| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
research-article |
Department of Physiology,3 University of Maryland, Baltimore, Maryland 21201
Department of Anatomy and Cell Biology,4 University of Kansas, Kansas City, Kansas 66160
Department of Epidemiology and Preventive Medicine,5 University of Maryland, Baltimore, Maryland 21201
Department of Biology,6 Syracuse University, Syracuse, New York 13244
Department of Veterinary Biosciences,7 University of Illinois, Urbana, Illinois 61802
ABSTRACT
It is believed that a finite pool of primordial follicles is established during embryonic and neonatal life. At birth, the mouse ovary consists of clusters of interconnected oocytes surrounded by pregranulosa cells. Shortly after birth these structures, termed germ cell cysts or nests (GCN), break down to facilitate primordial follicle formation. Tumor necrosis factor alpha (TNF) is a widely expressed protein with myriad functions. TNF is expressed in the ovary and may regulate GCN breakdown in rats. We investigated whether it participates in GCN breakdown and follicle formation in mice by using an in vitro ovary culture system as well as mutant animal models. We found that TNF and both receptors (TNFRSF1A and TNFRSF1B) are expressed in neonatal mouse ovaries and that TNF promotes oocyte death in neonatal ovaries in vitro. However, deletion of either receptor did not affect follicle endowment, suggesting that TNF does not regulate GCN breakdown in vivo. Tnfrsf1b deletion led to an apparent acceleration of follicular growth and a concomitant expansion of the primordial follicle population. This expansion of the primordial follicle population does not appear to be due to decreased primordial follicle atresia, although this cannot be ruled out completely. This study demonstrates that mouse oocytes express both TNF receptors and are sensitive to TNF-induced death. Additionally, TNFRSF1B is demonstrated to be an important mediator of TNF function in the mouse ovary and an important regulator of folliculogenesis.
apoptosis, follicle, follicular development, ovary
Although it has been challenged recently [1, 2], it remains widely accepted that mammalian females are endowed with a finite pool of primordial follicles within their ovaries [3–6]. During each reproductive cycle, a group of follicles is recruited from the dormant primordial follicle pool into the growing follicle pool, where one or several follicles will ovulate their oocyte for fertilization [7]. Once recruited, the follicle has only two possible fates: ovulation or death (atresia). Follicle recruitment continues only until the primordial follicle reserve is depleted. Thus, since new oocytes are not produced after birth, the size of the primordial follicle endowment at birth or shortly thereafter has tremendous consequences on the fertility and reproductive longevity of the female.
The primordial follicle population is established by a series of events that occur during embryonic life, and in rodents continue into neonatal life, collectively referred to as follicular endowment [7]. In mouse ovaries at birth, the ovary consists of a population of interconnected prophase I-arrested oocytes clustered together in discrete units called germ cell cysts or nests (GCNs), which are surrounded by pregranulosa cells [8]. Shortly after birth, oocyte death within the GCN leads to the disruption of connections between oocytes and the breakdown of the GCN [9]. This facilitates the invasion of surrounding granulosa cells into the nest and their enclosure of surviving oocytes to form the primordial follicles, a process completed by Postnatal Days (PN) 7–8 in mice [9].
GCN breakdown is critical for primordial follicle formation to occur, but how these two processes are regulated is largely unknown. Recent studies have identified several factors that are important for follicle formation and may also be important for the process of GCN breakdown. For example, Soyal et al. [10] identified the transcription factor known as factor in the germ line alpha (FIGLA) as being essential for follicle formation. In Figla-deficient (Figla–/–) females, normal oocytes are present at PN 1, but they do not become incorporated into follicles, and they rapidly die. Nerve growth factor (NGFB) appears important for GCN breakdown, and follicle formation as well, as evidenced by large numbers of oocytes remaining in nests rather than being incorporated into follicles in Ngfb–/– ovaries at PN 7 [11]. Steroid hormones also may be involved in regulating GCN breakdown and follicle formation. Using both in vitro and in vivo methods, Kezele and Skinner [12] demonstrated that progesterone has an inhibitory effect on follicle formation in neonatal rat ovaries, as in its presence, there is a significantly higher percentage of oocytes that remain within GCN.
