Involvement of the luteinizing hormone surge in the regulation of ovary and oviduct clock gene expression in mice
Abstract
Circadian dysfunction perturbs the female reproductive cycle. In particular, mice lacking the clock gene Bmal1 show severe infertility, implying that BMAL1 plays roles in ovulation and luteinization. Here, we examined temporal changes in clock gene expression in the ovary and oviduct before and during gonadotropin‐induced follicular growth, ovulation, and luteinization in sexually immature mice. While the oviduct did not show a drastic change in clock gene expression, Bmal1 expression in the ovary was higher than that in control mice during the period from 4 to 16 hr after human chorionic gonadotropin (hCG) administration. Bmal1 expression reached a maximum at 16 hr after hCG administration, when follicle luteinization occurred. In an interesting manner, administration of hCG to ex vivo‐cultured oviduct triggered a shorter circadian period and inevitably resulted in phase advance. Together, our present data suggest that LH surge induces continuous expression of BMAL1 in the mouse ovary and modulates circadian phase in the mouse oviduct.
1 INTRODUCTION
The mammalian circadian clock is driven by clock gene expression rhythms (Reppert & Weaver, 2002; Rosbash et al., 2007). Interlocked transcriptional feedback loops of clock genes generate circadian gene expression in a genomewide manner, which in turn leads to circadian oscillation in diverse behavior and physiology (Doherty & Kay, 2010; Mohawk, Green & Takahashi, 2012). The circadian clock enables maximum expression of genes at appropriate times of the day, allowing organisms to adapt to environmental rhythms generated by earth's rotation. Chronic desynchronization between internal circadian and external environmental rhythms carries a significant risk of diverse disorders, ranging from sleep disorders to diabetes, cardiovascular diseases, and cancer (Sahar & Sassone‐Corsi, 2012; Wijnen & Young, 2006).
The circadian clock and female reproductive cycle are strongly related to each other. For example, circadian patterns in physiology and behavior vary depending on rodent estrous or human menstrual stages (Boivin, Shechter, Boudreau, Begum & Ng Ying‐Kin, 2016; Takahashi & Menaker, 1980), suggesting that the female reproductive cycle affects the circadian clock. This theory is strengthened by the fact that estrogen receptors are expressed in the suprachiasmatic nucleus (Vida et al., 2008). On the other hand, it has been reported that circadian dysfunction perturbs the female reproductive cycle (Miller et al., 2004; Yoshinaka et al., 2017), indicating that the circadian clock in turn controls the reproductive cycle. In particular, mice lacking the clock gene Bmal1 show severe infertility (Alvarez et al., 2008), implying that BMAL1 protein plays roles in ovulation and ovarian luteinization. However, information is scarce about how Bmal1 is expressed before and during ovulation and ovarian luteinization, and what kind of factors induce this Bmal1 expression.
We hypothesized that healthy ovulation and ovarian luteinization require the induction of Bmal1 expression in response to the luteinizing hormone (LH) surge. Based on this hypothesis, we injected equine chorionic gonadotropin (eCG), which exhibits FSH (follicle‐stimulating hormone) activity and is commonly used because of its long half‐life, and human chorionic gonadotropin (hCG), which mimics LH, into 4‐week‐old, sexually immature female mice. We used sexually immature female mice to facilitate the interpretation of experimental outcomes by distinguishing the effect of exogenous hCG from that of endogenous LH. These hormone injections artificially induce ovarian follicular growth, ovulation, and luteinization in sexually immature mice. Under these conditions, we examined temporal changes in clock gene expression, including that of Bmal1, in the ovary and oviduct after hCG injection.
