Truth Seeker
New Member
Abstract
The mechanisms by which synchronized embryonic development to the blastocyst stage, preparation of the uterus for the receptive state, and reciprocal embryo-uterine interactions for implantation are coordinated are still unclear. We show in this study that preimplantation embryo development became asynchronous in mice that are deficient in brain-type (CB1) and/or spleen-type (CB2) cannabinoid receptor genes. Furthermore, whereas the levels of uterine anandamide (endocannabinoid) and blastocyst CB1 are coordinately down-regulated with the onset of uterine receptivity and blastocyst activation prior to implantation, these levels remained high in the nonreceptive uterus and in dormant blastocysts during delayed implantation and in pregnant, leukemia inhibitory factor (LIF)-deficient mice with implantation failure. These results suggest that a tight regulation of endocannabinoid signaling is important for synchronizing embryo development with uterine receptivity for implantation. Indeed this is consistent with our finding that while an experimentally induced, sustained level of an exogenously administered, natural cannabinoid inhibited implantation in wild-type mice, it failed to do so inCB1 −/− /CB2 −/−double mutant mice. The present study is clinically important because of the widely debated medicinal use of cannabinoids and their reported adverse effects on pregnancy.
Previous investigation suggested that cannabinoid exposure has adverse effects on pregnancy outcome (reviewed in Refs. 1-5). Cannabinomimetic drugs interact with two types of cannabinoid receptors, brain-type (CB1) and spleen-type (CB2) (6-9). These receptor subtypes are negatively coupled to adenylate cyclase and to N-type and P/Q-type calcium channels and positively coupled to mitogen-activated protein kinase and to A-type potassium channels through Gi/o proteins (7, 10). CB1 is expressed in brain and other peripheral tissues (7-11). CB2 is expressed primarily in immune tissues including the spleen, leukocytes, and tonsils (12). We have previously shown that CB1 is expressed in the preimplantation mouse embryo at much higher levels than in the brain (1, 2). The discovery of cannabinoid receptors led to the identification of endogenous cannabinoid ligands, N-arachidonoylethanolamine (anandamide), and 2-arachidonoylglycerol (13-17). We observed that anandamide is synthesized in the pregnant mouse oviduct and uterus (18) and that its levels are far higher in the uterus than in any other mammalian tissue examined (3). However, anandamide levels are significantly lower at implantation sites than at interimplantation sites, suggesting endocannabinoid ligand-receptor signaling during implantation (3). Indeed anandamide at higher levels adversely affects embryo development and implantation, whereas at lower levels it stimulates embryo growth and differentiation via CB1 (1-5). Interestingly, 2-arachidonoylglycerol is present in the mouse uterus at amounts similar to or lower than the lowest anandamide levels (1—5 nmol/g of tissue), and its level does not vary significantly between implantation and interimplantation sites.1 Using cannabinoid receptor mutant mice (19) and physiological approaches, we further defined the role of ligand-receptor signaling with cannabinoids in preimplantation embryo development and in implantation.
EXPERIMENTAL PROCEDURES
Animals
Adult CD-1 (Charles River Breeding Laboratories),CB1, and CB2 mutant mice on a C57BL/6J background or (LIF)2 mutant mice on a mixed background were housed in the animal care facility at the University of Kansas Medical Center. Story Landis (NINDS, National Institutes of Health) kindly provided LIF mutant mice that were originally generated by Philippe Brulet (Pasteur Institute). We have characterized these mice previously (20). Heterozygous breeding was performed to generate littermate (+/+), (+/−), and (−/−) mice of specific mutation. Homozygous males and females with specific null mutations were crossed to generate homozygous embryos with respective mutations or double mutation. Females were mated with fertile or vasectomized males of the same strain to induce pregnancy or pseudopregnancy (day 1 = vaginal plug), respectively. To induce conditions of delayed implantation, mice were ovariectomized on the morning (0800—0900) of day 4 of pregnancy and maintained with a daily injection (subcutaneously) of progesterone (P4, 2 mg/mouse) from days 5—7. To activate dormant blastocysts and initiate implantation, P4-primed delayed implanting pregnant mice were injected with estradiol-17β (E2, 25 ng/mouse) (21).