In the rat, tumor necrosis factor alpha (TNF) has been identified as a potential factor involved in GCN breakdown and follicle formation [13]. TNF is a widely expressed proinflammatory cytokine that exerts a variety of effects on cells depending on the cell type, concentration, and receptor type present [14]. TNF has a variety of functions within the ovary, including the inhibition of gonadotropin-induced steroidogenesis [14]; the promotion of granulosa cell proliferation [15]; and the induction of oocyte, granulosa, and luteal cell death [13, 16–18]. TNF is expressed in the ovaries of most species [14]. In the adult mouse ovary, Tnf mRNA is expressed in oocytes of follicles containing more than two layers of granulosa cells [19]. TNF immunoreactivity is observed in oocytes in neonatal rats, as well as in oocytes of all follicle stages in adult rat ovaries [20]. Additionally, both of its receptor types are present in rat ovaries beginning at Embryonic Day 19 [21]. To our knowledge, TNF receptor expression in neonatal mouse ovaries has not been reported.
TNF signals through two distinct receptors, the p55/p60 (type 1, TNFRSF1A, abbreviated as TNFR1) receptor, and the p75/p80 (type 2, TNFRSF1B, abbreviated as TNFR2) receptor [14]. These receptors have homologous extracellular domains, but they have nonhomologous intracellular domains that activate distinct signaling pathways [14]. TNFR1 binds soluble TNF with much higher affinity than does TNFR2, largely due to the fact that the interaction is much more stable between TNF and TNFR1 [22, 23]. TNF signaling through its receptors can stimulate apoptosis, but receptor binding also activates the nuclear factor kappa b (RELA) transcription factor, which can activate transcription of genes that block TNF-induced cell death [24, 25].
TNF has been shown recently to promote oocyte death in neonatal rat ovaries [13], suggesting that it may be involved in regulating GCN breakdown. Therefore, the objective of this study was to test whether TNF is involved in regulating GCN breakdown in neonatal mouse ovaries as well. Specifically, we tested the hypotheses that TNF is a positive regulator of GCN breakdown and follicle formation by promoting oocyte death, and that TNFR2 is important for follicular development.
Mice with a deletion in the Tnfrsf1a gene (further referred to as Tnfr1–/–) were in a C57BL/6 background and were obtained from Immunex (Seattle, WA) [26]. Mice with a deletion in the Tnfrsf1b gene (Tnfrsf1btm1Mwm/Tnfrsf1btm1Mwm, further referred to as Tnfr2–/–) were in a C57BL/6 background and were obtained from Jackson Laboratories (Bar Harbor, ME). C57BL/6 mice were obtained from either Charles River Laboratories (Wilmington, MA) for experiments involving ovary culture, Harlan Inc. (Indianapolis, IN) for use as controls for experiments involving Tnfr1–/– mice, or from Jackson Laboratories as controls for experiments involving Tnfr2–/– mice. In all experiments, the day of birth (DOB) was considered to be PN 0. Animals were housed in clear plastic cages and maintained on a 12L:12D cycle in a temperature-controlled room (24°C ± 1°C) with 35% ± 4% relative humidity. Mice were provided food and water ad libidum. The University of Maryland Institutional Animal Use and Care Committee approved all protocols involving mice.
Ovary cultures were performed as described in Morrison and Marcinkiewicz [13], with minor modifications. Briefly, ovaries were collected on the DOB in unsupplemented Waymouth medium MB752/1 (Sigma, St. Louis, MO) at 37°C. The ovarian bursa was removed, and ovaries were transferred to Corning culture plates containing Costar Transwell Membrane inserts (VWR International, West Chester, PA) in a drop of unsupplemented medium. DOB ovaries were cultured whole for 3 days in 1.8 ml Waymouth medium supplemented with 0.23 mM pyruvic acid, 50 mg/l streptomycin sulfate, 75 mg/l penicillin G, 3 mg/ml BSA, and 10% [v/v] fetal bovine serum. To test its effect on follicle numbers in vitro, recombinant mouse TNF (Sigma) was added to culture media at a concentration of 0, 0.1, 1, 10, or 50 ng/ml, and ovaries were cultured in a humidified incubator at 37°C, with 5% CO2 in air. Additional cultures were performed using recombinant human TNF (hTNF; Chemicon, Temecula, CA) treatment at a concentration of 50 ng/ml. Media were changed after 48 h. Following 3 days in culture, ovaries were collected and fixed in Kahle solution (4% formalin, 28% ethanol, and 0.34 N glacial acetic acid), where they were kept until processing for histology. For all experiments, at least three separate cultures were performed.
