Cell Tissue Res (1986) 245:629 633

and T'msue Resea v_h �9 Springer-Verlag 1986

Immunohistological localization of 17p-estradiol and testosterone in the ovary of the rainbow trout (Salmo gairdneri Richardson) during the preovulatory period Riidiger Schulz Ruhr-Universitfit Bochum, Fakultfit Biologic, AG Vergleichende Endokrinologie, Bochum, Bundesrepublik Deutschland

Summary. Antisera (AS) raised in rabbits against 17fl-estra- diol (E) and testosterone (T) were tested for their suitability to localize E and T on deparaffinized, rehydrated sections of preovulatory trout ovaries, using the unlabeled antibody technique.

Conventional control experiments demonstrated the specificity of the staining reactions. Furthermore, no stain- ing was observed after the removal of T-specific antibodies by affinity chromatography, or following gonadectomy when non-gonadal tissue sections of male trout were incu- bated with T-AS. Antiserum, raised against ll-oxotestos- terone and devoid of antibodies cross-reacting with T, did not stain ovarian sections.

The loci at which E and T are detected in the somatic compartment are consistent with the two-cell concept of estrogen synthesis, where aromatizable androgens are pro- duced in the thecal/interstitial layer and serve as substrates for estrogen synthesis in granulosa cells.

Both steroids were detected in yolk vesicles from the stage of endogenous vitellogenesis. T-AS showed affinity for nuclei of vitellogenic oocytes. Nucleoli were not stained.

Key words: Preovulatory follicles - Androgenic and estro- genic steroids - Immunohistological localization - Salmo gairdneri Richardson

17fl-Estradiol (E) and estrone are the main estrogens pro- duced in the ovary of salmonid fish (for review, see Fostier et al. 1983). In the monoestrous annual cycle of the rainbow trout, blood levels of these steroids are highest during rapid oocyte growth, i.e. exogenous vitellogenesis (van Bohemen and Lambert 1981). E synthesis in the amago salmon, On- corhynchus rhodurus, takes place in separate ovarian corn-

Send offprint requests to: Dr. R. Schulz, Ruhr-Universit/it Bochum, Fakultfit Biologie, AG Vergleichende Endokrinologie, ND 5/27, Universit/itsstr. 150, D-4630 Bochum t, Federal Republic of Ger- many

Acknowledgements. The study was supported by DFG grants BI 62/15-1 and 2. Many thanks are due to Priv. Doz. Dr. F. Falken- berg for the gift of goat anti-rabbit serum, and to Dipl. Biol. K.D. Schindel for the preparation of testicular and liver sections. The author is indebted to R. Oberstebrink-Scholl for technical help, and to I. Paas for photographical assistance

partments during this period (Kagawa et al. 1982; Young et al. 1982): gonadotropic hormone stimulates testosterone (T) production in special theca cells, and T is tranformed to E by granulosa cells; the same seems to be true for the rainbow trout (Nagahama 1983). A "two-cell concept" of estrogen synthesis is also well-known in higher verte- brates (e.g. Dorrington and Armstrong 1979; Huang and Nalbandov 1979).

The aim of the present study was (i) to determine wheth- er T- and E-antisera (AS) raised in a conventional way were suitable for the specific localization of the hormones using the unlabeled antibody (PAP) technique (Sternberger and Joseph 1979) in fixed, paraffin-embedded ovarian tissue samples, (ii) to determine whether the two-cell concept could be reconstructed morphologically, and (iii) to moni- tor the distribution of the hormones in the ovary during the preovulatory period in the rainbow trout.

Materials and methods

Experimental animals

Ovaries were collected from trout during the first reproduc- tive cycle and fixed and embedded as described previously (Schulz 1984).

Antisera

The production of AS against E and 11-oxotestosterone (OT) has been described previously (Schulz 1984). T-AS was produced using similar techniques.

