Androgens and Very Low Density Lipoprotein Are Essential for the Growth of Previtellogenic Oocytes from Japanese Eel, Anguilla japonica, In Vitro¹

Abstract

To investigate the regulation of lipid uptake into the eel oocyte in more detail, effects of 11-ketotestosterone (11-KT) and lipid transporters (lipoproteins) were determined in vitro. Ovarian explants from previtellogenic Japanese eels (Anguilla japonica) were incubated for 28 days with 11-KT and/or with very low density lipoproteins (Vldl), low density lipoproteins (Ldl), or high density lipoproteins (Hdl) purified from eel plasma. The androgen 11-KT induced notable increases in oocyte diameter, which were accompanied by the appearance of vacuoles rather than lipid. Ldl and Hdl increased oocyte diameters, whereas Vldl did not. However, coincubation of 11-KT and Vldl, but not of Ldl or Hdl, resulted in dramatic increases in oocyte size and lipid droplet surface area. Effects of both 11-KT (oocyte size) and Vldl (lipid droplet surface area) were dose-dependent between 1 and 100 ng/ml and between 0.5 and 5 mg/ml, respectively. Interestingly, abnormal oocyte cytology under conditions of coculture with 11-KT and Vldl could essentially be prevented if Vldl concentrations were high enough (≥ 5 mg Vldl/ml medium). Unlike 11-KT, estradiol-17beta had no effect on oocyte diameter or lipid droplet surface area. We conclude that Vldl is a key transporter of neutral lipids that accumulate into the eel oocyte during oogenesis and that Vldl-dependent lipid uptake is stimulated by the androgen 11-KT.

Introduction

Oocytes from oviparous vertebrates store large amounts of yolk material and oil (neutral lipids) in the cytoplasm during oogenesis [1]. In teleosts, the appearance of lipid droplets is generally first observed at the previtellogenic oocyte growth stage, and ongoing accumulation continues as development proceeds [2]. The lipids accumulated in the ooplasm are primarily metabolized for energy during embryogenesis [3]. Considerable variation occurs in the final amount of lipids in eggs among species. Thus, eggs from some species appear to be totally devoid of lipids, whereas those from others store large amounts [2, 3].

In teleosts, the origin of lipids in the lipid droplets is unclear; regardless, it has been suggested that lipoproteins play an important role in the accumulation of lipids into the oocytes [3, 4]. Lipoproteins are assemblies of lipids and apoproteins; the latter interact with the hydrophilic parts of the lipids to form a lipid-protein “shell” surrounding a hydrophobic core, allowing the transport of lipids from one tissue to another through the bloodstream. Lipoproteins are divided into several classes on the basis of their specific density following ultracentrifugation, yielding very low density lipoproteins (Vldl), low density lipoproteins (Ldl), and high density lipoproteins (Hdl). The different lipoprotein classes possess characteristic compositions of lipids and apoproteins, reflecting specific metabolic functions [5]. Vitellogenins (Vtg), which are recognized as very high density lipoproteins, are known for their role as yolk precursor proteins and their incorporation into the oocytes [4, 6, 7]. However, the accumulation of lipid droplets precedes the estrogen-mediated rise in circulating Vtg levels and the deposition of yolk globules. Furthermore, Vtg transports mainly phospholipids, which are considered to be polar lipids, whereas lipid droplets primarily consist of neutral lipids, such as triacylglycerols, and wax or steryl esters. Therefore, it seems unlikely that Vtg is the key lipoprotein transporting the neutral lipids found in the oil droplets. A much more likely transporter of these “lipid-droplet lipids” is Vldl, which is known to be rich in triacylglycerol [3, 4]. To our knowledge, however, this hypothesis has not been tested.

The involvement of the endocrine system, chiefly through hormonal mediators, including gonadotropins and sex steroids, in vitellogenic oocyte growth has been extensively studied (see, e.g., [8]). In contrast, information concerning the endocrine regulation and mechanisms of previtellogenic oocyte growth is very limited. Recently, 11-ketotestosterone (11-KT), a major androgen in teleosts, has been detected in blood of some female teleosts [9], suggesting that 11-KT plays a role in the biology of these females. Several studies regarding the effects of 11-KT have been conducted, especially in eels (Anguilla spp.), which have large amounts of lipids in the ooplasm. In temperate anguillids, serum 11-KT levels increase markedly with silvering; this event encompasses a transformation that prepares the fish for a long-distance spawning migration at a time roughly coinciding with the onset of puberty and, thus, with a shift in ovarian development from the previtellogenic to the vitellogenic stage [10, 11]. Moreover, in vivo treatment with exogenous 11-KT induced silvering-related changes, including effects on previtellogenic oocytes, as reflected in a greater abundance of lipid droplets in oocytes from short-finned eel (Anguilla australis) [12] and Japanese eel (Anguilla japonica) [13]. Furthermore, in vitro culture of ovarian fragments from short-finned eel showed that 11-KT induced an increase in the size of previtellogenic oocytes and promoted triacylglycerol accumulation into the ovary [14]. These studies suggest that 11-KT is involved in controlling aspects of previtellogenesis, but the mechanisms underlying this regulation in terms of previtellogenic oocyte growth and lipid droplet accumulation are still far from understood.

