|Home | About | Journals | Submit | Contact Us | Français|
The foundation for development of the male reproduction system occurs in utero, but relatively little is known about the regulation of primate fetal testis maturation. Our laboratories have shown that estrogen regulates key aspects of the physiology of pregnancy and fetal development. Therefore, in the present study, we characterized and quantified germ cells and Sertoli cells in the fetal baboon testis in late normal gestation (i.e., Day 165; term is 184 days) and in baboons administered the aromatase inhibitor letrozole throughout the second half of gestation to assess the impact of endogenous estrogen on fetal testis development. In untreated baboons, the seminiferous cords were comprised of undifferentiated (i.e., type A) spermatogonia classified by their morphology as dark (Ad) or pale (Ap), gonocytes (precursors of type A spermatogonia), unidentified cells (UI), and Sertoli cells. In letrozole-treated baboons, serum estradiol levels were decreased by 95%. The number per milligram of fetal testis (×104) of Ad spermatogonia (0.42 ± 0.11) was 45% lower (P = 0.03), and that of gonocytes (0.58 ± 0.06) and UI (0.45 ± 0.12) was twofold greater (P < 0.01 and P = 0.06, respectively), than in untreated baboons. Moreover, in the seminiferous cords of estrogen-deprived baboons, the basement membrane appeared fragmented, the germ cells and Sertoli cells appeared disorganized, and vacuoles were present. We conclude that endogenous estrogen promotes fetal testis development and that the changes in the germ cell population in the estrogen-deprived baboon fetus may impair spermatogenesis and fertility in adulthood.
The foundation for development of the male reproductive system occurs in utero; thus, improper maturation of the testis during fetal life may compromise reproductive function and fertility in adulthood. Although differentiation of the bipotential gonad into a testis and appearance of testicular cords (precursors of seminiferous tubules), Sertoli cells, and Leydig cells occur during the first trimester of human gestation [1, 2], significant development of the germinal and somatic cell components of the testis occurs during the second half of human pregnancy [3–5]. The adult complement of these cells, however, is not established in the human and other higher primates until after puberty [6, 7]. Based on nuclear morphology, undifferentiated type A spermatogonia cells have been characterized in the human and rhesus monkey as dark (Ad) or pale (Ap) spermatogonia [2, 8–11]. The Ad spermatogonia exhibit mitotic proliferation in the prepubertal , but not the adult , rhesus monkey and classically are considered to be a population of reserve spermatogonial stem cells [8, 11, 14]. The Ap spermatogonia divide in both the prepubertal and adult rhesus monkey  and classically are considered to be a population of renewing spermatogonial stem cells [8, 11, 14]. The Ap cells replicate and produce the first generation of differentiated, or type B, spermatogonia. The numbers of both Ad and Ap spermatogonia are increased several hundred-fold from birth to adulthood . Finally, gonocytes that appear to be in mitotic arrest and are considered to be the precursors of type A spermatogonia  also have been observed within the seminiferous cord of the fetal and neonatal testis [5, 16, 17].
Despite the importance of fetal testis maturation to fertility in the adult, little information is available about the development of spermatogonia and Sertoli cell populations in the nonhuman primate and human fetus. Our laboratories have shown that the baboon exhibits profiles of fetal development and patterns of hormone production during the reproductive cycle and pregnancy that are similar to those in the human, thus providing a valuable nonhuman primate translational model for the study of pregnancy and fetal development [18–20]. Therefore, the first aim of the present study was to use the baboon to characterize and quantify the germ cell and Sertoli cell populations within the fetal testis in late normal gestation. Previously, we have demonstrated that estrogen, the levels of which increase progressively during the second half of gestation, regulates key aspects of the physiology of pregnancy and fetal ovarian development [21, 22]. That estrogen might stimulate development of the male reproductive system  became apparent when it was shown that estrogen receptor beta (ESR2) was expressed by germ and Sertoli cells of the human [24, 25] and baboon  fetal testis. Moreover, male mice with targeted disruption of the P450 aromatase gene (Cyp19) exhibit a decline in sperm number in association with infertility [27, 28]. Estrogen receptor alpha (Esr1) knock-out mice exhibit altered morphology of the testis, efferent ductules, and epididymis as well as reduced sperm motility and also infertility [29–31], whereas ESR2 inactivation increased the number of gonocytes in neonatal mice . Therefore, as the second aim of the present study, we administered an aromatase inhibitor to baboons throughout the second half of gestation, when substantial germinal cell development occurs, to suppress the increase in production and levels of estrogen and, thus, to assess the impact of endogenous estrogen on seminiferous cord, germ cell, and Sertoli cell development in the fetal testis.