To assess the impact of TNF treatment and Tnfr1 and Tnfr2 deletion, we collected ovaries following in vitro culture from Tnfr1–/– mice at PN 7, and from Tnfr2–/– and wild-type (WT) mice at ages PN 7 and PN 80 (on the morning of estrus for adult animals), and subjected them to morphological assessment of follicle numbers. Ovaries were collected and fixed in Kahle solution for at least 24 h. Following fixation, ovaries were dehydrated through an alcohol series and embedded in Paraplast (VWR International). Ovaries were serially sectioned at 8-µm intervals, mounted on glass slides, stained with Weigert hematoxylin-picric acid methyl blue, dehydrated, and mounted in Permaslip (Alban Scientific Inc., St. Louis, MO). For cultured ovaries, the numbers of naked oocytes (i.e., those remaining in GCNs and not present within follicles), primordial follicles, and primary follicles were counted in every third section. In all other ovaries, the numbers of primordial, primary, preantral, and antral follicles were counted in every 10th section. For this analysis, primordial follicles were considered to be nongrowing follicles, whereas primary, preantral, and antral follicles were considered to be growing follicles. To avoid double counting, only follicles containing an oocyte with a visible nucleus were counted, and all counting was done without knowledge of treatment or genotype. Oocytes were counted as naked if they were present in clusters of at least two oocytes with an absence of intervening somatic cells. Follicles were counted as primordial if they contained an oocyte surrounded by flattened granulosa cells, or a mixture of fewer than seven flattened and cuboidal granulosa cells [27]. Follicles were counted as primary if they contained an oocyte surrounded by a single layer of seven or more cuboidal granulosa cells. Preantral follicles were those containing an oocyte surrounded by two to four complete layers of granulosa cells. Antral follicles were considered those that contained five or more complete granulosa cell layers with or without a visible antrum. Immature follicles were considered healthy if there was an absence of pyknotic granulosa cells and the oocyte was not malformed. If the follicle did not meet these criteria or there were granulosa cell structures without an oocyte, as was observed most frequently for primordial and very small primary follicles, the follicle was scored as atretic. Antral follicles were considered to be healthy if they had an intact oocyte and less than 10% pyknotic granulosa cells, whereas they were considered atretic if they contained more than 10% pyknotic granulosa cells or a degenerating oocyte. All counts are reported as the raw number counted per ovary without a correction factor applied.
Ovaries from C57BL/6 mice aged 2 or 7 days were collected and fixed in 4% paraformaldehyde for 2 h at room temperature, after which they were rinsed in PBS, dehydrated, embedded in Paraplast, serially sectioned at 5-µm intervals, and mounted on glass slides. Sections were deparaffinized and rehydrated. Antigen retrieval was performed in boiling citrate buffer (10 mM) for 5 min, and slides were cooled to room temperature. Slides were washed three times in PBS, incubated in a 3% H2O2 solution in PBS for 10 min, and washed three times in PBS. Sections were blocked with 5% heat-inactivated horse serum for 1 h, after which endogenous biotin sites were blocked using an avidin-biotin blocking kit (Vector Laboratories, Burlingame, CA) according to the manufacturer's protocol. Slides were washed three times in PBS and incubated with primary antibody (anti-TNF; Santa Cruz Biotechnology Inc., Santa Cruz, CA) diluted 1:50 in PBS for 48 h at 4°C. As a negative control, slides were incubated in PBS alone without primary antibody. Slides were washed three times in PBS and incubated with appropriate biotin-conjugated secondary antibodies diluted 1:250 in PBS for 1 h at room temperature. Slides were washed three times in PBS and incubated with the Vectastain Elite ABC reagent (Vector Laboratories) for 1 h at room temperature. Slides were washed three times in Tris buffer (pH 7.2) and then incubated in a 3,3' diaminobenzidine tetrahydrochloride (DAB; Sigma) solution in Tris buffer for 5–10 min at room temperature. Slides then were washed three times in Tris buffer, three times in PBS, counterstained with picric acid-methyl blue, and then dehydrated and mounted in Permaslip.
