INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIAL AND METHODS RESULTS DISCUSSION REFERENCES

IT IS THE CONCERN of all countries with fishery industries that their fishing practices should be efficient within the limits imposed by the maintenance of an ecological equilibrium in the exploited waters. Estimation of age and growth of fish is essential for fisheries management. The standard method of obtaining such information for teleosts (which are the predominant marketed fish) is by reading and interpreting calcified structures and most often otoliths. Otoliths are calcareous structures located in the inner ear and bathed in endolymph. They are involved in mechanoreception (6). Fish have three pairs of otoliths, the largest one being called sagittae. Otoliths grow throughout the life of the fish, without resorption (4), in a series of concentric layers of alternating high and low optical density zones that are assumed to record temporal events of seasonal (macroscopic) and daily (microscopic) periodicities (22) (Fig. 1A). In addition to age estimation, recent research has indicated that the chemical composition of otoliths could also provide information concerning fish movement and population dynamics (2, 14, 29, 33), because it is influenced by environmental parameters (temperature, salinity, pollution, etc.) and metabolic events in the life cycle (1, 15, 17, 27, 34, 35).

View larger version (38K): [in this window] [in a new window]   Fig. 1.   A: photograph of transverse section of turbot otolith. Arrows, sites of analysis by wavelength dispersive spectrometry (WD-EM) (results in Fig. 5). B: dorsal view of left saccule, showing disposition for endolymph sampling (sites 1-6).

Considerable progress has been made in improving the accuracy of age estimation, but the mechanisms of otolith formation are still poorly understood. During ontogenesis, variously directed growth gradients result in an elliptical, laterally compressed, distally concave otolith. Unlike most calcifying systems (vertebrate bones, enamel, mollusc shells, corals, etc.; Ref. 31), mineralization takes place in an acellular medium, the endolymph, which contains all the precursors of otolith formation. By comparison with plasma, the composition of fish endolymph is characterized by high K⁺, relatively low Na⁺, a total Ca²⁺ (totCa²⁺) content of ~1-2 mM, and a low protein concentration (2 g/l) (20, 23, 24). Unlike that of higher vertebrates, fish endolymph is more alkaline (pH 8) than plasma (pH 7.20-7.40), mainly due to a high total CO₂ (totCO₂) concentration (20, 23, 24). The ionic composition probably depends on the activity of ion-transporting cells called ionocytes or mitochondrial rich cells (18, 25, 32). In both trout and turbot, the distribution of ionocytes in the saccular epithelium is in two zones (18): the first consists of a ring of large ionocytes around the macula and the second is of smaller cubital ionocytes unevenly grouped at the opposite side of the macula. We thus hypothesized that the location of these two types of cell in relation to the orientation of the sagitta could result in a nonuniform composition of the endolymph. The microtechnique recently developed in trout and turbot (24) permitted the various chemical concentrations to be determined on single 4- to 5-µl samples of endolymph. We therefore selected two sampling sites (Fig. 1B): one distal, facing the smaller ionocytes (sites 1 and 2), and the other proximal, near the ionocytes around the macula (sites 4 and 5). In the present study we improved the technique to enable endolymph samples to be taken at various sites around the otolith and found a lack of uniformity in its composition. Therefore, we carried out microchemical analyses of otolith sections by wavelength dispersive spectrometry (WD-EM) to study whether the lack of uniformity of the endolymph induced a similar nonuniformity in the composition of the otolith.

	MATERIAL AND METHODS

TOP ABSTRACT INTRODUCTION MATERIAL AND METHODS RESULTS DISCUSSION REFERENCES

Fish handling. Trout (Oncorhynchus mykiss) of ~200 or 600 g body wt (depending on the experiment) were supplied by a local fish farmer near Nice, maintained in running tap water at ~14°C in circular tanks, and used for experiments at least 3 wk after transfer from the fish farm. Turbot (Psetta maxima) of 150 g or 2 kg body wt were reared at IFREMER Brest and kept in a running seawater temperature-controlled circuit (14°C). Both groups were fed every morning.