The present study was conducted on baboons housed at the University of Maryland School of Medicine and Eastern Virginia Medical School. Female baboons (Papio anubis) weighing 10–14 kg were housed individually in air-conditioned rooms under standardized conditions. Baboons received monkey chow (Teklad-Harlan) and fresh fruit twice daily, vitamins daily, and water ad libitum. Females were paired with male baboons for 5 days at the anticipated time of ovulation, which was determined by menstrual cycle history and perineal turgescence. Pregnancy was determined by palpation and ultrasonography, and Day 1 of gestation represented the day after ovulation (length of gestation, 184 days). Blood samples (2–4 ml) were obtained from a maternal saphenous (peripheral) vein at 1- to 3-day intervals between Days 100 and 165 of gestation after brief restraint and sedation with ketamine HCl (10 mg/kg body wt, i.m.) to assess estradiol levels.
Baboons were either administered 1 ml of sesame oil vehicle via s.c. injection (n = 11, aim 1) or the aromatase inhibitor letrozole (4,4′-[1,2,4-triazol-1-yl-methylene]-bis-benzonitrite; Novartis Pharma AG) daily on Days 100–164 of gestation at a dose of 115 μg (kg body wt)−1 day−1 s.c. in 1 ml of sesame oil (n = 13, aim 2). After maternal letrozole administration, serum letrozole levels as determined by HPLC were present in both the mother (4.12 ng/ml) and fetus (3.12 ng/ml), indicating that this aromatase inhibitor readily crossed the placenta to suppress potential sites of aromatization in the fetus as well. On Day 165, baboons were anesthetized with isoflurane, after which blood samples (1.0 ml) were removed from the umbilical artery (i.e., the fetus) and a uterine vein, the fetus and placenta delivered by cesarean section, and the fetus killed by i.v. injection of pentobarbital (100 mg/kg body wt; Euthasol, Vibrec, Inc.).
Animals were cared for and used strictly in accordance with U.S. Department of Agriculture regulations and the Guide for the Care and Use of Laboratory Animals prepared by the National Research Council (National Academy Press, 1996). The experimental protocol used in the present study was approved by the Institutional Animal Care and Use Committees of the University of Maryland School of Medicine and Eastern Virginia Medical School.
Serum estradiol and testosterone concentrations were determined by radioimmunoassay (RIA) via an automated chemiluminescent immunoassay system (Immulite; Diagnostic Products Corp.) as described previously .
Serum FSH and LH levels were measured by RIA, as described previously [34, 35], using polyclonal rabbit antisera to recombinant baboon FSH and LH purchased from the National Institutes of Health National Hormone and Peptide Program. Optimization was achieved through studies using various incubation schedules, times, and buffers. Specificity was confirmed by less than 0.1% cross-reactivity of recombinant baboon LH in the FSH RIA and 3.0% cross-reactivity of FSH in the LH RIA. The baboon FSH and LH RIA exhibited minimum detectable doses of 0.018 and 0.036 ng/tube and minimum effective doses of 0.125 and 0.60 ng/tube, respectively. The baboon FSH and LH RIA had intra-assay coefficients of variation of 5.4% and 6.2%, respectively, and interassay coefficients of variation of 6.9% and 4.5%, respectively.
One testis from each baboon fetus was weighed, and its volume was measured by displacement of distilled water. The testis was then fixed in Bouin's fixative, embedded in paraffin, and sectioned (thickness, 4 μm), after which alternating sections were stained with periodic acid Schiff-hematoxylin. Morphometric quantification of germ cells and Sertoli cells was performed by established methods as described by Marshall et al.  and Simorangkir et al. [12, 37]. The volume fractions of germ cell and Sertoli cell nuclei and seminiferous cords were determined by the point-counting method  using a grid of intersecting lines superimposed over the tissue sections. The number of intersections on the grid (test points) overlying the tissue component was counted, and the ratio of these points to the total number represented the volume fraction of the respective component. The lengths of the seminiferous cords were estimated from their absolute volumes and diameters. In total, 6000 test points were examined on randomly selected sections of each testis, and cell counts were corrected using the method described by Abercrombie . The total numbers of germ cells and Sertoli cells per testis were calculated in two different ways: 1) the product of total length and number of cells per cross section of the seminiferous cords, and 2) the absolute volume (i.e., volume fraction of cell nuclei × testis weight ÷ specific gravity of testis) of all nuclei of the particular cell type divided by the mean nuclear volume. The total numbers of cells per testis were then divided by whole-testis weight, which included the interstitial tissue and rete testis as well as the seminiferous cords.