Whole-Mount Immunohistochemistry
Ovaries were collected from CD1 mice (Charles River Laboratories) on the morning of PN 6 and were fixed in 5.3% formaldehyde overnight at 4°C. Ovaries then were washed twice briefly in 0.1% Triton X-100 in PBS (PT) and once for at least 30 min. Ovaries then were incubated in PT + 5% BSA for 30–60 min before being incubated with primary antibody (anti-TNFR1 [E20] or anti-TNFR2 [L20], Santa Cruz Biotechnology Inc.) diluted 1:50 in PT + 5% BSA overnight at 4°C. As a negative control, ovaries were incubated overnight in PT + 5% BSA without primary antibody. Ovaries then were washed in PT + 1% BSA for 30 min and treated with RNase A for 30 min. DNA was labeled with propidium iodide (Invitrogen Molecular Probes, Carlsbad, CA) for 20 min, and the ovaries were washed in PT + 1% BSA for 30 min. Fluorescein isothiocyanate-conjugated secondary antibody (Invitrogen Molecular Probes) was diluted 1:200 in PT + 5% BSA and incubated with the ovaries for 2 h at room temperature. Ovaries were then washed three times in PT + 1% BSA for 30 min, then briefly with PBS before being mounted in Vectashield (Vector Laboratories). Confocal analysis then was performed using a Zeiss LSM Pascal laser-scanning microscope.
Follicle counts from control groups were similar in all DOB ovary culture experiments, and so data from all experiments were combined, and the n is the number of ovaries in each group. A test of trend was used to analyze these data using a linear regression model to assess whether there was a dose effect of TNF on follicle numbers due to the fact that the numbers of ovaries for the two low-dose TNF groups were half those in the other groups. Follicle numbers between WT, Tnfr1–/–, and Tnfr2–/– ovaries at all time points were compared using Student t-test. P values less than 0.05 were considered to be statistically significant.
To our knowledge, studies have not reported whether TNF and its receptors are expressed in the neonatal mouse ovary. Thus, we first examined whether this is the case. Figure 1 shows that TNF and both receptors are expressed in neonatal mouse ovaries. TNF, TNFR1, and TNFR2 are expressed in the cytoplasm of naked oocytes, as well as in the cytoplasm of oocytes in primordial and growing follicles.
|
We next performed in vitro TNF treatment experiments on ovaries collected on the DOB and cultured for 3 days in vitro. Data from these experiments are shown in Figure 2. TNF treatment caused a dose-dependent reduction in the number of naked oocytes (P 
0.01, test for trend; n = 14 for control ovaries; n = 6 for 0.1 ng/ml TNF ovaries; n = 6 for 1.0 ng/ml TNF ovaries; n = 12 for 10 ng/ml TNF ovaries; n = 12 for 50 ng/ml TNF ovaries; Fig. 2). TNF did not affect the number of primordial follicles (P = 0.13, test for trend; Fig. 2). Similar to naked oocytes, TNF treatment significantly reduced the number of primary follicles (P 0.001, test for trend; Fig. 2). TNF treatment also caused a dose-dependent reduction in the total number of oocytes (P 0.05, test for trend).
