Introduction
A fascinating question arises when studying the development of the human inner ear – when does hearing commence? By 20 weeks, the human cochlea has achieved a developmental status comparable to that at which other mammals first respond to sound.Reference Pujol, Lavigne-Rebillard and Uziel1 Clinical data in humans suggest that some fetuses may respond to sound as early as 19 weeks of gestation.Reference Birnholz and Benacerraf2, Reference Hepper and Shahidullah3
Taking into account the unusual conditions of the foetal inner-ear environment (i.e. fluid in outer and middle ears, and the effect of surrounding maternal tissues and amniotic fluid), it would be expected that low frequency sounds would propagate more efficiently than high frequency ones. The fetal sound experience would therefore be limited to the low frequency components of the mother's speech and the physiological internal noise. It has been suggested that this may help fine-tune the developing auditory system and may influence postnatal speech perception.Reference Peck4 Brainstem auditory evoked response studies and otoacoustic emission recordings on preterm neonates suggest that, in the early stages of development, fetuses respond only to low frequencies.Reference Starr, Annlie, Martin and Sandes5–Reference Brienesse, Anteunis, Wit, Gavilanes and Maertzdorf8 Only later in development does the frequency response range expand to include higher frequencies.
To someone unfamiliar with the development of the cochlea, this would suggest that the apical turn should mature first, followed by the basal turn. The case is known to be the reverse, with the cochlea following a base to apex course of development and maturation. Animal studies suggest that during development the mechanics of the inner ear may change in such a way that the basally located hair cells respond to relatively low frequencies first, shifting to ever higher frequencies as development progresses.Reference Rubel9 This suggests that the ‘place code’ of the cochlea is not fixed, but changes during development of auditory function. This possibility will be explored later and related to our original observations.
Material and methods
In this study, sections of human temporal bones ranging in age from eight weeks of gestation to full term were examined. The material, obtained from perinatal deaths, was derived from the temporal bone collection of the Institute of Laryngology and Otology (now part of the University College London Ear Institute). All fetal specimens were decalcified, embedded and then cut horizontally, commencing at the upper surface of the temporal bone. The embedding medium in the majority of cases was low-viscosity nitrocellulose, a substance related to celloidin. In a small number of cases, paraffin wax was used. The nitrocellulose-embedded material was sectioned at 20 µm thickness, and each 10th section was mounted on a glass slide and stained with haematoxylin and eosin. Paraffin wax sections were cut at 7 µm.
From 208 human fetuses, 98 were initially selected. Inclusion criteria for the study included absence of evidence of congenital disease in the medical history, and a recorded age of gestation and/or a measured crown–rump length. Specimens observed to suffer from severe autolysis were deemed unsuitable for any meaningful observation and were thus excluded. Fetuses with obvious congenital abnormality were also excluded. Eighty-one temporal bones from 45 fetuses were finally studied. The temporal bones of one child and four adults were used as controls. In 41 fetuses, the age was estimated using the crown–rump length. In the remainder, the age was calculated from the first day of the last normal menstruation, as recorded in the medical notes.
Results
The caudal wall of the cochlear duct is the site of differentiation of the organ of Corti. At the end of the eighth week, the caudal wall was observed to consist of tightly packed layers of cells with darkly staining nuclei. A greater and a lesser ridge were identifiable on the caudal wall. There was no evidence of differentiation between supporting and sensory cells. The tectorial membrane primordium was beginning to emerge in the basal turn as a condensed, eosinophilic substance (Figure 1a). By the end of the 10th week, the tectorial membrane primordium could be traced even in the most apical turns. An area of clear cytoplasm was observed between the greater and lesser ridges. This area was replaced by a notch in the 60 mm fetus, which was constantly seen in our series of specimens.
Fig. 1 (a) Eight-week fetus; the greater and lesser epithelial ridges are identified at the caudal wall of the developing cochlear duct; arrow = lesser ridge (H&E; ×100) (b) Individual hair cells and supporting cells are distinguishable in the 14-week fetus; arrows = hair cells (H&E; ×250) (c) By the end of the 14th week, an oval space appears between the inner and outer hair cells. (H&E; ×100) (d) Higher magnification of previous specimen. (H&E; ×250) SV = stria vascularis; gr = greater ridge; RM = Reissner's membrance; CW = caudal wall; bv = blood vessel; SM = scala media; m = mesenchyme at developing basilar membrane; IHC = inner hair cells; OHC = outer hair cells; PC = pillar cells
As development progressed (i.e. the 14th week), an area of more lightly staining nuclei and cytoplasm was seen approximately midway in the caudal wall of the cochlear duct, and this could be traced up to the most apical turns (Figure 1b). At the basal turn, individual hair cells became identifiable in this area and could be distinguished from the underlying supporting cells. The latter were arranged in two layers. Medial to the sensory cells, a hollow groove was formed at the area of the future spiral sulcus (Kollicker's organ). This consisted of three to four layers of cells, with an area of more lightly staining cytoplasm apically. Interdental cells were observed medial to Kollicker's organ and were arranged in two to three layers of cells. A sharp angle was formed between the interdental region and the future Reissner's membrane. At this stage, the formation of the tectorial membrane was readily identifiable in the basal turn as a streak of darkly staining material overlying the interdental region. Lateral to the sensory cells and for the rest of the basilar membrane, there were two to three layers of undifferentiated, darkly staining cells (Figure 1b).
