Effect of maternal diet on ovarian development in potential replacement ewe lambs

Results

Fetal weight tended (P < 0.1) to be less in groups fed R diet (3,427 ± 192 g) than those fed M diet (3,860 ± 165 g). However, levels of Se in diet did not affect fetal weight, and there were no interactions between plane of nutrition and Se levels in diet. Dietary restriction in maternal diet affected fetal ovarian weight, which was less (P < 0.01) for groups fed R diet (66 ± 4 mg) than for those fed M diet (93 ± 9 mg). However, levels of Se in diet did not affect fetal ovarian weight, and there were no interactions between plane of nutrition and Se levels in diet. Fetal ovarian weight in groups fed R diet with ASe or HSe was less (P < 0.03) than in group fed M diet with HSe but similar to group fed M diet with ASe (Table 2). All fetal ovaries contained primordial, primary, and secondary follicles. Presence of antral follicles was detected in 72% ovaries (23 out of 32 ovaries analyzed). Maternal diet did not have any effect on incidence of antral follicles in fetal ovaries. Proliferating cells were detected in primordial, primary, secondary, and antral follicles, in stromal tissues and blood vessel of fetal ovaries. The effects of dietary restriction (P < 0.08) and Se (P < 0.07) levels but no interaction between plane of nutrition and Se in diet on proportion of proliferating primordial follicles were observed. In addition, the effects of dietary restrictions (P < 0.08) but not Se or interactions between plane of nutrition and Se in diet on LI of primordial follicles were observed. The proportion of proliferating primordial follicles was less (P < 0.08) in groups fed R diet than groups fed M diet (7.4 ± 1.2 vs.11.3 ± 1.7%), and was less (P < 0.07) in groups fed HSe than ASe (7.0 ± 1.0 vs.10.9 ± 1.6). The proportion of proliferating primordial follicles was greater (P < 0.05) in the group fed M diet with ASe than in other treatment group. Labeling index of primordial follicles tended (P<0.08) to be less in groups fed R than M diet (2.9 ± 0.4 vs. 4.9 ± 094). Labeling index tended (P=0.1) to be greater in group fed M diet with ASe that in any other group. Overall, LI was greatest (P < 0.001) in granulosa cells of secondary and antral follicles, less in primary follicles, and least in primordial follicles (19.4 ± 0.6 and 19.5 ± 0.6 vs. 10.4 ± 0.4 vs 4.6 ± 0.5%). Labeling index in granulosa and theca cells of secondary (granulosa, 21.1 ± 0.8% vs. theca, 17.8 ± 0.7%) and antral (granulosa, 19.4 ± 0.8% vs. theca, 19.7 ± 1%) follicles was similar. Therefore, data for granulosa and theca were combined and analyzed together within secondary and antral follicles.
For primary follicles, interactions (P < 0.02) between plane of nutrition and Se levels in diet on LI was observed. Labeling index in granulosa layer of primary follicles was less (P < 0.04) in group fed R diet with HSe than in groups fed M diet with HSe or R diet with ASe. For secondary follicles, the interactions (P < 0.09) between plane of nutrition and Se levels in diet, and the effects of level of Se in diet (P < 0.005) on LI in were observed. Labeling index was less (P < 0.005) in groups fed diet with HSe than with ASe (17.3 ± 0.8 vs. 21.0 ± 0.8%). Labeling index was less (P < 0.03) in the group fed M diet with HSe than in the groups fed M diet with ASe or fed R diet with ASe, but it was similar to group fed R diet with HSe. For antral follicles, LI was affected by level of Se in diet, which was less (P < 0.01) in groups fed with HSe than with ASe (17.3 ± 0.9 vs. 20.9 ± 0.8%). Labeling index was less (P < 0.002) in group fed M diet with HSe that in any other treatment group. For stromal tissues, interactions (P < 0.09) between plane of nutrition and Se levels and the effects of Se level in diet (P < 0.03) on LI were observed. Labeling index was less (P < 0.03) in groups fed diet with HSe than with ASe (3.6 ± 0.8 vs. 6.9 ± 1.1 %). Labeling index was greater (P < 0.05) in the group fed M diet with ASe than in any other treatment group. For blood vessels, the effects of plane of nutrition (P < 0.09) and Se level (P < 0.01) in diet on LI were observed. Labeling index tended (P < 0.09) to be less in groups fed R than M diet (7.9 ± 2.2 vs. 11.6 ± 2.7 %), and LI was less (P < 0.01) in groups fed diet with HSe than with ASe (6.2 ± 1.5 vs. 13.7 ± 0.7 %). Labeling index in blood vessels was greater (P < 0.04) in the group fed M diet with ASe than any other treatment group. In the majority of fetal ovarian tissue sections apoptotic cells were not detected. Only in a few tissues sections (5 out of 32), apoptotic cells were present occasionally in a granulosa layer or stromal tissues in fetal ovaries. Apoptotic cells were also detected in ovine luteal tissues serving as a control. Appearance of apoptotic cells in fetal ovaries was not affected by nutritional treatment. Blood vessels marked by factor VIII staining were detected in fetal ovaries. Smaller size (≤ 30 μm) blood vessels were present in the areas containing primordial to antral follicles, and larger (> 30 μm) and small blood vessels were detected in ovarian medulla and hilus. For antral follicles, a network of blood vessels was detected in the theca layer. However, areas containing primordial follicles were rather poorly vascularized. Distribution of blood vessels in the ovaries seemed to not be affected by nutritional treatment.