Rabbit control serum (RCS), steroid-AS, and goat anti- rabbit serum (GAR) were stripped of endogenous steroids by charcoal treatment and then incubated with a rainbow trout tissue powder prepared as described by Ortez et al. (1980), except that the preparation was further extracted three times with 5 vol of diethyl ether. After evaporation of residual ether, the preparation was ground to a fine powder. 4-5 mg were suspended in I ml of the diluted sera and incubated for 1 h at room temperature with constant shaking; the sera were cleared by centrifugation.

The (3H)steroid-binding activities of the AS were tested in radioimmunoassays (RIAs; see Schulz 1984, for experi- mental details) before and after the procedures described above.

630

Immunoh&tology

Sections (7 [am thick) were dried onto gelatine-coated glass slides. After removal of paraffin and rehydration with buffer (0.05 M Tris/HCl pH 7.6, 0.15 M NaC1, 0.1% RCS, 3 • 3 rain), they were incubated with serum dilutions (RCS 1/10-1/5000, E-AS 1/500-1/5000, T- and OT-AS 1/1000-1/5000) for 18 h at 4 ~ C, followed by incubation for 30 min at 37 ~ C with GAR diluted 1/25 and peroxidase anti-peroxidase complex (PAP) diluted 1/100. Finally, the sections were incubated in the dark for 7 min in a chilled DAB-solution (50mg 3,3-diaminobenzidine tetrahydro- chloride and 30 ~tl H202 per 100 ml buffer), washed under running tap water, cleared, and mounted.

Sections from each tissue sample were stained conven- tionally to monitor the stages of ovarian development, i.e. previtellogenesis, endogenous, and exogenous vitellogenesis (VG).

Specificity controls

Steroid-AS were replaced by RCS; GAR or PAP were re- placed alternatively by buffer; or DAB o r H20 2 were omit- ted.

20mg testosterone-3-oxime prepared according to Fieser and Fieser (1967) were coupled to AH-Sepharose 4B (500 mg dry weight) by the mixed anhydride procedure (Erlanger 1973), and used in a chromatography column (2 ml bed volume) equilibrated with Tris/HC1 buffer. 50 lal of E-, T-, and OT-AS free of endogenous steroids diluted 1/10 were applied to the column and incubated for 2 h at room temperature, before the column was eluted. The eluates were tested for their ability to stain trout tissue sections, and to bind (3H)steroids in RIAs.

Mature males were castrated according to Brown and Richards (1979). After 8 days, tissue samples of several organs of intact males and castrates were fixed and pre- pared for immunohistology. Plasma T levels of the castrates dropped from 30.2-57.8 ng/ml to levels below 2 ng/ml after 8 days.

Results

Specificity controls

Conventional control experiments gave no staining except for a few small spots in the zona radiata; these spots were also seen after treatment with DAB and H202 only.

Affinity chromatography of the T-AS abolished all staining and was accompanied by a loss of 94.5% of the (3H)T binding activity. RIA and immunohistological char- acteristics of the E-AS were not affected by passage through the column. Charcoal and tissue powder treatment had no effect on (3H)steroid-binding activities.

The effect of castration on the staining of non-gonadal tissues after incubation with T-AS can be seen in liver tissue. The granular staining of the nucleoplasm and the perinucle- ar staining on sections from intact males (Fig. 1 a) is absent in sections from castrates (Fig. 1 b).

Sections from ovaries accumulating exogenous yolk were incubated with OT-AS. The staining pattern was com- parable with that seen following incubation with T-AS. After passage of the OT-AS through the affinity column,

Fig. 1 a-d. Liver sections from control (a) and gonadectomized (b) males after incubation with T-AS (dilution 1/600). Peri- and endo- nuclear staining disappeared 8 days after castration, x 160. Testi- cular sections incubated with OT-AS (1/1000) before (e) and after (d) T-affinity chromatography. The typical pattern with labeled (arrowheads) and unlabeled (points) spermatocytes is not affected by removal of antibodies cross-reacting with T. x 40

the staining disappeared; this was accompanied by a loss of 66.2% of the (3H)OT binding activity. Testicular sec- tions, however, showed comparable staining locations and intensities (Fig. 1 c, d) after treatment with the same OT-AS preparations.

