There are a multitude of agents that activate the PPARs (Table 1). Many of these agents have well established roles in ovarian biology. For example, endogenous factors that have been shown to activate the PPARs that also impact ovarian function are fatty acids and prostaglandins, and exogenous activators include herbicides, industrial plasticizers, non-steroidal anti-inflammatory drugs (NSAIDs), fibrates (a class of drugs used to treat hyperlipidemia), thiazolidinediones (TZDs; hypoglycemia drugs), polycyclic aromatic hydrocarbons, organotin compounds, and traditional medicines 51415161718192021. An example of how these exogenous PPAR agonists impact the ovary is the inhibition of ovulation and 'reversible female infertility' caused by NSAIDs 22.

There is some specificity observed between ligands and the PPAR subtypes. For example, fibrates (i. e. WY-14,643, clofibrate) show a high affinity for PPARα, but at higher concentrations can also activate PPARγ 4. The thiazolidinediones (troglitazone, ciglitazone, pioglitazone, rosiglitazone) selectively activate PPARγ 423. Long chain fatty acids, particularly polyunsaturated fatty acids, preferentially activate PPARα 24, but are also capable of activating PPARδ and PPARγ 52325. Prostaglandins activate all PPAR family members, with PGA₁ and 15-deoxy-Δ^12,14-prostaglandin J₂ (PGJ₂) preferentially activating PPARδ and PPARγ, respectively 52526. Prostacyclin and its analogue, carbaprostacyclin, also binds to PPARδ (reviewed by 1627). Hydroxyeicosapentaenoic acids and leukotriene B4 are activators of PPARα 525. Interestingly, indomethacin and other NSAIDs that inhibit the production of prostaglandins, are also able to activate PPARα and PPARγ 28. Oxidized products of LDL (9-HODE and 13-HODE) are ligands for PPARγ (see 4).)29 for a review). Structural and amino acid differences in the binding pocket of the PPAR isoforms contribute to selectivity for ligand binding 30.

PPARs heterodimerize with 9, cis-retinoic acid receptors (RXRs) (Figure 2). PPAR interaction with RXRs can occur in the absence and/or presence of ligand 31. The heterodimer binds to a short sequence of DNA, a PPAR response element (PPRE), present in the promoter regions of target genes. The PPRE is a direct repeat of the sequence AGGTCA, separated by one nucleotide (a DR1 sequence; reviewed in 45). In addition to the PPRE, the 5' flanking region has been shown to be important for PPAR binding to DNA, especially PPARα binding. The binding affinity of the PPAR/RXR heterodimer is greatly enhanced if the nucleotide between the two hexamers is an adenine, and when there is an AA/TCT sequence 5' of the PPRE (reviewed in 4532). These DNA features result in a polarity to the bound heterodimer; PPAR binds to the upstream hexamer while RXR interacts with the lower, 3' hexamer 532. The integrity of the 5' sequence offers selectivity in binding for the PPAR isotypes.

Similar to other steroid hormone receptors, there are coactivators and corepressors that associate with the PPARs. Corepressors, such as nuclear receptor corepressor (NCoR) and silencing mediator for retinoid- and thyroid-hormone receptors (SMRT), dissociate from the receptor upon ligand binding (reported in 433). The conformational change that occurs upon ligand binding also facilitates the recruitment of coactivators. Two coactivators that have histone acetyltransferase activity, steroid receptor coactivator-1 (SRC-1) and CREB binding protein/p300 (CBP), can bind to PPARs in a ligand-dependent manner. The latter coactivators can also interact with PPARs in a ligand-independent manner, but only transiently (reviewed in 4). RIP140, ARA70, and members of the DRIP/TRAP family of coactivators also bind to PPARs (see 34 for a review). Other coactivators that have been identified to interact with PPARs are: PPAR interacting protein 33, PPARγ coactivator-1 (reviewed in 35), and PPAR binding protein (PBP; 36). Although these coactivators also bind other steroid receptors, deletion of the PBP gene in mice results in embryonic lethality due to placental insufficiency 37, the same results seen in PPARγ null mutants 38. These findings are consistent with the hypothesis that PBP is a required factor for PPARγ transcriptional activity. The regulated expression of these various corepressors and coactivators and their concentrations in tissues also offers selectivity in transcriptional regulation by the PPAR isotypes.

