JOURNAL OF MORPHOLOGY 212:99-107 (1992)

Relationship of Saccular Ultrastructure to Otolith Growth in the Teleost Oreochromis niloticus

Z. ZHANG School of Biological Science, University of Wales, Bangor, Gwynedd LL57 2UW, United Kingdom

ABSTRACT The sacculus of Oreochromis niloticus is anatomically separated from the utriculus and semicircular canals. The saccular wall is composed of the sensory epithelium, transitional epithelia, and squamous epithelium. Cellu- lar granules are abundant in the sensory and transitional epithelia but scarce in the squamous epithelium. Over the dorsal side of the dorsal transitional epithelium there exists an oval patch of cells with distinctive microvilli. New finding is a shallow groove which extends from the anterior end of the sensory epithelium approximately halfway down along the ventral perimacular transi- tional epithelium. Small vesicles, which appear "empty" under transmission electron microscopy (TEM), are aggregated in the posterior region of the groove. These small vesicles are also present in both the sensory and transi- tional epithelia. A second kind of vesicle is comparatively large and appears filled with stainable contents. These vesicles are restricted to the sensory region. Both kinds of vesicles appear to be involved in apical secretion and possibly provide the otolithic membrane with fibers. The otolithic membrane is composed of a gelatinuous layer and subcupular meshwork. The meshwork appears to contribute to the formation of the otolith. The small empty vesicles appear to originate in sensory and transitional epithelial cells and may form the subcupular meshwork. The larger filled vesicles are derived predominantly from sensory cells in the sensory epithelium and appear to contribute to the gelatinuous layer of otoliths. o 1992 Wiley-Liss, Inc.

Since the discovery of daily growth incre- ments in fish otoliths (Pannella, '71, '74), considerable knowledge on the general na- ture of these increments has been gathered. However, the cellular mechanisms of otolith formation remain largely unknown.

The saccular otolith (sagittal is enclosed in the sacculus, an endolymph-filled sac which forms part of the labyrinth (Saitoh and Ya- mada, '89). Otolith precursor materials present in the saccular endolymph are be- lieved to be secreted by the saccular epithelial cells. After intraperitoneal injection of 45Ca into rainbow trout, Salmo gairdneri, Mugiya ('74) found a heavy accumulation of 45Ca in the epithelial cells on the medial side of the sacculus and on the facing surface of the otoliths. The uptake of radiocalcium by oto- liths stripped of the saccular tissue was much less than that of those contained within the saccular tissue (Mugiya, '84, '861, suggesting that calcium for otolith growth is supplied, at least partially, by these macular cells. Gaul-

die and Nelson ('88) found a diurnal rhythm in the number of protein vesicles staining with hematoxylin and eosin within the macu- lar cells, corresponding with the daily cycle of otolith growth increment deposition. They suggested that proteins entering the saccular sac derive from the macular cells.

Despite the importance of saccular cellular activities to otolith growth, ultrastructural studies are limited. Most previous research on the fish labyrinth has been concentrated on its neurophysiological function (see Sai- toh and Yamada, '89).

Dunkelberger et al. ('80) presented some new aspects on the relationship between the ultrastructure of the otoliths and otolithic membrane, and suggested that the latter might contribute the organic matrix to the former. During the course of this study, a paper by Saitoh and Yamada ('89) on the

2. Zhang's present address is Pacific Biological Station, Nan- aimo, British Columbia V9R 5K6, Canada.

D 1992 WILEY-LISS, INC.

100 Z. ZHANG

ultrastructure of the saccular epithelium and the otolithic membrane in relation to otolith growth in Oreochromis niloticus was pub- lished, adding further knowledge about the involvement of the saccular epithelium in otolith growth.

In this paper, I report findings on the rela- tionship between specific cell types of the saccular epithelium and the secretory pro- cesses required for otolith growth in Oreo- chromis niloticus.

