Учебники / Genetic Hearing Loss Willems 2004

synapses between e erent endings and OHC are seen at 9 dab, but it is not until about 20 dab that OHC e erent synapses are mature. (See Table 3.)

III.FUNCTIONAL DEVELOPMENT OF THE EAR

A.Human

The development of function in the ear has not been extensively studied in humans, primarily because all of the relevant events occur in utero, well before the normal time of birth (44). However, studies of the responsiveness of fetuses to sound delivered to the abdomen of the mother indicate that the

ear becomes functional beginning at about the nineteenth week, when reactions to low frequencies are observed (108). Reponses to higher frequencies are observed by 33–35 weeks (47). By birth, the human inner ear functions at essentially an adult level. Although the middle ear is frequently fluid-filled, when this clears in the first days after birth, the middle ear is also fully functional. Reflecting these changes in middle ear fluid status, otoacoustic emissions increase by 5–10 dB in the first 2–3 dab (94).

B.Mouse

The onset of cochlear function has been extensively studied in the mouse and other altricial rodents such as the rat and gerbil. An important determinant of early function is the increased potassium concentration of endolymph, and the appearance of the endocochlear potential (EP). Before birth in the rat and mouse, endolymph and perilymph have similar ionic characteristics. However, beginning at birth the potassium concentration of endolymph increases rapidly, and the sodium concentration falls, to reach adult-like values by about 10 dab. The EP lags this potassium increase, appearing at 3–4 dab in the mouse, reaching adult values by 14 dab (5,6,20,123,102). The maturation of the EP closely follows the morphological development of the stria vascularis.

The cochlear microphonic (CM) can first be recorded from the mouse cochlea in response to very intense stimuli at around 7 dab. CM thresholds improve rapidly until about 15 dab (110). The compound action potential (CAP) shows similar characteristics (45). The maturation of these evoked responses lags EP maturation, suggesting that other factors contribute to the course of their development. Maturation of the HCs themselves is one such factor. In the gerbil the negative EP, observed during anoxia and thought to reflect ion flux through hair cells, develops after the positive EP (123). This and the morphological data reviewed above suggest that HC maturation also lags that of the EP. In the gerbil, direct stimulation of the stapes with a piezoelectric driver elicited CM about 2 days prior to acoustic stimulation, and improved developing thresholds up to 40 dB depending upon frequency (125). This suggests that middle ear immaturity also plays a significant role in delaying the functional development of the cochlea.

While cochlear thresholds appear to mature by 15 dab, the full dynamic range of cochlear responses does not develop until several days later (126). This presumably reflects maturation in the capacity of cochlear cells to produce high levels of output. Also, CAP and VIIIth nerve fibers can respond to higher repetition rates as thresholds and dynamic range improve.

The frequency tuning of the inner ear changes during the period of threshold improvement. The characteristic frequency of tuning curves re-

corded from the same location in the spiral ganglion of the gerbil increases by more than one octave, while tuning curves become dramatically sharper (32). Tonotopy in the cochlear nucleus also shifts, presumably reflecting events in the cochlea (100). This presumably results from a combination of events. The organ of Corti has more mass early in development, which would tend to lower the frequency at which it would be maximally responsive. At the same time, the maturation of the cochlear amplifier acts to shift the best frequency upward.

The development of OHC motility, which is thought to represent the major component of the active cochlear amplifier, has not been studied in the mouse. However, in a similar altricial rodent, the gerbil, it appears by 7– 8 dab, and reaches mature levels by 14–17 dab (46). This coincides with the period of OHC synaptic remodeling, and also with the maturation of CAP parameters, such as the ‘‘S’’ shape of the input-output curves and the sharpening of CAP tuning curves (93), that are considered to reflect the activity of the cochlear amplifier. (See Table 4.)

IV. MOLECULAR DEVELOPMENT OF THE EAR

As noted above, the formation of the ear is mediated by the coordinated expression of many developmental genes. A complete review of such expression is beyond the scope of this chapter. Moreover, the developmental expression of a number of genes in which mutations lead to hearing loss will be described in later chapters of this book. However, there are general patterns of gene expression that underlie critical events in the morpho-

genesis and functional onset of the ear. A brief review of some of these patterns is presented below.

