The ovarian interstitial tissue consists of cellular clusters or nests, which are scattered in the cortical stroma. In mammals, it has been described for many species (Brambell, 1966), although the amount and distribution of interstitial cells vary greatly in the different animals as well as during the ovarian cycle of the same species (Guraya, 1991). Most of the investigations indicate that its constitutive cells present the features usually associated with the steroid secretion: abundant lipid droplets, well-developed agranular endoplasmic reticulum, and numerous mitochondria (Al-Mehdi, 1979). Consequently, this compartment has been implicated in the production of steroid hormones, principally androgens (Gore-Langton and Armstrong, 1984).
Some authors have observed morphological characteristics that suggest steroidogenic activity during pregnancy in the interstitial tissue (Lawrence et al., 1977; Garcia et al., 1984). Furthermore, there is evidence that the gonadotrophins, principally the luteinizing hormone (LH), may modulate the development and steroidogenic activity of this ovarian compartment (Carithers and Green, 1972; Capps et al., 1978; Clarke and Brook, 2001). Despite the potential relevance of this structure in the ovarian function, its importance has been frequently overlooked and its biological significance remains to be perfectly established.
The viscacha (Lagostomus maximus maximus), the largest member of the Chinchillidae family, inhabits the southern hemisphere from Paraguay through central Argentina (Redford and Eisenberg, 1992). Our laboratory has extensively studied the reproduction of male viscacha (Fuentes et al., 1991, 1993; Muñoz et al., 2001; Aguilera-Merlo et al., 2005), but the investigations about females are recent (Gil et al., 2005). In its natural habitat, this rodent is a seasonal breeder. A large number of pregnant animals are present in winter; although, pregnant females have been found in other seasons (Jackson, 1989). The gestation period lasts approximately 154 days (Weir, 1971a), after which two well-developed offspring are usually born. The general aspects of the ovarian morphology have been described by Weir (1971b); however, there are not detailed studies about the ovary of Lagostomus. Moreover, there are no investigations describing the endocrine variations related to the ovary physiology in female viscachas. Because of these antecedents, the aim of the present study was to examine the morphological and endocrine aspects of the ovarian interstitial tissue of adult female viscachas to establish the probable function and the biological significance of this compartment in this rodent.
MATERIALS AND METHODS
Pregnant and nonpregnant adult female viscachas, weighing 2–4 kg, were captured in their habitat near San Luis, Argentina (33° 20′ south latitude, 769 m altitude) during 2002 and 2004. In San Luis, in summer, there is up to 14 hr of light daily with an average temperature of 25°C. In winter, the light phase decreases to 10 hr and average temperature to 10°C. The rainfall is 206 mm in summer and 18 mm in winter.
In this investigation, 12 pregnant and 6 nonpregnant animals were used for the morphological study and 4 pregnant and 4 nonpregnant viscachas for the hormonal assays. The reproductive condition of the viscachas was carefully assessed on the following basis: (1) the examination of the uterine horns, and (2) the analysis of the ovaries by light microscopy. The age of animals was estimated by (1) body weight (Llanos and Crespo, 1954, Branch et al., 1993) and (2) lens weight (Jackson, 1986). The number and size of the embryos and fetuses were used to evaluate the pregnancy degree of progress. In agreement with these criteria, the pregnant animals were classified in three lots: initial pregnancy (four to six embryos measuring 1–3 cm), mid-pregnancy (two fetuses measuring 9–14 cm), and advanced pregnancy (two fetuses measuring more than 15 cm).
The animals were anesthetized with Nembutal (pentobarbital), and blood samples were obtained by cardiac puncture. The serum was separated by centrifugation, transferred to glass vials and stored at −20°C until assaying. After that, the animals were killed by decapitation. The abdominal cavity was exposed and the ovaries were quickly removed and used for light microscopy, electron microscopy, and biochemical assays. Animals were handled according to the procedures approved in the UFAW Handbook on the Care and Management of Laboratory Animals, Vol. 1: Terrestrial Vertebrates (Poole, 1999), and the Guide for Animal Use and Handling of the National University of San Luis.
For routine microscopy, the ovaries were fixed in Bouin's fluid, embedded in paraffin wax, serially sectioned at 5 μm and stained with hematoxylin and eosin and Masson's trichrome.
