In the mammalian ovary, follicular development, ovulation, and corpus luteum (CL) functions are governed by a multitude of signals including gonadotropins, steroid hormones, growth factors, prostaglandins, cannabinoids, and other lipids and cytokines. These signals generate different types of intracellular signals, which have to be integrated through cross-talking networks. One of the important intracellular messengers generated within cells is cyclic adenosine 3′,5′-monophosphate (cAMP), which plays important roles in ovarian physiology (Richards, 2001). The complex mechanisms by which cells regulate the concentration of cAMP by integrating the signals transduced by a diversity of receptors and relay the information downstream are only now beginning to be understood. Accumulating evidence suggests that in addition to the canonical signaling from the plasma membrane, G protein-coupled receptors present within the nuclei and other intracellular sites are capable of generating cAMP (Gobeil et al., 2006; Zhu et al., 2006; Padmanabhan et al., 2009). Moreover, distinct localization of protein kinase A (PKA) isoforms, A kinase anchoring proteins, and phosphodiesterases appear not only to create cAMP gradients and compartments in cells but also generate unique downstream signaling (Baillie et al., 2005; Di Benedetto et al., 2008), including the oocyte (Burton and McKnight, 2007). In addition, the PKA holoenzyme is also localized to the nucleoplasm, including the nucleolus (Zippin et al., 2004). The major determinant of cellular cAMP concentration is the activity of the enzyme adenylyl cyclase (AC), which catalyzes the formation of cAMP from ATP. To date, nine transmembrane ACs (and several splice variants) and a single soluble AC have been discovered (Buck et al., 1999; Sunahara et al., 2002). All nine isozymes are activated by Gsα and inhibited by adenosine. However, their basal activities and their sensitivity to Giα, protein kinases A and C, calcium-calmodulin kinases, calcium and so forth vary widely (Defer et al., 2000; Kuznetsova, 2002 and references therein).
Existence of multiple isoforms of ACs has been previously shown in human granulosa cells (Asbóth et al., 2001), CL of the cow (Mamluk, et al., 1999), and cumulus cells (Lastro et al., 2006). Transcripts for various ACs in the oocytes and cumulus cells and the expression of ACIII protein have been detected in Bouin's-fixed and paraffin-embedded rodent ovaries (Horner et al., 2003). However, to our knowledge, the in situ distribution of multiple ACs has not been reported in the rat ovary. Therefore, the objective of this study was to examine the distribution of multiple membrane-bound ACs by indirect immunofluorescence microscopy in frozen sections of both pregonadotropin and gonadotropin-primed rat ovaries. It has been suggested that ACs, acting as coincidence detectors, are able to integrate incoming signals within cells (Anholt, 1994). Thus, identifying the distribution of various ACs in the ovary in time and space is a necessary first step before future physiological studies can be carried out on the integrated roles of these enzymes in ovarian physiology.
MATERIALS AND METHODS
All of the experiments described in this study conform to the guide for the care and use of laboratory animals, published by the National Research Council (Publications No. 0-309-05377-3, 1996) and were approved by the Kent State University Animal Care and Use Committee.
Induction of Follicular Development and CL Formation
Twenty-two to twenty-five day old immature female rats were injected subcutaneously with 15 international units (IU) of equine chorionic gonadotropin (eCG; Sigma-Aldrich, St. Louis, MO) in 100 μL of phosphate buffered saline (PBS) containing 0.2% bovine serum albumin (BSA). Control rats were injected with vehicle alone. Two days later, the rats were injected with 5 IU of human chorionic gonadotropin (hCG) (Sigma-Aldrich, St. Louis, MO) in 100 μL of the same buffer. The animals were sacrificed 1–2 days after eCG injection and 4–5 days after hCG injection. The eCG causes the synchronized development of preovulatory follicles in the immature ovary. Subsequently, hCG induces the ovulation of these follicles and the formation of CL. The ovaries were removed and processed for immunofluorescence microscopy as described below. Observations were made from ovaries obtained from at least three animals for each treatment at each time point.