MATERIALS AND METHODS

One to two month old Oreochromis niloti- cus (Teleostei: Cichlidae), reared in the labo- ratory at 26°C with a photoperiod of 16L:8D and fed on trout pellets once a day to satia- tion, were used. The heads of fish were fixed in 5% glutaraldehyde in a phosphate buffered saline overnight. To study the histological structure some heads were decalcified in 2% EDTA which was changed daily for three days. They were then dehydrated in a graded alcohol series and embedded in LKB his- toresin. Serial sections were cut on an LKB microtome in three planes, horizontal, trans- verse, and longitudinal, so that the labyrinth can be studied from different angles. The sections were mounted on slides, stained with toluidine blue, and examined under a light microscope (LM).

For scanning electron microscope (SEMI study, some of the heads were dehydrated in up to 70% ethanol after the fixation. During the dehydration a small amount of tissue with the sacculus attached was dissected out under a stereo microscope. Tissues were then transferred to 1:1,3:7, and then 1:9 mixtures of ethanol and acetone, leaving them in each mixture for about 10 minutes. Finally they were transferred to pure acetone which was changed twice, before being critical point dried in a Polaron E 3000 Series I1 critical point dryer using liquid COa. After drying they were mounted on SEM stubs prior to dissection or preparation to reveal the inner area to be studied. Specimens on the SEM stubs were coated with gold in a Polaron Sputter Coater E5000, before being exam- ined under an Hitachi S520 SEM.

For transmission electron microscope (TEM) study, some of the fixed heads were washed in the saline buffer, during which time the sacculi (with otolith) together with a small amount of surrounding tissue were gently dissected out under a stereo micro- scope and post-fixed in 2% osmium tetroxide for one hour a t 4°C. They were then washed

in the same buffer prior to block-staining in 2% filtered uranyl acetate overnight. The otoliths were decalcified in the acidic uranyl acetate. The block-stained sacculi were dehy- drated in graded concentrations of ethanol up to pure propylene oxide, which was changed hourly for three times to ensure that they were thoroughly dehydrated. Fi- nally they were embedded in Spurr resin. A block containing a sacculus was sawed out of the polymerized Spurr resin and trimmed using an LKB pyramitome. Ultrathin sec- tions were transversely cut 60 nm on an LKB ultratome 3 with a diamond knife. They were mounted on 100 mesh copper grids coated with Pioloform, stained with lead citrate for 15 minutes, and examined on a Corinth 275 TEM.

RESULTS Composition of the sacculus

Histological observations show that the lab- yrinth can be divided into two parts, one composed of the sacculus and the lagena and the other composed of the utriculus and three semicircular canals. The two parts do not have fluid communication and are linked only by connective tissue. There is an opening between the sacculus and lagena, and the three semicircular canals have openings at either end where they join the utriculus.

The sacculus has a very narrow but ex- tended dorsal projection, which ends blindly (Fig. 1). The saccular wall consists of the sensory epithelium, the transitional epithe- lia, and the squamous epithelium (Fig. 2). The sensory epithelium forms an elongated patch and lies approximately in a vertical position on the medial side of the saccular wall, extending nearly from the posterior end to the anterior end of the sacculus (Fig. 3). Cells are flatter towards the dorsal and ven- tral peripheral regions, forming two transi- tional epithelia. On the lateral sides of the saccular wall is the squamous epithelium (Fig. 2). There is an oval patch of distinctive cells with abundant microvilli over the dorsal side of the dorsal transitional epithelium around the middle section of the sacculus (Figs. 3,4).

There are two types of cells in the sensory epithelium, sensory and supporting cells, which are connected to each other apically by well developed desmosomes. The transitional epithelia contain a single layer of non-sen- sory cells (Fig. 5) and the squamous epithe- lium is composed of very flattened cells. Cel- lular organelles are abundant in the sensory

SACCULAR ULTRASTRUCTURE IN 0. NILOTICUS 101

Fig. 1. Oreochromis niloticus. LM micrograph of a transverse section across the sacculus (S), showing the narrow and extended dorsal projection (arrow), which has a blind end. (L, lateral side; M, medial side; D, dorsal side; V, ventral side). x 110.