17fl-estradiol

The zona granulosa was consistently stained following incu- bation with E-AS. During pre-VG and the beginning of endogenous VG, the precipitate was concentrated on that side of the granulosa neighboring the oocyte (Fig. 2a, c). Later on, it was distributed throughout the zona granulosa (Fig. 2d f). In all cases, the staining was located at the periphery of granulosa cells and/or in the intercellular spaces between them, the latter being most obvious during exogenous VG (Fig. 2b, d-f). Each ovary contained follicles in which different parts of the zona granulosa showed dif- ferent labeling intensities (Fig. 2b, d, e). The granulosa of immature follicles in vitellogenic ovaries showed the same type of staining as immature ovaries.

The staining of thecal and interstitial tissues was vari- able between individuals and within the same ovary (Fig. 2a). Different intensities could not be related to other observations. During exogenous VG, the staining (if pres- ent) was mostly restricted to a band of interstitial tissue separated from adjacent granulosa layers by unstained or less intensely stained thecal areas (Fig. 2d, e).

The ooplasm always showed a faint staining. Germ-cell nuclei were not labeled. Vitellogenic oocytes carried small, labeled vesicles (Fig. 2d-f) with constant diameters; these vesicles showed an affinity for light-green in conventionally stained adjacent sections.

631

Fig. 2. Sections from ovaries after E-AS incubations during previtellogenesis (a, x 160), endogenous vitellogenesis (b and c, x 160), and exogenous vitellogenesis (d, x 40; e and f, x 160). The DAB-precipitate accumulated predominantly in granulosa cells of immature follicles close to the ooycyte (a, c), and in advanced follicles between the cells (see tangential sections b, d, f). Note the different labeling intensities in individual granulosa (b, d, e) and thecal/interstitial areas (a). Abbreviations, see Fig. 3

Testosterone

T-AS incubations always resulted in a prominent staining of thecal and/or interstitial areas (Fig. 3a, b, e), whereas the granulosa was very weakly labeled in immature ovaries (Fig. 3a). The cytoplasm of many granulosa cells in vitello- genic follicles showed a label contrasting with their nuclei (Fig. 3b, e). Some granulosa cells also showed a distinct nuclear label, mainly during exogenous VG (Fig. 3e). Inter- cellular spaces were always unlabeled. Rarely, cells with a more intense cytoplasmic staining than the immediate surrounding were found in the thecal and/or interstitial layer (Fig. 3c). Beginning with endogenous VG, the nuclei of advanced germ cells developed an increasing affinity for T-AS, and were clearly labeled during exogenous VG (Fig. 3d). Small spots at the periphery of the nuclei were not labeled. A faint staining of the ooplasm, appearing more intensely in smaller oocytes (compare left and right germ cell, Fig. 3 b), the presence of stained, peripheral vesi- cles during VG (Fig. 3d), and a staining restricted to a band of interstitial tissue in exogenous VG (Fig. 3e) were observed as in sections treated with E-AS. In contrast to

E-AS incubations, the latter was constantly found after T- AS treatments.

Discussion

The control experiments demonstrate that specific staining depends upon the presence of steroid antibodies and of the respective antigens in the tissue section. Under RIA- conditions, the OT-AS shows only 3.4% cross-reaction with T (Schulz 1984), but two thirds of the (3H)OT binding capacity are lost after passage through a column carrying immobilized T. Possible explanations for this discrepancy are the excess of cross-reacting steroid in the solid phase and the absence of homologous steroid. These conditions may lead to a more extensive binding to the heterologous steroid, and to the staining of ovarian sections. Thus, it cannot be excluded that the staining seen after incubations with untreated AS directed against C-19 steroids (i.e. un- treated by chromatographical steps), does not discriminate strictly between different androgens. Estrogens and andro- gens, however, are clearly differentiated. Thompson et al. 1985) have reported that in rat uteri preineubated with