Collection and analysis of blood and endolymph samples. After blood sampling by puncturing the caudal artery, the sample was centrifuged and the separated plasma was kept on ice until analysis. Then endolymph was sampled as already described by Payan et al. (24). Briefly, after the fish were decapitated, the operative field was washed and dried, the saccular wall was incised, and the endolymph was sampled under a stereomicroscope with a capillary tube connected to a micromanipulator and withdrawal pump. In fish (trout or turbot) weighing ~150-200 g, distal samples were withdrawn from sites 1 and 2 and proximal samples from 4 and 5 (see Fig. 1B). In turbot of ~2 kg, six sampling sites (1-6) could be exploited per saccule.

The volume of sampled endolymph was calculated from the length of the fluid in the calibrated capillary as measured with a slide caliper. Protein, Na⁺, Cl, and K⁺ concentrations were determined on individual samples diluted in 0.5 ml of distilled water. totCa, totCO₂, PO³₄, and Mg²⁺ content measurements required the whole sample to be directly introduced into the measurement vial containing the reagent.

For both plasma and endolymph, Na⁺ and K⁺ were determined by flame photometry (Eppendorf); totCa, totCO₂, Cl, PO³₄, and Mg²⁺ were measured by spectrophotometry using Sigma kits; and protein content was determined using Coomassie blue with BSA as a standard.

Preparation and analysis of section of otolith turbot for microchemistry assay. After extraction, the otolith was cleaned in millipore deionized water (MilliQ). Otoliths were then dried at 40°C. After drying, the entire otoliths (sagittae) were embedded in Araldite epoxy resin and polymerized at 25°C for 24 h. Transverse and frontal sections (200 µm thickness), containing the nucleus, were obtained from the otolith with a low-speed saw. Sections were then embedded on glass slides with epoxy resin. After polymerization, they were ground down to the nucleus with successive grades of carborundum paper (800-1200 grade) and polished with successive diamond sprays (9, 3, 1, 0.25 µm) and the sections were cleaned in an ultrasonic bath in deionized water (MilliQ) (13). The usual precautions were taken to avoid chemical contamination of the otoliths and to obtain perfect surface conditions as required in microprobe analyses. These analyses were carried out on a Camebax SX 50 WD-EM microprobe fitted out with five spectrometers used to assay calcium (Ca), potassium (K), sodium (Na), strontium (Sr), and phosphorus (P). Operating conditions were as follows: spot size 5 µm, voltage 10 kV, beam current 15 nA, and peak counting time 40 s for Ca and P and 120 s for Na, K, and Sr.

Five transverse sections from five different otoliths (left sagittae sampled from 5 different turbots) were prepared, and eight sampling sites of analysis zones by WD-EM were defined on each section as indicated in Fig. 1A: proximodorsal, dorsal, distodorsal, distal, distoventral, ventral, proximoventral, and sulcus. Three analyses were carried out at each otolith site. The limits of detection (LD) were 158 parts/million (ppm) for K, 109 ppm for Na, 328 ppm for Sr, and 520 ppm for Ca. The LD of K/Ca, Na/Ca, and Sr/Ca ratios were estimated using the ratio between the K, Na, and Sr LD and the mean Ca concentration.

In the same way, 5 frontal sections from five different otoliths (left sagittae sampled from 5 different turbots) were prepared, and eight sampling sites of analysis zones by WD-EM were defined on each section: proximoposterior, posterior, distoposterior, distal, distoanterior, anterior, proximoanterior, and sulcus. LD were 153 ppm for K, 103 ppm for Na, 312 ppm for Sr, and 546 ppm for Ca. The LD of ratios were estimated as mentioned above.