In vitro approaches have proven to be very useful in examining the effects of various factors regulating gametogenesis. In Japanese eel, long-term organ culture systems for testis [15, 16] and ovary [17] under serum-free conditions have been developed, and in vitro Vtg incorporation and yolk globule formation can be demonstrated using such culture systems [17]. Therefore, we employed in vitro methodologies in the present study to investigate lipid droplet accumulation into previtellogenic oocytes of Japanese eel. We focused on lipoproteins and 11-KT as the factors involved in lipid droplet accumulation, and four experiments were designed to evaluate the effects of these factors. In experiment 1, the effects of prevailing lipoproteins (Vldl, Ldl, and Hdl) and 11-KT were examined. Further analyses were carried out to investigate the effects of various concentrations of Vldl in experiment 2 and those of 11-KT in experiment 3. In addition, experiment 4 was conducted to compare the effects of 11-KT and estradiol-17β (E₂).

Materials and Methods

Animals

All animal care and use in the present study were conducted in accordance with the animal experimentation guidelines of our university (Hokkaido University Animal Care Committee). All in vitro culture experiments were conducted using ovarian tissue from previtellogenic female Japanese eels. The animals had been wild-captured in the glass eel stage and feminized with E₂ (10 mg/kg feed) for 5 mo, from March to August, in circulating freshwater tanks with a capacity of 1000 L. Animals were subsequently reared on a standard diet in 4500-L tanks for an additional 10 or 11 mo, from August to June or July in the next year, until use for experimentation. Fish were maintained at a water temperature of 26 ± 5°C under natural photoperiod in breeding facilities at Hokkaido University, Faculty of Fisheries (41°N, 140°E; Hakodate, Japan).

Isolation of Lipoproteins

Male eels (body weight, 300–400 g), purchased from a commercial eel supplier (Daigotsusho Co., Ltd., Shizuoka, Japan), were used for isolation of lipoproteins. Eels were euthanized in 2-phenoxyethanol, and after tail removal, blood was collected from the caudal vessels in heparinized tubes containing a final concentration of 1 mM ethylenediaminetetra-acetic acid, 1 mM PMSF, and 5 μg/ml of leupeptin. The blood was centrifuged at 850 × g for 20 min at 4°C. The supernatant was collected as plasma and immediately used for lipoprotein isolation. Lipoproteins were obtained by sequential ultracentrifugation techniques [18] at 16°C using a Beckman Optima TL ultracentrifuge with a TLA 100–3 rotor. The density (d) range of eel lipoproteins has been previously reported [19] and was applied in our experiments to isolate the different lipoprotein classes (Vldl, d < 1.006 g/ml; Ldl, 1.006 g/ml < d < 1.085 g/ml; Hdl, 1.085 g/ml < d < 1.210 g/ml). The density was adjusted by the addition of NaCl-NaBr solution. After centrifugation at d = 1.006 g/ml and 410 000 × g for 3 h, Vldl was concentrated in a supernatant layer. Ldl was first recovered from the infranate of the Vldl fraction after further centrifugation for 4 h at d = 1.085 g/ml and 410 000 × g, followed by Hdl after centrifugation at d = 1.210 g/ml and 410 000 × g for 8 h. These isolated lipoproteins were dialyzed against Dulbecco phosphate-buffered saline, concentrated by ultrafiltration, and immediately used in ovarian culture experiments. The protein concentration of each lipoprotein fraction was determined using the Micro BCA Protein Assay Kit (Pierce) with bovine serum albumin as a standard. In the present study, the amount of each lipoprotein is expressed on the basis of protein concentration. The purity of each lipoprotein fraction was checked by polyacrylamide gel electrophoresis under both native and denaturing conditions.

Organ Culture Techniques

Ovarian culture was conducted according to the method of Kayaba et al. [17]. In brief, the basal culture medium consisted of Leibovitz L-15 medium (Gibco) supplemented with 0.5% bovine serum albumin fraction V (Sigma), 0.16 μM bovine insulin (Sigma), and 10 mM Hepes (adjusted to pH 7.4 with 1 M NaOH). Freshly removed eel ovaries were washed and cut into fragments of 2–3 mm in each dimension in basal culture medium. The ovarian fragments were placed on floats of elder pith (pretreated with 100% ethanol, autoclaved, dried, and covered with a nitrocellulose membrane) in 24-well, plastic tissue-culture dishes (Iwaki). Five hundred microliters of culture medium were added to each well, and three replicate incubations were used per treatment group. Tissues were cultured for 28 days at 20°C in 100% humidified air, and the culture medium was changed every 7 days during the culture period.