632

Fig. 3a-e. Sections from ovaries after T-AS incubations during previotellogenesis (a, x 160), endogenous vitellogenesis (b, x 160; c, x 400), and exogenous vitellogenesis (d, x 40; e, x 160). The DAB-precipitate accumulated predominantly in extracellular thecal and/or interstitial areas (a, b, and e). Rarely, the cytoplasm of interstitial or thecal cells was stained more intensely than their surroundings (e). Nuclei of vitellogenic oocytes showed a fine-granular staining, from which small spots in the periphery (d, arrowheads) were excluded. Granulosa cells of follicles in endogenous or exogenous vitellogenesis often showed cytoplasmic staining (b and e, points). During exogenous vitellogenesis, some nuclei were also labeled in granulosa cells (e, arrowheads); g granulosa, i interstitium, n nucleus, th theca, yv yolk vesicles

(14C)E, about 13% of the radioactivity persists after com- parable histological procedures. They argue that E is detect- able because it is linked to fixative-labile proteins in vivo. Ortez et al. (1980) have come to the same conclusion regard- ing the detectability of cyclic nucleotides in the retinas of the goldfish, Carassius auratus. Taken together, the above considerations suggest that, in the trout, estrogens and an- drogens are present and can be specifically localized in fixed, paraffin-embedded samples using the PAP-technique.

The sites at which the steroids are detected can be pre- dicted from the two-cell concept of estrogen synthesis (Ka- gawa et al. 1982; Young et al. 1982). The immunohistologi- cal information supports in-vitro data (Nagahama 1983) that estrogen production also follows this scheme in the rainbow trout. The above cited authors separate theca and granulosa cells for in-vitro experiments, but there is no indi- cation that theca cells are separated from interstitial tissue. In the amago salmon, only special theca cells show the ultrastructural characteristics of steroidogenic cells (Naga- hama 1983). In the rainbow trout, Upadhyay et al. (1978) and van den Hurk and Peute (1979) describe cells having this typical organelle equipment in thecal and interstitial areas, which also show 3fl-hydroxysteroid-oxidoreductase activity. Thus, the situation in the trout might be slightly different from the amago salmon; this is supported by the present data.

The restricted distribution of estrogens in all stages of development in the trout ovary indicates that aromatizing enzymes are expressed early (van den Hurk et al. 1982) and typically reside in granulosa cells. Nagahama et al. (1978) and Kagawa etal. (1981) describe intercellular spaces filled with amorphous material between granulosa cells of follicles during exogenous VG in three salmonid species. These spaces correspond to the predominant loca- tion of E and the heterogeneity of the staining here might

reflect the presence of different amounts of E. The fact that E is barely detectable in granulosa cells indicates that it is washed out, possibly because it is free in solution at the moment of fixation, or that it rapidly leaves the cells and transiently accumulates between them. However, the labeling intensity is strongest during exogenous VG; this correlates well with the high estrogen blood-levels in trout at this stage of development (van Bohemen and Lambert 1981 ; Schulz 1984).

Whereas the staining for estrogens in thecal and intersti- tial areas may be caused by an intrinsic hormone produc- tion, the cytological characteristics indicative for steroid synthesis starting from cholesterol are absent in the granu- losa (for review, see Nagahama 1983). The cytoplasmic staining, following incubation with T-AS during VG, thus may be related to the concomitant maximum activity of the aromatizing enzymes (van Bohemen and Lambert 1981), and may represent T at sites of its metabolism. One of the limiting factors for estrogen synthesis presumably is the quantity of the respective enzymes. If their expression is regulated by androgens, this may provide an explanation for the nuclear staining in granulosa cells.