Phosphorus was undetected in all the analyses and thus rejected from the subsequent data analysis. Data analysis was performed using the mean value of the three assays of each otolith site.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIAL AND METHODS RESULTS DISCUSSION REFERENCES

Ionic, totCO₂, and protein concentrations of plasma and saccular endolymph in trout and turbot are given in Table 1.

Lack of spatial uniformity of ion and protein in endolymph. In both trout and turbot, proteins, PO³₄, and Mg²⁺ were significantly more concentrated in the proximal than in the distal zone, whereas the opposite was the case for K⁺ and totCO₂. [Na⁺] was 20% higher in the proximal zone than in distal in trout, but in the turbot there was no difference. Total Ca²⁺ and Cl contents were similar whatever the sampling zone in the two species. These results clearly demonstrate that the composition of the endolymph surrounding the otolith is not uniform.

Thanks to our microtechnique, Na⁺, K⁺, Cl, and protein concentrations were measured in each sample and thus their interrelationships could be studied. In one series of experiments, different samples of endolymph were withdrawn at different sites inside the saccule. Results are presented in Fig. 2, A (trout) and B (turbot), in which ionic concentrations are plotted as functions of protein concentrations. In both species, similar significant negative relationships exist between the concentrations of K⁺ and protein: a high protein content (proximal zone) corresponds to a low K⁺ content. In some high K⁺ samples, proteins were not even detectable by the Coomassie blue method used. There were certain differences between the trout and turbot values. In the trout, the negative K⁺-to-protein relationship was largely counterbalanced by a positive Na⁺ to protein one and Cl showed no correlation. Thus the estimated osmolarity (sum of the 3 main monovalent ions) was constant. In the turbot, the protein range was about half that of trout, there was no relationship between Na⁺ and protein, but there was a negative relationship between Cl and protein. Consequently, the sum of the three monovalent ions was negatively correlated to protein. The origin of these differences between trout and turbot needs further investigation.

View larger version (24K): [in this window] [in a new window]   Fig. 2.   Relationships between concentrations of monovalent ions (Na+, K+, and Cl) and protein (Prot) concentrations in endolymph of trout (A) and turbot (B). Regression lines are shown together with their equations and significances.

Lack of spatial uniformity of totCO₂ in endolymph. Experiments were performed to check the spatial distribution of the totCO₂ content within the endolymph. In three trout, each weighing ~600 g, six endolymph samples (8-10 µl) were taken and each sample was divided into two aliquots for determination of totCO₂ and protein concentrations. Figure 3A shows that there is a significant negative relationship between totCO₂ and protein content: a high protein content (proximal zone) corresponds to a low totCO₂ content.

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Total CO2 (totCO2) distribution within the endolymph. A: relationship between totCO2 and protein concentrations in trout endolymph. Equation of line is presented and is significant at P < 0.05. B: spatial differences in totCO2 content within turbot endolymph. Means ± SE of 11 determinations from 6 turbots. Six samples taken from sites 1-6 (see Fig. 1B) were removed from each saccule. * P < 0.001 by comparison to site 1.

In 11 turbot, each weighing ~2 kg, six 4- to 5-µl samples were taken from each saccular compartment at the six different sites shown in Fig. 1B. Figure 3B shows that the lowest values of totCO₂ were from proximal sites near the macula (sites 4 and 5) and the highest at the opposite side (site 1). It took ~10 min to remove the six samples from each saccule, during which time exchanges of CO₂ between endolymph and the atmosphere could have occurred. We thus evaluated the effect of the sampling time on the endolymphatic CO₂ concentrations. Two series of samplings were made, alternating the initial sacculus (left or right) to be sampled. In each series, two orders of sampling were used inside the sacculus: 1) a clockwise order in which the first sample was taken in area 1 and the last in area 6 and 2) a counterclockwise order in which the first sample was taken in area 3 and the last in area 4 (see Fig. 1B). ANOVA for repeated measurements indicated that totCO₂ concentrations were significantly different with the otolith sites (P < 0.001) but that sampling order and sacculus position (right or left) made no difference.

