Ontogeny of oestrogen-sensitive mesenchymal cells in the bursa of Fabricius of the chick embryo. An immunohistochemical study on progesterone receptor

The bursa of Fabricius is a maturation site for B lymphocytes in birds (Cooper, Lawton & Kincade, 1972). The bursal primordium appears on the 4–5th day of incubation as a thickening of the cloacal epithelium (Metcalf & Moore, 1971). Formation of bursal lumen starts at 6–7 days. The most-active growth and remodelling of the organ takes place between 8 and 15 days (Valinsky, Reich & Le Douarin, 1981). At 10–11 days, the bursa starts to form plicae inside the lumen. At this age, the bursa consists of three layers: the innermost is stratified columnal or cuboidal epithelium; a thin outer layer of serosa and muscle; between these two layers is a layer of mesenchymal cells, tunica propria. The bursal primordium is surrounded by a loose cloacal mesenchyme. At 12–13 days, groups of epithelial cells start to proliferate forming epithelial buds that project into the underlying tunica propria. Mesenchymal cells around the buds form a dense mesenchymal sheet under the developing bud. The epithelial bud will develop into the medulla of lymphoid follicles and the mesenchymal sheet into the cortex. At 9 days of incubation, haemopoietic precursor cells begin to colonize the bursa. Some of the haemopoietic precursor cells migrate into the epithelium and differentiate into B lymphocytes. The haemopoietic cells that remain in the mesenchyme surrounding the follicles differentiate into granulocytes. Interaction between the epithelial and stromal cells is vital for bursal development. If the interaction is prevented the haemopoietic cells cannot colonize the bursa. (Metcalf & Moore, 1971; Le Douarin et al. 1982).

It was first shown by Kirkpatrick & Andrews (1944) that steroids affect the development of the bursa of Fabricius. Further studies showed that exogenous sex steroids affect bursal growth when given in ovo (Erickson & Pincus, 1966; Norton & Wira, 1977). These works suggest that the sex steroids may affect bursal development in vivo during embryonic life. The results, however, only suggest the steroid sensitivity, they tell nothing of the in vivo effects of the endogenous steroid or the mechanism through which the steroids mediate their effects.

In the present work, we studied the presence of PR during bursal embryogenesis as a marker for progesterone sensitivity and its induction as a marker for oestradiol action, since PR is known to be an oestradiol-induced protein. A functional oestrogen-receptor system has been shown to appear in the embryonic chicken liver between days 10 and 12 (Elbrecht, Williams, Blue & Lazier, 1981). In the cloacal area (Gasc & Stumpf, 1981a), the gonads (Gasc, 1980) and the bursa of Fabricius (Gasc & Stumpf, 1981b), embryonic cells concentrate oestradiol in their nuclei, which is evidence for the presence of oestradiol receptors. Whether or not such receptors are functional, however, remained uncertain. There is detectable amounts of oestradiol in the plasma of chick embryos at least from 7–8 days of incubation. Oestradiol concentration is somewhat increased during embryonic life, but more pronounced increase takes place during sexual maturation (Woods & Brazzill, 1981; Tanabe, Nakamura, Tanase & Doi, 1981). It is not known whether the concentration of oestradiol during embryonic life is high enough to affect bursal development. We report here that mesenchymal cells in the bursa of Fabricius of the chicken embryo possess a functional effector mechanism for oestradiol very early in the embryonic development. We also suggest, however, that the actual effects of endogenous oestrogen and progesterone on BF are minimal during embryonic life.

Chick embryos (White Leghorn), 6–16 days of incubation, received oestradiol benzoate (100 μg per embryo, in sesame oil) or the vehicle only in the allantoic sac. The embryos were killed 3 days after the oestradiol administration. The dissected cloacal areas were fixed with Bouin’s fixative (1 part acetic acid, 5 parts formalin (38%) and 15 parts saturated picric acid (1·2 %)) or with Baker’s fixative (4 % paraformaldehyde, 1 % CaCl₂ in water, pH6-7) for 1·5–2 h at 4 °C.