An androgen-specific staining more pronounced on, or restricted to, thecal and interstitial cells is observed occa- sionally, but a uniform staining of large, extracellular areas is more often the typical pattern for androgens and, al- though more variable, for estrogens too. Since the steroids that are free in solution at the moment of fixation are prob- ably the major fraction lost during tissue processing, ster- oids detectable at extracellular sites may have been asso- ciated with fixative-labile proteins. The intercellular fluid bathing steroidogenic cells in mammalian gonads contains steroid-binding proteins derived from the blood (Waites and Gladwell 1982). Sex-steroid-binding plasma proteins with similar characteristics have also been described in re-

633

male trout (Fost ier and Breton i975). In this context, the differential staining of thecal and interstit ial areas (most prominent during exogenous VG) is of interest. Avai lable data indicate that the theca is the pr imary source of andro- gens, a l though in the trout, the steroidogenetic capaci ty of the interst i t ium cannot be est imated correctly. In any case, a poor staining of thecal areas despite high steroid concentrat ions in the immediate ne ighborhood would be expected if the theca is the predominant drainage route for steroids leaving the ovary via the intercellular fluid. The intersti t ium and the zona granulosa (in the case o f E) may then represent an in t ragonadal steroid reservoir that can be regulated, e.g., by changing the velocity of flow and/or the volume of the interstit ial fluid.

Small vesicles in the cytoplasm of follicles during VG show affinity for both AS. The time of their appearance, their size, location, and affinity for light-green indicate that they are yolk vesicles. Since they give rise to the cortical alveoli, whose components are released during insemina- tion, these steroids are not necessarily involved with endo- crine events during oogenesis.

Dur ing VG, the germ-cell nuclei develop an increasing affinity for T-AS. The fine-granular staining covers the complete nuclear area except for small vesicles (most likely nucleoli) at the periphery. S tumpf and Sat (1982) point out that receptors cannot be positively identified using ster- o id-AS only. However, nothing is known about the quanti- ties and retention times of androgen-receptor complexes in these nuclei. Moreover, the staining does not necessarily represent androgen- recep to r -DNA complexes. Therefore, it it possible that the staining represents androgen-receptor complexes. The presence of steroid receptors in germ cells has been demonst ra ted in male rats (Frankel and Chapman 1984) and possibly also in two elasmobranch species (Cal- lard and Mak 1985). In addit ion, the results of Gronemeyer et al. (1981) indicate that ecdysteroids bound to high-affini- ty, low-capaci ty receptor proteins are detectable at genomic sites by immunohistological techniques. Since the detectabi- lity of nuclear androgens is related to the onset of VG in a given germ cell, androgens probably have other func- tions than serving as substrates for estrogen synthesis.

In general, the distr ibution of androgens in thecal and interstitial areas does not show major variations, al though the distr ibution of estrogens in granulosa cells or of andro- gens in nuclei are characteristic of the stage a given follicle has reached, irrespective of the other follicles in the same ovary. This indicates that every follicular envelope and the enclosed germ cell should be regarded as an individual unit with special microenvironments adapted to its individual needs.

References

Brown LA, Richards RH (1979) Surgical gonadectomy of fish: a technique for veterinary surgeons. Vet Rec 104:215

Bohemen ChG van, Lambert JGD (1981) Estrogen synthesis in relation to estrone, estradiol, and vitellogenin plasma levels dur- ing the reproductive cycle of the female rainbow trout, Salmo gairdneri. Gen Comp Endocrinol 45:! 05-114

Callard GV, Mak P (1985) Exclusive nuclear location of estrogen receptors in Squalus testis. Proc Natl Acad Sci USA 82:1336-1340

Dorrington JH, Armstrong DT (1979) Effects of FSH on gonadal functions. Recent Progr Horm Res 35:301-342

Erlanger BF (1973) Principles and methods for the preparation

of drug protein conjugates for immunological studies. Pharma- col Rev 25:271 280

Fieser LF, Fieser M (1967) Reagents for organic chemistry. John Wiley and Sons, New York London Sydney, pp 410~11

Fostier A, Breton B (1975) Binding of steroids by plasma of a teleost: the rainbow trout, Salmo gairdneri. Steroid Biochem 6: 345-351