The morphology of the seminiferous cords was assessed in hematoxylin and eosin-stained sections (thickness, 4 μm) of the fetal baboon testis. Seminiferous cords were considered to be normal in appearance if the basement membrane was well defined, intact, and in close contact with the peritubuler myocytes; if germ cells and Sertoli cells were in close contact with the basement membrane for at least 75% of the circumference of the seminiferous cords; and if vacuoles were absent. The percentage of abnormal-appearing (vs. normal-appearing) seminiferous cords was quantified in a minimum of 100 randomly chosen cords per fetal baboon testis.
Paraffin-embedded fetal baboon testis sections (thickness, 4 μm) were boiled in 0.01 M sodium citrate, treated with Protease (Biomeda) for 5 min at room temperature, incubated in H2O2 to inhibit endogenous peroxidase, and blocked with serum-free protein block (DAKO Corp.). Tissues were incubated overnight at 4°C with rabbit polyclonal primary antibody to caspase 3, which recognizes the long and short active cleaved forms of caspase 3 (final dilution, 1:500; Abcam, Inc.) or mouse monoclonal antibody to MKI67 (final dilution, 1:250; DakoCytomation). Tissues were then incubated for 1 h at room temperature with biotinylated anti-rabbit or anti-mouse secondary immunoglobulins (1:500 each; Vector Laboratories, Inc.) and for 1 h with an avidin-biotin-peroxidase complex (ABC Elite; Vector Laboratories). Tissue sections were developed using diaminobenzidine (Sigma) and lightly counterstained with hematoxylin. Negative controls included omission of the primary antibody or substitution of the secondary immunoglobulin with one raised against a different species (DAKO).
Baboons were randomly assigned to the treatment groups, with data expressed as the mean ± SEM. Statistical analysis of the data was performed by ANOVA with post hoc comparisons of the means by Student-Newman-Keuls test, Student unpaired t-test, or Mann-Whitney test using statistical analysis software (SAS/STAT, Version 8.0, Cary, NC).
Maternal peripheral serum estradiol levels in untreated baboons gradually increased from approximately 1.00 ng/ml at midgestation (i.e., Day 100) to approximately 2.50 ng/ml late in gestation (i.e., Day 165) (Fig. 1). Within 1–2 days of the onset of letrozole administration, maternal peripheral serum estradiol concentrations declined to, and then remained throughout the remainder of gestation, at a level (0.10 ± 0.00 ng/ml) that was less then 5% (P < 0.001) of that in the untreated animals (Fig. 1).
Serum estradiol levels on Day 165 in the umbilical artery (i.e., in the fetus, 0.51 ± 0.09 ng/ml) (Fig. 2A) and maternal uterine vein (4.42 ± 0.47 ng/ml) (Fig. 2B) of untreated baboons were decreased (P < 0.01) to approximately 5% of normal by administration of letrozole. In contrast, fetal serum testosterone concentrations on Day 165 of gestation in untreated baboons (0.95 ± 0.32 ng/ml) were increased (to 3.52 ± 0.58 ng/ml, P < 0.01) by letrozole administration (Fig. 2C) as a result of the block in aromatization of C19-steroid precursors within the placenta and, possibly, the fetal testes.
Fetal serum FSH and LH concentrations in untreated baboons on Day 165 (1.98 ± 0.12 and 30.6 ± 3.4 ng/ml, respectively) were not significantly altered by letrozole treatment (Fig. 3).
Placental weight as well as fetal body and testis weights were similar in untreated and letrozole-treated baboons (Table 1).