Fig. 2. Oreochromis niloticus. LM micrograph of a transverse section of a sacculus, showing the sensory epithelium (S), transitional epithelia (curved arrows), and squamous epithelium (straight arrows). Note the visible gelatinous layer (open arrow). X 276.

Fig. 3. Oreochromis niloticus. SEM micrograph of the medial inner surface of a sacculus, showing the sensory epithelium (S), which extends nearly from the posterior end (P) to the anterior end (A) of the sacculus. Also note an oval patch of distinctive cells (arrow), dorsal to the dorsal transitional epithelium (D). ~ 3 9 5 .

Fig. 4. Oreochromis niloticus. Higher magnification of the patch of cells in Figure 3, showing abundant microvilli on the upper surface of the cells. ~ 3 , 9 5 0 .

102 Z. ZHANG

Figures 5-8

SACCULAR ULTRASTRUCTURE IN 0. NILOTICUS 103

and transitional epithelia. Mitochondria are the most abundant, especially around the nucleus of the sensory cells; Golgi complexes and smooth endoplasmic reticulum are also frequently observed. Rough endoplasmic re- ticulum is also present, but is not as common as the other three organelles. In contrast, organelles are scarce in the squamous epithe- lium. Intracellular granules, about 0.15 pm in diameter, are apparent in the apical areas of the sensory and transitional epithelia (Fig. 6) and a few such granules are also observed in the squamous epithelium.

Otolithic membrane The otolith is attached to the medial epithe-

lium by the otolithic membrane which con- sists of a gelatinous layer and a subcupular meshwork. The former is composed of a tightly arranged fibrous material and is con- fined to the sulcus acousticus region above the sensory epithelium. The subcupular meshwork is composed of relatively loose fi- bers and covers the whole of the sensory and transitional epithelia, extending from the sur- face of the transitional epithelia to the me- dial surface of the otolith (Fig. 7). Over the sensory epithelium these fibres lie beneath the gelatinous layer. In sections from most of the otoliths, the gelatinous layer is visible under LM after staining with toluidine blue (Fig. 2). The subcupular meshwork is never observed with the LM. Some fibers of the subcupular meshwork continue into the oto- lith. Indeed, the organic meshwork is more concentrated in the otolith than in the subcu- pular meshwork (Fig. 8).

Fig. 5. Oreochromis niloticus. TEM micrograph, showing the transition from the sensory to the transi- tional epithelium. Cells change from a tall columnar form in the sensory epithelium (S) to a cuboidal form in the transitional epithelium (T). Also note conspicuous mi- crovilli (arrow) on the apical surface of the transitional epithelium. (Black line, junction of the sensory and tran- sitional epithelia.) ~2,040.

Fig. 6. Oreochromis niloticus. TEM micrograph, showing intracellular granules (arrows) in the apical area of the transitional epithelium. ~34,000.

Fig. 7. Oreochromis niloticus. TEM micrograph, showing the otolith membrane, which is composed of subcupular meshwork ( S ) and gelatinous layer (G) . ~34,000.

Fig. 8. Oreochromis niloticus. TEM micrograph, showing continuation (arrows) of fibers in the subcupu- lar meshwork ( S ) into the otolith (0). ~5 ,525 .

Vesicles Between the sensory epithelium and the

transitional epithelium on the ventral side, there is a shallow groove which extends from the anterior end of the sensory epithelium approximately half way down the sacculus (Fig. 9). Microvilli are scarce in the groove, dense in the sensory epithelium, and much denser in the transitional epithelia (Fig. 10).