Experimental Design

Experiment 1: Effects of different lipoprotein classes and 11-KT on lipid droplet accumulation

Previtellogenic ovarian fragments removed from one female Japanese eel were incubated in basal culture medium treated with or without each isolated lipoprotein (Vldl, Ldl, or Hdl; 5 mg/ml) in the presence or absence of 11-KT (10 ng/ml; Sigma).

Experiment 2: Effects of various Vldl concentrations on lipid droplet accumulation

Ovarian fragments from one previtellogenic female eel were exposed to Vldl (0, 0.5, 1, 2.5, or 5 mg/ml) in the presence or absence of 11-KT (10 ng/ml).

Experiment 3: Effects of various 11-KT concentrations on lipid droplet accumulation

Different concentrations of 11-KT (0, 0.1, 1, 10, or 100 ng/ml) were added to cultures of previtellogenic ovarian fragments supplemented with or without Vldl (5 mg/ml).

Experiment 4: Comparison of effects between 11-KT and E₂ on lipid droplet accumulation

Ovarian fragments from the same eel used in experiment 3 were cultured in basal medium or in basal medium supplemented with Vldl (5 mg/ml), 11-KT (10 ng/ml), or E₂ (Sigma; 10 ng/ml) or cotreated with Vldl and either 11-KT or E₂.

Histological Analysis

Freshly dissected and incubated ovarian fragments were fixed in 2% paraformaldehyde and 2.5% glutaraldehyde in 0.1 M phosphate buffer (pH 7.2) for 24 h at 4°C. Samples were postfixed for 2 h in 1% osmium tetroxide in the same buffer and embedded in EPON 812 (TAAB) according to standard procedures. Semithin sections (thickness, 1 μm) were cut and stained with toluidine blue or methylene blue-azure II for light microscopic examination using a Nikon ECLIPSE 80i microscope.

Image Analysis

To evaluate the effects of various culture conditions on previtellogenic oocyte growth and lipid droplet accumulation, the surface areas of oocytes and oil droplets in cross sections were measured. Because considerable differences in the features of oocytes were found, even within the same ovarian fragment, the following method was employed to obtain the largest oocyte group with respect to oocyte size and the surface area of lipid droplets: From each of three replicate incubations in each treatment group, 20 oocytes with a nucleus (germinal vesicle) visible in the section were randomly examined, and the data from the 10 highest-ranking oocytes were retained. Accordingly, data from 30 oocytes were obtained per treatment group with regard to both oocyte size and the amount of lipid droplet. Microscope images were captured using a Nikon video camera (DXM 1200F) and analyzed using Image J software (National Institute of Health).

Statistics

Data from the 10 highest-ranking oocytes were averaged to yield a single value for each replicate. Values for oocyte and lipid droplet surface areas for each of the three replicates from each treatment group were subsequently expressed as the mean ± SEM. Data were log-transformed, if necessary, and then analyzed by Student t-test (fresh, initial data compared to data from control incubations) or ANOVA to test whether treatment means differed significantly. Multiple post hoc comparisons were done using Tukey test. Differences between means were considered to be significant at P < 0.05.

Results

Experiment 1: Effects of Different Lipoprotein Classes and 11-KT on Lipid Droplet Accumulation

Photomicrographs of ovarian fragments from the various treatment groups are shown in Figure 1, whereas surface areas of oocytes and lipid droplets in the largest oocyte group are presented in Figure 2. In control ovaries cultured only with basal medium, oocytes tended to be larger than oocytes in freshly dissected fragments (F_1,4 = 7.162, P = 0.055), whereas lipid droplet surface area was not affected (F_1,4 = 2.751, P = 0.173). Addition of lipoproteins did not affect oocyte surface area compared with controls; nonetheless, a treatment effect (F_3,16 = 3.409, P = 0.043) was seen, which was attributable to a significant difference in oocyte size between Vldl- and Ldl-supplemented cultures (Tukey post hoc comparisons). This difference was not reflected in lipid droplet surface area, which was comparable between explants from different treatments (presence or absence of lipoprotein) (Fig. 1, A–E). Cosupplementation of incubation media with 11-KT resulted in marked increases in oocyte size (F_1,16 = 42.712, P < 0.001) and a notable reduction in lipid droplet surface area (F_3,16 = 8.320, P < 0.001) except in explants that received both 11-KT and Vldl; this latter treatment dramatically increased (one-way ANOVA: F_3,8 = 9.555, P = 0.005) the amount of lipid (Fig. 1H). Interestingly, solo treatment with 11-KT resulted in abnormal cytology, notably the appearance of vacuoles in the ooplasm (Fig. 1F), which was prevented in explants cotreated with lipoprotein and 11-KT (Fig. 1, J and L). Meanwhile, differences were found in oocyte cytology, even among oocytes within the same treatment group. Thus, in all 11-KT treatment groups, normal oocytes not exhibiting significant change were encountered (Fig. 1, G and K), whereas cotreatment with Vldl and 11-KT yielded some oocytes with only small amounts of lipid droplets (Fig. 1I).