Gradually (i.e. by the end of the 14th week), the formation of small, clear spaces was observed between the outer hair cells and the supporting cells, which were now arranged in a single layer. A small but well formed oval space was observed between the inner and outer hair cells in the basal turn, and was just beginning to form in the more apical turns. This did not correspond to the tunnel of Corti, as the individual pillar cells could not be identified at this stage. Clear spaces were also forming between the inner hair cells and Kollicker's organ. The organ of Corti still lay on a thick mesenchymal layer, as the scala tympani had not yet formed. The tectorial membrane crossed over Kollicker's organ to reach the inner hair cells. Kollicker's organ (the future internal sulcus) formed a deeper groove of two to three layers of cells. The interdental cells were single-layered in the basal turn and a sharp angle was formed at the region of the future lip of the limbus (Figures 1c and d). In the more apical turn, the interdental cells were still two-layered.
Between the 14th and 15th weeks, Hensen's cells were readily recognised for the first time, with small, dark, elongated nuclei. The rest of the basilar membrane consisted of a single layer of cuboidal cells. The lip of the limbus was more prominent at this stage (Figures 2a and b). Differentiation of Hensen's cells could also be seen in the more apical turns, but the cells lateral to them were still columnar in two layers. At a slightly later stage (the 15th week), the tectorial membrane seemed to have increased in thickness and to stain more darkly. It had a feather-like appearance, which could be seen at higher magnification (Figure 2c). The interdental region had also increased in size and the underlying stroma had become less cellular.
Fig. 2 (a) Differentiation of Hensen's cells is observed between the 14th and 15th weeks of development. (H&E; ×100) (b) Higher magnification of previous specimen. (H&E; ×250) (c) 15-week fetus; tectorial membrane has increased in thickness. (H&E; ×250) (d) 17-week fetus; Kollicker's organ becomes less cellular, and pillar cells (arrow) are recognised for the first time. (H&E; ×100) KO = Kollicker's organ; tm = tectorial membrane; bm = basilar membrane; IC = interdental cells; IHC = inner hair cells; OHC = outer hair cells; DC = Deiters' cells; HC = Hensen's cells
Gradually (i.e. the 17th week), the Kollicker's organ medial to the inner hair cells became less cellular (Figure 2d). Cells were arranged in two layers, while their nuclei were more basally placed. A pair of darkly staining cells, possibly pillar cells, appeared in the oval space between the outer and inner hair cells. By the end of the 17th week, the spiral limbus had increased in size and there was more abundant stroma and fewer cells, compared with previous specimens. Individual pillar cells were now readily identifiable and began to develop rapidly. A small space (space of Nuel) was now readily identifiable between the first row of outer hair cells and the outer pillar cells (Figures 3a and b). The lip of the limbus was also more prominent. The basilar membrane beneath the organ of Corti consisted of four to five layers of cells.
Fig. 3 (a) By the end of the 17th week, pillar cells are readily identifiable in all turns. (H&E; ×250) (b) By the end of the 17th week, the space of Nuel is beginning to form in the basal turn. (H&E; ×250) (c) 20-week fetus; tunnel of Corti is identified as an opening between inner and outer pillar cells. (H&E; ×250) (d) 20-week fetus (same as previous specimen); Kollicker's organ has regressed to form the inner sulcus, and the lip of the limbus is quite prominent. (H&E; ×250) OHC = outer hair cells; HC = Hensen's cells; DC = Deiters' cells; PC = pillar cells; sN = space of Nuel; TC = tunnel of Corti; IS = inner sulcus; IC = interdental cells; ll = lip of limbus; tm = tectorial membrane
At a somewhat later stage, both the outer and inner hair cells appeared more slender, with spaces forming between outer hair cells. Short projections, possibly stereocilia, could be identified on the apical ends of outer hair cells. The tectorial membrane had a laminated appearance with a hook-like projection at the end. The tunnel of Corti had not yet opened and pillar cells were seen in close contact with one another. Lateral to Hensen's cells, two distinct populations of cells could be seen. The first was arranged immediately laterally and consisted of a single layer of cuboidal cells with darkly staining cytoplasm and nuclei. The second extended short of the spiral prominence and consisted of more oval-shaped, lighter staining nuclei with lighter staining cytoplasm.