|
As TNF was able to kill naked oocytes, and naked oocyte death is necessary for follicle formation, we next investigated whether TNF is involved in regulating GCN breakdown and follicle formation in vivo by examining the effect of Tnfr deletion on follicle formation. Tnfr1 deletion had no effect on primordial, primary, or preantral follicle numbers in both ovaries at PN 7 compared with WTs (n = 5 WT ovary pairs; n = 7 Tnfr1–/– ovary pairs; P = 0.29; Fig. 3A). Similarly, Tnfr2 deletion did not affect the number of primordial follicles in both ovaries of animals at PN 7 (n = 4 WT ovaries; n = 9 Tnfr2–/– ovaries; P = 0.35; Fig. 3B). These data suggest that since Tnfr deletion does not affect primordial follicle numbers, follicle formation is unaffected, and TNF is not involved in GCN breakdown in vivo.
|
Interestingly, however, primary follicle numbers were increased in Tnfr2–/– (762.7 ± 21.1) compared with WT (388.3 ± 21.3) ovaries (P 0.001). In addition, preantral follicles were significantly increased in Tnfr2–/– (7.9 ± 1.6) compared with (0.0 ± 0.0) WT ovaries (P 0.005; Fig. 3C). The percentage of follicles growing was also significantly increased in Tnfr2–/– (33.3% ± 1.4%) compared with WT (22.0% ± 1.1%) ovaries (P 0.001; Fig. 3D). Representative photomicrographs demonstrate the increased size of Tnfr2–/– compared with WT ovaries at PN 7 (Fig. 3, E and F). These data suggest that TNF signaling through TNFR2 may regulate initial follicle recruitment.
We next tested the hypothesis that TNF inhibits initial follicle recruitment via TNFR2. To do this, we first performed ovary culture experiments using DOB Tnfr2–/– ovaries (Fig. 4A). TNF had no effect on naked oocyte, primordial follicle, or primary follicle numbers in Tnfr2–/– ovaries compared with untreated Tnfr2–/– ovaries (n = 12 control ovaries; n = 12 TNF-treated ovaries; P = 0.28). Further, hTNF, which binds only to murine TNFR1 and not to TNFR2, had no effect on follicle numbers in DOB WT ovaries in vitro (n = 9 control ovaries; n = 9 hTNF ovaries; P = 0.42; Fig. 4B). These data suggest that TNFR2 may be involved in mediating TNF inhibition of follicle recruitment. Importantly, these data also demonstrate that TNFR2 is required for TNF-induced oocyte death.
|
To further examine the hypothesis that TNF regulates follicle recruitment through TNFR2, we investigated the effect of Tnfr2 deletion on follicle numbers in adult ovaries (Fig. 5). Surprisingly, the number of primordial follicles at PN 80 was significantly greater in Tnfr2–/– (314.75 ± 32.3) compared with WT (136.4 ± 16.5) ovaries (n = 5 WT ovaries; n = 4 Tnfr2–/– ovaries; P 0.001). This was not simply due to an increased size of the ovary, as the density of primordial follicles is significantly greater in Tnfr2–/– (14.4 ± 0.6 follicles per section) compared with WT (7.5 ± 0.7 follicles per section) ovaries (P 0.001). Similarly, primary follicle numbers were greater in Tnfr2–/– (102.3 ± 11.5) compared with WT (59.6 ± 3.7) ovaries (P 0.05). Preantral follicles also were elevated in Tnfr2–/– (64.0 ± 1.7) compared with WT (38.8 ± 6.2) ovaries (P 0.01). No difference was observed in the number of antral follicles (P = 0.2).
|
The follicle count data at PN 80 argue against the hypothesis that increased follicle numbers in Tnfr2–/– ovaries are due to an abrogation of TNF inhibition of follicle recruitment. Thus, we next tested the hypothesis that the increased number of growing follicles in Tnfr2–/– ovaries is due to decreased follicle atresia in the absence of TNF signaling through TNFR2. Figure 6 shows the amount of atresia occurring in vivo at PN 80 in WT and Tnfr2–/– ovaries. There was no difference in the number of atretic primordial (P = 0.37), primary (P = 0.7), preantral (P = 0.1), or antral (P = 0.8) follicles observed in Tnfr2–/– compared with WT ovaries.