The fixatives used for the histochemical studies of lipids included formaldehyde–calcium with and without postchroming and 10% neutral formalin. For the study of lipids, postchromed samples embedded in paraffin wax and serially sectioned at 5 μm or samples embedded in gelatin and sectioned at 10 μm in a cryostat were used. The sections were colored with acetylated Sudan black B and mounted in neutral glycerin jelly. The presence of cholesterol and its esters was tested by the Liebermann–Schultz reaction.
For electron microscopy, the ovaries were fixed in situ with formaldehyde–glutaraldehyde in phosphate buffer (Karnovsky, 1965) for 10 min, removed, and placed in the same fixative for an additional 6-hr period at room temperature, postfixed in cold 2% OsO4 for 12 hr, dehydrated in acetone, and embedded in Spurr's resin. One-micrometer-thick sections were obtained with a Porter Blum ultramicrotome and dyed with toluidine blue for the light microscopy. Ultrathin sections were stained with uranyl acetate and lead citrate (Milloning, 1961) and were observed under a Siemens Elmiskop I electron microscope.
A computer-assisted image analysis system was used to measure the area of lipid droplets from ovaries at different levels in the four stages of the reproductive state: nonpregnancy, initial pregnancy, mid-pregnancy, and advanced pregnancy. One-micrometer-thick sections dyed with toluidine blue were used for the study. The fixation and coloration provide a good contrast of the lipid droplets. The system consisted of a Nikon Optiphot-2 binocular stereomicroscope (magnification, × 400), interfaced with a host computer, image processing, and recording system. The images were captured by a Panasonic GP-KR222 camera and processed with Image Pro Plus 5.0 software under control of a Pentium II computer. The software allowed the following processes: image acquisition, automatic analogous adjust, thresholding, background subtraction, distance calibration, area measuring, and diskette data logging. Of the seven parameters, which could be estimated by the software, relative area of lipid droplets is the focus of attention here. The image was displayed on a color monitor, and the areas were measured with the image analysis system. Fields not entirely occupied by the tissue were not analyzed. The light microscope images of 80 fields per section (three to five serial sections per animal) were evaluated. The results were expressed as percentage of the standard area and presented as means ± SEM. The area of interstitial tissue and the area of lipid droplets within it were measured for 200 areas of each reproductive state, clearly showing the lipid droplets, for observation under ×40 objective. The relative areas of lipid droplets were calculated by a proportion quotient (PQLD) according to this formula: area of lipid droplets/area of interstitial tissue containing the lipid droplets × 100.
Progesterone and androstenedione in serum were measured by specific radioimmunoassay (RIAs) using antisera as previously described (Bussmann and Deis, 1979; Forneris and Aguado, 2002). Assay sensitivities were less than 5 ng/ml and 0.02 ng/ml, respectively. For both hormones, the inter- and intra-assay coefficients of variation were less than 10%. The progesterone and androstenedione contents in ovaries were measured with prior extraction according to the methodology previously described (Sanchez-Criado et al., 1992; Tellería et al., 1995). Briefly, the ovaries were thawed and homogenized in 1 ml of 100% ethanol and centrifuged for 10 min at 2,500 g, and the pellet was extracted twice with 500 μl of acetone. The combined supernatants were evaporated to dryness and redissolved in 1 ml of phosphate-buffered saline (0.01 mol/L, pH 7.0) containing 0.15% (w/v) gelatin. Thereafter, the progesterone and androstenedione contents were measured by RIA.
Means and standard errors were calculated for all data sets. Differences between two groups (nonpregnant vs. pregnant) were analyzed with Student's t-test. For morphometric analysis, differences between groups were evaluated using one-way analysis of variance (ANOVA) followed by Tukey–Kramer multiple comparison test. A probability of less than 0.05 was assumed to be significant.
In viscacha, the interstitial tissue is well developed, arranged in clusters, and scattered homogenously in the ovarian cortex. The clusters are delimited by connective cells and fibers, and they are located near blood vessels. They are often closely associated with atretic preantral follicles. In this rodent, during the atretic process of these follicles the oocyte and the zona pellucida regress, but the granulose cells remain, thus constituting the interstitial tissue (Fig. 1A).