The immunoperoxidase technique used in this study was the same as that described by Gasc et al. (1984). Briefly, the fixed tissue pieces were dehydrated and embedded in paraffin. An immunoperoxidase technique using a biotinylated secondary antibody and avidin–biotin–peroxidase complex (Vectastain reagents, Vector Laboratories, Burlingame, CA) was employed. We used three different PR antibodies: IgG-RB, which is an IgG fraction of the rabbit antiserum described in detail by Tuohimaa et al. (1984); IgG-G3, which is an IgG fraction of the goat antiserum described in detail by Renoir, Radanyi, Yang & Baulieu (1982); IgG-RB₂ (unpublished), which is raised in rabbit and resembles the IgG-RB in its specificity. In a separate paper, we have shown that IgG-RB recognizes the PR also in the BF and by using a combined technique of autoradiography and immunohistochemistry we have shown that the antibody detects the progestin-concentrating cells in the BF (Ylikomi, Isola, Gasc & Tuohimaa, 1987; Gasc. Ennis. Baulieu & Stumpf, 1986). Presaturation experiments were carried out by incubating the specific antibody with affinity-purified chick oviduct PR overnight at 4°C. Nonimmune and preimmune sera were used on the control sections. As a control antibody we also used anti-ovalbumin IgG, which was raised and purified in a similar manner as the IgG-RB. Ovalbumin is a secretory protein in the chick oviduct and is not expressed in the BF.

Between 9 and 11 days of incubation, the BF primor-dium consists of a pouch of epithelial structure with a wide lumen at the centre surrounded by mesenchymal cells. At 9 days of incubation, no PR-immunoreactive cell is observed in the absence of oestradiol treatment. After oestradiol induction, there are a few weakly positive cells in the surrounding cloacal mesenchyme. The epithelium and the tunica propria are PR-negative. Since it is difficult to distinguish the margin of the bursal serosa and the cloacal mesenchyme clearly it cannot be decided definitely whether some of the serosal cells express the PR (Fig. 1). In the same cells, a weak PR immunoreactivity is detectable at 11 days in untreated embryos (Fig. 2A). After oestradiol treatment, the serosal and muscular layer and surrounding cloacal mesenchyme is very rich in PR-positive cells. A very few weakly stained cells are seen also in the tunica propria. The epithelium is negative (Fig. 2B,C). At 16 days, the plicae have grown in height and follicle formation has started. PR-immunoreactive cells are not present without oestradiol treatment either in the bursal epithelium or in the stroma (Fig. 3A). After oestradiol treatment, the PR-expressing cells are seen both in the tunica propria and in the serosal and muscular layer. In the tunica propria, the positive cells are seen in the centre of the plicae, but not in mesenchymal sheet around the developing epithelia) buds (Fig. 3B). Similar PR induction was also obtained in both male and female chicks. The bursal epithelium never displayed any immunoreactivity. The staining was abolished after presaturation of the antibody with purified PR and was not obtained with preimmune sera or with the control seras (Fig. 2D). All three PR antibodies gave similar results.

On day 9–10 of incubation a cluster of cells showing peroxidase staining with IgG-RB is seen on each side of the bursa, both in the oestradiol-treated and untreated chicken. Then the cells start to colonize the plicae of the bursa and become very numerous (Figs 2, 3, arrowheads). The positively stained cells are large and round in shape, thus resembling the haemopoietic cells, which are known to enter the BF during the same time period (Le Douarin et al. 1982). Fig. 3 shows that the cells are seen around the follicular medullas, but not entering them. The cells thus resemble granulocytes, which are known to differentiate in the mesenchyme around the follicular medullas (Le Douarin et al. 1982). Staining is also seen with the other antibodies, but with various intensities and also with some batches of biotinylated goat anti-rabbit antibodies (Fig. 3C). It was also fixation dependent, since it was not seen in the Baker-fixed sections. The staining was not modified when the IgG-RB was presaturated with purified PR under conditions that completely abolished the immunostaining of PR (Fig. 2D). The staining was not due to endogenous avidin binding seen in some lymphoid tissues (Banerjee & Pattit, 1984), since staining was not seen when both the primary and secondary antibodies were omitted.