Discussion

The results of this experiment show that both maternal dietary restrictions and level of Se in the diet differentially affected cell proliferation depending on fetal ovarian tissue compartment. This indicates that plane of nutrition and Se in the maternal diet are involved in the regulation of early folliculogenesis and ovarian tissue growth. In fact, dietary restriction tended to suppress cellular proliferation in primordial follicles and blood vessels but high Se levels suppressed cellular proliferations in primordial, secondary and antral follicles, stromal tissues and blood vessels. These observations are novel, since very limited information is available concerning maternal diet on cellular processes within specific compartments of fetal ovaries including developing follicles and supporting tissues such as stroma and blood vessels. In addition, these data emphasize importance of maternal diet for fetal tissue and organ growth, which is a central concept of fetal/developmental programming (Nathanielsz, 2006; Barker, 2007).

Dietary restrictions, but not Se, in the maternal diet tended to affect fetal weight of the female lambs in our study. However, inconsistent results of the effects of nutrient restrictions on total fetal weight (i.e. difference due to fetal sex were not reported) at the end of pregnancy in sheep were reported, showing decrease or no effects on fetal weight (Osgerby et al. 2002; Rae et al. 2001; Redmer et al. 2004; Lea et al. 2006; Luther et al. 2007). Similar to our findings, a lack of effect of the maternal dietary restriction or environmental pollutants on ovine fetal ovarian weight was reported by others (Rae et al. 2001; Osgerby et al. 2002; Murdoch et al. 2003; Fowler et al., 2008). Thus, the effects of maternal diet restriction on fetal growth seem to depend on the level and/or length of restriction, and the time of restriction implementation. In addition, it seems that level of Se in maternal diet composition does not affect fetal ovarian growth.

By the end of pregnancy, fetal ovaries may contain all types of follicles including primordial, primary, secondary and antral follicles, as observed in this and other studies in sheep and cattle (Wandji et al. 1992; Lundy et al. 1999; Bodensteiner et al. 2000; Rae et al. 2001). Growth and development of fetal ovaries is regulated by numerous factors of fetal and maternal origin including FSH, LH, estrogens, activin, c-kit and its ligand stem cell factor, enzymes controlling steroidogenesis, growth differentiation factor 9, epidermal growth factor and many other factors (McNatty et al. 2000; Tanaka et al. 2001; Juengel et al. 2002; Sawyer et al. 2002; Byskov & Westergaard 2004; van der Hurk & Zhao 2005; Pepe et al., 2006). Since we rarely detected apoptotic cells and atretic follicles in fetal ovaries, it seems that by the end of pregnancy, selected ovarian compartments are rather in the growing phase, and the apoptotic process is minimal. In fact, our observations are in agreement with previously published data demonstrating very low apoptosis in ovine fetal ovaries on day 99 of pregnancy, but enhanced apoptosis in the earlier stages (days 58 to 73) of pregnancy (Murdoch et al. 2003; Qi et al. 2008). Although altered expression of genes Bax and Mcl-1 which are associated with apoptosis, in fetal ovaries on day 110 of pregnancy was reported for underfed sheep, it is unclear if changes in gene expression were accompanied with appearance of apoptotic cells/bodies (Lea et al. 2006). Thus, these data indicate that apoptosis in fetal ovaries is negligible during late pregnancy.