2 RESULTS
Before investigating the effect of hCG on clock gene expression, we examined the effect of eCG‐induced sexual maturation on ovarian clock gene expression (Figure 1). Four‐week‐old, sexually immature female ICR mice were intraperitoneally administered eCG (5 IU) at zeitgeber time 8 (ZT8). Control mice were treated with saline. The ovaries and livers of four mice at each time point were harvested at 4‐hr intervals after eCG administration, and total RNA was extracted and reverse‐transcribed. The resulting cDNA was subjected to quantitative RT‐PCR. Expression levels of the clock genes Bmal1 and Per2 were measured and normalized with those of β‐actin. As it was expected that eCG does not act on the liver, the diurnal pattern of liver clock gene expression was similar before and after eCG administration (Figure 1a). Given that the diurnal pattern of peripheral clock gene expression differs between newborn and adult mice (Polidarova et al., 2014; Sladek, Jindrakova, Bendova & Sumova, 2007), a bimodal pattern of Bmal1 expression in the liver indicates that peripheral clocks are still developing at the age of 4 weeks. In contrast, the diurnal pattern of Per2 expression in the ovary likely differs before and after eCG administration, from a bimodal to monomodal pattern (Figure 1b). However, given the absence of antiphasic expression rhythms of Per2 and Bmal1, which is common in adult mouse tissues, ovarian clocks may be still developing at this age. Before the circadian oscillator develops and matures, clock gene products might play noncircadian roles in sexually immature female mice.
Next, to examine the possible effect of LH surge on ovarian clock gene expression, 4‐week‐old, sexually immature female ICR mice were intraperitoneally administered hCG (7.5 IU, ZT 8) 48 hr after injection of eCG (7.5 IU, ZT 8; Figure 2). Control mice were treated with saline using the same procedure and at the same time points. The ovaries and oviducts were harvested from four mice at a 4‐hr interval after hCG administration, and total RNA was extracted and reverse‐transcribed. The resulting cDNA was subjected to quantitative RT‐PCR. Expression levels of the clock and clock‐related genes Bmal1, Per2, Rev‐erbα, and Dbp were measured and normalized with that of β‐actin. To ensure successful ovulation by hCG treatment, ovarian luteinization was visually confirmed 16 hr after hCG administration, when control mice treated with vehicle did not show ovarian luteinization. In the ovary of hCG‐treated mice, Bmal1 gene expression was higher than that of control mice during the period from 4 to 16 hr after hCG administration, with statistical significance. Given that ovulation occurs 12–14 hr after hCG administration, these data are consistent with our hypothesis that healthy ovulation requires the induction of Bmal1 expression in response to the LH surge. The data also imply that BMAL1 may be involved in ovarian luteinization because Bmal1 expression reached maximum 16 hr after hCG administration, when follicle luteinization occurred.
Although the oviduct did not show a drastic change in clock gene expression compared to the ovary, transient activation of Per2 transcription was detected 4 hr after hCG administration (Figure 3). Given that this transient activation is sufficient to shift circadian phase in fibroblasts (Sato, Murakami, Node, Matsumura & Akashi, 2014), hCG may play a role in circadian phase control of the oviduct. Although the present data are not sufficient to make conclusions, a statistically significant reduction in expression levels of E‐box‐regulated genes such as Dbp and Rev‐erbα 16 hr after hCG administration may indicate circadian phase advance in the oviduct. Likewise, in the ovary, we detected a drastic reduction in expression levels of E‐box‐regulated genes (Dbp, Rev‐erbα, and Per2) 16 hr after hCG administration, but saw no transient activation of Per2 transcription. These data suggest that LH surge may play a role in resetting the circadian clock of the ovary and oviduct. Although continuous monitoring of clock gene expression for several days may enable the drawing of conclusions about hCG‐induced circadian phase shift, a limitation of the present experimental model using 4‐week‐old, sexually immature mice is that it is difficult to evaluate circadian phase shift based on the unclear circadian rhythms of clock gene expression produced by immature peripheral clocks.
To compensate for this technical limitation, we also aimed to investigate the effect using tissue explants of sexually mature mice having mature circadian clocks under ex vivo conditions free from endogenous LH (Figure 4). The oviduct but not ovary was used for ex vivo culture due to technical hurdles in organ culture. In a specific manner, ex vivo culture of the oviduct was performed using sexually mature female Period2Luciferase (Per2Luc) knock‐in mice, and clock gene expression was monitored as bioluminescence in real time. Clear circadian rhythms in bioluminescence were detected from ex vivo‐cultured oviducts, confirming that the autonomous circadian oscillator was mature in the oviduct of sexually mature mice (Figure 4a). In an interesting manner, administration of hCG to culture medium triggered a shorter circadian period in a statistically significant manner and inevitably resulted in phase advance (Figure 4a,b), demonstrating that LH surge potentially adjusts the circadian phase of the oviduct in vivo. hCG‐induced transient activation of Per2 expression in the oviduct was observed in vivo as shown in Figure 3, but was not detected in this ex vivo experiment. This may be because transient changes in bioluminescence that were not drastic were likely masked due to the high background bioluminescence.