Fig. 1

Photomicrographs of toluidine blue-stained sections (thickness, 1 μm) of Japanese eel (Anguilla japonica) ovarian fragments before (A) and after culture for 28 days in basal culture medium alone (B) or with 5 mg/ml of Vldl (C), 5 mg/ml of Ldl (D), 5 mg/ml of Hdl (E), 10 ng/ml of 11-KT (F and G), 5 mg/ml of Vldl and 10 ng/ml of 11-KT (H and I), 5 mg/ml of Ldl and 10 ng/ml of 11-KT (J and K), or 5 mg/ml of Hdl and 10 ng/ml of 11-KT (L). Arrows indicate lipid droplets, and arrowheads indicate vacuoles. Bar = 50 μm.

Fig. 2

Surface areas of oocytes (A) and their lipid droplets (B) in Japanese eel (Anguilla japonica) ovarian explants cultured in vitro for 28 days. Measurements were made on the largest oocytes in cross sections of ovarian fragments before and after culture in basal culture medium treated with or without isolated lipoprotein (Vldl, Ldl, or Hdl; 5 mg/ml) or 11-KT (10 ng/ml) or cotreated with each lipoprotein and 11-KT. Bars represent the mean ± SEM (n = 3). Different letters above bars represent significant differences between treatment groups of lipoproteins plus 11-KT (Tukey, P < 0.05). Asterisks represent a significant main effect of 11-KT (***P < 0.001).

Experiment 2: Effects of Various Vldl Concentrations on Lipid Droplet Accumulation

The surface area of oocytes tended to increase (F_1,4 = 4.309, P = 0.107), and that of lipid droplets was decreased (F_1,4 = 20.031, P = 0.011), after incubation of ovarian fragments in basal culture medium alone and when compared to preincubation controls (initial) (Fig. 3, A and B). Vldl had no effect on oocyte size (F_4,20 = 0.521, P = 0.721) (Fig. 4A), but its addition resulted in increased surface areas of lipid droplets in a dose-dependent manner (F_4,20 = 59.421, P < 0.001) (Figs. 3, C and D, and 4B). Dose-dependent effects of Vldl on the surface area of lipid droplets were dramatically enhanced when explants were cotreated with 11-KT (Fig. 3, F–H), yielding significant differences in the amount of intra-oocytic lipids between many of the Vldl treatment regimes. Moreover, addition of 11-KT to Vldl-containing media induced a 30% increase in oocyte surface area compared to Vldl-supplemented incubations at the same concentrations (F_1,20 = 71.582, P < 0.001) (Fig. 4A). In contrast, varying the Vldl concentrations induced only minor changes in the surface area of oocytes. Treatment with 11-KT in the absence of Vldl induced many abnormal cytological features (Fig. 3E), and normal oocytes were almost never seen. In addition, notable decreases in the surface area of lipid droplets were observed in 11-KT-only incubations compared to the control. However, following cotreatment with Vldl, the incidence of abnormal oocytes was reduced, and the amount of lipid increased as Vldl concentrations rose. In addition, when cotreating ovarian explants with 5 mg/ml of Vldl plus 10 ng/ml of 11-KT, abnormal oocytes were not found at all (Fig. 3, G and H). Furthermore, oocytes with greater abundance of lipid droplets were observed (Fig. 3G), leading to a remarkable increase in the surface area of lipid droplets compared to single treatment with either 5 mg/ml of Vldl or 10 ng/ml of 11-KT. However, in this treatment group, an increase in the amount of lipid droplets was not induced in all oocytes, and oocytes with few lipid droplets could also be observed (Fig. 3H).

Fig. 3

Photomicrographs of methylene blue-azure II-stained sections (thickness, 1 μm) of Japanese eel (Anguilla japonica) ovarian fragments before (A) and after culture for 28 days in basal culture medium alone (B) or with 1 mg/ml of Vldl (C), 5 mg/ml of Vldl (D), 10 ng/ml of 11-KT (E), 1 mg/ml of Vldl and 10 ng/ml of 11-KT (F), or 5 mg/ml of Vldl and 10 ng/ml of 11-KT (G and H). Arrows indicate lipid droplets, and arrowheads indicate vacuoles. Bar = 50 μm.

Fig. 4

Surface areas of oocytes (A) and their lipid droplets (B) in Japanese eel (Anguilla japonica) ovarian explants cultured in vitro for 28 days. Measurements were made on the largest oocytes in cross sections of ovarian fragments before and after culture in basal culture medium treated with or without Vldl (0.5, 1, 2.5, or 5 mg/ml) in the presence or absence of 11-KT (10 ng/ml). Bars represent the mean ± SEM (n = 3). Different lowercase letters above bars represent significant differences between solo treatment groups of Vldl, whereas different uppercase letters represent significant differences between treatment groups of Vldl plus 11-KT (Tukey, P < 0.05). Hash symbols (#) represent significant differences (P < 0.05) in means between fragments before culture (initial) and those cultured in basal culture medium alone, and asterisks represent a significant main effect of 11-KT (***P < 0.001).