In samples dyed with Sudan black B, the interstitial tissue of nonpregnant and pregnant females is very notable because of the numerous sudanophilic lipid droplets (Fig. 1B). The Liebermann–Schultz reaction was markedly positive in the interstitial tissue of nonpregnant viscachas (Fig. 1C). In contrast, the same reaction was approximately 90% negative in the interstitial tissue of pregnant females (Fig. 1D). No reaction was observed in the other ovarian structures (follicles and corpus luteum) in pregnant as well as nonpregnant females. These results showed only the presence of cholesterol and esters in the interstitial tissue, being these substances are abundant in nonpregnant females and scarce in the pregnant animals.
The ultrastructural features of interstitial cells varied during the reproductive and gestation stages.
The nuclei presented irregular shapes and abundant clumps of dense heterochromatin. Mitochondria and endoplasmic reticulum were scarcely developed. Numerous lipid droplets occupied a great part of the cytoplasm (Fig. 2A).
The ultrastructural appearance of interstitial cells was similar to that in nonpregnancy (Fig. 2B).
The more conspicuous changes were observed during this state. A few cells showed the ultrastructural features described above; the most interstitial cells presented an appearance related to the highest biosynthetic activity. The nuclei were regular, and the heterochromatin was scanty. The cytoplasm contained numerous mitochondria associated often with endoplasmic reticulum. The amount of lipid droplets was notably decreased (Fig. 2C).
The population of interstitial cells was moderately heterogeneous. They exhibited an aspect similar to mid-pregnancy, but the amount of lipid droplets increased (Fig. 2D).
In the interstitial cells, the cytoplasmic structures related to steroidogenesis, such as lipid droplets, mitochondria, endoplasmic reticulum, were arranged in close relation (Fig. 3A). This special arrangement was frequently observed in the sections from female viscachas at mid-pregnancy.
In the mid-pregnancy, the relative area of lipid droplets, estimated by proportion quotient (PQLD), was lower than in nonpregnancy and other periods of gestation (one-way ANOVA; Fig. 4).
Figure 5 shows that the progesterone concentrations in ovarian content (a) and in serum (b) in pregnant viscacha were significantly higher than those in nonpregnant animals. Figure 6 shows that the androstenedione concentrations in the ovaries (1) and in serum (2) in pregnant viscacha were significantly higher than those in nonpregnant animals.
A notable characteristic of the ovarian interstitial tissue is the massive accumulation of lipids (Carithiers and Green, 1972), which is related to the high or low secretory activity. The steroid-secreting cells show scarce lipid droplets constituted principally by phospholipids. In these cells, the cholesterol and its esters are absent or scanty, which leads us to postulate an active utilization for steroidogenesis. Similarly to other reports (Guraya and Greenwald, 1964; Brook and Clarke, 1989), we verified the presence of lipids and cholesterol and its esters in the ovarian interstitial tissue of viscacha. Furthermore, the lipids were demonstrated by ultrastructural evidence. However, in interstitial tissue of viscacha, the quantity and quality of lipids vary according to the reproductive condition. In nonpregnant females, there was abundant cholesterol and its esters; in contrast, these substances were scarce during the pregnancy, which suggests that the cholesterol and its esters may be stored before the pregnancy, and then mobilized and used during pregnancy. Moreover, the amount of lipid inclusions was smaller during mid-pregnancy.
The morphological changes of interstitial tissue during pregnancy have been little studied. In 1976, Lawrence and Burden (1976) investigated the autonomic innervation of the interstitial gland of the rat ovary during gestation. They observed that the number and intensity of interstitial fluorescent adrenergic nerves increased as pregnancy progressed. Subsequently, Lawrence et al. (1977) analyzed the fine structure of rat interstitial cells during the same reproductive condition. This time, they observed ultrastructural features that suggest higher steroidogenic activity during the first half of pregnancy and regressive activity during the second half. Similarly, Garcia et al. (1984) found ultrastructural characteristics of interstitial cells of bat (Myotis myotis) during gestation that may indicate steroidogenic activity by these elements. In viscacha ovary, the interstitial cells presented the features of steroid-secreting structures that varied according to the reproductive and pregnancy stages. The most notable differences were observed during mid-pregnancy. In this period, the interstitial cells showed numerous mitochondria and developed endoplasmic reticulum and scarce lipids. When the lipids are present, they are often in close relation with mitochondria and endoplasmic reticulum. This particular arrangement has been proposed as an indicative functional interrelation to contribute to the synthesis of steroids (Guraya, 1991). All these aspects suggest steroidogenic activity, which may be indicating the participation of interstitial tissue in the ovarian function while gestation occurs.