At 20 weeks, a small opening (the tunnel of Corti) was readily identified between the inner and outer hair cells for the first time. The space of Nuel was also well formed. Stereocilia were readily identifiable atop hair cells. Hensen's cells had increased in size, with a lighter staining cytoplasm, but had not increased in height. The inner sulcus had deepened considerably and become less cellular, with only one layer of cells medially. The lip of the limbus was well formed and the underlying stroma was also less cellular. The interdental cells near the lip had lighter staining cytoplasm and their nuclei had reduced in size. The thickness of the spiral limbus had increased (Figures 3c and d).
From 21 to 25 weeks, the tunnel of Corti progressively opened from its base upwards. The nuclei of inner and outer pillar cells were far apart, but their cytoplasmic projections were still in close contact. The inner sulcus cells near the lip of the limbus also attained a lighter staining cytoplasm. Pillar cells were now recognisable even in the most apical turns.
At 26 weeks, the inner sulcus consisted of a single layer of flattened cells. Supporting cells had increased in size and become columnar in appearance. Hensen's cells had become taller and were seen making contact with the tectorial membrane. Claudius cells had also increased in height. Similar changes were seen in the most apical turns, although the inner sulcus was more cellular.
In later stages, the tunnel of Corti continued to develop further, increasing in size even in the most apical turns. Hensen's cells were also seen to develop further. Data are summarised in Table I.
Table I Cochlear development in the human foetus
Discussion
Development of the organ of Corti
The organ of Corti develops from the caudal wall of the cochlear duct. The first sign of cell differentiation in the caudal wall of the developing cochlear duct is a modification of the homogenous pattern of epithelial cells between the greater and lesser ridges, at the end of the 12th week. By the 14th week, hair cells are clearly differentiated from the underlying supporting cells in all turns. Hensen's cells are readily identifiable between the 14th and 15th week. The tectorial membrane is identified long before any evidence of cell differentiation, by the end of the eighth week, over the future interdental region. It is logical to assume that this is the secretory site, as suggested by other studies.Reference Sanchez-Fernandez, Rivera and Macias10 The transformation of Kollicker's organ to inner sulcus is a late event (after the 20th week). It is suggested that apoptosis may play a role in its regression.Reference Hinojosa11
A frequent misconception in the literature regards the development of the tunnel of Corti.Reference Bast and Anson12 A small, oval space develops between the inner and outer hair cells by the end of the 14th week. According to our findings, this does not correspond to the tunnel of Corti, as the pillar cells have not yet differentiated. Pillar cells are clearly distinguishable after the 17th week, whilst opening of the tunnel of Corti is not observed until the 20th week of gestation. This is of particular interest, as there are animal studies that correlate the beginning of auditory function with the development of the tunnel of Corti.Reference Pujol and Hilding13 The opening of the tunnel gradually enlarges and continues to develop even in near-term fetuses.
According to the literature, the cochlea reaches its adult size by approximately the 25th week of gestation.Reference Bredberg14 This does not mean that the cochlea stops developing at this stage. As discussed above, the tunnel of Corti keeps developing up to near-term. The basilar membrane gradually thins, while Hensen's and Claudius cells are becoming high columnar. The transformation of Kollicker's organ to inner sulcus is also a late event. It can thus be assumed that the mechanical properties of the oscillating part of the basilar membrane/organ of Corti/ tectorial membrane complex constantly change during development. This has important implications for the development of the auditory function, as discussed below.
Correlation between auditory function and anatomical structure
When one tries to correlate auditory function with the anatomical development of cochlear structures, it is difficult to assess which factors are responsible for specific functional characteristics, or, indeed, the time of onset of auditory function.
The central auditory system is probably capable of transmitting information before the cochlea or the VIIIth nerve are able to transmit to the central nervous system. Experimental studies in rats have shown that central auditory tracts and secondary auditory centres are capable of transmitting impulses directly to the cortex, three to four days before the appearance of cochlear action potentials.Reference Uziel, Romand and Marot15 Studies in cats and pigs have shown that myelination is already in process around the lamina spiralis fibres before the appearance of action potentials.Reference Pujol and Hilding13, Reference Anniko16 Moreover, the morphological maturation of synapses on cochlear hair cells precedes the functional ability of the cochlea in animals.Reference Pujol and Hilding13, Reference Pujol17 Thus, one can deduce that the limiting factor for the onset of auditory function is localised in the periphery – i.e. the end organ itself. Onset of function probably coincides with the growth of pillar cells and the opening of the tunnel of Corti, the regression of Kollicker's organ, and the formation of the inner spiral sulcus and separation of the tectorial membrane, as suggested by animal studies (Table II).