Experiment 3: Effects of Various 11-KT Concentrations on Lipid Droplet Accumulation

Culture of ovarian explants in basal medium resulted in increased oocyte surface area compared to preincubated tissue (F_1,4 = 21.618, P = 0.010), whereas lipid droplet area decreased, albeit not significantly (F_1,4 = 4.729, P = 0.095) (Figs. 5, A and B, and 6, A and B). Furthermore, the surface area of oocytes treated with 11-KT increased in a dose-dependent manner (main effect: F_4,20 = 27.246, P < 0.001) (Figs. 5, C and D, and 6A) and was greater at 10 or 100 ng/ml of 11-KT compared to 0, 0.1, or 1 ng/ml (Tukey post hoc comparisons). Some cytologically abnormal oocytes were first found in the 1 ng/ml of 11-KT group (Fig. 5C). The relative abundance of these abnormal oocytes increased following treatment with 11-KT at 10 or 100 ng/ml (Fig. 5D), leading to a decreasing trend in surface area of lipid droplets compared to controls. Supplementation with Vldl resulted not only in increased oocyte size (F_1,20 = 14.371, P = 0.001) but also in greater abundance of lipid droplets in the ooplasm (F_1,20 = 216.996, P < 0.001) (Figs. 5, E–G, and 6B); furthermore, Vldl significantly interacted with 11-KT dose (F_4,20 = 4.301, P = 0.011), as reflected in superficially reversing trends for lipid droplet surface area when comparing the dose-response curves in the absence and presence of Vldl. Unexpectedly, however, some oocytes were observed to have small amounts of lipid droplets, and few abnormal oocytes were seen following incubation with 11-KT at 10 or 100 ng/ml, even in the presence of Vldl (Fig. 5H).

Fig. 5

Photomicrographs of toluidine blue-stained sections (thickness, 1 μm) of Japanese eel (Anguilla japonica) ovarian fragments before (A) and after culture for 28 days in basal culture medium alone (B) or with 1 ng/ml of 11-KT (C), 100 ng/ml of 11-KT (D), 5 mg/ml of Vldl (E), 5 mg/ml of Vldl and 1 ng/ml of 11-KT (F), or 5 mg/ml of Vldl and 100 ng/ml and 11-KT (G and H). Arrows indicate lipid droplets, and arrowheads indicate vacuoles. Bar = 50 μm.

Fig. 6

Surface areas of oocytes (A) and their lipid droplets (B) in Japanese eel (Anguilla japonica) ovarian explants cultured in vitro for 28 days. Measurements were made on the largest oocytes in cross sections of ovarian fragments before and after culture in basal culture medium treated with or without Vldl (5 mg/ml) in the presence or absence of 11-KT (0.1, 1, 10, or 100 ng/ml). Bars represent the mean ± SEM (n = 3). Different lowercase letters above bars represent significant differences between solo treatment groups of 11-KT, whereas different uppercase letters represent significant differences between treatment groups of Vldl plus 11-KT (Tukey, P < 0.05). Hash symbols (#) represent significant differences (P < 0.05) in means between fragments before culture (initial) and those cultured in basal culture medium alone, and asterisks represent a significant main effect of Vldl (***P < 0.001).

Experiment 4: Comparison of Effects Between 11-KT and E₂ on Lipid Droplet Accumulation

Results concerning effects of 11-KT and Vldl were as reported above (experiment 3). Thus, 11-KT increased the surface area of oocytes, decreased that of lipid droplets, and led to abnormal cytology (Figs. 7, A–C, and 8), whereas increases in oocytes and lipid droplet size were seen following cotreatment with Vldl and 11-KT (Fig. 7E). In contrast, E₂ did not induce alterations in cytology (Fig. 7D), oocyte size, or lipid droplet abundance (Tukey post hoc comparisons). Cotreatment with Vldl and E₂ (Fig. 7F) resulted in a slight, but nonsignificant, reduction in oocyte size compared to either Vldl or E₂ (Fig. 8A). The surface area of lipid droplets was increased compared to the E₂-only treatment, reflecting a main effect of Vldl supplementation (F_1,12 = 85.812, P < 0.001) (Fig. 8B).

Fig. 7

Photomicrographs of toluidine blue-stained sections (thickness, 1 μm) of Japanese eel (Anguilla japonica) ovarian fragments after culture for 28 days in basal medium alone (A) or incubated with 5 mg/ml of Vldl (B), 10 ng/ml of 11-KT (C), 10 ng/ml of E₂ (D), 5 mg/ml of Vldl and 10 ng/ml of 11-KT (E), or 5 mg/ml of Vldl and 10 ng/ml of E₂ (F). Arrows indicate lipid droplets, and arrowheads indicates vacuole. Bar = 50 μm.

