Control of Murine Primordial Follicle Growth Activation by IκB/NFκB Signaling

Introduction

A key determinant of the length of time that ovaries function during adult life is the rate at which primordial follicles leave their dormant state and begin to grow (primordial follicle growth activation, PFGA)[1, 2]. The ovaries of women in their thirties usually contain tens of thousands of primordial follicles (e.g., the ovarian “reserve”). A small number of primordial follicles undergoes PFGA daily, supplying the ovary with its population of growing follicles [3, 4]. Once committed to growth, the vast majority of follicles die in a process referred to as atresia [5, 6]. This results in the single dominant follicle (in “monovular” species including humans) or, follicles (in “polyovular” species including mice), that survive to ovulate a mature, fertilization-competent egg(s).

Exhaustion of the primordial reserve results in the cessation of ovarian function, and in humans, this is termed the menopause. The median menopausal age in women is 51 years [7]. When ovaries begin to fail before age 40, it is termed “primary ovarian insufficiency (POI)” [7, 2, 8]. POI affects approximately 1 in 250 women under age 35, and 1% of all women under age 40. Approximately 5% of women experience ovarian failure between ages 40 and 45 years, while outside the definition of POI [8]. Beyond the loss of fertility, ovarian demise results in significant and costly health and well-being compromises. Cardiovascular health [9–11], bone density, muscle mass, body fat composition, and self-reported well-being measures [12] all worsen upon ovarian failure. Specifically with regards to cardiovascular health, lower age of menopause and time since the start of menopause are independently associated with poorer carotid artery thickness and aortic stiffness [10, 11]. If afflicted with POI, women face years and possibly decades of increased risk for these problems and their treatment limitations [13]. A major goal of women’s health research is to develop interventions that slow the rate of PFGA and thereby delay or prevent POI.

A key known negative physiological regulator of PFGA is Antimüllerian hormone (Amh), a protein that can also be used to indirectly estimate the ovarian reserve [14, 15]. Amh knockout mice [14] were found to display an acceleration of PFGA, characterized by a more rapid decline in primordial follicle number and an increase in growing follicles at different postnatal time points. More recently and conversely, Amh delivery to mice as purified recombinant protein, or, using a gene therapy approach, resulted in reduced PFGA and in the latter case, ovarian “quiescence” [16]. Another negative regulator of PFGA is the lipid phosphatase Phosphatase and tensin homolog deleted on chromosome ten (Pten) [17–19]. Blocking Pten activity accelerates PFGA in mouse models in vivo and also in human primordial follicles in experimental clinical interventions [20, 21]. Pten is a key upstream negative regulator of pro-growth nutrient-responsive Akt/mTOR signaling, which has been shown to regulate PFGA at two levels. First, Akt has been shown to phosphorylate the transcription factor Forkhead box protein O3 (FoxO3), and prevent its nuclear localization [22, 23]. FoxO3 has been shown to be a powerful negative regulator of the rate of PFGA [22], and its inactivation by Akt is thus consistent with growth. Second, mTOR action upon downstream factors 4-EBP and p70S6K result in positive regulation of protein synthesis, and the cell cycle, respectively [17, 23]. Pten’s action upstream of each event favors blocked PFGA.

Tnfα signaling through the receptor Tnfr2 has shown to be a key positive regulator of PFGA ][24, 25]. While the primordial reserve of follicles did not differ between Tnfr2 knockout (Tnfr2KO) animals and wild-type (WT) controls early in postnatal life, mutant ovaries harbored significantly greater numbers of primordial (and small growing) follicles later in life. That outcome can be interpreted as resulting from a slowed rate of PFGA, revealing the positive regulatory action of Tnfr2 and Tnfα. Since these initial reports, no further mechanistic information about Tnfα:Tnfr2 action has been produced. These data are again relevant due to a recent GWAS analysis of nearly 40,000 patients, where the Nuclear factor kappa-light-chain-enhancer of activated B cells (NFκB) signaling pathway was identified as strongly associated with the timing of menopause in women [26]. Because TNFα is known to stimulate NFκB signaling, and TNFα is a well-recognized NFκB target gene, it is reasonable to hypothesize that NFκB signaling may play a role in PFGA.

The signaling pathway leading to NFκB activation following TNFα has been well characterized. In quiescent cells, members of the IκB family of inhibitory proteins (IκBα and IκBβ bind to NFκB dimers in the cytoplasm, preventing their nuclear translocation and subsequent DNA binding [27]. The expression level of the IκB family of inhibitory proteins is the predominant force preventing NFκB nuclear translocation and subsequent target gene expression. TNFα binding to its receptor leads to the activation of I kappa B kinases (IKK), resulting in IκB phosphorylation [28]. This signaling event inactivates the inhibitory potential of the IκBs by targeting them for degradation, allowing NFκB subunit release to the nucleus and subsequent DNA binding and altered gene transcription [27].