In untreated baboons near term, the seminiferous cords of the fetal testis were comprised of undifferentiated (i.e.. type A) spermatogonia that were classified as follows: Ad, with a darkly stained, spherical nucleus and very dense, fine, homogenous chromatin as well as a clear retraction area inside the nuclear membrane, one or more dark, round nucleoli surrounded by a clear space, and eosinophilic material in the cytoplasm (Fig. 4); and Ap, with a spherical or slightly ovoid, lightly stained nucleus as well as coarse, granular, heterogenous chromatin and elongated nucleolus (Fig. 4). Many of the type A spermatogonia were situated close to or on the basement membrane, although some were located in central regions of the cord as well. Gonocytes also were observed in the fetal baboon testis, typically being located in the center of the seminiferous cords (Fig. 4, insert). Gonocytes were very large, round cells having a spherical nucleus, with dispersed homogenous chromatin and a central filamentous nucleolus. Another cell type also existed in the fetal baboon testis with morphological characteristics that did not support assigning this type to any germ cell described in the literature. These cells, termed unidentified cells (UI) (Fig. 4), were larger than Ap or Ad spermatogonia and had a prominent nucleolus and clumps of dark-staining heterochromatin. Differentiated spermatogonia were not observed in the fetal baboon testis. Sertoli cells were much more numerous than spermatogonia, and their nuclei exhibited an irregular ovoid shape with one or more small, densely stained nucleoli typical of an immature type (Fig. 4). Although many fibroblasts populated the interstitium, Leydig cells were relatively low in number (not shown).
In the fetal testis of untreated baboons, many of the Sertoli cells and germ cells were closely aligned against the basement membrane, and the basement appeared well defined, intact, and completely surrounded the seminiferous cords (Fig. 5A). In contrast, in the fetal testis of baboons treated in utero with letrozole, the seminiferous cords often appeared abnormal (i.e., the basement membrane appeared fragmented and crenated, the germ cells and Sertoli cells appeared disorganized, and vacuoles often were present) (Fig. 5B). The percentage of abnormal-appearing (vs. normal-appearing) seminiferous cords in the fetal testis of estrogen-deprived baboons (30% ± 7%) was fivefold greater (P = 0.03) than that in estrogen-replete animals (6% ± 1%).
The volume fractions (number per mg testis × 104) of type A spermatogonia, gonocytes, UI, and Sertoli cells in the fetal baboon testis are presented in Table 2. The number of Ad spermatogonia in letrozole-treated baboons (0.42 ± 0.11) was approximately 45% lower (P = 0.03) than the respective value in untreated animals. In contrast, the number of gonocytes (0.58 ± 0.06) was twofold greater (P < 0.01), and the number of UI (0.45 ± 0.12) 2.5-fold greater (P = 0.06), in baboons deprived of estrogen in utero than in untreated baboons. Consequently, an inverse relationship was found between the volume fractions of Ad spermatogonia with gonocytes and UI in estrogen-replete and estrogen-deprived baboons (Fig. 6). The number of Ap spermatogonia in untreated baboons (4.18 ± 0.66), however, was not significantly different than that in baboons treated in utero with letrozole (Table 2). Moreover, the number of Sertoli cells and the ratio of type A spermatogonia to Sertoli cells (Table 2), as well as seminiferous cord volume and diameter (Table 1), were similar in untreated and letrozole-treated baboons.
Caspase 3 protein was localized by immunocytochemistry on Day 165 within Ad spermatogonia (Fig. 7A), Ap spermatogonia (Fig. 7C), and UI (Fig. 7E) of the fetal testis of untreated and letrozole-treated (Figs. 7, B, D, and F, respectively) baboons. The gonocytes and Sertoli cells, however, showed no caspase 3 immunostaining. The number of Ad and Ap spermatogonia and UI, collectively, that exhibited caspase 3 immunoreactivity, quantified in a minimum of 100 randomly chosen seminiferous cords per testis, was similar in untreated (1.12 ± 0.13 cells/mm2 seminiferous cord area) and letrozole-treated (0.98 ± 0.10 cells/mm2 seminiferous cord area) baboons.
Germ cells and Sertoli cells within the seminiferous cords of the fetal testis of untreated (Fig. 8A) and letrozole-treated (Fig. 8B) baboons on Day 165 of gestation showed no MKI67 immunolabeling. As a positive control, however, MKI67 protein was expressed within the nuclei of spermatogonia of an adult male baboon testis (Fig. 8C).