Fig. 8

Surface areas of oocytes (A) and their lipid droplets (B) in Japanese eel (Anguilla japonica) ovarian explants cultured in vitro for 28 days. Measurements were made on the largest oocytes in cross sections of ovarian fragments before and after culture in basal culture medium treated with or without Vldl (5 mg/ml), 11-KT (10 ng/ml), or E₂ (10 ng/ml) or co-treated with Vldl and either 11-KT or E₂. Bars represent the mean ± SEM (n = 3). Different lowercase letters above bars represent significant differences between solo treatment groups with either 11-KT or E₂, whereas different uppercase letters represent significant differences between cotreatment groups of Vldl and either 11-KT or E₂ (Tukey, P < 0.05). Hash symbols (#) represent significant differences (P < 0.05) in means between fragments before culture (initial) and those cultured in basal culture medium alone, and asterisks represent a significant main effect of Vldl (***P < 0.001).

Discussion

In the present study, we investigated the effects of lipoproteins and 11-KT on lipid droplet accumulation into previtellogenic oocytes of Japanese eel in vitro using a long-term organ culture system developed previously for eel testis [15, 16] and ovary [17]. In a preliminary experiment, we examined the effects of culture duration (14 and 28 days) on lipid droplet accumulation. Differences were not seen between the initial control and the 11-KT-Vldl cotreatment group after 14 days of culture, whereas markedly increased lipid droplets were found in oocytes of the 11-KT-Vldl cotreatment group after 28 days (data not shown). Thus, we conducted subsequent in vitro experiments for 28 days. Experiment 2 yielded dose-dependent effects of Vldl on lipid accumulation, most markedly in the presence of 11-KT. Moreover, in experiment 3, lipid accumulation was promoted with rising concentrations of 11-KT in the presence of Vldl. These results strongly indicate that lipids in the lipid droplets originate from Vldl and that 11-KT plays an important role in their transfer and/or accumulation into oocytes.

Lipid droplet surface areas were greater when incubations were supplemented with 11-KT and Vldl compared to addition of other lipoprotein preparations, again suggesting that Vldl is the main source of lipids in the lipid droplets. Ldl had no effect on lipid droplet surface area, although it did inhibit the formation of the vacuoles induced by 11-KT. In contrast, Hdl contributed to increased amounts of lipids, but to a lesser extent than Vldl. Therefore, Hdl may also be a significant source of lipid droplets accumulating in the oocytes. However, the mechanisms controlling this accumulation may differ from Vldl-dependent lipid uptake, because a decrease in lipid droplet abundance was observed using cotreatment with 11-KT.

The composition of plasma lipoproteins from Japanese eel has been reported by Ando and Matsuzaki [19]. Vldl consists of 91% lipid and 9% protein, and the main lipid component of Vldl is triacylglycerol. Ldl contains 33% protein, and lipids are primarily made up of triacylglycerols, phospholipids, and cholesterol esters. In contrast, about half of Hdl is protein, and phospholipids and cholesterol esters are its main lipid constituents. In teleosts, lipids in lipid droplets of oocytes consist mainly of neutral fats, such as triacylglycerols, and wax or steryl esters [3]. A comparison of lipid constituents between lipoproteins and lipid droplets identifies Vldl, rich in neutral fats and containing a high ratio of lipids, as the most plausible lipoprotein responsible for delivering the lipids found in the lipid droplets.

In the present study, the concentration of three different lipoproteins was standardized over protein. Hence, the relative amount of Vldl added to the culture medium was approximately fourfold higher than that of Ldl and approximately sixfold higher than that of Hdl, because the protein ratio in Vldl is 9%, compared to 33% in Ldl and 50% in Hdl. However, the amounts added are a direct reflection of the component characteristics of each lipoprotein; because the function of the lipoproteins depends on the apoprotein composition, the key role of Vldl in supplying the fats for the oocyte lipid droplets is further emphasized.

The cellular pathways that are used for uptake of lipid from Vldl into oocytes have not been conclusively identified in teleost fish. Two possible pathways have been proposed to mediate lipid uptake into teleost oocytes [20]: First, Vldl is incorporated into the oocyte by receptor-mediated endocytosis, after which Vldl is processed to generate free fatty acids in the cell. Second, Vldl-associated triacylglycerols are hydrolyzed by an enzyme, such as follicular (non-oocytic) lipoprotein lipase (Lpl), and resulting fatty acids then enter oocytes as a substrate for biosynthesis of neutral lipids. Either of these pathways, or both, may be involved in lipid uptake into oocytes in the eel.