In a previous study, we have shown the presence of oestradiol-inducible PR in the interfollicular cells of the BF after hatching (Ylikomi et al. 1985. 1987). In the present study, we extend our observations to the chick embryo at early stages of development. The inducibility of PR in mesenchymal cells surrounding the bursa at 9 and 11 days of incubation demonstrates the presence of functionally active oestradiol receptors. At the same stages of development, apolipoprotein II and vitellogenin II are also induced in the liver by oestradiol (Elbrecht, Lazier, Protter & Williams, 1984). In the course of the present work, we also observed the inducibility of PR in mesenchymal cells of the cloacal area at an even earlier stage (Gasc et al. unpublished data). This result is in agreement with the observation by Gasc & Stumpf of cells in the cloacal area concentrating radioactively labelled oestradiol as early as at 6 days of incubation (Gasc & Stumpf, 1981a).

Though oestrogen receptor is actually present in the cloacal area and then in the bursal mesenchyme, the expression of PR in the bursa, either constitutively or induced by endogenous hormones, remained uncertain. In the absence of oestradiol treatment, the only PR-immunoreactive cells are located at the periphery of the bursa. At 11 days of incubation, these cells are scarce and weakly stained, and it is unclear whether they belong to the bursa or the cloacal area (urogenital sinus). At 16 days, these cells no longer form an anatomically defined layer discernible from the surrounding cloacal area, which contains numerous PR-immunoreactive cells even in the absence of oestradiol treatment.

If PR is not spontaneously present inside the bursa, although it is inducible, through which mechanism(s) can exogenous progesterone affect bursal development (Norton & Wira, 1977)? An indirect mechanism via the surrounding mesenchymal cells of the cloacal area is possible. Alternatively, an effect of progesterone through the androgen receptor is possible since the receptor is present early in development in many cells of the bursa and is able to bind progesterone, although with weak affinity (Gasc & Stumpf, 1981a,b;Le Douarin, Michel & Baulieu, 1980).

Oestradiol has been shown to suppress bursal growth (Norton & Wira, 1977). In the present work, we demonstrate that stromal cells in the bursa are indeed able to react to oestrogen in a very specific manner, by expressing PR. These stromal cells most probably mediate the effects of exogenous oestrogen on bursal development, either directly or through some stromal–epithelial interaction. However, the fact that endogenous oestrogens are not able to induce the PR concentration to detectable levels implies that, although BF is oestrogen-sensitive, the actual effects of endogenous oestrogens during bursal embryogenesis are minimal or negligible unless, as already suggested, oestradiol acts through the surrounding cloacal mesenchyme. In the BF, progesterone receptor becomes spontaneously detectable well after hatching (Ylikomi et al. 1987). This dissociation between inducibility and natural expression of genes has also been detected in other systems. For instance, vitellogenin (a precursor for different egg yolk proteins) is oestradiol inducible in Xenopus laevis larval liver during metamorphosis. The natural expression of the protein takes place only in sexually mature females, probably induced by endogenous oestrogens (Tata, 1976; May & Knowland, 1981).

We observed an intense peroxidase staining in the haemopoietic stem cells, possibly granulocytes, colonizing the BF at the age of 9–10 days with different antibodies. After hatching, the staining is not detectable in the BF (Ylikomi et al. 1985, 1987). The staining is not due to PR immunoreactivity since it is not presaturable with purified PR and is detectable with an intensity that varied with each antibody. We think the staining may come from the binding of a subclass of immunoglobulins more represented in some antibodies than in others.

In the present work, we identified early in development stromal cells in the bursa sensitive to oestrogen, characterized by the expression of PR under oestradiol induction. The oestrogen-responsive cells, however, form only a subset of bursa mesenchymal cells and endogenous oestradiol levels are not able to induce PR to a detectable level. This implies that, despite oestradiol sensitivity, the actual direct effects of endogenous oestradiol and progesterone on the BF during embryonic development are minimal.