Some information is available about NFκB pathway action in cells of the ovary and in ovarian physiology. In an early study, Xiao et al. showed that NFκB activity, as measured by nuclear NFκB subunit translocation, induced expression of the anti-apoptotic gene Xiap in rat granulosa cells.[29] This pro-survival NFκB response occurred in the presence of Tnfα. Signaling through EGR1 (early growth response 1) and its impact on NFκB activation has also been implicated in granulosa cell survival [30]. Similarly, NFκB activation was again found to regulate cell survival in primary granulosa cells collected from large bovine follicles [31]. In that case, NFκB activity was shown to correspond to the amount of estradiol, where high estradiol was associated with higher amounts of NFκB activity and fewer apoptotic granulosa cells. Further linking sex hormones and ovarian NFκB signaling, repression of estrogen receptor beta gene transactivation by NFκB was reported using granulosa cell lines [32]. To date, NFκB signaling has mostly been evaluated in terms of its action regulating granulosa cell death (and thereby follicle atresia), and regulating gene expression in coordination with steroid hormones [32, 33, 31] in order to guide follicle development and survival [34].

In this study, we have evaluated the expression of key NFκB pathway members in the mouse ovary, with special emphasis on potential expression within primordial follicles. We show that key NFkB pathway proteins are expressed in primordial and also growing follicles, We have also evaluated the ovarian (and ovarian aging) phenotype of “A Knockout B Knock-In” (AKBI) mice that are resistant to pro-inflammatory stress induced NFκB activation [35, 36]. This included a comparison of the expression of Amh and Pten, between AKBI ovaries and wild-type controls. . Furthermore, we demonstrate that AKBI mice have a similar ovarian phenotype as reported for the Tnfr2 knockout mouse [24]. This manifests in an increased primordial follicle “reserve” at multiple postnatal time points versus WT controls. Overall, our data fill in gaps in our knowledge of signaling that occurs downstream of Tnfr2 activation that favor PFGA, and suggest that selective modulation of IκB activity may be therapeutically useful in the treatment of POI.

Materials and Methods

Results

Key members of the NFκB signaling cascade are expressed in the mature ovary

Our first objective was to evaluate NFκB pathway expression in adult (60–65 day old) WT ovaries by Western blot (Fig. 1) . We used liver tissue lysates as positive controls. We found that the ovary expressed the NFκB subunits cRel, p65 and p50 at levels similar to or even exceeding hepatic expression. Additionally, the we detected ovarian expression of the activating kinases IKKα and IKKβ, as well as the cytosolic inhibitory protein IκBα, IκBβ, and IκBε. Expression of IKKα was lower in ovarian compared to liver lysates, but longer exposure confirmed presence of IKKα in ovarian lysates (Fig. 1, inset). These results demonstrate that the key components of the NFκB signaling cascade are expressed in the ovary.

Our next objective was to localize NFκB pathway expression in adult (60–65 day old) WT ovaries by colorimetric immunohistochemical staining (IHC). Critically, the key members of NFκB signaling are expressed by primordial oocytes and pregranulosa cells within primordial follicles. (IκBα and IκBβ shown in Fig. 2A, B; p50, p65, shown in 2C, D; negative control where primary antibody was omitted shown in Fig. 2E). Because pregranulosa cells are so closely apposed to their enclosed primordial oocyte, we compared NFκB pathway staining to that of FoxO3, which is oocyte-specific in murine primordial follicles (inset, Fig. 2E) [39]. We found that the nucleo-cytoplasmic signal detected in pregranulosa cells for the NFκB pathway proteins (insets, Fig. 2A–D) was lacking in sections stained for FoxO3 (inset, Fig. 2E), supporting bona fide IκBα/β, p65 and p50 expression in those cells as well as primordial oocytes. These results demonstrate that the key components of the NFκB signaling cascade are expressed in follicles, including within arrested primordial follicles.