Chemical composition of otolith sections by WD-EM microprobe. On frontal and transverse sections, a variation in K/Ca and Na/Ca ratios was observed from zone to zone, whereas Sr/Ca ratios remain roughly uniform. Multivariate ANOVA tests for repeated measurements showed significant differences (P < 0.001) between otolith sites for both K/Ca and Na/Ca ratios (Fig. 4). On the transverse sections, the average K/Ca ratio of the distal zone was twice that of the proximal, whereas the Na/Ca ratio was only 20% higher. On the frontal sections, the K/Ca and Na/Ca ratio differences between the two zones were less marked, being 27 and 12%, respectively. In fact, data from frontal sections showed greater variability both within and between samples, although the chemical patterns are very similar in the two types of sections. The electron microprobe technique is considered to be one of the best for measuring Na and K ions (5). Actually, in this experiment, all K⁺ concentrations measured in the proximal zone in transverse and frontal sections are below or very close to the K⁺ limit of detection. This makes it difficult to interpret the slight differences between the two types of sections.

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Chemical composition of turbot otoliths analyzed by WD-EM. Graphs represent distribution of K/Ca, Na/Ca, and Sr/Ca ratios (means and SE) on left sagittae sections. A: transverse sections from 5 turbots. B: frontal sections from 5 turbots. Comparisons between otolith sites were made with a multivariate ANOVA for repeated measurements, and significance was accepted for P < 0.05 (*). Mean of measurements of proximodorsal zone was reference in A, and in B, reference was proximoposterior zone. PD, proximodorsal; P, posterior; DP, distoposterior; Di, distal; DA, distoanterior; A, anterior; PA, proximoanterior; S, sulcus.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIAL AND METHODS RESULTS DISCUSSION REFERENCES

Lack of uniformity within the endolymph. The present study confirms that the endolymph is a K⁺- and totCO₂-rich fluid (see Table 1) and also brings to light a hitherto unknown and exceptional feature of this fluid in that its main constituents are spatially distributed in a nonuniform way, creating important increasing or decreasing proximodistal (P-D) gradients, a phenomenon that has a physiological relevance to the biomineralization process of the otolith.

In the inner ear of higher vertebrates, Sterkers et al. (30) reported that endolymph was variable in composition. There are small but significant differences in the ionic concentrations between the cochlea and the vestibule and also along the length of the cochlea from base to apex. However, these differences only represented ~8% of the maximum K⁺ and Cl values, and no associated physiological significance was proposed.

The very high K⁺ concentration of the vertebrate inner ear endolymph is unique among extracellular fluids, and its functional significance in mammals is generally related to electrophysiological events in the macula (21, 30). K⁺, being the ion with the highest concentration in the endolymph around the hair cells, normally carries most of the transduction current (21). However, in the present study, the lowest K⁺ concentration was measured near the macula area of the sacculus, which contains the sensory hair cells. However, transduction channels are relatively nonselective cation-passing pores (16) and, because Na⁺ is present at a higher concentration than K⁺ near the macula in fish, this ion would assume a role in the mechanically sensitive transduction mechanism.