3 DISCUSSION
It has been reported that conditional knockout of the Bmal1 gene in ovarian steroidogenic or theca cells results in the failure of ovulation and implantation (Liu et al., 2014; Mereness et al., 2016). Consistent with these previous reports, our present data indicate that LH surge induces continuous expression of the transcription factor BMAL1 in the ovary, likely leading to activation of the genes required for ovulation. This interpretation explains why Bmal1‐deficient mice show infertility. Although a previous study reported that hCG induces Bmal1 transcription in a hypophysectomized, immature rat model (Karman & Tischkau, 2006), which is similar to our present study, this rat model is highly invasive, and consistency with and comparison to previous results obtained using knockout mice are limited by the species difference. Our present study using a noninvasive immature mouse model reproduces these previous results using an invasive rat model. In addition, strong elevation of Bmal1 expression 16 hr after hCG administration suggests a potential role of BMAL1 protein in ovarian luteinization. Supporting a role of BMAL1 in ovarian luteinization, it was previously reported that knockdown or knockout of the rat or mouse Bmal1 gene results in reduced expression of progesterone and prostaglandin biosynthesis‐related genes in luteinizing granulosa or corpus luteum cells (Chen et al., 2013; Ratajczak, Boehle & Muglia, 2009).
Although BMAL1 is well known as a major circadian component, its involvement in healthy ovulation and ovarian luteinization suggests its functional versatility. Indeed, several noncircadian roles for BMAL1 have been reported. For example, BMAL1 plays important roles in the regulation of adipose differentiation and lipogenesis in mature adipocytes (Shimba et al., 2005). In addition, BMAL1 is involved in pancreatic islet growth and development (Marcheva et al., 2010). Moreover, Bmal1‐deficient mice display early aging phenotypes with increased levels of reactive oxygen species in some tissues (Kondratov, Kondratova, Gorbacheva, Vykhovanets & Antoch, 2006). BMAL1 belongs to the family of basic helix‐loop‐helix (bHLH)‐PAS domain‐containing transcription factors and forms a heterodimer with other members of this protein family, not only with the circadian transcription factor CLOCK or NPAS2 but also with hypoxia‐inducible factor 1 (HIF‐1), HIF‐2, or aryl hydrocarbon receptor (Hogenesch, Gu, Jain & Bradfield, 1998; Hogenesch et al., 1997). Noncircadian roles of BMAL1 may be exerted through dimerization with these noncircadian partners. In particular, HIFs, which regulate neoangiogenesis by activating the vegf gene in a wide range of biological processes (Doedens & Johnson, 2007; Jain, Nikolopoulou, Xu & Qu, 2018), may play indispensable roles in follicular development and ovulation (Kim, Bagchi & Bagchi, 2009; Tam et al., 2010). In particular, HIF‐1α and HIF‐2α protein levels increase in the ovary during ovulation and ovarian luteinization (Kim et al., 2009), when Bmal1 expression peaks in the ovary according to our present data.