A recent intriguing finding is that the association of corepressors with PPARδ can inhibit the activity of PPARs α and γ. Shi et al. (2002) demonstrated that PPARδ repressed PPARα and γ-mediated gene transcription. This repressive activity of PPARδ involved DNA binding and association with the corepressor SMRT 39. The authors of this study concluded that the levels of each PPAR isotype, as well as the ratio of PPARs α and γ to PPARδ in a particular tissue influences the activity of each isotype.

The activity of PPARs are modified not only by ligand binding, but also by phosphorylation, nitration, and ubiquitination. Phosphorylation sites have been identified on both PPARs α and γ. The impact of phosphorylation on the activity of PPARs depends on: 1) the residue being phosphorylated, and 2) the kinase cascade that was activated (reviewed in 40). A modification of PPARγ that influences its activity is nitration of tyrosine residues. Shibuya et al. (2002) demonstrated that nitration of tyrosine residues in PPARγ inhibited the translocation of PPARγ from the cytosol to the nucleus 41, thus reducing its potential to influence gene transcription. PPARs can also be ubiquitinated. Ligand binding to PPARγ induces ubiquitination of the receptor 42, and therefore its degradation. In contrast, ligand binding to PPARα stabilizes the receptor by decreasing its rate of ubiquitination 4344.

All three PPAR subtypes have been detected in ovarian tissue. In the rat ovary, the expression of mRNA for PPARα is found primarily in the theca and stroma, whereas mRNA for PPARδ is found throughout the ovary (Figure 3). The expression of these two PPAR isotypes remains steady throughout follicular development and the ovarian cycle in the rat 6162.

PPARγ has been more extensively studied in ovarian tissue than the other two family members. It has been detected in the mouse 63, rat 4962, pig 64, sheep 65, cow 6667, and human 55 ovary. Using RT-PCR, PPARγ was detected in granulosa cells collected during oocyte aspiration from women undergoing treatment for in vitro fertilization 68, and in porcine theca and granulosa cells 64. This PPAR isotype has also been reported to be in oocytes from cattle 69, zebrafish 3, and Xenopus laevis (trace amounts; 70). In cycling rats and sheep, the expression of PPARγ is restricted primarily to granulosa cells in developing follicles 616265. However, unlike the steady expression of PPARs α and δ, the expression of PPARγ is down-regulated in response to the LH surge (Figure 4) 6265. The expression of PPARγ decreases only in follicles that have responded to the LH surge 71. In the rat, expression of PPARgamma is low in newly forming luteal tissue, and higher in luteal tissue present from previous ovulations 61. Because PPARγ is primarily expressed in granulosa cells, it may influence development of these cells and their ability to support normal oocyte maturation. PPARγ could also potentially affect somatic cell/oocyte communication not only by impacting granulosa cell develpment, but by direct effects on the oocyte. Disrupting the expression of PPARγ in the ovary therefore, could potentially affect oocyte developmental competence.

Results from a study by Cui et al. (2002) indicate that PPARγ plays an important role in normal ovarian function. Using cre/loxP technology, the expression of PPARγ was disrupted in the ovary, rendering 1/3 of the females sterile, and the remaining females sub-fertile 63. Females that were sub-fertile took longer to conceive and had smaller litters. There were no differences found in the number of primordial, primary, or preantral/antral follicles, size of copora lutea, or response to exogenous gonadotropins between control animals and those with PPARγ disrupted in the ovary. On the day of estrus, levels of progesterone in animals with PPARγ disrupted in the ovary were half that found in controls. However, the differences in circulating progesterone were not significantly different between the two groups, most likely due to the small sample size (n = 4/group). Implantation sites (6) were only observed in the uterus of one of three females examined with PPARγ disrupted in the ovary, compared with 5 and 7 implantation sites observed in two control females, respectively. Because the expression of PPARγ was not disrupted in the uterus of these transgenic females, the lesion responsible for the sub- and infertility most likely lies within the ovary. The authors concluded that "...ovarian function might not be sufficient to induce implantation" 63. The insufficient ovarian function may relate to the ability of the corpus luteum to produce enough progesterone, or produce enough progesterone in a timely manner, to support the establishment of pregnancy. In addition, estradiol production by the ovary around day 4 post-coitum is also an important player in preparing the uterus for implantation. Impaired production of estradiol by ovarian cells in the transgenic females during this critical period may also lead to reduced implantation. Although not tested in this study, the competence of the oocyte to undergo fertilization and support embryonic development might also be compromised in these genetically altered mice. Further study into the role of PPARγ in ovarian steroidogenesis and somatic cell/oocyte interactions is needed to determine the cause of the fertility problems in females with reduced ovarian PPARγ expression.