The current study showed that near term, the fetal baboon testis was comprised of highly organized seminiferous cords that contain Ad and Ap spermatogonia, gonocytes, an unidentified type of cell, and immature Sertoli cells but lacked differentiated spermatogonia. The morphology of the different type A spermatogonia in the fetal baboon, based primarily on the pattern of nuclear staining, appeared similar to that reported in the human  and rhesus monkey  neonates. The ratio of number of Ap spermatogonia to Ad spermatogonia in the testis of the near-term baboon fetus, however, was greater than that observed in the 1- to 2-day-old rhesus monkey neonate , although there was considerable difference in developmental age between the late-gestation baboon fetuses of the present study and the monkey newborns.
The present study further showed that compared with untreated animals, the number of Ad spermatogonia was significantly lower and the numbers of gonocytes and UI greater, whereas the number of Ap spermatogonia and Sertoli cells was unchanged, in the testis of fetuses delivered late in gestation to baboons in which the level of estrogen was suppressed by administration of the aromatase inhibitor letrozole through the second half of gestation. Moreover, changes in the morphology of the seminiferous cords, including disruption of the basement membrane, cell disorganization, and presence of vacuoles, often were observed in the fetal testis of baboons deprived in utero of estrogen. Recently, we have shown that cells within the seminiferous cords of the fetal baboon testis near term express ESR2 , as shown in the human fetal testis [24, 25]. We suggest, therefore, that estrogen has a role in promoting in utero development of the germ cells within the seminiferous cords of the primate fetal testis.
The potential impact of the estrogen-dependent changes in testis maturation shown in the fetal baboon on reproductive function in adulthood requires further study. Classically, however, Ad spermatogonia are considered to constitute a reserve spermatogonial stem cell population [8, 11], whereas gonocytes are considered to be the precursors of type A spermatogonia . It is conceivable, therefore, that the decline in Ad spermatogonia and increase in predecessor gonocytes in the fetal testis of estrogen-deprived baboons reflected a delay in maturation and that this would disrupt the normal progression of germ cell formation and maturation that occurs during postnatal and pubertal development. Consistent with the proposed impact of estrogen deprivation in utero on reproductive function in adulthood, Cyp19-null [27, 28] and Esr1-null [29–31] mice display a reduction in sperm numbers, altered morphology of the testis, and/or infertility. In the few cases reported, men with an ESR1  or CYP19 [41, 42] defect exhibited infertility, small testes, and/or azoospermia. In vitro studies have shown that estrogen acts as a cell survival factor by preventing apoptosis in adult human germ cells  and that estradiol stimulates the renewal of spermatogonial stem cells in the rodent testis [44, 45]. Recent studies also have shown that estradiol and the ESR2-selective agonist 5α-androstane-3β,17β-diol, but not the potent androgen 5α-dihydrotestosterone, stimulated DNA synthesis by and proliferation of rat type A spermatogonia . Based on the present and previous studies, we propose that the changes in morphological development of the primordial germ cell population and the seminiferous cords observed in the estrogen-deprived baboon fetus near term would impair spermatogenesis and, consequently, fertility in adulthood.
The mechanisms underlying the change in numbers of Ad spermatogonia, gonocytes, and UI and the integrity of seminiferous cords in estrogen-suppressed fetal baboons are unknown. Because the expression of caspase 3, a primary mediator of the intrinsic mitochondrial pathway of apoptosis , was similar in the near-term fetal testis of untreated and letrozole-treated baboons, the alteration in seminiferous cord cell numbers in estrogen-deprived baboons does not appear to reflect a change in the amount of programmed cell death unless this process occurred earlier in gestation. Interestingly, because caspase 3 was expressed by type A spermatogonia in the baboon fetus, it appears that turnover of these undifferentiated germ cells occurs during intrauterine development. Because MKI67, a protein present during and thought to be important for progression through the G1, S, G2, and M phases of the cell cycle , was not expressed within cells of the seminiferous cords of either untreated or letrozole-treated baboons on Day 165 of gestation, the change in fetal testis cell numbers induced by estrogen suppression also does not seem to have resulted from a difference in the rate of cell proliferation at this time in development. Both germ cells and Sertoli cells, however, did exhibit MKI67 immunoreactivity within the fetal baboon testis on Day 148 of gestation (not shown), and the rate of mitosis has been shown to decrease in late gestation in the human fetal testis [49, 50]. Therefore, changes in cell proliferation may have occurred in the fetal testis of letrozole-treated baboons earlier during gestation to account for the differences in germ cell numbers in the animals of the present study. Because the UI population did not exhibit a level of caspase 3 or MKI67 immunolabeling that was unique or different from that of Ad or Ap spermatogonia, expression of cell-lineage markers or other approaches are needed to identify this particular cell in the fetal baboon testis. Moreover, further study—for example, of the expression of collagen, laminin, or other seminiferous cord basement membrane components , or of the impact of peritubular myoid cells, which express the androgen receptor —is needed to establish the molecular/biochemical mechanisms underlying disruption of the seminiferous cords in estrogen-suppressed fetal baboons.
