivered by an adeno-associated virus (AAV) vector and of the neurotrophin gene by an adenoviral (Adv) vector was able to maintain long-term effects in cochlear cells and protect against deafness (14,18). However, target gene introduction by a Herpes simplex virus (HSV) vector showed poor transfection efficacy in the inner ear (19,20).

The safety of gene transfection should also be taken into consideration. AAV is an ideal vector due to its low antigenicity. Hydroxyapatite (HAT) nanoparticles also represent a vector free from the risk of biological disaster (21). By contrast, the application of HSV and Adv vectors has been restricted due to their potential immunogenicity and cytotoxicity (22,23). In addition, the injection procedure was shown to negatively affect the cochlea. Injection by cochleostomy into the scala media or scala tympani causes irreversible electrophysiological and morphological damage to the cochlea (24,25). The methods used to examine cochlear characteristics pre- and post-gene transfection are presented in Table I.

Table I Methods used to examine the characteristics of the cochlea pre- and post-gene transfection.

Table I

Methods used to examine the characteristics of the cochlea pre- and post-gene transfection.

Examination	Methods	Ref.
Electrophysiology
Outer hair cell (OHC)	Cochlear microphonic (CM), otoacoustic emission (OAE)	(21,22)
Inner hair cell (IHC)	Compound action potential (CAP)	(23,24)
Spiral ganglion neurons (SGNs)	Compound action potential (CAP)	(23)
Auditory pathways	Auditory brainstem responses (ABR), electrical auditory brainstem responses (EABR)	(25,26)
Single neuron	Compound action potential (CAP)	(27)
Staining
Hair cells	Silver nitrate, hematoxylin-eosin (HE), phalloidin-TRITC and DAPI staining	(25,28)
SGNs	Prussian blue and DAPI staining	(25)
Nerve fibers	Anti-neurofilament 200 (NF200) antibodies	(29)
Observation (microscopy)
Light	Surface preparation, paraffin section, semithin section	(30)
Fluorescence	Surface preparation, frozen section	(31)
Laser scanning confocal (LSCM)	Surface preparation, frozen section	(32,33)
Scanning electron	Sample coated by conducting material	(28)
Transmission electron	Ultrathin section	(32)

[i] DAPI, 4′-6-diamidino-2-phenylindole.

Approximately 45 deafness-related genes have been identified for non-syndromic hereditary hearing loss, and another 30 genes associated with syndromic hearing loss have been identified (26,27). The expression of exogenous genes may be used to rescue, replace or silence mutant loci (28,29). The introduction of genes such as glial cell line-derived neurotrophic factor (GDNF), Bcl-2 and XIAP was shown to have protective effects against ototoxic insults, including chemical- and noise-mediated injury (30–32). Moreover, mouse atonal homolog 1 (Math 1) and human atonal homolog 1 (Hath 1) were shown to induce the regeneration of hair cells in the organ of Corti and utricular maculae (16,33).

In addition, the efficacy of transfection for various vectors in subjects should be considered. Previous studies showed that most transfected cells were located in the basal turn of the cochlea or in the basilar membrane (34,35).

3. Viral vectors

Adv vectors

The first studies of cochlear gene transfection utilized an Adv vector composed of a linear DNA molecule ~35 kb in length. Adv vectors may be generated at high titers and are able to accommodate large (8-kb) DNA inserts. They have the advantage of not requiring cell division for transfection and may be used successfully for the transfection of the terminally differentiated cells in the mammalian inner ear. Adv vectors have been shown to effectively transfect hair cells, and a transfection efficacy of ~90% in the inner hair cells (IHCs) and 50% in the outer hair cells (OHCs) and SGNs has been observed in vivo and in vitro(36–39).

First-generation E1−, E3− replication-deficient Adv vectors showed cytotoxicity and the induction of an immune response to cochlear hair cells within 8–10 days of transfection (23,39,40). The hair cell lesions induced by E1−, E3− replication-deficient Adv vectors are inhibited by immunosuppression using glucocorticoids, while morphological evidence indicates that hair cells remained intact following the injection of glucocorticoids prior to Adv vector treatment (41). Second-generation Adv vectors with deletions in the E1, E2b and E3 regions were introduced into trials of cochlear injection and functional lesions in the cochlea were delayed until 28 days following transfection of the virus. This was suggested to be due to a delayed immune response, which was eliminated using gutted adenoviral vectors with deletions in E1, E2b, E3 and/or pol or another locus (39). The current focus of Adv vector development is on the elimination of excess viral genes to minimize host immune responses and cytotoxicity.