Our results also suggest that LH surge may reset the circadian phase of the ovary and oviduct in mice. Given the temporal inconsistency that Bmal1 expression in the ovary peaked 16 hr after hCG administration to sexually immature mice, this expression peak may be not a direct outcome of the hCG stimulation but a de novo circadian peak caused by hCG‐induced phase shift. This hypothesis is also confirmed by the fact that the expression level of Rev‐erbα, a potent transcriptional repressor of Bmal1, was drastically decreased in an antiphasic manner 16 hr after hCG administration. To compensate for the limitation that peripheral circadian clocks are still developing in the sexually immature mice used in the present in vivo experimental model, we also demonstrated hCG‐induced circadian phase advance in the oviduct of sexually mature mice with mature peripheral clocks, under ex vivo conditions free from endogenous LH. Although the present data indicate that hCG likely phase advances circadian rhythms of the mouse ovary and oviduct in vivo and ex vivo, the effect might vary in a phase‐responsive manner; in other words, hCG administration at other time points might cause a phase delay. Related to our hypothesis that LH surge resets the circadian phase of the ovary and oviduct in mice, a study using sexually immature rats demonstrated that hCG stimulation triggers and initiates circadian rhythms in ovary clock gene expression (Gras, Georg, Jorgensen & Fahrenkrug, 2012). In addition, a study using chicken ovaries indicated that the expression of ovarian circadian clock genes may be influenced by the LH surge in vivo (Tischkau, Howell, Hickok, Krager & Bahr, 2011). Because the successful establishment of pregnancy requires functional coordination of various reproductive processes in multiple organs, including ovulation, fertilization, implantation, and luteinization, autonomous circadian rhythms in each organ likely need to be synchronized with each other to ensure a high pregnancy success rate (Sellix, 2013; Sellix & Menaker, 2010). LH surge‐induced resetting of circadian rhythms of female reproductive organs may therefore play a role in temporal coordination of reproductive processes in multiple organs. In light of this theory, the hCG‐induced resetting of ovary circadian gene expression shown in the present study may reflect the idea that LH surge enables an increase in BMAL1 protein and downstream gene products at the time of ovulation and luteinization. In addition, LH‐induced resetting of oviduct circadian gene expression may be required for ovulated oocytes to undergo normal reproductive processes at appropriate circadian timings.
Although our data suggest that hCG administration induces circadian phase shift in both the ovary and oviduct, the early response of clock gene expression to hCG differs between these organs. In a specific manner, immediate early expression of the Bmal1 or Per2 gene was observed in the ovary or oviduct, respectively. cAMP and MAPK signaling pathways function downstream of LH/hCG receptors (Li, Zhang, Peng, Wang & Zhu, 2014), and both pathways are involved in immediate early and transient expression of Per2 in fibroblasts (Akashi & Nishida, 2000; Travnickova‐Bendova, Cermakian, Reppert & Sassone‐Corsi, 2002). This Per2 expression likely triggers a phase shift in oviduct circadian rhythms, given that various extracellular stimuli induce circadian phase shift accompanied by immediate early and transient expression of Per2 in fibroblasts (Balsalobre, Marcacci & Schibler, 2000) and that this transient activation of Per2 transcription is sufficient to phase shift circadian rhythms in fibroblasts without other intracellular signaling (Sato et al., 2014). Although the physiological role and intracellular signaling for immediate early expression of Bmal1 are poorly understood, it was reported using cultured chicken granulosa cells that intracellular signaling pathways similar to those for Per2 are involved in hCG‐induced activation of Bmal1 transcription (Li et al., 2014).
Taken together, our present data indicate that LH surge induces BMAL1 expression in the mouse ovary and that this may be required for healthy ovulation and luteinization. These results are consistent with those previously obtained using different species with different experimental approaches. Further, our present data suggest that LH surge adjusts circadian phase in the mouse ovary and oviduct and that this adjustment may be required for functional coordination of various reproductive processes in multiple organs.
4 EXPERIMENTAL PROCEDURES
4.1 Animals
Mice were maintained on a 12‐hr light–dark cycle and allowed ad libitum access to food and water. All protocols for animal experiments were approved by (a) the Committee on the Ethics of Animal Experiments of Obihiro University of Agriculture and Veterinary Medicine and (b) the Animal Research Committee of Yamaguchi University. Animal studies were performed in compliance with (a) the Guiding Principles for the Care and Use of Research Animals Promulgated by Obihiro University of Agriculture and Veterinary Medicine and (b) the Yamaguchi University Animal Care and Use guidelines. Four‐week‐old, sexually immature female ICR mice (CLEA Japan) were intraperitoneally administered with eCG (5 IU, Asuka Pharma) to examine the effect of eCG‐induced sexual maturation or with hCG (7.5 IU, Asuka Pharma) 48 hr after injection of eCG (7.5 IU) to examine the possible effect of LH surge on clock gene expression. Control mice were treated with saline under the same procedure and at the same time points. To ensure successful ovulation by hCG treatment, ovarian luteinization was visually confirmed 16 hr after hCG administration, when control mice treated with vehicle did not show ovarian luteinization. 13‐week‐old, proestrus Period2::Luciferase (Per2Luc) knock‐in mice, a kind gift from Dr. Joseph Takahashi (Yoo et al., 2004), were used for ex vivo culture experiments.