Fostier A, Jalabert B, Billard B, Breton B, Zohar Y (1983) The gonadal steroids. In: Hoar WS, Randall DJ, Donaldson EM (eds) Fish physiology. Vol IX, Part A. Academic Press, New York, pp 277-372

Frankel AI, Chapman JC (1984) Nuclear androgen binding sites in the male rat - [If. Late spermatids and spermatozoa in the testis, with an introdution to epididymal spermatozoa. J Steroid Biochem 20:I301-1311

Gronemeyer H, Hameister H, Pongs O (1981) Photoinduced bond- ing of endogenous ecdysterone to salivary gland chromosomes of Chironimus tentans. Chromosoma 82 : 543-559

Huang ES, Nalbandov AV (1979) Synthesis of sex steroids by cellu- lar components of chicken follicles. Biol Reprod 20:454-461

Hurk R van den, Peute J (1979) Cyclic changes in the ovary of the rainbow trout, Salmo gairdneri, with special reference to sites of steroidogenesis. Cell Tissue Res 199:289 306

Hurk R van den, Lambert JGD, Peute J (1982) Steroidogenesis in the gonads of rainbow trout fry (Salmo gairdnerO before and after the onset of gonadal sex differentiation. Reprod Nutr Dev 22:413~25

Kagawa H, Young G, Adachi S, Nagahama Y (1982) Estradiol-17fl production in arnago salmon (Oncorhynchus rhodurus) ovarian follicles : role of the thecal and granulosa cells. Gen Comp En- docrinol 47 ; 440M48

Kagawa H, Takano K, Nagahama Y (1981) Correlation of plasma estradiol-17fl and progesterone levels with ultrastructure and histochemistry of ovarian follicles in the white-spotted char, Salvelinus leucomaenis. Cell Tissue Res 218:315-329

Nagahama Y (1983) The functional morphology of teleost gonads. In : Hoar WS, Randall D J, Donaldson EM (eds) Fish physiolo- gy. Vol. 1X Part A. Academic Press, New York, pp 223-276

Nagahama Y, Clarke WC, Hoar WS (1978) Ultrastructure of puta- tive steroid-producing cells in the gonads of coho (Oncorhyn- chus kisutch) and pink salmon (Oncorhynchus gorbuscha). Can J Zool 56:2508 2519

Ortez RA, Sikes RW, Sperling HG (1980) hnmuno-histochemical localization of cyclic GMP in goldfish retina. J Histochem Cy- tochem 28:263 270

Schulz R (1984) Serum levels of 11-oxotestosterone in male and 17fl-estradiol in female rainbow trout (Salmo gairdneri) during the first reproductive cycle. Gen Comp Endocrinol 56:111-120

Sternberger EA, Joseph SA (1979) The unlabeled antibody method. J Histochem Cytochem 27 : 1424-1429

Stumpf WE, Sat M (1982) Histochemical approaches for the local- ization of steroid hormone receptors. Acta Histochem Cyto- chem ! 5:56(~571

Thompson SA, Radde L, Farley DM, Rasazza JP, van Orden DE (1985) Immunocytochemica[ localization of tissue-bound estradiol in rat paracervical ganglion. Histochem J ! 7:493-506

Upadhyay SN, Breton B, Billard R (1978) Ultrastructural studies on experimentally induced vitellogenesis in juvenile rainbow trout Salmo gairdneri. Ann Biol Anim Biochim Biophys 18:1019-1026

Waites GMH, Gladwell RT (1982) Physiological significance of fluid secretion in the testis and blood-testis barrier. Physiol Rev 62:624-671

Young G, Kagawa H, Naghama Y (1982) Secretion of aromatiz- able A4-androgens by thecal layers during estradiol-17fl produc- tion by ovarian follicles of amago salmon (Oncorhynchus rho- durus) in vitro. Biomed Res 3:659 667

Accepted March 31, 1986