In our study, 8-13% of primordial follicles had at least one granulosa cells that stained positively for PCNA. In contrast, Lundy et al. (1999) demonstrated that 98.7% primordial follicles contained at least one proliferating granulosa cell in fetal ovaries on day 135 of gestation in sheep. These differences are likely due to different techniques used to determine proportion of proliferating cells even though both studies used PCNA as a marker of proliferating cells. We used 5 μm histological sections to determine differences in proportion of proliferating primordial follicles in between nutritional treatments. However, Lundy et al. (1999) used a nucleator technique to determine a proportion of proliferating follicles within a follicle type which allowed for determination of the absolute number of granulosa cells in the individual follicle.

The present study demonstrated that the LI was less in primordial than primary follicles, and than in secondary and antral follicles (5% vs 10% vs. 19%). Similar to our study, LI was less in primary follicles than secondary and antral follicles in adult rat ovaries (Gaytan et al. 1996). Since a longer period of time is required for a new growing follicle to reach the preantral stage than for a transition from the preantral stage to antral follicle (Gougeon 2004), we hypothesize that differences in LI in follicle types may reflect the timing of growth from primordial to primary, from primary to secondary follicle stage, or from secondary to antral follicle stage. However, this subject requires additional study, since very little information is available concerning timing of the follicular growth in fetal ovaries.

The proliferation in granulosa and theca cells of antral fetal follicles (17-20%) was within a range of LI (17-22%) reported for small follicles from adult ovine ovaries (Jablonka-Shariff et al. 1994, 1996). This indicates that cellular proliferation within small antral follicles remains at similar level in fetal ovaries and ovaries from adult animals. In addition, the LI in stromal cells (3-9%) and in blood vessels (9-20%) were also very high. These great levels of cellular proliferation clearly demonstrate that growth of ovarian follicles is accompanied by growth of surrounding stromal tissues and angiogenesis in fetal ovaries at the end of pregnancy. In addition to providing the scaffold for growing follicles and blood vessels, stromal tissues are the source of matrix metalloproteinases and tissue inhibitors which likely play a role in fetal gonadal development (Robinson et al. 2001). Stromal tissues also express estrogen, androgen and progesterone receptors, which are likely involved in the regulation of fetal ovarian growth (Juengel et al. 2002). Vasculogenesis and angiogenesis, marked in this study with high proliferation in blood vessels and a dense network of blood vessels in medulla, hilus, and areas containing primordial to antral follicles, are critical for supporting fetal organ growth (Augustin 2000). In addition, vascular endothelial growth factor protein, the potent angiogenic factor, is expressed in ovine fetal ovaries (Grazul-Bilska, unpublished). Although angiogenesis in adult ovaries has been studied extensively (Redmer & Reynolds 1996; Grazul-Bilska et al. 2001; Fraser 2006), little is known about the angiogenic process in fetal ovaries, and therefore, it needs further investigation. Nevertheless, a high density of blood vessels in ovarian medulla and hilus, and high proliferation in blood vessels and stromal tissues observed in this study indicates that fetal ovaries are fast growing and differentiating. In fact, it seems that by the end of pregnancy, fetal ovaries are approaching a very fast growth stage, as ovaries increase more than 10-fold (from ~80 mg to 1 g), while offspring weight increases only 3-fold (from ~3.5 to 10 kg) from day 135 of gestation to day 20 postpartum (Grazul-Bilska, Vonnahme et al., unpublished). Thus, it is reasonable to postulate that high cellular proliferation in stroma and blood vessels observed in this study is critical for ovarian growth support. Moreover, a complex vascular network in the mouse fetal ovaries has been recently described by Bullejos et al. (2002) which emphasize the importance of blood supply to developing organs. Cellular proliferation in stromal tissues and blood vessels in fetal ovaries are likely controlled by growth and other factors. However, very little information is available concerning the expression and function of growth factors in fetal ovaries (McNatty et al., 2000), and mechanism of regulation of fetal ovarian growth remains to be elucidated.