|
TNF is a ubiquitously expressed protein that has a number of ovarian functions, including the modulation of steroidogenesis and folliculogenesis [14]. Here we demonstrate that, in the neonatal mouse ovary, it also can promote oocyte death in vitro, and that TNFR2 is an important mediator of TNF activity both in vitro and in vivo. Using an in vitro culture system of neonatal ovaries, we observed that TNF has the ability to kill oocytes, as assessed by its ability to significantly reduce the number of oocytes within treated ovaries. This is consistent with a report by Morrison and Marcinkiewicz [13], in which they observed that TNF promotes apoptosis of oocytes in neonatal rat ovaries in vitro; however, the dose response that we observed is in contrast to their findings. In this study, we observed that TNF led to a dose-dependent reduction in oocyte number, whereas Morrison and Marcinkiewicz [13] observed that the only effective dose for TNF-induced oocyte death in neonatal rat ovaries was 1 ng/ml. The difference between these studies could be due to differential sensitivity of the oocytes to TNF, which was due potentially to differential expression patterns of its receptors among species. We observed that both TNF receptors are expressed in mouse oocytes (Fig. 1). While both receptor types are expressed in rat ovaries as early as Embryonic Day 19 [21], their localization has not been reported.
We observed that TNF and both receptors have staining patterns that include the cytoplasm of the oocyte. Immunoreactive TNF has been reported previously in the cytoplasm of rat oocytes [20]. Similarly, immunoreactive TNFR1 has been observed intracellularly in other cell types localizing, for example, to structures including the Golgi apparatus and tubular structures associated with the endoplasmic reticulum, and changes in subcellular localization of the receptor appear to be important for controlling activation of downstream signaling pathways [28, 29]. Importantly, TNFR1 can localize to lipid rafts, and its association with lipid rafts appears to affect whether it signals death or survival [30]. Lipid rafts are present not just in the plasma membrane, but in intracellular membranes as well, including the Golgi apparatus, late endosomes, and recycling endosomes [31–34]. Additionally, TNFR2, when targeted to intracellular lipid rafts, is able to interact with its signaling molecules and effectively transduce signals [35, 36]. It has been proposed that when TNF and its receptors are concurrently expressed within a cell, they may interact to initiate signaling intracellularly [36]. Thus, it is possible that since we observed immunoreactive TNF along with its receptors within the cytoplasm of oocytes, an intracellular signaling pathway is active, and this may help determine oocyte fate.
Interestingly, we observed that TNF-induced death of oocytes requires TNFR2. This was demonstrated by the fact that TNF did not kill oocytes in Tnfr2–/– ovaries, and that hTNF, which only binds to murine TNFR1, did not kill oocytes in WT ovaries. This was surprising because in many cell types TNFR1 mediates TNF-induced apoptosis. This is largely due to the fact that the intracellular portion of TNFR1, but not TNFR2, contains a death domain [37]. However, TNF-induced death of mitotic hematopoietic progenitor cells is dependent on TNFR2 [38], and thus it is possible that TNFR2 mediates TNF-induced oocyte death as well.
It is also possible that TNFR2 is indirectly involved in promoting oocyte death. For example, TNF has been shown to potentiate FASL-induced apoptosis through TNFR2 signaling [39–41]. Further, there are reports of cooperativity between TNFR1 and TNFR2 signaling. Tartaglia et al. [42] described a situation in which TNF signaling via TNFR1 is enhanced by ligand passing from TNFR2. They showed that TNF associates 20 times more rapidly with TNFR2 than with TNFR1 when the receptors are isolated. When the receptors are present together, TNFR2 greatly enhances TNF association with TNFR1. TNF association with TNFR2 is not stable, and thus it appears that when the receptors are present together TNFR2 concentrates TNF at the membrane, thus increasing its local concentration and facilitating interaction with TNFR1, with which TNF binds much more stably [22, 23, 42]. Further, it has been shown that TNF induces the formation of transient heterocomplexes between TNFR1 and TNFR2 [43]. Studies suggest that these heterocomplexes could facilitate ligand passing of TNF or cooperativity in signaling initiated by the two receptors by the aggregation of their intracellular domains. In fact, Fang et al. [44] demonstrated that TNFR2 enhances TNFR1-induced apoptosis. Therefore, as both receptors are present in mouse oocytes, TNFR2 may indirectly promote oocyte death by facilitating TNF-induced apoptosis via TNFR1, and in the absence of TNFR2, TNFR1-mediated death could be attenuated.