Tissue Processing and Immunofluorescence Microscopy
The ovaries were frozen fresh in O.C.T. compound (Lab-Tek Products, Naperville, IL) in ethanol-dry ice and kept at –70°C until use. Subsequently, 8-μm thick sections were cut in a cryostat set at −15°C. Immunostaining was performed as before with minor modifications (Bagavandoss and Grimshaw, 2010). Briefly, the sections were fixed in cold acetone for 10 min and air-dried. Subsequently, the sections were washed in PBS containing 0.2% BSA and 0.05% Tween-20 (PBST) and incubated with PBST containing 10% BSA or 4% donkey serum in PBST for 30 min at 25°C. After removing the buffer, the sections were incubated for 2 hr at 25°C or overnight at 4°C with affinity-purified rabbit polyclonal antibodies specific for AC isoforms ACI (sc-586), ACII (sc-587), ACIII (sc-588), ACIV (sc-589), and ACV/VI (sc-590) and affinity-purified goat antibodies specific for AC7 (sc-1966), AC8 (sc-1967), and ACIX (sc-8578). All antibodies were used at a dilution of 1:100 (1 μg/mL). These antibodies (manufactured by Santa Cruz Biotechnology, in Santa Cruz, CA) were generously donated to us by Dr. William Baldridge of Dalhousie University, Halifax, Canada. After three 5-minute washes in PBST, the sections were incubated with either 1 μg/mL of Alexa Fluor 488-conjugated goat anti-rabbit antibody (Molecular Probes, Eugene, OR) or 1:100-diluted fluorescein isothiocyanate-conjugated donkey anti-goat antibody (Santa Cruz Biotechnology, CA) for 45 min at 25°C and washed and mounted in Vectashield® mounting medium (Vector laboratories, Burlingame, CA). Control incubations were performed with antibodies that had been preincubated with excess of (2 μg/mL) respective blocking peptides (Santa Cruz Biotechnology, CA). The sections were viewed and photographed using an Olympus IX inverted microscope equipped with Olympus FluoView, 3-laser confocal microscopy system. The intensity settings were kept constant between all sections and also between sections treated with and without blocking peptides. Using National Institutes of Health's ImageJ software, the color micrographs were converted to 8-bit gray scale images and the regions of interest were measured from five to eight regions of the sections with and without blocking peptides (background). Background fluorescence for each measurement was corrected as follows: Corrected fluorescence = Integrated Density–(area of measurement X mean fluorescence of background measurements).
For single comparison (± blocking peptide; pregonadotrophin and postgonadotropin), corrected fluorescence data were analyzed by two-tailed Student's t-test and the values were considered significant at a P value of <0.05 (indicated by asterisks). Data are expressed as mean relative fluorescence ± SEM (n = 3 animals). We do, however, recognize that the demonstration of statistical significance or lack thereof, based on the analysis of descriptive data is not necessarily a demonstration of functional biological significance.
Expression of ACI
In the immature rat ovaries, before eCG administration (pregonadotropin ovaries), ACI was distinctly localized to the nucleus of the oocytes within the preantral and antral follicles along with weak immunoreactivity in granulosa and thecal cells (Fig. 1A,B). Immunoreactivity was also detectable in the cells scattered throughout the ovarian stroma (Fig. 1A,B). The tissue immunoreactivity was completely abolished when the antibody was preadsorbed with ACI blocking peptide (Fig. 1C). Subsequent to eCG administration, the nuclei of the oocyte in the developing preantral and antral follicles continued to show bright ACI immunoreactivity (Fig. 1E,F). Whereas the granulosa cells showed weak staining before eCG administration, they exhibited strong immunoreactivity 48 hr post-eCG treatment. The cumulus cells also reacted with ACI antibody (Fig. 1F). However, after eCG treatment, no ACI staining was observed in the thecal layer of follicles. Within the arterioles, ACI staining was restricted to the endothelium (Fig. 1D). As the follicles began to transform into CL in response to hCG, a peripheral ring of staining in each cell, which we have designated “plasma membrane-type staining,” was observed in the luteal cells (Fig. 1G). ImageJ analysis of ACI (Fig. 2) indicated the existence of significant staining in the oocyte of pregonadotropin ovaries and in the oocyte, granulosa cells, and CL of the postgonadotropin ovaries. In addition, the staining in granulosa cells significantly increased subsequent to eCG treatment. No significant staining was present in the theca.
