It Takes Two to Tango but Four for the Finale

In many nonprimate species, regression of the corpus luteum (CL) is initiated by the uterine release of prostaglandin F2α (PGF2α). In fact, this knowledge has been exploited for decades in development of estrous synchronization protocols. When a fully functional CL is present, a single, large-dose injection of PGF2α will result in luteolysis followed by the onset of estrus and ovulation, allowing a renewed opportunity for breeding. This procedure has been extensively utilized to study the events involved in the regression of the CL. At the time of natural luteolysis, PGF2α is released from the uterus in a pulsatile fashion, and evidence suggests that the pattern of decline in progesterone differs between animals that receive a single dose of PGF2α compared to repeated doses [1, 2]. The problem has been that it is impossible to predict the precise timing of the onset of uterine release of PGF2α and luteal regression in a natural cycle, making it difficult to study this stage of the estrous cycle experimentally. As reported in this issue of Biology of Reproduction, Atli et al. [3] used intrauterine infusions of PGF2α delivered at 6-h intervals, based on the work of Ginther et al. [1], to mimic the physiological concentration and pulsatile release of PGF2α that occurs during natural luteolysis. They combined this with a unique procedure developed in Dr. Milo Wiltbank's laboratory—namely, repeated ultrasound-guided needle biopsies of the CL. While one might think that this physical insult would induce luteal prostaglandin release and initiate luteolysis itself, the procedure has been previously validated, and it was clearly demonstrated not to be the case [4]. Also, in the present study, sequential luteal biopsies were performed in cows that received saline rather than PGF2α infusions, and no indication of luteal regression was observed in these animals. Sayre et al. [5] demonstrated that two injections of PGF2α were more effective than one injection in increasing IGFBP1 mRNA in the CL, but because the CL were removed, the progression of luteolysis could not be assessed in that study. Atli et al. [3] included a treatment group in which cows received two infusions of PGF2α (2X), separated by a saline infusion, rather than the four infusions of PGF2α administered to a separate treatment group. This not only allowed the investigators to assess acute versus chronic effects of PGF2α in the CL but, as with many great discoveries that have occurred by serendipity, to observe that luteolysis occurred in one-half of the animals in the 2X treatment group. As a consequence, direct comparisons of gene expression in the regressing compared to nonregressing CL could be made. This is the first study to utilize a physiological model of luteolysis that allowed precisely timed, repeated collection of luteal tissue samples. The animal part of these experiments was no doubt extremely difficult and time-consuming, but the excellent experimental design provided definitive evidence for changes in luteal mRNA concentrations that relate to life-or-death decisions in the CL. Despite decades of work on PGF2α-induced luteolysis in ruminants, the exact mechanism by which PGF2α brings about the demise of the CL is still not understood. Early in vitro experiments using short-term incubations of luteal cells left researchers perplexed, because treatment with PGF2α resulted in an increase in progesterone synthesis [6–8]. Once longer-term culture of primary luteal cells was developed, it became clear that chronic treatment with low concentrations (10 ng/ml) of PGF2α could inhibit gonadotropin-stimulated progesterone production [9], thus providing a model to study the antisteroidogenic effect of PGF2α in bovine luteal cells. Numerous studies using in vitro luteal cells as well as ruminant and rodent animal models were conducted in an attempt to localize the cellular site of action of PGF2α that resulted in the loss of steroid synthesis. Hypotheses for PGF2α action in luteal cells included downregulation of luteinizing hormone (LH) receptors, inhibition of adenylate cyclase, and prevention of lipoprotein-associated cholesterol uptake, but most experimental data did not support these hypotheses. When PGF2α binds to its receptor on luteal cells, it activates the protein kinase C signaling pathway and induces expression of the immediate early genes, FOS and JUN [10]. However, the exact pathways that link expression of these genes to the loss of progesterone synthesis have remained elusive. In cultured luteal cells, PGF2α prevented association of cholesterol with the cytochrome P450 side-chain cleavage