The data showing that TNF stimulated oocyte death in vitro supported the hypothesis that TNF might be involved in the regulation of GCN breakdown and follicle formation. However, neither receptor knockout model affected primordial follicle numbers, suggesting that follicle formation occurs normally in their absence and that in vivo TNF is dispensable for GCN breakdown and follicle formation. It is possible, however, that the TNFR1 and TNFR2 receptors could have overlapping functions in mediating TNF signaling during these processes, and thus there would be no observable phenotype following single receptor deletion. To better assess this, double-receptor knockouts or TNF-deficient mice should be examined to see whether there is any effect on GCN breakdown and follicle formation. There did appear to be more naked oocytes in WT compared with Tnfr2–/– ovaries following 3 days of culture in vitro (compare the naked oocyte numbers in control groups from WT animals [Figs. 2a and 4b] to those in Tnfr2–/– animals [Fig. 4a]), suggesting that Tnfr2 deficiency led to differences in follicle distribution during follicle formation. However, if differences in follicle distribution do exist between WT and Tnfr2–/– animals in vivo during follicle formation, they are rectified by the completion of follicle formation.
Interestingly, Tnfr2–/– ovaries at PN 7 exhibited greatly increased follicle growth compared with WT ovaries. This fact, together with the data demonstrating that TNF reduces primary follicle numbers in cultured neonatal WT but not Tnfr2–/– ovaries, led us to hypothesize that TNF regulates initial follicle recruitment through TNFR2. To test this hypothesis, we performed morphometric analysis of follicle numbers in ovaries from adult animals. The phenotype of older Tnfr2–/– ovaries, in which there is an expansion of the primordial follicle population concomitant with an increase in the size of the growing follicle population, suggests that TNF does not regulate follicle recruitment. If it did regulate follicle recruitment, then the increased numbers of growing follicles observed in Tnfr2–/– ovaries would be accompanied by a decrease in the size of the primordial follicle population rather than an increase, as is seen following the deletion of other recruitment inhibitory factors, such as anti-Müllerian hormone (Amh) [45].
The increased number of growing follicles is difficult to reconcile with an expanding primordial follicle population. However, as primordial follicles are lost not just by growth but also by death [46, 47], and as TNF kills immature oocytes via TNFR2, we hypothesized that Tnfr2 deletion abrogates TNF-induced atresia of immature follicles, leading to an expansion of the primordial follicle population over time. However, this was not the case, as we did not observe reduced atresia of any follicle class in adult Tnfr2–/– ovaries. However, we were most interested in whether primordial follicle atresia was abrogated by Tnfr2 deletion. While some studies claim to have observed large numbers of atretic primordial follicles [1, 48, 49], others suggest that it is a rare event [50–52]. The belief that primordial follicle atresia is rare may be due to the fact that it is very difficult to detect morphologically due to the follicles' small size, rapid clearance, and absence of visible pyknotic granulosa cells [52, 53]. Therefore, due to the difficulties associated with morphological assessment of primordial follicle atresia, it is not possible to say definitively that it is unaffected by Tnfr2 deletion. Future studies should further address this question.
Another potential reason for the expansion of the primordial follicle population could be increased postnatal oogenesis. Johnson et al. [2] proposed a model in which germline stem cells (GSCs) residing in bone marrow send germ cell precursors through the bloodstream to the ovary to produce new oocytes in adult animals. It could be possible that oocyte production is enhanced in postnatal Tnfr2–/– females, but we do not feel this is likely. The GSC model goes against the central dogma of female reproductive biology [3] and has been met with a great deal of skepticism [4, 5]. Additionally, a recent paper refutes the claim that the bone marrow or the bloodstream is a source of new oocytes, at least ovulation-competent ones [6]. Thus, it remains predominately accepted that postnatal oogenesis does not occur, and we therefore do not think that this is an explanation for the observed expansion of the primordial follicle population.