Expression of ACII
In the pregonadotropin ovaries, variable and diffuse ACII staining was observed in the oocyte, with no evidence of nuclear staining (Fig. 3A,C). A weak and diffuse membrane-type ACII staining was present in the granulosa, thecal, and stromal cells of the follicles along with staining in the endothelial cells of the arterioles (Fig. 3A,C). In addition, bright immunoreactivity for ACII was observed in the ovarian surface epithelium (OSE), which can be blocked with blocking peptide (Fig. 3C,D). Therefore, the observed staining of the OSE, which defines the border of the ovary, is not simply due to edge effect. Subsequent to eCG administration, no marked changes in oocyte staining were observed, while the granulosa and the cumulus cells showed increase in the intensity of apparent plasma membrane-type staining (Fig. 3E). No changes were observed in the theca or stroma (not shown). Subsequent to hCG injection, luteal cells of the CL showed a weak but specific non-nuclear immunoreactivity for ACII (Fig. 3G). The tissue immunoreactivity was completely abolished when the antibody was preadsorbed with ACII blocking peptide (Fig. 3B,D,F). ImageJ analysis of ACII (Fig. 4) indicated the existence of significant staining in the oocyte, granulosa cells, and theca of pregonadotropin ovaries and in the oocyte, granulosa cells, theca, and CL of the postgonadotropin ovaries. However, in response to eCG treatment, a significant increase in granulosa cell staining and a significant decrease in thecal staining were observed in the postgonadotropin ovaries.
Expression of ACIII
In the pregonadotropin ovaries, only faint ACIII staining was present in the oocytes and granulosa cells (Fig. 5A). Bright ACIII staining, however, was observed in the theca and the interstitium (Fig. 5A). Subsequent to eCG administration, ACIII immunoreactivity was observed in the oocyte (Fig. 5B, inset). Although no change in staining was observed in the granulosa cells, ACIII showed predominant localization in the thecal and interstitial vessels (Fig. 5B,C). Subsequent to hCG administration, immunoreactivity for ACIII was observed in the CL (Fig. 5D,E). In the arterioles, ACIII was localized to the smooth muscle cells (Fig. 5D). The tissue immunoreactivity was markedly reduced when the antibody was pre-adsorbed with ACIII blocking peptide (Fig. 5F). ImageJ analysis of ACIII (Fig. 6) indicated the existence of significant staining in the theca of pregonadotropin ovaries and in the theca and CL of the postgonadotropin ovaries. Further, in both the pregonadotropin and postgonadotropin ovaries, no significant ACIII staining was observed in granulosa cells and oocyte (Fig. 6).
Expression of ACIV
In the pregonadotropin ovaries, ACIV staining was present in the granulosa, thecal, and stromal cells, with predominant localization in their nuclei (Fig. 7A). ACIV was also present in the OSE (Fig. 7A). The tissue immunoreactivity was completely abolished when the antibody was preadsorbed with ACIV blocking peptide (Fig. 7B). Subsequent to eCG administration, immunoreactivity in the nucleus was observed in the oocyte, granulosa, thecal and stromal cells, (Fig. 7C–E) and OSE (not shown). Subsequent to hCG administration, luteal cells of the CL showed plasma membrane-type staining (Fig. 7F,G). The ACIV staining in the OSE was still present in the luteal phase. ImageJ analysis of ACIV (Fig. 8) indicated the existence of significant staining in the granulosa cells of pregonadotropin ovaries and in the oocyte, granulosa, theca, and CL of postgonadotropin ovaries. In addition, following eCG treatment, a significant increase in ACIV staining was observed in the oocyte and theca. However, following eCG treatment, a significant decrease in granulosa cell staining was observed (Fig. 8).
Expression of ACVIII
In the pregonadotropin ovaries, ACVIII showed weak staining in the oocyte and granulosa cells with little staining in the oocytes, theca, or elsewhere within the ovary (Fig. 9A). Subsequent to eCG administration, granulosa cell staining became more intense, without any associated increase in thecal or interstitial staining (Fig. 9C). Subsequent to hCG injection, luteal cells exhibited bright staining for ACVIII, whereas the staining in granulosa cells of the follicles decreased (Fig. 9D). The tissue immunoreactivity was completely abolished when the antibody was preadsorbed with ACVIII blocking peptide in the untreated ovaries (Fig. 9B), although some background staining persisted in the eCG plus hCG-injected ovaries (Fig. 9E). ImageJ analysis of ACVIII indicated no significant staining in the pregonadotropin ovaries, whereas a significant staining was observed in the granulosa cells and CL of postgonadotropin ovaries. In the postgonadotropin ovaries, a significant increase in granulosa cell staining in response to eCG was observed (Fig. 10).