Much of our understanding of the molecular control of ear development is based on information obtained initially in other organs, in which the patterns of gene expression that lead to organogenesis have been explored in some detail. With respect to the general morphogenesis of the outer, middle, and inner ears, the boundary model developed by Meinhardt (72) to explain pattern formation in the vertebrate limb is instructive. According to this model, determination of the limb requires a primary patterning event along the main axis of the embryo. This primary pattern results in areas with di erent gene expression separated from each other by defined borders. The model proposes that the interaction between at least two di erently determined types of cells results in the production of a morphogen that di uses into the surrounding cells, creating a graded morphogen distribution. This is called positional information. A pattern then becomes determined, and a secondary axis can be formed by interpretation of this positional information. The result of Meinhardt’s model is the formation of ‘‘organizing regions’’ that serve to direct development along the di erent axes in the limb—proximodistal, dorsoventral, and anteroposterior.

As with the inner ear, vertebrate limb initiation takes place in a specific location along the anteroposterior and dorsoventral axes of the developing embryo. This placement appears to rely on the di erential and coordinated expression of the Hox genes within the lateral plate mesoderm (78). In addition to placement of the limbs, the Hox genes are involved in the regulation of later stages of limb development, particularly cartilage proliferation and di erentiation.

Limb bud formation is initiated by fibroblast growth factor (FGF) and Wnt signaling, originating in the mesenchyme of the limb-forming region. FGF-10 in the mesenchyme activates expression of FGF-8 in the overlying ectodermal cells, forming the apical ectodermal ridge (AER). This is a thickened rim of ectoderm at the tip of the limb bud that serves as an organizing region for limb outgrowth (116). Furthermore, this ridge arises at a cell lineage-restricted boundary between the Engrailed 1 expressing ventral ectodermal compartment and the radical fringe/Wnt7a expressing dorsal ectodermal compartment (56,62). Radical fringe may in turn interact with the Notch/Delta signaling pathway to maintain the AER, similar to the process in Drosophila (18). The interaction of the genes mentioned above may also specify patterning along the dorsoventral axis. Formation of the anteroposterior axis of the limb is mediated by sonic hedgehog (Shh) and bone morphogenic protein (Bmp) expression in the zone of polarizing activity, a small region of mesenchyme cells in the posterior margin of the developing bud (31).

Morphogensis of the ear involves processes of gene expression and pattern formation similar to that of the limb. The normal positioning and development of the ear along the anteroposterior axis of the developing embryo requires the expression of Hox1a and Hox1b in the hindbrain (40). The entire inner ear forms from a simple epithelium that invaginates and separates from the surface ectoderm adjacent to rhombomeres 5 and 6. As in limb bud initiation, studies suggest a role for mesoderm tissue in otic induction, and it has been proposed that FGF-19 expression in the paraxial mesoderm induces Wnt-8c expression in the overlying neural plate, and that both factors act cooperatively in initiating otic development (59). The activity of these two genes results in the expression of some otic placode marker genes like Pax2 Dlx5, Nkx5.1, and SOHo. In addition, these two signals cooperate to induce thickening of the otic placode ectoderm.

The invagination of the otocyst is also induced by the adjacent mesenchymal tissue and neural tissue of the rhombencephalon (120). Based on the expression of FGF-3 in the appropriate region of the rhombencephalon, Represa et al. (96) treated early chick embryos with anti-FGF-3 antisense otligonucleotides and observed repression of otocyst formation. However, when Mansour et al. (68) deleted the fgf-3 gene in mice, they observed normal otocyst formation. This suggested that FGF-3 was uninvolved, or that functional redundancy existed, presumably with another FGF-family member. Recently Maroon et al. (69) found that FGF-3 and FGF-8 cooperate to induce otocyst formation in the chick, confirming that two members of this gene family participate in otocyst formation.

Specific models for the development of the otocyst into the labyrinth have been developed. Fekete (35) proposed a boundary model for the development and patterning of the otocyst that takes into account the formation of separate gene expression compartments, similar to the boundary model for limb formation. Compartments are defined by broad gene expression domains that will specify di erent structural parts of the ear such as semicircular canal, saccule, cochlea, utricle, and endolymphatic duct. These compartments arise partly in response to di usable factors that originate from external sources such as the rhombencephalon and the notochord. According to the model, cells located at the boundary of two or more compartments are induced to secrete a morphogen that will di use into the surrounding cells, directing development and pattern formation among the cells that respond to it. The borders may specify location of sensory organs, emigration of neuronal precursors, or endolymphatic duct outgrowth (17). According to the Fekete model, the developing otocyst is divided into eight compartments as a result of the intersection of three boundaries: the dorsoventral boundary, the mediolateral boundary, and the anteroposterior boundary. According to the model, linear structures such as