Vesicles with a diameter of 0.5-1.5 pm are apparent on the surface of the sensory and transitional epithelia between the microvilli, but they are most concentrated in the groove approximately from the middle part down to the posterior end of the groove (Figs. 10 , l l ) . Attachment of a few vesicles to the apical ends of microvilli is observed (Fig. 12). This kind of vesicle appears empty under TEM. The vesicles are often present among the subcupular meshwork over the transitional epithelia (Fig. 13). A few appear to be par- tially embedded in the epithelium, in which vesicles resembling those attached to the api- cal surface are also observed (Fig. 14). A second kind of vesicle is comparatively

large with a diameter of about 2.5 pm. Its distribution is restricted to the sensory epi- thelium. Some of these have a smooth sur- face and others appear rough (Fig. 15). These larger vesicles appear to have stainable con- tents under TEM and extrusions of such vesicles are observed on sensory cells near the cilium. Vesicle contents resemble the ad- jacent apical cytoplasm (Fig. 16). Very few vesicles are seen on the side of the squamous epithelium.

DISCUSSION

The teleostean labyrinth is enclosed in the skull and lacks an external opening. The fish labyrinth can be physiologically divided into a pars superior, made up of the utriculus and semicircular canals, and a pars inferior includ- ing the sacculus and lagena. Anatomically separation of the two parts is considerable (Lowenstein, '71). There is only a narrow canal joining the two parts in the minnow, Phoxinus (Lowenstein, '361, and in the bichir, Polypterus bichir (Popper, '78); the two parts are completely separated in Gobius jozo (Lo- wenstein, '71), in Nandus nandus (John and Nair, '84), and in the rainbow trout, Salmo Gairdnerii (see Mugiya, '86). The two parts are also separate in Oreochromis niloticus. Therefore, otolith (sagitta) growth in Oreo- chromis niloticus is presumably not associ-

104 2. ZHANG

Fig. 9. Oreochromis niloticus. SEM micrograph of the inner surface of a sacculus, showing the groove (GI, extending from the anterior end (A) to the middle of the sensory epithelium (S) between the sensory epithelium and the ventral transitional epithelium (V). (p, posterior end of the grove.) x 790.

Oreochrornis niloticus. Higher magnification of the posterior area of the groove in Figure 9. Microvilli (arrowheads) are scarce in the groove, dense in the sen- sory epithelium ( S ) , and denser in the transitional epithe-

Fig. 10.

lium (TI. Vesicles (long arrow), however, are dense in the groove. x 3,950.

Oreochromis niloticus. SEM micrograph of the surface of sensory epithelium, showing vesicles (*) between the microvilli. ~ 7 9 0 .

Fig. 11.

Fig. 12. Oreochromis niloticus. SEM micrograph of vesicles (*) attached to the microvilli (arrowhead and arrow). x 15,800.

SACCULAR ULTRASTRUCTURE IN 0. NILOTICUS 105

Fig. 13. Oreochromis niloticus. TEM micrograph, showing empty vesicles (arrow) in the subcupular mesh- work (S) between the transitional epithelium (E) and the otolith (0). ~ 7 , 9 6 8 .

Fig. 14. Oreochromis niloticus. TEM micrograph, showing empty vesicles (*) partially embedded in the transitional epithelium (E). Also note the resemblance of the vesicle (solid arrow) at the apical region of the epithe- lium to that (open arrow) attached to the apical surface of the epithelium. x 79,680.

Fig. 15. Oreochromis niloticus. SEM micrograph of the surface of the sensory epithelium, showing two filled vesicles near cilia (0, one has a rough surface (R) and the other has a relatively smooth surface (S). x 12,450.

Fig. 16. Oreochromis niloticus. TEM micrograph, showing extrusion of a filled vesicle (thick black arrow) from a sensory cell, near the desmosome (white arrow) and cilia (thin black arrow). Note the resemblance of the contents of the vesicle to the apical cytoplasm. x 19,920.

106 2. ZHANG

ated with the utriculus or the semicircular canals.