Origin of ionic gradients within the endolymph. The presence of these gradients in the endolymph implies separate zones of secretion and reabsorption, and its maintenance depends on several parameters, including the ratio between transepithelial flux and pool, diffusional characteristics in the endolymph, and the presence of chelating agents that have not as yet been studied. In higher vertebrates, it has been clearly established that the secretion of endolymph depends on Na⁺-K⁺-ATPase activity in several areas of the inner ear (30). These are zones of specialized epithelial cells, called dark cells, concentrated in relatively small areas around the sensory hair cells. These cells are osmiophilic and produce various enzymes, such as Na⁺-K⁺-ATPase, Ca²⁺-ATPase, and carbonic anhydrase generally involved in ionic transport and energy metabolism (30). In teleosts, recent studies (25, 32) have demonstrated that saccular ionocytes are mitochondrial-rich cells (DASPMI probe), exhibiting high Na⁺-K⁺-ATPase (antroylouabain) and carbonic anhydrase II activities. In trout and turbot, Payan and coworkers (24) demonstrated that energy-dependent mechanisms are responsible for the maintenance of the K⁺ and H⁺ levels in endolymph. Therefore, from our results in fish and by analogy with the situation in higher vertebrates, we propose that the presence of chemical gradients of monovalent ions in the endolymph is the result of the uneven distribution of ionocytes in the saccular epithelium. The chemical gradients, whether positive or negative, are oriented along a P-D axis related to the distribution of the two ionocyte populations. As yet, it is not possible to assign specific roles to each type of ionocyte present in this epithelium, but we suggest that these specialized cells are the sites of active transfers of K⁺ and H⁺ (or related HCO₃) as illustrated schematically in Fig. 5A.

View larger version (29K): [in this window] [in a new window]   Fig. 5.   A: hypothetical model of protein, K+, and H+ transports across saccular epithelium. Representation of a transverse section of a saccule. Note that overall movement of H+ results in net excretion of H+ as measured by Payan et al. (24) using a titration technique with isolated saccules mounted as closed sacs. H+ arrow signifies titratable acidity and corresponds either to efflux of H+ or influx of HCO3. Influx and efflux of K+ have to be identical to satisfy ionic equilibrium. Shaded areas represent increasing chemical gradient of H+ (or decreasing gradient of HCO3) and protein between distal and proximal zones and increasing chemical gradient of K+ between proximal and distal zones. B: schematic representation of chemical gradients within endolymph. Macula (proximal zone) is on left and distal zone on right of graph with otolith in middle. y-Axis gives concentration of parameters, x-axis shows proximal-distal (P-D) axis. Protein, totCa2+, and HCO3 concentrations were directly measured, whereas Ca2+ and pH concentrations were deduced (see DISCUSSION).

Origin of protein gradient within the endolymph. The sensory area of the vertebrate inner ear contains supporting cells displaying a secretory phenotype that are thought to be involved in the formation of tectoria associated with the macula (16). In fish, the secretory activity of the macula was demonstrated by Gauldie and Nelson (12) with hemotoxylin and eosin. More recently, Davis et al. (7, 8), using a molecular biology approach, reported that supporting cells located at the outer perimeter of the saccular macula were the site of collagen secretion. These observations and our results suggest the involvement of the macula as the region of endolymphatic protein production. A limited zone of production followed by dilution within the endolymph could explain the establishment of the protein gradient (Fig. 5, A and B). Furthermore, the shape and location of the sagitta would act as a physical barrier to the diffusion of the locally produced ions and proteins and would thus help to maintain the chemical gradient (see Fig. 5A).

Are the chemical gradients significant to otolith growth? Otolith mineralization consists of the deposition of CaCO₃ on a protein matrix (9), thus the principal substances involved in otolith growth are proteins, ionized calcium (Ca²⁺), and bicarbonate ions (HCO₃). Growth is not continuous, because inorganic deposition must necessarily follow the formation of the organic matrix (22) and it is believed that, although proteins typically represent only a few percent of the total otolith weight (0.2-10%, Ref. 9), they determine the structural organization and properties of the overall mineralized structure (22), as in other calcified systems (31).