the organ of Corti are the result of the activity at the boundary between two compartments. Punctate structures would arise at the intersection of al least three di erent compartments, for example the cristae. In support of this model, the development of the endolymphatic duct has been mapped to a putative boundary between medial and lateral zones in the developing mouse ear. In the mouse, the expression of Pax2 and EphA4 has been found in the medial wall of the otic vesicle, including the endolymphatic duct region and the enlarging cochlear duct. Immediately adjacent to this domain is a region of SOHo expression. Work by Brigande et al. (17) has shown that the endolymphatic duct arises in the area of the medial-lateral boundary, defined by the intersection of Pax2/EphA4 and SOHo expression. By labeling individual cells in the developing otocyst and following them by time-lapse photography, Kil and Collazo (25) have found that most otocyst cells move freely between the di erent regions of the otocyst. This suggests that they may express and be exposed to di erent morphogens at di erent times in their development.

The formation of the sensory epithelia has also been studied by Morsli et al. (76), who followed the expression of genes associated with sensory epithelia, HCs, and supporting cells in the developing mouse otocyst. They found that Bmp4 was an early marker of the epithelia of the semicircular canal cristae from 11.5 to 12 dpc, while luntatic fringe was an early marker for developing macular and cochlear epithelia from 12 to 13 dpc.

Once the inner ear is patterned, the development of specific cell types follows templates that are common to many cell types in the body. The development of sensory and neuronal cell types appears to follow patterns similar to those previously delineated in brain. In particular, proneural genes such as the bHLH transcription factor Math1 have been shown to be required for the specification of the neural cell fate in neural progenitor cells. Once a neural fate is established, a number of genes are involved in neuronal di erentiation. For example, reciprocal Notch/Delta signaling is important for di erentiation of neuronal from nonneuronal cell phenotypes. In addition, members of the Brn-3 family of POU-domain transcription factors are involved in the early di erentiation of many di erent sensory neurons, with di erent members of the family expressed in di erent neuronal populations (71).

In the inner ear, Math1 is expressed in the prosensory regions of the inner ear. It is required for HC formation, and exogenous expression can induce HC formation in nonsensory regions of the organ of Corti (15). Math1 expression is observed at 12.5 dpc in vestibular epithelia and 13.5 dpc in the cochlea. The Delta ligand Jagged 2 is expressed in HC precursors following HC fate determination. Deletion of the Jagged 2 gene results in HC duplications (60). The Brn-3 family POU-domain transcription factor

Brn-3.1 (Brn-3c, POU4f3) is also expressed in HC precursors immediately after the commitment step occurs, at 12.5 dpc in the vestibular epithelia and 13.5 dpc in the cochlea. Brn-3.1 expression continues throughout life (34,101), and is required for HC di erentiation and survival (34,129,118).

Once the cells and tissues of the cochlea are formed, cochlear function depends on the expression of genes that mediate the transduction of sound and the transmission of information into the central auditory system (30,39, 48,128). For example, strong expression of Na,K-ATPase isoforms in the stria vascularis develops in parallel to the appearance of the endocochlear potential (128). The initial expression of the OHC motor protein prestin is correlated with the development of electromotility in these cells (14), while alpha-9 acetylcholine receptor expression precedes e erent innervation of the hair cells (65). The development of glutamate receptor expression in spiral ganglion neurons is related to the functional maturation of the synapse between HCs and spiral ganglion neuron a erent dendrites, as shown in Figure 5 (66).

V.DISCUSSION

The normal development of the external, middle, and inner ears provides the background against which genetic disorders occur. Many forms of inherited hearing loss involve defects in genes that are expressed during development and that are critically involved in regulating the ontogeny of the ear. For this reason, knowledge of normal ear development is helpful in understanding the e ects of mutations in genes associated with hearing loss. The reverse relationship is, if anything, more important. The coordination of the development of the components of the ear means that many changes occur in parallel, with the result that temporally related processes are not

Figure 5 Expression of mRNA encoding a gene required for inner-ear function, the GluR2 AMPA glutamate receptor, in the developing rat cochlea. Embryonic development of the inner ear in the rat lags that of the mouse by about 4 days. The left panels show autoradiographs corresponding to the tissue illustrated in the right panels. Labeled GluR2 cRNA riboprobes were hybridized to tissue sections from E- 18 (17.5 dpc) to adult inner ear. GluR2 mediates neurotransmission between hair cells and the a erent dendrites of spiral ganglion neurons. GluR2 mRNA is expressed in the spiral ganglion well in advance of the onset of auditory function, which occurs at about 10 dab. The function of strong GluR2 expression in the cells of the spiral limbus (Lim) is unknown. P = postnatal day; SL = spiral ligament; SV = stria vascularis; SG = spiral ganglion. (From Ref. 66.)

necessarily causally related. Because virtually everything is changing at the same time, all processes correlate to some degree during development. For this reason, correlational studies in development are often minimally informative regarding causation. More definitive information about causal relationships in development has come from the study of natural and induced mutations, as described in subsequent chapters.