enzyme (CYP11A1) [11]. When the steroidogenic acute regulatory protein (STAR) was discovered as a key component of the rate-limiting step of steroidogenesis [12], Juengel et al. [13] and Pescador et al. [14] provided evidence that STAR mRNA declined during luteal regression, but the decrease in STAR mRNA occurred after the initial drop in progesterone. Despite numerous studies [15–18], the mechanism of action of PGF2α action within the CL retains its mystery. Curiously, PGF2α does not initiate apoptosis in cultured luteal cells. This led investigators to consider whether other types of cells were missing from the steroidogenic cell cultures (i.e., endothelial cells [18–20] and immune cells [21], and the results of many in vitro experiments suggested that the luteolytic effect of PGF2α likely involved communication among these heterogeneous cells. Support for a critical role of the endothelial cell-derived protein, endothelin 1, in luteal regression was also obtained from in vivo experiments [22–25]. In addition to the novel in vivo model, the study by Atli et al. [3] is the first to show temporal changes in a comprehensive set of genes known to be involved in luteal function, with the direct comparison of CL that regress to those that do not. As expected, an acute effect of PGF2α was to increase concentrations of the immediate early genes, JUN, FOS, NR4A1, and EGR1, and without repeated exposure to PGF2α, concentrations of these mRNAs declined, providing the first clue that repeated exposure to PGF2α is necessary for continuance of luteolytic pathways within the CL. Two mRNAs, STAR and VEGFA, increased in response to the first exposure to PGF2α but were decreased following the second exposure. These two genes, classified as steroidogenic and angiogenic genes, respectively, may play key roles in the continued progression of luteolysis, because they did not decline in the CL that were maintained, which may provide additional support for the importance of repeated PGF2α pulses. Interestingly, the apparent continued capacity of regressing luteal cells to synthesize progesterone, as evidenced by unchanged concentrations of CYP11A and HSD3B7 proteins, helps to explain why luteal cells obtained from regressing CL survive and produce progesterone in response to LH when removed from the animal and placed into culture [26]. The most compelling differences between CL that regressed and those that did not were changes in mRNAs consistent with the concept of autoamplification of luteal prostaglandin synthesis and in immune response genes. This study by Atli et al. [3] supports the notion that luteal prostaglandin synthesis is an obligate component of the luteolytic process, first suggested by Milvae [27] and further defined by Silva et al. [28] and Niswender et al. [29], and it adds to the growing amount of evidence that immune cells within the CL mediate some of the luteolytic effects of PGF2α [18, 30]. Most importantly, this work clearly demonstrates that the process of luteolysis involves multiple pathways and their repeated activation to commit the cells to a pathway that will lead to regression. As the authors eloquently state, “[R]egulatory pathways in the CL [are] increasingly shunted into a luteolytic direction with increasing numbers of low-dose PGF2α pulses.” The key discovery reported by Atli et al. [3] is the temporal progression of intraluteal changes in mRNA and protein concentrations that are directly correlated with the necessity for repeated pulses of PGF2α to ensure luteal regression. Figure 1 provides a simplified illustration of the pathways that are activated during determination of the fate of the CL. Future studies will likely focus on which of these genes are necessary for the CL to continue down the pathway of cell death and tissue destruction and which are coincident with the loss of structure and function. Exposure to a single dose of PGF2α results in acute responses in the CL that may trigger the beginning of a luteolytic cascade (immediate early genes, prostaglandin synthesis), but some genes are induced that are inconsistent with the onset of luteolysis (STAR, HDPG, VEGFA). If repeated pulses of PGF2α are absent, luteal cells reset to a state in which prostaglandin synthesis is not excessively stimulated, PGF2α is inactivated by HDPG, and steroidogenesis is supported. With repeated exposure to PGF2α, intracellular prostaglandins and cytokines increase in concentration, and steroidogenic enzymes are inhibited. The keys to the decision to proceed down a pathway leading to luteolysis versus maintenance may lie in the luteal concentration of prostaglandin and the activation of immune cells.