AAV vectors

AAVs consist of a single-stranded DNA parvovirus capable of transfecting pre- and post-mitotic cells with no requirement for actively dividing cells. AAVs accommodate DNA inserts 3.5–4.0 kb in length (42). AAVs integrate into the host cell's DNA, usually on chromosome 19, following induction by the viral rep gene. It was reported previously that transfected AAVs may be retained for 6 months and induce stable transgene expression in cochlear cells (43). Moreover, compared with Adv vectors, AAV vectors are based on a non-pathogenic human virus that has not been associated with disease and shows less ototoxicity (38).

There are at least 10 AAV serotypes based on amino acid sequence differences in their respective capsid proteins. Of these, AAV serotypes 1–5 are useful for gene therapeutic applications due to their typical tropism and profile in vivo(44). AAVs of serotype 2 have commonly been used to drive the expression of genes in several cochlear cell types, including hair cells, especially OHCs, and supporting cells in the organ of Corti, SGNs in the Rosenthal's canal, cells of the spiral limbus and spiral ligament and sensory and supporting cells of the crista ampullaris (45–47). AAVs of serotype 5 exhibited high transfection efficacy in SGNs, but failed to transfect hair cells, while AAV1 and AAV7 showed good transfection efficacy in cells of the spiral ligament, and expression driven by AAV5 and AAV8 was especially apparent in Claudius cells (48). These phenomena may be closely correlated with the expression of co-receptors to various virus serotypes on the target cell surface (49,50). However, it is difficult to produce AAV vectors in high titers.

Additionally, the application of mutant AAVs needs to be examined. A recombinant AAV vector with mutations in capsid surface-exposed tyrosine residues showed a 10-fold increase in transfection efficacy in HeLa cells, and a 30-fold increase in murine hepatocytes in vitro compared with tyrosine-phosphorylated AAV vectors (51). This study indicates potential for the development of a high-efficacy transfection system at a low virus dose that is also an ideal candidate for use in human gene therapy (Fig. 1).

Figure 1 (A) Mut-aav2-GFP (3×1012 PFU/ml) showed a high transfection efficacy in both the OHCs and IHCs of guinea pigs by basilar membrane preparation. (B) aav2-GFP (3×1012 PFU/ml) showed a high transfection efficacy in IHCs but a lower efficacy in OHCs. Arrows and arrowheads indicate fluorescent signals specifying expression in OHCs and IHCs, respectively. OHCs, outer hair cells; IHCs, inner hair cells.

HSV vectors

HSV is a DNA virus with a 152-kb double-stranded DNA genome, which may be used to transfect cells of neuronal origin. This type of vector is easy to produce and capable of carrying large DNA inserts. HSV type I has generally been used to transfect cochlear cells. HSV-1-linked NT-3 has been used successfully for the stable transfection of SGNs and for protection against the ototoxicity of cisplatin (52). However, HSV vectors are not the preferred approach for cochlear transfection due to their low transfection efficacy in cochlear cells (20). In addition, there are concerns about the apparent immune response and inflammation induced by viral infection of the inner ear (53,54). Investigation of HSV for cochlear applications remains in the developmental stage, yet these vectors show potential for promoting the survival of neural and neural-derived cells.

Lentivirus vectors

Lentiviruses, such as human immunodeficiency virus, may be used as retroviral vectors to transfect dividing as well as non-dividing cells (55). Although stable protein expression mediated by a lentivirus vector was reported in rat brain without observable toxicity for 6 months (56), the efficacy of transfection into cochlear cells was rather poor in vivo and in vitro. Lentivirus vectors show a narrower distribution throughout the cochlea compared with AAV and Adv vectors. The results of previous in vivo and in vitro studies indicated an intact cellular and tissue cytoarchitecture within lentivirus-infused cochlea and an absence of inflammation or pathological changes (57). Lentiviruses may be suitable as vectors for the transfection of neurotrophins and other protective factors into the cochlea.