It also is possible that the expansion of the primordial follicle population is due to a slowing of follicular growth, resulting in a follicle backup. Increased follicle numbers in Tnfr2–/– ovaries are observed only in the small growing follicle stages, not in antral follicles. The transition of follicles to the antral stage is exquisitely regulated by a feedback loop involving estradiol, inhibin, and FSH [7]. While it is not clear how antral follicle numbers are equilibrated at this stage in Tnfr2–/– ovaries, it does not appear to be due to increased atresia of preantral or antral follicles, nor is it due to altered estradiol production, as serum estradiol levels are not altered in Tnfr2–/– mice (data not shown). A possible way in which follicle numbers could be equilibrated at the transition from preantral to antral follicles is by an inhibition of granulosa cell proliferation and a slowing of growth of follicles into the antral pool. While TNF may stimulate granulosa cell proliferation [15, 54], it can also inhibit FSH-stimulated granulosa cell proliferation independently of altered cell viability [54]. Thus, it is possible that in the absence of Tnfr2, TNF inhibition of the FSH-stimulated proliferation of granulosa cells is enhanced without an effect on atresia. This could cause a slowdown of follicle growth, leading to an accumulation of immature follicles and an expansion of the primordial follicle pool.
In conclusion, we have demonstrated that TNF is cytotoxic to small, immature mouse oocytes, and that TNFR2 is an important mediator of TNF activity in the mouse ovary. The absence of Tnfr2 results in expansion of the primordial follicle pool and altered folliculogenesis. It is not clear whether this primordial follicle population expansion is due to a slowing of follicular growth, or whether it is due to reduced primordial follicle atresia. This is an important distinction that should be addressed by future studies. However Tnfr2 deletion affects follicle numbers, it does not appear to affect female fertility, as antral follicle numbers and steroidogenesis are unaltered, supporting previous work showing that Tnfr2–/– females produce similar numbers of litters compared with WT females [55]. TNF action and signaling pathways are very complex and may vary based on ovarian cell type and follicle stage, as well as the subcellular localization of each receptor. Future studies should continue to further analyze the involvement of each receptor type in the myriad activities of TNF within the ovary.
FOOTNOTES
1Supported by National Institutes of Health HD38955, T32HD07170, and CA50616, and a grant from the Women's Health Research Group at the University of Maryland. 
Correspondence: 2FAX: 217 244 1652; e-mail: jflaws{at}uiuc.edu
Received: 11 July 2006.
First decision: 16 August 2006.
Accepted: 16 October 2006.
REFERENCES
, a germ cell-specific transcription factor required for ovarian follicle formation. Development 2000; 127:4645–4654[Abstract] enhances oocyte/follicle apoptosis in the neonatal rat ovary. Biol Reprod 2002; 66:450–457 in follicular development, ovulation, and the life span of the corpus luteum. Domest Anim Endocrinol 1997; 14:1–15[CrossRef][Medline] (TNF) increases granulosa cell proliferation: dependence on c-Jun and TNF receptor type 1. Endocrinology 2004; 145:1218–1226 gene expression in mouse oocytes and follicular cells. Biol Reprod 1993; 48:707–714[Abstract]
B-mediated X-linked inhibitor of apoptosis protein expression prevents rat granulosa cells from tumor necrosis factor -induced apoptosis. Endocrinology 2001; 142:557–563 B-mediated induction of Flice-like inhibitory protein prevents tumor necrosis factor -induced apoptosis in rat granulosa cells. Biol Reprod 2002; 67:436–441 mediates both apoptotic cell death and cell proliferation in a human hematopoietic cell line dependent on mitotic activity and receptor subtype expression. J Biol Chem 1999; 274:9539–9547
, tumor necrosis factor-, and cycloheximide. Endocrinology 1998; 139:4860–4869. Acta Pharmacol Sin 2001; 22:1039–1044[Medline]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