Embryo Culture
Two-cell embryos recovered on day 2 were cultured in 25 μl of Whitten's medium in groups of 5—10 under silicon oil in an atmosphere of 5% CO2 in air for 72 h with or without anandamide (7.5 nM) (1, 2). The control cultures contained the same concentration of the vehicle. The embryos were observed every 24 h to monitor their development. At the termination of culture the number of embryos forming blastocysts was recorded.
Anandamide Binding
Autoradiographic ligand binding in blastocysts was performed as previously described (1, 2). Blastocysts were incubated with 4.5 nM [3H]anandamide (specific activity 221 Ci/mmol, PerkinElmer Life Sciences) in the presence or absence of a 500-fold molar excess of unlabeled anandamide. After incubation, blastocysts were washed, fixed in 2% paraformaldehyde, cytospun onto slides, air-dried, and subjected to autoradiography using a liquid emulsion. Blastocysts were poststained in hematoxylin. The autoradiographic signals were quantitated under dark field conditions using an OPTIMA II program with an image analysis system. Statistical analysis was performed using one-way analysis of variance followed by a Newman-Keuls test.
CB1 Immunostaining
Blastocysts were cytospun onto slides, fixed in 2% formalin for 15 min, washed in phosphate-buffered saline, and incubated in blocking solution for 10 min. They were then incubated with an affinity-purified rabbit antipeptide antibody specific to CB1 (1 μg/ml) overnight at 4 °C (2). Immunostaining was performed using a Zymed Laboratories Inc. Histostain-SP kit (2). Blastocysts were counterstained with hematoxylin. Red deposits indicated the sites of immunoreactive CB1 protein. Control blastocysts were incubated parallel with the antibody that was preneutralized with excess antigenic peptide.
Analysis of Anandamide and Other N-Acylethanolamines
Extraction, isolation, and analysis ofN-acylethanolamine (NAE) were done essentially as previously described (3). Uteri were homogenized in chloroform/methanol (2:1), and the homogenate was partitioned against 2.5% aqueous NaCl. Internal standards (N-acyl-1,1,2,2-2H4-ethanolamines with chain lengths of 16:0, 17:0, 18:0, 18:1n-9, 18:2n-6, and 20:4n-6; d 4 NAEs, 0.1 μg each) were added to the extract. The NAEs were isolated using solid phase silica extraction columns (Alltech Associates). Neutral lipids were eluted with 4 ml of chloroform followed by elution of the NAEs with 4 ml of chloroform:methanol (98:2). The NAE fractions were taken to dryness and treated with 50 μl oftert-butyldimethylchlorosilyl/imidazole reagent (Alltech) at 80 °C for 1 h. Thetert-butyldimethylchlorosilyl/imidazole derivatives were extracted into 50 μl of hexane for analysis by gas chromatography-mass spectrophotometry in the selected ion monitoring mode. Endogenous NAEs were quantitated by comparing the peak size of their (M − 57) ions to those of the corresponding stable isotope-labeled internal standards (22). Anandamide accounted for over 90% of total NAEs.
Administration of Drugs
(−)-Δ9-Tetrahydrocannabinol ((−)-THC) or its inactive stereoisomer (+)-THC was delivered at a constant rate via miniosmotic pumps (Alzet Corp., Palo Alto, CA) to study their effects on implantation in wild-type andCB1 −/− /CB2 −/−pregnant mice. Miniosmotic pumps (mean pumping rate of 0.52 μl/h and fill volume of 96 μl) containing either (−)THC or (+)THC were placed under the back skin of mice on day 2 (1000 h) of pregnancy and continued through day 5. These pumps released THC at a rate of 20 μg/h. Cytochrome P450 inhibitors, metyrapone and clotrimazole (50 mg/kg body weight), were first injected intraperitoneally 2 h before the installation of pumps and injected twice daily until day 4 of pregnancy (4). On day 5 (1000 h), implantation sites were determined by the blue dye method (21). If implantation sites were absent, uterine horns were flushed with saline to recover unimplanted blastocysts. Mice without implantation sites or blastocysts were excluded from the experiments.