Using an immature rat model for the induction of ovarian development, we have studied the distribution of membrane-bound ACs with affinity-purified antibodies against I, II, III, IV, and VIII. We carried out the localization in the presence or absence of blocking peptides for the respective ACs. We do recognize that, although powerful, immunocytochemistry protocols can have limitations even if specificity is demonstrated in the presence of blocking peptides. For example, one cannot completely rule out the possibility that the antibodies could still be potentially recognizing unrelated proteins even in the presence of blocking peptides (Burry, 2011). We were unable to get a good signal-to-noise ratio with antibodies to ACV/V1, ACVII, and ACIX. Therefore, further investigations were not conducted with these antibodies. In this animal model, before gonadotropin administration, the ovaries primarily contain small preantral follicles with oocytes, granulosa, and thecal cells. Administration of eCG induces the development of large antral follicles within 2 days. Subsequently, hCG administration induces the resumption and completion of meiosis I in oocytes, followed by ovulation, and CL formation. By the fourth day post-hCG, CL begins to secrete increasing amounts of progesterone (Horikoshi and Wiest, 1971). Therefore, this animal model is suitable for exploring the temporal changes in the expression of biomolecules within the ovary during defined stages of development. Using this model, we show a unique and differential distribution of the above ACs in the developing rat ovary in time and space. The results of our findings are summarized in Table 1.
Table 1. Distribution of adenylyl cyclases in the rat ovary
Before and after eCG injection, ACI and ACIV were preferentially localized to the nucleus of the oocyte in association with the prominent nucleolus, a characteristic of the oocytes (Albertini, 1984; Zucker et al., 2000). The development of preantral follicles takes place in the absence of gonadotropins (Robker and Richards, 1998 and references therein). Therefore, the presence of these cyclases in the immature ovary suggests that the expression of these enzymes is governed by nongonadotropins. Nevertheless, as this study was conducted in intact animals (non-hypophysectomized), exposure of the preantral follicles to endogenous gonadotropins cannot be completely ruled out. The presence of ACs within the oocyte suggests a potential for the local generation of cAMP, which has been known to be the mediator of maintenance of meiotic maturation (Cho et al., 1974; Conti et al., 2002a; Mehlmann, 2005). It has been previously demonstrated that oocytes obtained from eCG-injected rat ovaries not only contain predominantly ACIII but also its deficiency in ACIII-null mice leads to the resumption of meiosis in more than 50% of the oocytes (Horner et al., 2003). However, in our study, before eCG injection, we were unable to observe any ACIII in the oocytes of the immature rat ovaries and only weak staining was observed after eCG injection. Further, unlike Horner et al. (2003), who were unable to find any ACI or ACIV transcripts in the rat ovary, we found immunoreactivity for these enzymes. The reason for this discrepancy is not fully clear. However, a recent study examining the relationship between RNA expression and immunolocalization reveals that 13% of the immunolocalized proteins could not be detected at the transcript level (Klevebring et al., 2010). Two possibilities have been suggested: either these genes have a very low expression but are efficiently translated into stable products, or the antibodies are cross-reactive providing a false-positive result (Klevebring et al., 2010). Additional studies are needed to determine these possibilities. If ACI and ACIV are indeed functional in the oocyte of immature rat ovary at a time when meiotic competence begins (Tsafriri and Pomerantz, 1984), these cyclases could play a potential role in the maintenance of meiotic arrest in preantral follicles.
It is now known that constitutively active G protein-coupled receptors GPR3 in the mouse (Mehlmann et al., 2004; Freudzon et al., 2005; Hinckley et al., 2005; Ledent et al., 2005; Vaccari et al., 2008) and human oocytes (DiLuigi et al., 2008) and GPR12 in the rat oocytes (Hinckley et al., 2005) maintain meiotic arrest by generating cAMP. However, in Gpr3−/− cycling mice, only about one-third of the oocytes in antral follicles undergo meiotic resumption. Thus, the remaining two-thirds of the oocytes are maintained in meiotic arrest by unknown mechanisms (Ledent et al., 2005). In both mouse (Mehlmann et al., 2002; Freudzon et al., 2005) and human (DiLuigi et al., 2008), cAMP in oocytes is generated through Gs G protein. Although these receptors are present at both the plasma and perinuclear membranes in cultured neurons (Tanaka et al., 2007), in the oocyte, GPR3 is restricted to the plasma membrane and Gs is present both in the cytoplasm and the plasma membrane but not in the nucleus (Freudzon et al., 2005). If this is also the case in the rat oocytes, how could nuclear ACs generate cAMP in the absence of Gs G protein? Of the two ACs we find in the rat oocyte's nucleus, ACI is directly activated by calcium and calmodulin to generate cAMP even in the absence of Gs (Tang et al., 1991). Stimulation of ACI even in the presence of Gs also requires calcium-calmodulin kinases (Wayman et al., 1994). As both calmodulin (Liao et al., 1999) and calcium (Shirakawa and Miyazaki, 1996) diffuse into the nucleoplasm, one can hypothesize that cAMP generated at the oocyte plasma membrane can potentially generate calcium oscillations in the oocyte leading to the activation of ACI in the nucleus. Moreover, it is becoming increasingly clear that cAMP generated in plasma membrane microdomains adjacent to ACs is spatiotemporally restricted due to diffusion limitations and is unlikely to result in elevation of global cytosolic cAMP (Rich et al., 2000, 2001; Zaccolo and Pozzan, 2002; Karpen and Rich, 2004; Willoughby and Cooper, 2007). However, it is not inconceivable that GPR12-mediated increase in cAMP and Ca2+ (Uhlenbrock et al., 2002) at the oocyte plasma membrane could potentially activate the nuclear ACI through diffusible calcium-calmodulin kinase pathway leading to the generation of a high concentration of cAMP in the nucleoplasm. Further, functional ACs have been previously demonstrated in other cellular systems, including the rat and mouse ventricular cardiac myocytes (Boivin et al., 2006).