4. Non-viral vectors

Liposomes

Positively charged (cationic) liposomes coupled with a negatively charged (anionic) integrated target gene are able to bind the plasma membrane of target cells and release the gene into the cytoplasm (58). The genes delivered by liposomes have been shown to be incorporated into the genome of the host, with the encoded protein expressed for only 14 days in the neurosensory epithelia and surrounding tissues of the cochlea in guinea pigs (59). Studies of cationic liposomes have demonstrated a wide distribution of the reporter gene in hair cells and supporting cells in guinea pigs, and in the spiral ligament, Reissner's membrane and SGNs in mice (59,60). However, liposomes may affect the physiological activity of cells, such as inhibition of the mitochondrial inner membrane, protein kinase C and ATPase activity, resulting in cytotoxicity (61–63). Thus, liposome vectors may be suitable for gene transfection when expression is required only for a short time.

HAT

Polylactic/glycolic acid was the first nanoparticle vector used to deliver materials to the cochlea (64). However, HAT was the first nanoparticle vector used successfully to transfect cochlear cells (15). The infusion of HAT particles 40–50 nm in length into the cochlea resulted in no significant damage. In addition, the HAT-mediated gene transfection of NT-3 has been shown to have a protective effect against the excitotoxicity of kainic acid in SGNs. Thus, HAT may be a useful candidate for cochlear transfection if the low transfection efficacy (16% in HeLa cells) associated with these nanoparticles can be improved (21).

Hemagglutinating virus of Japan envelope (HVJ-E)

HVJ-E vector is a non-viral second generation HVJ vector, which was first used for gene transfection into the central nervous system (65). It is relatively easy to produce HVJ-E and these vectors show higher fusion activity compared with first-generation HVJ-liposome vectors. An HVJ-E vector combining hepatocyte growth factor (HGF) injected through the cerebrospinal fluid was shown to approach the cochlea and the expression of HGF prevented kanamycin-induced hearing loss (66). These results indicate that HVJ-E vectors are more efficient compared with other non-viral vectors and safer compared with viral vectors. Thus, they represent another potentially useful therapeutic approach to sensorineural hearing impairment. Table II briefly compares the advantages and disadvantages of the various vector types described above.

Table II Overview of the vectors for gene transfection.

Table II

Overview of the vectors for gene transfection.

Vectors	Advantages	Disadvantages
Adv	Ability to transfect post-mitotic cells Easy to produce Large insert size	Limited duration of transgene expression Immunogenic
AAV	Ability to transfect post-mitotic cells Long-term and stable expression Nonpathogenic virus	Variable transfection efficiencies Low gene capacity Unable to pass freely through the round window membrane Hard to produce in high titers
HSV	Neurotrophic Easy to produce Large insert size Stable expression	Immunogenic Variable transfection efficiencies
Lentivirus	Stable expression	Risk of insertional mutagenesis Low transfection efficiencies Cell division required
Liposomes	Easy to produce Large insert size Nonpathogenic	Low transfection efficiencies
HAT	Easy to produce Large insert size Nonpathogenic	Low transfection efficiencies
HVJ-E	Easy to produce Large insert size Nonpathogenic	Variable transfection efficiencies

[i] Adv, adenoviral; AAV, adeno-associated virus; HSV, herpes simplex virus; HAT, hydroxyapatite; HVJ-E, hemagglutinating virus of Japan envelope.

5. Injection approaches

As the cochlea is surrounded by a bony wall and is isolated due to the blood-labyrinth barrier, direct infusion into the cochlea is usually necessary to achieve transgene expression in cells within this structure (Fig. 2). This is an ideal approach into the cochlea and does not cause damage or at least functional impairment, and which allows easy and convenient manipulation. There are three main approaches to injection into the cochlea: the scala media, the semicircular canals and the scala tympani.

Figure 2 Dissection of a guinea pig acoustic bulla in vitro. A large bulla space is advantageous for operating on the inner ear in cases of cochlear injury. The arrow indicates the malleus and incus; the arrowhead shows the round window niche; the dotted line indicates the spiroid canal.