Additional studies have shown that endogenous PPARγ is active in the ovary. Granulosa cells from rats and sheep were transiently transfected with reporter constructs whose expression was driven by PPREs. Both in the absence and presence of agonists for PPARγ, there was an increase in reporter activity 6572. PPARγ in rat granulosa cells was also shown to bind DNA 73. These findings demonstrate that PPARγ is functional in granulosa cells, and that endogenous ligand is also present within these cells.

One way PPARs may influence ovarian function is by modifying the ability of estradiol to elicit cellular responses. PPARs are able to bind to estrogen response elements – EREs, 7475, and can act as competitive inhibitors 74. PPARγ can also stimulate ubiquitination of estrogen receptor α, leading to its degradation 76.

The synthesis and metabolism of estradiol is also affected by the PPARs. PPARγ can inhibit the expression of aromatase, the rate limiting enzyme for the conversion of androgens to estradiol by disrupting the interaction of NF-κB with the aromatase promoter II 77. Activation of PPARα decreased the expression and activity of aromatase in granulosa cells 7879. In cultured human granulosa-luteal cells 68, and granulosa cells from eCG-primed immature rats 78, activation of PPARγ reduced the expression of aromatase. PPARγ was also shown to partially mediate the suppressive effects of phthalates on ovarian estradiol production 78. However, using a different strain of rat and culture model, agonists of PPARγ were shown to increase estradiol secretion by granulosa cells collected from gonadotropin-primed immature rats 62. Reduced levels of aromatase in granulosa cells after activation of PPARγ was also reportedly due to increased turnover in conjunction with decreased transcription 80. We reported previously that there was no correlation between the expression of mRNAs for PPARγ and aromatase in granulosa cells during folliculogenesis or the periovulatory period 71. PPARs may also limit the synthesis of estradiol by reducing production of androgenic precursors by theca cells. PPARγ is expressed in the theca 6164, primarily in the theca externa and in an inconsistent pattern 61. Both endogenous (PGJ₂) and exogenous (troglitazone) agonists of PPARγ reduced basal and LH-stimulated thecal androgen production in vitro 6481. One study reported that troglitazone increased mRNA for CYP17, but not the corresponding protein 64, whereas a second study showed no effect of the PPARγ agonists on mRNA for CYP17, but a decrease in its phosphorylation 81. In both granulosa 78 and liver cells 82, agonists of PPARα stimulated the expression of 17β-hydroxysteroid dehydrogenase type IV, an enzyme that oxidizes estradiol into the less active estrone. The expression of PPARα in granulosa cells is very low 6162 and therefore may be unlikely to modify estradiol metabolism under normal physiological conditions. Taken together, these data indicate that PPARs are able to influence estradiol production, and that age and the endocrine environment may influence how these transcription factors impact ovarian steroidogenesis.

The activation of PPARγ can also influence progesterone production by ovarian cells. In cultured human granulosa cells, activators of PPARγ inhibited basal and gonadotropin-stimulated progesterone production 83. However, activators of PPARγ stimulated progesterone secretion by granulosa cells obtained from eCG-primed immature rats 62. When porcine theca cells were treated with synthetic and natural ligands for PPARγ, progesterone production increased 64. Progesterone production by bovine luteal cells treated with the endogenous ligand for PPARγ, PGJ₂, increased progesterone production over a 24 hour culture period 67. Our previous work has shown that there is an inverse relationship between the expression of mRNA for PPARγ and P450 side chain cleaveage, the rate limiting enzyme in progesterone synthesis, in granulosa cells and luteal tissue from naturally cycling and gonadotropin-treated rats 7184. Therefore, the effect of PPARγ on progesterone production may depend on the cell type, stage of differentiation, stage of the cycle, and/or the species studied.

PPARs regulate the expression and activity of proteases involved in tissue remodeling and angiogenesis which are critical processes for follicular and luteal development. Plasminogen activators (PA) and matrix metalloproteinases (MMPs) are proteolytic enzymes involved in ovarian tissue remodeling and angiogenesis 858687. Activation of PPARα and PPARγ decreases MMP-9 expression and its activity 88899091. The promoters for MMP-3 92 and MMP-9 93 contain a PPRE, indicating that transcription of these proteases is likely directly regulated by PPARs. PPARγ activation can also reduce expression of MMP-13 and MMP-1 by interfering with AP-1 activation 949596. PPARγ negatively affects plasminogen activator by inhibiting its expression 97 and increasing the expression of plasminogen activator inhibitor-1 9798. However, there are also reports of troglitazone treatment reducing the expression of plasminogen activator inhibitor-1 99100. These findings indicate that the PPARs are capable of modulating the balance of proteolytic enzymes and their inhibitors, thereby altering tissue remodeling events. Whether PPARs regulate these processes in ovarian cells, particularly at the time of ovulation when MMP and PA activities must be tightly regulated, is an important area of investigation.