Although ACI was primarily present in the oocytes of preantral follicles, subsequent to eCG injection, granulosa cells also expressed this enzyme, which suggests that ACI expression in these cells is regulated by gonadotropins. No ACI expression was observed in thecal cells. The observed expression of ACI, ACIV, and ACVIII in this study is consistent with that seen in cultured human granulosa cells obtained from antral follicles (Asbóth et al., 2001). However, unlike in the cultured human granulosa cells (Asbóth et al., 2001), where no staining for ACII was observed, we observed specific ACII staining in granulosa cells, which appear to increase in intensity in response to eCG administration. As the stimulation of ACII is primarily regulated by the simultaneous presence of both Gsα and Gβγ subunits (Gao and Gilman, 1991), G protein-coupled receptor agonists might play an important role in the regulation of this enzyme in granulosa and luteal cells. In contrast to observations in human granulosa cells, we did not see a prominent staining for ACIII in the rat granulosa compartment; rather a weak staining was observed. The affinity-purified antibodies used here were also used in the human granulosa cell studies although they utilized the biotin-streptavidin alkaline phosphatase Fast Red substrate protocol. Perhaps, the observed variations in ACII and ACIII expression between human and rat could be due to differences between species, differences in methodology, or both. One other possibility is that during cell culture for 48 hr (Asbóth et al., 2001), the expression of these enzymes could have been affected. ACIII is primarily associated with the thecal cells and in the smooth muscle cells of the blood vessels. Similarly, utilizing the affinity-purified antibody to ACIII from the same source as ours, staining in the blood vessels was previously reported in the rat ovary (Horner et al., 2003). Smooth muscle cell contraction and relaxation are regulated by intracellular calcium and protein kinase C (Clinton, 2003 and references therein), which regulate the activity of ACIII (Defer et al., 2000 and references therein). Therefore, the presence of ACIII in the smooth muscle cells of the vasculature suggests that this enzyme could play a potential role in regulating blood flow to the follicle and CL as cAMP induces relaxation of vascular smooth muscle cell (Lincoln and Cornwell, 1991). We also observed ACIII in the luteal cells, which is consistent with a previous study showing its presence in both rat and bovine luteal cells in which it is regulated by calcium ions (Mamluk et al., 1999). We also observed ACs I and II staining in the arterioles, where the staining is restricted to the endothelial cells. Smooth muscle cell relaxation of blood vessels often requires nitric oxide (NO)-derived from endothelial cells, whose synthesis requires the rise in endothelial cell calcium. Therefore, the presence of calcium-calmodulin activated ACI and protein kinase C activated ACII (Halls and Cooper, 2011 and references therein) suggests a potential role for these enzymes in the regulation of blood flow to the ovary. The presence of calcium-calmodulin activated ACI and ACVIII in granulosa cells of antral follicles, luteal cells, and OSE suggests potential regulation of these enzymes by gonadotropins. For example, luteinizing hormone not only activates the cAMP-PKA pathway but also increases the intracellular calcium in granulosa cells (Conti, 2002b and references therein), which could subsequently regulate the ACs I and VIII.
In summary, our investigation illustrates the presence of redundancy in the distribution of ACs within the oocyte, granulosa, and luteal cells as well as the OSE. The results of our study also raise the potential for the generation of cAMP within the nuclear compartment. Finally, our study reveals the potential for the regulation of arteriolar smooth muscle cell function by ACIII, endothelial cell function by ACs I and II, and OSE function by ACs II and IV.
The authors thank Dr. William Baldridge of Dalhousie University, Halifax, Canada for kindly providing us with the AC antibodies. They also thank Drs. Bruot, Kline, and Vijayaraghavan for sharing their laboratory facility to perform this research.