Table II Correlation of histological and electrophysiological findings in animal models
Summarised from Pufol and Hilding13 CM = Cochlear microphine; AP = action potential
As can be seen, it is difficult to relate the onset of auditory function to a single anatomico-physiological event. It seems rather that there is a temporal coincidence of different developmental events that are responsible for early foetal audition.Reference Romand, Despres and Giry18 According to our own original observations, these changes appear around the 20th week. Other processes involved may include: thinning of basilar membrane; development of endocochlear potential; maturation of ciliary structure; and myelination of ganglion cell bodies and distal processes.
We have previously published observations on the anatomical maturation of various important cochlear structures in the human foetus.Reference Bibas, Liang, Michaels and Wright19, Reference Bibas, Liang, Michaels and Wright20 Functional data concerning the early stages of auditory development in human fetuses are still incomplete because of obvious difficulties in fetal recordings. However, studies of preterm neonates allow insight into some aspects of the maturation of the auditory function. Table III summarises findings from three studies of auditory maturation in the human fetus.
Table III Auditory development in foetuses and pre-term neonates
Cumulative data.6–8 ABR = Auditory Brainstem response; OAE = Otoacoustic Emissions
It can be seen that neonates start to respond to sound as early as the 19th week. Initially, the fetus responds to a restricted frequency range (around 500 Hz). This range expands to include higher frequencies during the course of development. These results should be interpreted with caution, as the stimulus is rather vibro-acoustic and not purely auditory.
As discussed in the introduction, animal studies indicate that the frequency organisation of the cochlea shifts during development. Thus, the regions responsive to high frequencies in mature animals are maximally sensitive to lower frequencies in fetuses, indicating that the place code is not fixed until later in development.Reference Rubel, Lippe and Ryals21 Clinical data from preterm infants show a spontaneous otoacoustic emission frequency shift towards higher frequencies during development, which cannot be explained by changes in middle-ear characteristics (i.e. the presence of fluid).Reference Brienesse, Anteunis, Wit, Gavilanes and Maertzdorf8
What is responsible for this frequency shift in the developing cochlea? Possible causes are: changes in the mass and stiffness of the vibrating part of the basilar membrane during development; establishment of efferent innervation; or changes in the fine structure of hair cells.Reference Rubel9 There is as yet no evidence in the literature to suggest any of the above. Our own observations suggest that the fine structures of the organ of Corti continue to develop well after the 25th week, resulting in changes in the mass and stiffness of the vibrating parts of the basiliar membrane/organ of Corti/tectorial membrane complex. One can postulate that the peak of the travelling wave for each frequency will also shift during development, accounting for the observed frequency shift. Furthermore, as the outer hair cells become active and the efferent innervation supplying them starts to function, so the cochlear amplifier will be activated and this will also alter the cochlear micromechanics.
The question of functional maturation of the cochlea has important clinical implications, apart from pure academic interest. The sensitivity of the human foetus to low frequency sounds means that the fetus will be exposed to maternal sounds of speech (the fundamental frequency for the female voice is around 225 Hz), which may be important for the development of language and of maternal bonding. Low frequencies may be potentially more harmful to the developing fetal ear than high frequencies, and exposure to intense sound should probably be avoided. Finally, the ability to record fetal behavioural responses to auditory stimuli makes the idea of prenatal hearing testing rather attractive.Reference Granier-Deferre, Lecannet, Cohen and Busnel22 Further research is needed to determine whether prenatal screening is feasible and practical.
• Opening of the tunnel of Cortu is not observed until the 20th week and not before as previously reported in the literature
• The fine structure of the organ of Corti continue to develop well after the 25th week with resulting changes in cochlear mechanics, which may account for the frequency shift observed in preterm infants
Conclusions
The small oval space that develops between the inner and outer hair cells by the end of the 14th week of gestation does not correspond to the tunnel of Corti, as the pillar cells have not differentiated at this stage. This is frequently misinterpreted in the literature. Pillar cells are clearly distinguishable after the 17th week, while opening of the tunnel of Corti is not observed until the 20th week of gestation.
The limiting factor for the onset of auditory function is localised to the cochlea. There is no single event, but a temporal coincidence of different developmental events that are responsible for early fetal audition, such as the growth of pillar cells and opening of the tunnel of Corti, the regression of Kollicker's organ, and the formation of the inner spiral sulcus and separation of the tectorial membrane. These changes appear around the 20th week.
The fine structures of the organ of Corti continue to develop well after the 25th week, resulting in changes in the mass and stiffness of the vibrating part of the basilar membrane/organ of Corti/tectorial membrane complex, which may account for the frequency shift observed in preterm infants. These changes will have to be taken into account in the development of prenatal hearing screening tests.
Acknowledgements
This study was supported by the Alexander S Onassis' Public Benefit Foundation (Hellenic Section of Scholarships and Research).