The present results clearly demonstrated that maternal diet affects cellular proliferation in fetal ovaries. The similar pattern of cellular proliferation decrease in primordial, secondary or antral follicles, stromal tissues and blood vessels was observed in groups fed R diet or in groups fed high Se comparing with groups fed M diet with ASe in our study. However, differential effects of maternal diet on cellular proliferation in follicle types (e.g., primary vs. secondary or antral) were observed in this study. These differences are likely due to a specific stage of follicular development regulation by factors controlling ovarian growth and function (McNatty et al. 2000; Erickson, 2001; Sawyer et al. 2002; Byskov & Westergaard 2004; van der Hurk & Zhao 2005). In fact, the diverse effects of maternal diet on ovine fetal ovaries including ovarian weight, delayed ovarian growth and development, and altered ovarian cell proliferation and apoptosis have been reported by others (Borwick et al. 1997; Rae et al. 2001; Da Silva et al. 2002, 2003; Osgerby et al. 2002). In addition, maternal undernutrition has been demonstrated to reduce the number of follicles beyond the primordial stage in sheep (Rae et al. 2001). This may be associated with reduced cellular proliferations in primordial follicles observed in our study, which likely progressed at slower rates to primary and secondary follicle stage in underfed compared to normally fed animals.

In contrast to our observations, Lea et al. (2006) reported enhanced cellular proliferation in granulosa layer of fetal follicles on day 110 of gestation in underfed sheep. In addition, undernutrition did not change cellular proliferation in primordial follicles (Lea et al. 2006). These differences are likely due to different length of nutritional restrictions and different gestational time point for tissue collection.

In our study, high Se in the maternal diet decreased cellular proliferation in primordial, secondary and antral follicles, stromal tissues and blood vessels. Interestingly, high Se in the maternal diet also decreased jejunal proliferation in female and male fetuses in a similar study (Carlson et al. 2008). In fact, it has been demonstrated that Se is involved in the regulation of cell proliferation (Zeng et al. 2002) and angiogenesis inhibition in mammary cancer (Jiang et al. 1999). Furthermore, high Se concentration can decrease cell proliferation through inhibition of DNA synthesis and induction of apoptosis (Salbe et al. 1990; Yeh et al. 2006; Zeng & Combs 2008). Since, we rarely detected apoptotic cells in fetal ovaries in this study, we hypothesize that high Se in the maternal diet suppresses cell proliferation through affecting DNA synthesis and cell cycle events. However, additional studies should be undertaken at the cellular and molecular levels to determine the mechanism of Se regulation of cell proliferation in fetal ovaries.

The role of Se in regulation of ovarian function in the fetus is unclear. However, for adult ovaries in animal models, it has been demonstrated that Se may modulate granulosa cell proliferation and estradiol-17β synthesis in vitro, and Se administration may affect the ovulation process, implantation and the number of live embryos (Parshad 1999; Basini & Tamanini 2000).

We have clearly demonstrated that maternal diet affected fetal ovarian function marked in this study by changes in cellular proliferation. In fact, the effects of maternal diet and other environmental factors on fetal and postnatal organ growth and function have been investigated for several models including humans (Rhind et al. 2003; Redmer et al. 2004; Luther et al. 2005; Barker 2007; Gardner et al., 2008). It has been demonstrated that environmental factors including maternal diet composition affect fetal development through many different mechanisms including alternations of gene expression or changes in organ structure and physiology (Rhind et al. 2003). However, the mechanism of the effects of undernutrition and Se in the maternal diet on fetal ovarian function remains to be elucidated.

In summary, maternal dietary restriction and/or high level of Se in diet decreased cell proliferation in primordial, secondary and/or antral follicles, stromal cells and blood vessels in fetal ovaries. Moreover, the LI was greater in secondary and antral than in primordial and primary follicles, and LI was similar for granulosa and theca cells. Thus, dietary restriction and Se in maternal diet affect fetal ovarian growth at the end of pregnancy. Furthermore, apoptosis was minimal and a dense network of blood vessels is present in fetal ovaries by the end of pregnancy.