In general, lipoproteins are incorporated into target cells by receptor-mediated pathways. In several oviparous animals, the lipoprotein receptor on the oocyte membrane has been characterized and found to be homologous to mammalian VLDL receptors, belonging to the LDL receptor family [21–26]. In birds, it has been reported that a single oocyte lipoprotein receptor serves as both VLDL and VTG receptor [21, 27–29]. In contrast, in rainbow trout (Oncorhynchus mykiss), the ovarian Vtg receptor appears to be highly specific for Vtg, binding Vldl only very weakly, whereas somatic lipoprotein receptors primarily bind Vldl, Ldl, and Hdl [30]. Furthermore, two cDNAs encoding rainbow trout lipoprotein receptors were isolated. One of the cDNAs was exclusively expressed in the ovary, whereas expression of the second cDNA was detected in both the ovary and somatic tissue. Therefore, the latter lipoprotein receptor may be a Vldl receptor in the rainbow trout ovary [24]. Unlike rainbow trout, white perch (Morone americana) expresses only one lipoprotein receptor in the ovary. Regardless, compelling evidence in teleosts exists to link the uptake of lipids from Vldl into oocytes to a lipoprotein receptor [20, 26], and a lipoprotein receptor also appears to be involved with the incorporation of Vtg. In the domestic hen (Gallus gallus domesticus), it has been shown that the VTG/VLDL receptor recognizes apolipoprotein B (apoB) and VTG as ligands [27, 31]. In tilapia (Oreochromis aureus), Vtg has a similar receptor-binding region to mammalian apoB [25]. In teleost fish, apoB is a component of Vldl and Ldl [32], implying that the oocyte lipoprotein receptor may be concerned with the uptake of Ldl as well as Vldl.

Lipoprotein lipase is a key enzyme that hydrolyzes the triacylglycerols in VLDL to provide fatty acids to the underlying tissues [33]. In rainbow trout, high enzymatic activity of Lpl was found in vitellogenic ovaries [34], and lpl transcripts were detected throughout oogenesis, their abundance being especially high during the late vitellogenic stage [35]. In European sea bass (Dicentrachus labrax L.), high Lpl activity was seen at the beginning of vitellogenesis and subsequently maintained throughout oogenesis [36]. Moreover, lpl mRNA levels were high during the mid- and late-vitellogenic stage, and transcripts were localized in the follicle cells in the ovary of European sea bass [36]. These studies suggest that fatty acids originating from Vldl are taken up into the oocytes by an Lpl-mediated pathway. More recently, up-regulation of ovarian lpl mRNA expression by 11-KT, both in vivo and in vitro, has been observed in short-finned eel [37]. Furthermore, androgen receptors (Ars), which have a high affinity for 11-KT [38], are mainly expressed in the follicle cells and the epithelial cells of the ovigerous lamellae in the ovary of Japanese eel [39]. Thus, these results strongly suggest that an enhancement of lipid droplet accumulation by 11-KT in the present study may well result from up-regulation of Lpl in the ovary.

It has been documented that sex steroid hormones are involved in lipid metabolism in mammals [40]. Thus, dihydrotestosterone (DHT) augmented the expression of LPL in adipose tissue via an AR-mediated pathway [41]. In addition, DHT led to increases in mRNA abundance of not only LPL but also diacylglycerol O-acyltransferase 1 and glycerol-3-phosphate dehydrogenase 1, which are involved in triacylglycerol synthesis, in adipose tissue [42]. In contrast, estrogens generally decrease the production of LPL in adipose tissue of mammals [40]. Thus, it is possible that the suppression of Vldl effects by additional E₂ treatment may be caused by decreasing Lpl production in the fish ovary as well.

It has been reported that the Ar is expressed in the teleost ovary [14, 38, 43–48]. In the Japanese eel, two ar genes (ara and arb) have been identified [38], and the transcripts encoding the ara subtype, rather than the arb subtype, were predominantly expressed in the eel ovary [14, 38]. Moreover, as mentioned above, both ar subtype transcripts are mainly expressed in the somatic cells in the ovary of Japanese eel [39]. Thus, 11-KT may act on the ovarian somatic cells through the Ara subtype and regulate the growth of the previtellogenic oocytes in eels, although the target genes of this steroid in the ovary have not been identified.

Treatment with 11-KT not only led to increased lipid droplet accumulation but also, in the presence of Vldl, to a significant increase in the size of oocytes. In contrast, treatment with E₂ did not yield these changes, and the effects of Vldl were suppressed. Therefore, it seems reasonable to conclude that the stimulatory effects of 11-KT on oocyte growth and lipid droplet accumulation in previtellogenesis are androgen specific. In mammals, androgens have been implicated in promoting follicular growth in the early stages of folliculogenesis, from primordial to small antral follicles [49, 50]. Preantral follicles have been likened to previtellogenic follicles of fish [14], and growth-promoting effects of androgens on oocytes have, in fact, been described in teleosts [14, 51, 52].

In primate follicles and oocytes, mRNA levels of the genes encoding insulin-like growth factor 1 (IGF1) and IGF1 receptor were increased by androgens [53, 54]. In coho salmon (Oncorhynchus kisuch), Igf1 is thought to be involved in regulation of previtellogenic oocyte growth [55]. Moreover, in vitro culture of ovarian tissue from New Zealand short-finned eel demonstrated that recombinant human IGF1 increased the size of previtellogenic oocytes [14]. In addition, the granulosa cells are the main site of Igf1 production in the previtellogenic ovary of red seabream (Pagrus major) [56] and tilapia [57]. Therefore, it is plausible that androgen-induced previtellogenic oocyte growth in teleosts may be mediated by production of ovarian Igf1 and/or its receptor.