In the present study, the decreasing P-D protein gradient clearly matches the growth axes of the otolith. Actually, the proximal zone facing the macula corresponds to the convex shape of the otolith where the thickness of the increments is greater than on the concave side. Thus the endolymph fraction with the highest protein concentration bathes the side of the otolith characterized by the highest growth. The decreasing P-D gradient of protein has a further significance in relation to the fact that in all biological fluids proteins are strong chelators of Ca²⁺. Mugiya (19) found 72% of diffusible calcium in the endolymph of trout and flatfish. Because the distal endolymph protein concentration is only half that of the proximal, whereas the totCa²⁺ remains constant (Table 1), a P-D increase of ionized calcium is created (Fig. 5B). We also recorded a twofold increase of totCO₂ from proximal to distal. Inasmuch as CO₂ is a highly diffusible gas, the partial pressure of CO₂ is probably constant throughout the endolymph and, with a pH of ~8.0 (20, 23, 24), most of the totCO₂ would be as bicarbonate (HCO₃). The constants of the Henderson-Hasselbach equation (pK and solubility coefficient) are not known for endolymph, so the absolute pH value cannot be calculated. However, on the assumption that these parameters are constant, a difference of 0.3 U pH between proximal and distal zones can be estimated, and this corresponds to a P-D gradient of ~10 mM of totCO₂ (Table 1). Gauldie and Nelson (12), on the basis of crystalline deposition analysis, already proposed the presence of a pH gradient between the sulcus and otolith edge. In conclusion, our results suggest that increasing spatial Ca²⁺, HCO₃, and pH gradients occur from the proximal to the distal zones in the endolymph (Fig. 5B). As Ca²⁺ and HCO₃ combine to form CaCO₃, the presence of these gradients would favor the formation of CaCO₃ along the P-D axis. These results do not agree with the otolith growth gradients (see above). To explain this apparent contradiction, we propose the following hypothesis. The net reaction of Ca²⁺ with HCO₃ generates both CaCO₃ and the hydrogen ion H⁺ according to the equation HCO₃ + Ca²⁺  CaCO₃ + H⁺, and this H⁺ ion must be removed if calcification is to proceed. In most biocalcifying systems, this removal is carried out by specialized cells in close contact with the mineralized structure. However, as no cells are present on the otolith surface to pump out the H⁺, we propose that the H⁺ must diffuse away from its production site toward a zone of lesser concentration along the P-D axis to where it is locally buffered by HCO₃ to form CO₂ gas and water (H⁺ + HCO₃ CO₂ + H₂O). The ionic gradients (Ca²⁺ and HCO₃), which are in favor of CaCO₃ formation along the P-D axis within the endolymph, should help mineralization to progress all around the otolith. In brief, we propose that the gradient of protein concentration explains the differential growth of the otolith and that the deposition of CaCO₃ all around the otolith is brought about by two driving forces: 1) the increasing P-D pH gradient, which allows the proton to move away from the zone of CaCO₃ precipitation, and 2) the presence of increasing P-D ionic gradients (Ca²⁺ and HCO₃), favoring the availability of the ions necessary for this precipitation. This hypothesis supports the assumption that the otolith biocalcification process is initiated in the macula zone (22, 12).

Relationship between the chemical compositions of endolymph and otolith. The chemical composition of the otolith depends on exogenous factors (chemical constituents of the water, food, pollution, factors causing stress, etc.), and endogenous factors (ontogeny, metabolism) (26). As yet, the mechanisms of incorporation of chemical substances are not well known. A narrow relationship probably exists between the chemical compositions of endolymph and otolith, and it was therefore important to study whether the lack of uniformity of the endolymph induces a similar nonuniformity in the otolith.

Variations of the Sr/Ca, Na/Ca, and K/Ca ratios were highly correlated to [Sr], [Na⁺], and [K⁺] variations. The 50% difference of the [K⁺] value between the proximal and distal zones of the otolith is reflected by a similar difference in the endolymph, which clearly indicates a close relationship between endolymph [K⁺] and otolith [K⁺]. The relationship between endolymph [Na⁺] and otolith [Na⁺] has still to be investigated. The fact that, in the endolymph samples, the photometric measurements give the total concentration of ions (i.e., both free and protein linked) may explain the apparent lack of correlation between the two media. Actually, it is the concentration of free ions that is more likely to have a significant correlation with the otolith ionic composition. Further studies of the distributions of free and protein-linked ions, including trace elements used in microchemical studies (3, 10, 11), must also be investigated.

Perspectives