The scala media approach involves injection into the endolymphatic system using a vector able to transfect sensory cells in the organ of Corti (Fig. 3A). However, this pathway is difficult for clinical application due to the complexity of manipulation and possible surgical side-effects, such as disruption of the cochlear structure (the stria vascularis and the spiral ligament) and hearing loss (67,68). Animal studies indicated that surgical exposure from the level of the mandible to the acoustic bulla is required, which may increase the risk of functional impairment, and serious threshold shifts and hair cell loss were observed following surgery in animal models (67,69). Thus, this is not the preferred option for the introduction of transgenes into the cochlea from the viewpoint of functional recovery.

Figure 3 Surgery of needle insertion into the cochlea. (A) Injection into the scala media with the needle crossed the hole in the second turn of the cochlea (arrowhead). (B) Injection into the scala tympani, the puncture site to the first turn of the cochlea was below the RW niche (arrow). (C) Injection into the scala tympani using the trans-RWM approach in a neonatal mouse. The RW niche (arrowhead) and stapedial artery (arrow) were exposed simultaneously. (D) Injection into the scala media by canalis semicircularis posterior in an adult mouse (arrowhead). The arrow shows the canalis semicircularis lateralis. RWM, round window membrane.

Conversely, the scala tympani approach, represented by cochleostomy and the trans-round window membrane (RWM) technique, is a comparatively convenient and simple method for animal experiments as well as clinical application. Cochleostomy (Fig. 3B) may be better than the trans-RWM technique (Fig. 3C) for the administration of accurate volumes and to prevent potential fluid leakage. In addition, there is no evidence of threshold shifts by ABR tests following surgery for cochleostomy infusion (70,71), however, studies of histopathological changes in cochlear cells following cochleostomy should be conducted to ensure the protection of function (24).

Several vectors, including liposomes, Adv vectors and HSV, have been shown to travel into the cochlea from the middle ear space via the RWM, resulting in cochlear cell transfection (34,72). The inflammatory response induced by macrophages and precipitated in the inner ear to such vectors probably enhances their ability to enter the inner ear space due to an increase in permeability of the RWM (73). Conversely, AAVs are unable to traverse the RWM freely without specific treatments to enhance the permeability of the membrane (74). Methods to alter the permeability of the RWM may also be used to enhance the rate of passage of other vectors.

Based on the structure of the basilar membrane, which lacks tight junctions, but consists of fibrils, a homogeneous ground substance and mesothelial cells, vectors in the perilymph space are able to cross the basilar membrane to approach hair cells (75,76). Another pathway into the cochlea involves passage through the habenula perforata to enter the organ of Corti (77).

The semicircular canal approach, also termed canalostomy (Fig. 3D), is used to introduce vectors into the cochlea as well as the vestibular system (78). The semicircular canal is relatively simple to expose and the procedure carries little risk of injuring the cochlea and surrounding blood vessels compared with cochleostomy. However, the main disadvantage of the semicircular canal approach is that it is impossible to determine whether the tip of the needle opens into the endolymphatic compartment or into the perilymph during surgery, and the seal between the bone and the tube is usually ruptured, resulting in leakage of the inner ear fluid and vector suspension. Furthermore, canalostomy was shown to be associated with temporary vestibular function disorders in mice, including adverse effects on circling behavior, head tilt and swimming ability, but they recovered within 2 weeks following surgery (79).

6. Future directions

The future of gene transfection is likely to include improving the properties of vectors to achieve a higher transfection efficacy and cell targeting, refining the methods of gene delivery to minimize lesions to the cochlea, while confirming widespread transfection throughout the cochlea or localized transfection within specific areas.

In addition to treating chemical- and noise-induced hearing loss, gene therapy may be used to improve cochlear implant function. Neurotrophins promote the survival of and delay the degeneration of SGNs. Neurotrophin gene transfection performed in conjunction with cochlear implant surgery may enhance neurite growth to the cochlear implant. The development of cochlear implants with improved performance would improve the quality of life for a number of deaf children and elderly people.

The genes and factors involved in cell fate determination in the sensory epithelium of the inner ear have been explored. Math 1 and Hath 1 have been shown to drive regeneration of hair cells posterior to the lesions induced by ototoxic factors. Further basic investigation of drug delivery in fetal and neonatal animals is likely to facilitate the development of novel methodologies for the effective treatment of genetic diseases.

Acknowledgements

This study was supported by the China National Funds for Distinguished Young Scientists (grant no. 30925035).

References