Along with the proteases, vascular endothelial growth factor (VEGF) and its receptors (Flt-1, -2) are important players in new blood vessel formation in the ovary 101102. The activation of PPARγ with PGJ₂ inhibited the expression of Flt-1 and Flt-2 in human umbilical vein endothelial cells 97. Activation of PPARγ with its endogenous and exogenous ligands has also been shown to inhibit VEGF-stimulated endothelial cell proliferation and migration (reviewed in 103). However, Yamakawa et al. (2000) reported that activating PPARγ in vascular smooth muscle cells results in an increase in the expression of VEGF 104. Therefore, the effect of PPARγ on angiogenesis may depend on agonist used, experimental model, and/or cellular differences in cofactor availability 103. Besides its effects on angiogenesis, PPARγ may influence the ovarian vasculature by its ability to regulate endothelin-1 (ET-1) and nitric oxide synthase (NOS). ET-1 is a potent vasoconstrictor and recent studies have shown that it is also an important player in ovarian physiology, especially luteal function (reviewed in 105). NOS synthesizes nitric oxide, a vasodilator, from arginine. Nitric oxide has been implicated as a player in luteolysis 106, ovarian cyclicity 107, ovulation 107108109, oocyte maturation 108, and follicular development 110111. PPARγ decreases the secretion of ET-1 from endothelial cells 112, and also inhibits the expression of NOS in macrophages 90 and vascular smooth muscle cells 113.

The ability of PPARs to affect tissue remodeling could alter folliculogenesis and luteal development, and impact ovulation. Ovarian tissue is constantly changing to accommodate the dynamic geometry of growing follicles which increase in size exponentially from the primordial to preovulatory stage. For successful release of the oocyte at ovulation, the granulosa cell layer, follicular basement membrane, theca interna and externa, ovarian stroma, tunica albuginea, and surface epithelium need to be traversed. In addition, the tissue remodeling involved in developing the increased vasculature required to support follicular development and luteal formation requires protease activity. The ability of PPARs to regulate the expression of proteases and angiogenic factors, and the fact that they are expressed in the ovary and in the case of PPARγ, modulated during the periovulatory period encompassing ovulation and luteal formation, warrant further study into how the PPARs may influence these aspects of ovarian biology.

PPARs are important mediators of inflammatory responses (reviewed in 27114115116). The process of ovulation has been likened to an inflammatory response 117 and prostaglandins, major regulators of inflammation, have well documented roles in ovulation as well as luteal function (see 118 for a review). The rate-limiting enzyme in prostaglandin production is cyclooxygenase-2 (COX-2). The promoter region of COX-2 contains a response element for the PPARs 119, indicating that PPARs can directly influence transcription of this gene. However, there are reports of PPARγ both stimulating 119 and inhibiting 120121 the expression of COX-2. In rat granulosa cells, the expression of COX-2 is stimulated within 4 hours of the ovulatory gonadotropin surge 122, however, PPARγ is significantly reduced in this same time frame 62. This inverse relationship between the expression of PPARγ and COX-2 has also been observed in the placenta 123. The variability in reported effects of PPARγ on COX-2 expression could result from: 1) the use of different cell-types, 2) transfection with COX-2 promoter constructs that did 119 or did not 124 contain the PPRE, 3) the ability of PPARs to influence COX-2 expression by binding to its promoter region, and/or 4) by PPARγ interfering with activation of NF-κB 121. The periovulatory expression pattern of PPARγ suggests it plays an inhibitory role in COX-2 expression in ovarian cells in vivo.

Not only can PPARs regulate COX-2 expression, but as discussed earlier, prostaglandins themselves are endogenous ligands that can activate PPARs. In addition, PGF_2α can activate kinase cascades resulting in the phosphorylation of PPARγ and inhibiting its activity 125. Cumulatively, these findings imply that there is a cyclic relationship between the presence of prostaglandins, activation and/or inhibition of PPARs and feedback to the prostaglandin synthesizing enzyme – COX-2.