The middle section of the sacculus of Ore- ochromis niloticus bears an oval patch of cells which display distinctive microvilli. The cytological characteristics of these cells were not studied here and their function, if any, in relation to otolith growth is not clear. Saitoh and Yamada ('89) reported that at the ante- rior to the middle part of the lateral saccular wall in the same species there is an oval patch of cells, which have abundant mitochondria. Unlike the cells described here, those cells do not have microvilli on the luminal surface. The abundance of mitochondria in cells of the sensory and transitional epithelia is pos- sibly associated with the transport of calcium ions to the lumen. This region was also be- lieved to be responsible for supplying calcium for otolith growth (Mugiya, '74). Numerous mitochondria were also reported in the sen- sory epithelium of the sacculus in sea lam- prey, Petromyzon marinus (Popper and Hox- ter, '87). Saitoh and Yamada ('89) also found prominent mitochondria in the transitional epithelia. Golgi complexes, together with smooth endoplasmic reticulum and rough en- doplasmic reticulum, in the sensory and tran- sitional epithelia of Oreochromis niloticus may be responsible for the production of oto- lith protein precursors. Golgi complexes and rough endoplasmic reticulum have also been reported in the cells of the labyrinth sensory epithelium (Sokolowski, '86) and in sacculus transitional epithelium (Saitoh and Yamada, '89). The evident intracellular granules near the apical regions are probably the products of these organelles and perhaps the contents of these granules are secreted into the lu- men. Intracellular granules (electron dense bodies) were found present in the squamous epithelium by Saitoh and Yamada ('89).

Continuation of fibers from the subcupu- lar meshwork into the decalcified otolith is observed. Dunkelberger et al. ('80) also re- ported incorporation of the fibers of the sub- cupular meshwork into the otolith of Fundu- lus and suggested that these fibers contribute to the formation of the organic matrix. This is probably the case for Oreochromis niloti- cus as well.

The grooved area observed under SEM has not been described before. This zone may be mainly responsible for the production of the apparently empty vesicles, as they are so concentrated there. Empty vesicles over the sensory and transitional epithelia in the sac-

culus of fish have been described in other studies (Dale, '76; Dunkelberger et al., '80; Saitoh and Yamada, '89). They are also present in the otocyst of Oreochromis niloti- cus embryos (Zhangand Runham, '92). Their distribution is limited but persistent. Fur- thermore, the present study reveals some vesicles which are partially "embedded" in the epithelium, suggesting that they are prob- ably secretory in nature. These empty vesi- cles, which have also been labelled veils or vesicular dilations, were suggested to be pos- sibly secreted by microvilli in the sacculus of cod and Oreochromis niloticus by Dale ('76) and Saitoh and Yamada ('891, respectively. In the current study, only a few vesicles appear on the distal end of microvilli. The evidence that they are secreted by microvilli is not strong. Furthermore, the presence of numerous empty vesicles over the scarce mi- crovilli in the grooved region is evidence against their secretion by microvilli. My ob- servations on filled vesicles agreed with those of Saitoh and Yamada ('89) in that they are extruded from sensory cells near the junc- tional complex and that their contents are identical to the adjacent apical cell cyto- plasm.

The otolithic membrane of Oreochromis niloticus has two distinctive layers, the gelat- inous layer and the subcupular meshwork. In contrast to the subcupular meshwork, the gelatinous layer stains with toluidine blue. The subcupular meshwork is not visible un- der LM with toluidine blue staining, perhaps because its fibers are too loosely organized to be stained. Nevertheless, the two zones have different structures; the gelatinous layer is composed of tightly packed fibers while the subcupular meshwork has a reticular struc- ture made up of loose fibers. In addition, the initial development of the gelatinuous layer is later than the subcupular meshwork (Sokolowski, '86; Zhang and Runham, '92). The two layers might originate from dif- ferent sources. The distribution of empty vesicles above both the sensory and transi- tional epithelia coincides with that of the subcupular meshwork. In addition, these ves- icles are often observed within this mesh- work, suggesting that they might actually be incorporated into the subcupular meshwork. Dale ('76) also found that the cupular zone I1 of the otolithic membrane, which is equiva- lent to the subcupular meshwork, seems to consist of the same sort of material as the "veils" which are equivalent to the empty

SACCULAR ULTRASTRUCTURE IN 0. NILOTICUS 107

vesicles. The filled vesicles, which are limited to the area of the sensory epithelium under- neath the gelatinous layer, might provide materials to this layer, as also suggested by Saitoh and Yamada ('89). Nevertheless, it cannot be ruled out that empty vesicles could be the artifacts of fixation and filled vesicles may just be blown-out regions of sensory cells.