Although it is proposed that the normal levels of endogenous estrogen to which the fetus is exposed in utero are required for optimal testis development, the administration of supraphysiological levels of estrogen or environmental endocrine disruptors, particularly early in gestation, impairs male reproductive potential . For example, in utero exposure of the rodent fetus to abnormally high levels of estrogen or xenoestrogens caused a reduction in testis size and germ cell numbers [53, 54]. Thus, a window of sensitivity to exogenous estrogen appears to exist in which inappropriate exposure during early fetal life alters reproductive function . We suggest that either above-normal or below-normal levels of estrogen disrupt testis development in the fetus and that optimal development of the normal complement of spermatogonial stem cells in the primate fetal testis requires physiologically normal levels of endogenous estrogen.
Pituitary FSH and LH , as well as testosterone locally within the testis , promote and maintain spermatogenesis in the adult rhesus monkey. In the human  and rhesus monkey [58, 59], fetal anencephaly, which results in a loss of pituitary gonadotropin secretion, is associated with a reduction in testes weight, germ cell numbers, and testosterone production. These findings, plus the observation that the human fetal testis expresses the receptors for FSH  and LH , support the suggestion that FSH and LH regulate fetal testis development . Mitotic activity of type A spermatogonia, however, appears to be gonadotropin independent in the juvenile rhesus monkey testis . Moreover, in the current study, serum FSH and LH levels were not altered and testosterone levels were elevated, not decreased, in baboon fetuses in which testicular germ cell numbers and morphology were disrupted by estrogen suppression. Additionally, germ cells and/or Sertoli cells express the estrogen receptor, but not the androgen receptor, in the human  and baboon  fetal testis. We suggest, therefore, that the changes in germ cell numbers and disruption in seminiferous cord morphology induced in fetal baboons by administration of letrozole resulted from a decrease in the levels and action of estrogen, but not gonadotropin or testosterone. Because fetal serum testosterone concentrations were elevated in letrozole-treated baboons, however, it is possible that the observed changes in fetal testis development reflected an excess in androgen.
In summary, the present study shows that Ad and Ap spermatogonia, gonocytes, a UI type, and immature Sertoli cells, but not differentiated spermatogonia, were present in the testis of the baboon fetus near term. Moreover, the number of reserve spermatogonia stem cell (type Ad) was reduced, the number of spermatogonial precursor gonocytes and UI were increased, and the morphology of the seminiferous tubules was disrupted near term in the fetal testis of baboons in which endogenous estrogen levels were suppressed by administration of an aromatase inhibitor throughout the second half of gestation. We conclude that endogenous estrogen promotes the development of undifferentiated type A spermatogonia and seminiferous cords of the primate fetal testis and that the perturbations observed in the germ cell population and the integrity of the seminiferous cords in the estrogen-deprived baboon fetus may impair spermatogenesis and fertility in adulthood.
The authors greatly appreciate the secretarial assistance of Mrs. Wanda James with preparation of the manuscript and assistance of Graham Aberdeen, Ph.D., and Thomas Bonagura, Ph.D., with the baboon experimentation. The authors gratefully acknowledge Novartis Pharma AG for generously providing the letrozole employed in the present study.
1Supported by NICHD/ NIH through Cooperative Agreements U54 HD36207 to the University of Maryland and U54 HD01860 to the University of Pittsburgh as part of the Specialized Cooperative Centers Program in Reproduction and Infertility Research.