4.2 Peripheral clock gene expression
Beginning at the indicated time point after gonadotropin treatment, four mice each were humanely killed and tissues were harvested at 4‐hr intervals and frozen until gene expression analysis. Total RNA was extracted using TRIZOL (Life Technologies). After quality and concentration of total RNA was checked using NanoDrop (Thermo Fisher Scientific), samples were used to semi‐quantify mRNA levels. In brief, total RNA was reverse‐transcribed using a Prime‐Script RT Reagent Kit (TAKARA BIO), and the resulting cDNAs were subjected to real‐time reverse‐transcription PCR using a QuantiTect SYBR Green PCR Kit (Qiagen). Data were obtained using iQcycler (Bio‐Rad Laboratories). The sequences of the primers and probe for the Bmal1, Per2, Rev‐erbα, Dbp, and β‐actin genes are as follows.
| Genes | Primer sequence | Annealing temperature (°c) |
|---|---|---|
| Bmal1 | Forward: GGAGAAGGTGGCCCAAA | 58 |
| Reverse : AGGCGATGACCCTCTTA | ||
| Per2 | Forward : GGCACATCTCGGGATCG | 58 |
| Reverse : GAGCAGAGGTCCTCGCC | ||
| Rev‐erb α | Forward : CCCTGGACTCCAATAACAACACA | 58 |
| Reverse : GCCATTGGAGCTGTCACTGTAG | ||
| Dbp | Forward : GGAACTGAAGCCTCAACCAAT | 58 |
| Reverse : CTCCGGCTCCAGTACTTCTCA | ||
| β‐actin | Forward : CACACCTTCTACAATGAGCTGC | 58 |
| Reverse : CATGATCTGGGTCATCTTTTCA |
Expression levels of the clock and clock‐related genes Bmal1, Per2, Rev‐erbα, and Dbp were normalized with that of β‐actin.
4.3 Ex vivo culture
After mice were humanely killed, oviduct explants were prepared from 13‐week and proestrus Per2Luc knock‐in mice in ice‐cold Hank's balanced salt solution (Life Technologies) supplemented with 10 mM HEPES (Sigma‐Aldrich). Each explant was placed on a culture membrane (Millicell‐CM, Millipore) in a covered and sealed 35‐mm dish containing Dulbecco's modified Eagle's medium supplemented with penicillin‐streptomycin, 10% FBS and 0.1 mM D‐luciferin, as detailed in previous studies (Sato et al., 2014). Bioluminescence was measured in real time with photomultiplier tubes (LM2400; Hamamatsu) and integrated for 1 min at 15‐min intervals. Data sets were detrended by subtracting the 24‐hr running average. For statistical evaluation of the effect of hCG (2.5 IU/ml) on clock gene expression, circadian period length during the exposure of explants to hCG was obtained using the software Cosinor, kindly provided by Dr. Refinetti.
ACKNOWLEDGMENTS
We thank Atsuhiro Nishida, Yoshiki Miyawaki, and Rie Okamitsu for their expert technical assistance. We also acknowledge the support of fellowships from the Japan Society for the Promotion of Science.
CONFLICT OF INTEREST
The authors declare no competing financial interests.
AUTHOR CONTRIBUTIONS
T.S. and M.A. conceived and supervised the project. M.A. analyzed data and wrote the manuscript. M.K., K.W., R.M., and N.A. performed experiments and analyzed data. A.M., H.M., and K.M. provided general supports and gave conceptual advice.