The present study further clarifies the im- portance of the saccular epithelium to otolith development. The vesicles seem to play an important role in the otolith growth. Determi- nation of the chemical natures of these struc- tures and the otolith membrane could con- firm the relationships between vesicles and otolith growth. Furthermore, comparison of the chemical composition of the protein fi- bers in the otolith and the subcupular mesh- work, might help us do understand the contri- bution of the latter to the former.

ACKNOWLEDGMENTS

I want to express thanks to Dr. N.W. Run- ham for his valuable advice on this work. I also appreciate the comments by two anony- mous referees, which improved the quality of this paper.

LITERATURE CITED Dale, T. (1976) The labyrinthine mechanoreceptor or-

gans of the cod Gadus morhua L. (Teleostei: Gadidae) A scanning electron microscopical study, with special reference to the morphological polarization of the sen- sory cells. Norweg. J . Zool. 24:85-128.

Dunkelberger, D.G., J.M. Dean, and N. Watabe (1980) The ultrastructure of the otolithic membrane and oto- lith in the juvenile mummichog, Fundulus heterocli- tus. J. Morphol. 163:367-377.

Gauldie, R.W., and D.G. Nelson (1988) Aragonite twin-

ning and neuroprotein secretion are the cause of daily growth rings in fish otoliths. Comp. Biochem. Physiol. 9Ok.501-509.

John, K.C., and N.B. Nair (1984) On the inner ear (membranous labyrinth) of Nandus nandus (Hamil- ton) (Nandidae-Perciformes). J . Anim. Morphol. Phys- iol. 31:155-158.

Lowenstein, 0. (1936) The equilibrium function of the vertebrate labyrinth. Biol. Rev. 11:113-145.

Lowenstein, 0. (1971) The labyrinth. In W.S. Hoar and D.S. Randell (eds): Fish Physiology. New York: Aca- demic Press, vol. 5. pp. 207-240.

Mugiya, Y. (1974) Calcium-45 behavior at the level of the otolithic organs of rainbow trout. Bull. Jpn. SOC. Sci. Fish. 40:457-463.

Mugiya, Y. (1984) Diurnal rhythm in otolith formation in the rainbow trout, Salmo gairdneri: seasonal reversal of the rhythm in relation to plasma calcium concentra- tions. Comp. Biochem. Physiol. 78A:289-293.

Mugiya, Y. (1986) Effects of calmodulin inhibitors and other metabolic modulators on in vitro otolith forma- tion in the rainbow trout, Salmo guirdneri. Comp. Biochem. Physiol. 84k57-60.

Pannella, G. (1971) Fish otoliths: daily growth layers and periodical patterns. Science 173:1124-1127.

Pannella, G. (1974) Otolith growth patterns: an aid in age determination in temperate and tropical fishes. In T.B. Bagenal (ed): Ageing of Fish. London: Unwin, pp. 28-39.

Popper, A.N. (1978) Scanning electron microscopic study of the otolithic organs in the bichir (Polypterus bichir) and shovel-nose Sturgeon (Scaphirhynchus platoryn- chus). J . Comp. Neurol. 181:117-128.

Popper, A.N., and B. Hoxter (1987) Sensory and nonsen- sory ciliated cells in the ear of the sea lamprey, Petromy- zon marinus. Brain Behav. Evol. 30r43-61.

Saitoh, S., and J . Yamada (1989) Ultrastructure of the saccular epithelium and the otolithic membrane in relation to otolith erowth in tilauia. Oreochromis nilo- ticus (Teleostei: Cychlidae). Trins.' Am. Microsc. SOC. 108:223-238.

Sokolowski, B.H.A. (1986) Development of the otolith in embryonic fishes with special rei'erence to the toadfish, Opsanus tau. Scanning Electron Microsc. 4: 1635- 1648.

Zhang, Z., and N.W. Runham (1992) Initial development of Oreochromis niloticus Otolith. J. Zool. (in press).