Induction of vacuoles by 11-KT was suppressed at high lipoprotein concentrations, suggesting that vacuole formation was caused by a lack of lipoproteins. It is not clear what the vacuoles represent: Are they empty spaces generated during intra-oocytic lipid processing? Although possible, the number and size of the vacuoles induced by solo treatment with 11-KT, but not with E₂, did not correspond to the number and size of lipid droplets found in oocytes incubated with both Vldl and 11-KT. A role for the neutral lipids in the lipid droplets in energy metabolism during embryogenesis is well established [3]. However, it is possible that neutral lipids are also metabolized during oocyte growth, especially under conditions of in vitro culture in a chemically defined, and probably incomplete, medium. The present findings of increased oocyte size but decreased lipid droplet quantity after incubation in basal culture medium alone support this hypothesis. Indeed, it appears that the oocytes continue to grow using neutral lipids even when not supplied with sufficient neutral lipids from lipoproteins and that they exhibit abnormal cytology because of addition of 11-KT or some unknown factor induced by 11-KT. Oocytes with abnormal cytology, as observed in the present study, have not been found during normal oogenesis of Japanese eel by electron or light microscopic observations [58]. Atresia, the degeneration of ovarian follicles during oogenesis in teleosts, is a normal physiological event. Atretic ovarian follicles are characterized histologically by the disintegration of the nucleus, the breakdown of the vitelline envelope, and the increase in number and size of follicle cells. These hypertrophied follicle cells enter the oocyte to phagocytize materials [59]. Follicular atresia can be induced when ovarian development fails, for instance, because of nutrient deficiency and unusual hormone balance. In Senegalese sole (Solea senegalensis), ovarian transcript levels of fatty acid-binding proteins, involved in fatty acid uptake, transport, and metabolism, had a significant, positive correlation with the percentage of follicles undergoing atresia, suggesting that lipid metabolism may be involved in the process of atresia [60]. Furthermore, it has been reported in mammals that follicular atresia can be induced by androgens [49, 50]. Features of follicular atresia were not seen in abnormal oocytes analyzed by histological methods in the present study, so it is unclear whether our observations reflect atresia-related events or compromised, but not apoptotic, cell structure because of insufficient lipids.

Considerable differences in the effects of Vldl and 11-KT were seen among oocytes even in the same ovarian fragments. Some oocytes did not respond to Vldl and/or 11-KT. Moreover, the incidence of cytologically abnormal oocytes differed substantially between incubations from different individuals. These observations imply the existence of factors other than 11-KT that regulate previtellogenic oocyte growth and lipid droplet accumulation.

In Japanese eel, serum 11-KT levels increase gradually during vitellogenesis in artificially maturing females [61]. In addition, the predominant localization of ar mRNAs in follicle cells and their high levels in Japanese eel ovary toward late stages of ovarian development have been shown [39]. Lipid droplet accumulation continuously occurs during vitellogenesis in the eel [62]. Furthermore, ovarian lpl mRNA levels increased as oogenesis advanced and rose rapidly during midvitellogenesis, corresponding to increases in ovarian lipid levels and Lpl activity in the short-finned eel [37]. Thus, these findings and the present results strongly suggest that a major role of 11-KT on oocyte growth is the control of lipid droplet accumulation into oocytes in eels.

In conclusion, the present study demonstrated the accumulation of lipid droplets in previtellogenic oocytes of Japanese eel in vitro. Cotreatment with Vldl and 11-KT resulted in significant lipid droplet accumulation, indicating that the origin of lipids in the lipid droplets is mainly Vldl and that 11-KT, but not E₂, plays an important role in their transfer and/or accumulation into oocytes. Treatment with 11-KT led to an increase in the size of oocytes but also yielded abnormal oocytes under conditions of insufficient Vldl. This result suggests that Vldl and 11-KT are essential factors for previtellogenic oocyte growth and lipid droplet accumulation in the eel.

Acknowledgments

We are grateful to Dr. N. Hiramatsu, Hokkaido University, for his useful advice and helpful discussion. We also thank the eel research group in our laboratory for maintenance of fish.

Author notes

1

Supported by Grants-in-Aid from the 21st Century Center of Excellence Program (to T.T., S.A., and K.Y.); Grants-in-Aid for Scientific Research (18580172 to T.T.) from the Ministry of Education, Culture, Sports, Science, and Technology, Japan; and Grants-in-Aid from the Ministry of Agriculture, Forestry, and Fisheries, Japan (to S.I. and S.A.).

3

Current address: Marine Biological Technology Center, Aquaculture Business Promotion Office, Nippon Suisan Kaisha, Ltd., 508-8, Ariakeura, Tsurumi, Saiki, Ohita 876-1204, Japan.

4

Current address: South Ehime Fisheries Center, Research Group for Reproductive Physiology, Ehime University, Ainan, Ehime 798-4293, Japan.