In mammals, normal gestation requires progesterone mainly from the corpus luteum during all or part of the process, depending on the species. Weir (1971b) has suggested that the corpora lutea of Lagostomus persist throughout the pregnancy, and the ovary seems essential for the maintenance of the pregnancy. In this study, the levels of both ovarian and serum progesterone were higher in pregnant females. These high levels of progesterone may be related to the long gestation in this specie.
The corpus luteum is a transient endocrine gland that develops from the ruptured follicles at ovulation. This structure presents a functional life span and then regress. The maintenance of the corpus luteum involves the interplay of anterior pituitary, ovarian, and placental hormones, and the regression involves the discontinued production of progesterone (functional regression) and the structural involution associated with the programmed death of the luteal cells (Tellería et al., 2001). Today, data are available to consider that ovarian paracrine or autocrine factors modulate the maintenance or regression in the corpus luteum. Among these modulator factors, androstenedione has been proposed to present a luteotrophic effect (Casais et al., 2006). In fact, several investigations have demonstrated that androstenedione is able to stimulate the luteal progesterone secretion (Carrizo et al., 1994; Tellería et al., 1995) and to inhibit the luteal apoptosis (Goyeneche et al., 2002). In our investigation, according to the increase of progesterone, we observed the highest levels of androstenedione in pregnant viscachas. These high levels of androstenedione may play a role in the maintenance of the corpus luteum and contribute to its survival throughout the long pregnancy in this rodent. At present, the hormones produced by interstitial tissue are not clearly established. Most investigations indicate that the principal steroids are androgens (Gore-Langton and Armstrong, 1994). In viscacha, during pregnancy, the interstitial cells showed ultrastructural features of steroidogenic activity, which suggests that the interstitial tissue may be a source of androstenedione. However, this suggestion needs more evidence to be confirmed.
The development and differentiation of ovarian interstitial tissue has been related to the follicular atresia. In most species, the interstitial tissue originates from the theca interna of atretic follicles. However, in viscacha ovary, this component arises principally from the granulosa cells of atretic preantral follicles. During the atresia of preantral follicles, the oocyte and the zona pellucida regress, constituting a small mass with a colloidal appearance, and the granulosa becomes interstitial tissue. Harrison and Weir (1977) have pointed out that, in some species, the source of interstitial tissue was problematic. In some mammals, the theca interna is not well developed as in viscacha. These authors proposed that the granulosa cells might become interstitial cells. Furthermore, we observed that the ultrastructural features of interstitial cells, especially during the mid-pregnancy, are similar to the granulosa cells (results not shown). Thus, we suggest that the principal source of interstitial cells is the granulosa membrane in this rodent.
In this study, we used animals captured in their natural habitat. Consequently, our results described the reproductive physiology of viscacha and constitute an approach to the processes that occur in life in the wild life. On the other hand, there are no detailed studies about the ovary of female viscachas. Likewise, due to its features, we suggest that the viscacha ovary may be a good model to study the interstitial tissue. The lipid variation was the most significant change showed by interstitial tissue. Thus, we suggest that this compartment may be a storage of precursors for the steroidogenesis, and then, the precursors are principally used during pregnancy by other ovarian compartments (e.g., corpus luteum), when the endocrine requirements are high. Nevertheless, the observed ultrastructural features of interstitial cells suggest steroidogenic activity; thus, these cells may contribute to the total endocrine production synthesized by the ovary. Then, the interstitial tissue of viscacha may play a role as a source of steroid precursors during pregnancy and, probably, further in its maintenance. However, further studies are necessary to confirm these roles.
We thank Mrs. A. Bernardi, Mr. J. Arroyuelo, and Mr. N. Perez for their technical participation. E.G. thanks Facultad de Ciencias Humanas (UNSL) for the additional economic funding.