Discussion

Key findings about the role of NFκB signaling in early mouse follicle development are summarized as follows. The ovary expresses the key components of the NFκB signaling cascade. Inhibitory proteins IκBα and IκBβ, as well as NFκB subunits p50 and p65 are expressed in pregranulosa cells and primordial oocytes within primordial follicles. The levels of the NFκB inhibitory proteins IκBα and IκBβ are dynamic over life (Fig. 4), decreasing as the animal ages. These changes are blunted in the AKBI mouse ovary, and consistent with decreased NFκB activity we found that Tnfα levels are significantly lower in AKBI ovaries compared to WT controls (Fig. 5). A summary cartoon of these findings is in Fig. 9. In the ovary of AKBI mice, there is relative overexpression of IκBβ when compared to WT control (Fig. 3 and Fig. 9A). This is due to the IκBα gene being replaced by IκBβ cDNA in AKBI mice [37]. Our data suggest that as the WT animal ages, NFkB signaling becomes permissive as total ovarian levels of IκB decrease (Fig. 9B). This is consistent with increased TNFα and accelerated PFGA. In contrast, and due to blunted signaling resulting from IκBβ overexpression, the increase in TNFα in the AKBI ovary is attenuated, leading to delayed PFGA (Fig. 9B). Because NFκB signaling can both drive TNFα expression and can be activated by TNFα[27, 41, 42], these data are also consistent with the phenotype of the Tnfr2 knockout mouse [24]. Blinded histomorphometric analysis of juvenile, adult, and aged ovaries showed that PFGA is significantly slowed during the reproductive period, at least through 6 months of age (Fig. 7). Together, these data begin to fill in the gap in our understanding of regulation of the rate of PFGA downstream of Tnfα/Tnfr2.

Because Pten has been shown to be an important negative regulator of Akt/mTOR signaling and PFGA [18, 19], we interpret the nearly halved net downstream Akt activity in mutant ovaries (Fig. 6) as the key measure consistent with slowed PFGA. Interestingly, reduced Pten stability has been associated with increased Pten enzyme activity [43], and this may explain when Pten levels are lower in AKBI whole ovaries than WT controls (Fig. 6). Because AMH levels were unchanged in mutant ovaries (Fig. 6), we can attribute the slowed rate of PFGA to signaling independent of, or, downstream of that well-known negative regulator. More work is needed to determine how the involved signals are integrated so that PFGA can occur i) acutely in individual primordial follicles and ii) across the entire population of primordial follicles at a rate that supports ovarian function across a female’s reproductive lifespan. In addition, more work is needed to determine why the primordial pool of follicles “normalizes” relative to WT controls at the latest time point evaluated. While the PFGA rate is slowed during mid-life (at least through 6 months), a net gain of follicles later in life was not maintained. Altered NFκB signaling during aging is at least responsible for an altered pattern of ovarian aging, if not an increased pool of primordial follicles after 48 weeks.

While our data are in accordance with human GWAS (genome-wide association study) data that identified NFκB signaling as being involved in ovarian aging [26], and we have shed some light on the role of the IκB proteins in this process in our mouse model, there are some limitations of the work to consider. First, while NFκB subunits p50 and p65 were consistently noted in the nuclei of primordial follicles (both primordial oocytes and pregranulosa cells, Fig. 2), whether these factors engage in their expected action upon gene expression in primordial follicles remains unknown. Indeed, the AKBI ovarian phenotype with increased numbers of primordial follicles does not exactly phenocopy the Tnfr2 knockout [27], where increased numbers of primordial and small growing follicles were found during adulthood. This may be related to a second limitation of our study, that there are many physiological factors that can stimulate, or, block NFκB activation;[27] Tnfα/Tnfr2 interaction is only one example. The physiological factors that modulate NFκB action and also causally impact PFGA also remain unknown. Third, while the AKBI mouse has a reproducible and significant phenotype, it is unclear whether it is the loss of IκBα or the gain of the additional IκBβ gene (under the control of the IκBα locus), or both, that leads to altered ovarian aging. Last and perhaps most importantly, we did not identify the factor(s) responsible for initiating NFκB signaling. These experiments are in progress. Future experiments using whole body IκBβ knockout mice [37], and eventually, oocyte- or granulosa cell-specific IκB knockout models can be used to help resolve these questions. Despite these limitations, our data add NFκB machinery to the list of gene products that influence the rate of PFGA during reproductive life within the mouse model used. Experiments using human ovarian tissue are needed to determine whether the NFκB pathway functions similarly, and thus test whether these data are relevant to the reproductive aging of women.

PFGA involves the coordination between multiple identified pathways [reviewed in 2]. Our murine data suggest that one tractable intervention that might slow PFGA and extend ovarian function is the stabilization of IκBβ. Excitingly, the NFκB pathway and IκB proteins specifically are increasingly being shown to be clinically “druggable” targets [44, 45, 4], and the effect(s) of IκBβ stabilization can be tested using pharmacological agents in the aforementioned human in vitro studies. Increasing numbers of women are learning that they are at increased risk for the very common (above, Introduction) condition POI before their ovaries fail. This is due to their participating in assisted reproduction treatments where indirect measures of ovarian reserve are determined [2, 8]. There is therefore a window of treatment for these women where slowing the rate of PFGA could extend their ovarian function, staving off the mentioned significant health compromises.