For the purposes of this review, we have concentrated on the human and the mouse. Although ontogeny occurs at a much more rapid rate in the mouse, the development of the ear in the two species appears to occur via very similar processes, This means that the mouse is a good model for molecular research that is targeted at understanding human mutations. Thus we can learn a great deal about genetics in the ear of the human by assessing and manipulating the appropriate genes in the mouse. Of course it is also much easier to obtain information about development in an experimental animal than it is in humans. However, it is important to recognize that there are significant di erences between these two species that can influence genetic outcomes. For example, the life span of the mouse is far less than that in the human. Since many mutations in humans develop phenotypes in childhood or even adulthood, it is possible that mice would never live long enough for similar defects to appear. Also, the middle ear of the human is larger and includes structures such as the mastoid air spaces that are not present in the mouse. Genes causing defects in these structures would necessarily have di erent e ects in mice. It should also be noted that almost all of the development of the ear occurs in utero in humans. Virtually all of inner-ear ontogeny including functional onset is complete by the twenty-sixth fetal week, while the middle-ear structures achieve their adult configuration by the thirtieth week. While the initial stages of ear development also occur prior to birth in the mouse, significant development and functional onset occur postnatally. For research purposes this is a benefit. However, middle and external ear maturation in the mouse is likely to be influenced by exposure to the external environment, including air, at a much earlier developmental stage than in the human.

As noted above, the development of the inner ear is an exquisitely coordinated process, involving the precisely timed formation and di erentiation of millions of cells. The patterns of gene expression that underlie this process are as yet poorly understood. However, it is clear that thousands of genes are expressed including potentially several hundred that encode developmental regulatory molecules. It is therefore not surprising that mutations can result in abnormal development of the ear and result in inherited hearing loss. In fact, given the complexity of normal ear development, it is perhaps more surprising that the number of genes involved in inherited hearing loss, currently estimated to be a few hundred (90), appears to be relatively small.

This is consistent with the observation that many knockout studies of genes that are expressed in the ear during development and are suspected of an important role in ear ontogeny fail to produce a phenotype (77).

Why do mutations in genes that participate in inner ear development not produce inherited hearing loss? The e ects of such mutations may be compensated by redundancy, especially when several genes of the same family are expressed in the ear. Alternatively, mutations may a ect the structure of the inner ear, but in relatively minor ways that are not reflected in function. Mutations in still other genes may be lethal during early embryogenesis, so that a potential inner ear phenotype never develops. Genes whose mutations do produce defects in the ear and hearing loss often appear to occupy critical positions in developmental hierarchies. Such genes are sometimes referred to as master regulatory genes. Such genes tend to influence key decisions regarding tissue and cell phenotype, and their alteration has profound influences on subsequent events. Many models of development, examples of which are described above, attempt to characterize critical regulatory events that may be mediated by such master regulatory genes.

The identification of the genes that are involved in inherited hearing loss represents only the beginning of understanding these conditions. Characterization of events downstream from expression of a mutant protein, or lack of expression in null mutations, is required to determine the proximate causes of hearing loss, and to determine means by which such losses might be prevented. Since the majority of deafness genes a ect the development of the ear, the events outlined in this chapter are the backdrop against which mutation-induced changes must be viewed.

The fact that many mutations resulting in hearing loss influence inner ear development has profound implications for the development of potential interventions. At present, there are no genetically based treatments for inherited forms of hearing loss. However, there is intense interest in the potential for gene-based therapeutic interventions. A number of experimental studies of gene therapy have been performed in the postnatal rodent ear (51). The formation of the inner ear begins early in human development, during the third week of gestation, and is largely complete by the twenty-sixth week. Therefore, intervention to prevent defects that arise during this period must be e ective in utero. Treatment of the gamete or blastocyst is di cult, but allows manipulation of the very earliest stages of the embryo, long before the ear begins to form. However, after implantation intervention becomes much more problematic. Recent developments in the treatment of the human fetus suggest that it may eventually be possible to deliver therapy to the developing inner ear. In this case, a precise understanding of the structural and molecular development of the ear will be critical.