RESULTS
Embryo Development Is Asynchronous in Cannabinoid Receptor-deficient Mice
Because synchronized embryo development and uterine preparation are important for successful implantation, we first compared the in vivo developmental potential of cannabinoid receptor mutant embryos with wild-type embryos. We observed that the development of CB1 −/−,CB2 −/−, orCB1 −/−/CB2 −/− double mutant embryos recovered from the oviduct on day 3 (TableI) and from the uterus on day 4 (TableII) of pregnancy was asynchronous. On the morning of day 3, about 84% of the wild-type embryos recovered were at the 8-cell stage. In contrast, only about 24% ofCB1 −/−, 42% ofCB2 −/−, and 50% of theCB1 −/−/CB2 −/− double mutant embryos were at the 8-cell stage, resulting in an increased population of 2-, 4-, and 6-cell embryos (Table I). On the morning of day 4, about 98% of the wild-type embryos were blastocysts. However, only about 62% of CB1 −/−, 71% ofCB2 −/−, or 61% ofCB1 −/−/CB2 −/− embryos were at the blastocyst stage; a considerable number of embryos were at the morula stage. However, these morulae were apparently viable and normal because they rapidly developed into blastocysts in culture. We speculated that developmentally retarded embryos eventually form blastocysts and implant. Thus, we next compared the status of implantation in the wild-type and mutant mice on day 5 of pregnancy. Increased localized endometrial vascular permeability at the site of the blastocyst is one of the early markers of implantation (23). This can be monitored by an intravenous injection of a blue dye (Chicago blue B) solution which specifies implantation sites as distinct blue bands along the uterus (4, 21, 23). We observed that whereas all of the vaginal plug-positive wild-type or CB2 −/− mice (n = 7) showed an average of 9 or 7 implantation sites, respectively, 6 of 10 (60%) CB1 −/− and 14 of 18 (78%)CB1 −/− /CB2 −/−plug-positive mice showed an average of 7 and 9 implantation sites per mouse, respectively. These results suggest that the loss of CB1 had modest, if any, adverse effects on implantation. This is perhaps because of the implantation of blastocysts that eventually formed from the slowly growing embryos. Our next objective was to examine whether the mutant embryos are responsive to an endocannabinoid exposurein vitro.
Mutant Embryos Are Resistant to Anandamide
Two-cell wild-type or mutant embryos were cultured in the presence or absence of anandamide as previously described (1, 2). As observed previously (1,2), most (>79%) of the 2-cell wild-type embryos failed to develop to the blastocyst stage in the presence of anandamide. In contrast, more than 80% of the CB1 −/− orCB1 −/− /CB2 −/− double mutant embryos developed into blastocysts in the presence of anandamide (Fig. 1). However, in vitrodevelopment of CB2 −/− embryos, like wild-type embryos, was severely compromised in the presence of anandamide. These results provided genetic confirmation to our previous findings that CB1, but not CB2, in wild-type blastocysts responds to cannabinoid agonists in a stage- and dose-dependent manner in vitro (1-5) although these embryos express CB2mRNA (1). Whereas the developmental response ofCB2 −/− embryos to anandamide is similar to that of wild-type embryos in vitro (1, 2), early asynchronous development of CB2 −/− embryosin vivo, like CB1 −/− orCB1 −/− /CB2 −/−embryos, is surprising. It is possible that the reproductive tract environment differentially modulates the development of mutant embryos in response to endocannabinoids in vivo but not during their development in vitro. In this respect, the mouse uterus expresses the CB1 but not the CB2 gene (24), and thus uterine CB1 may influence embryonic development in vivoin the absence of embryonic CB2. Another possibility could be that uterine anandamide levels in mutant mice are different from those of the wild-type mice. However, uterine anandamide levels inCB1 −/− /CB2 −/− double mutant mice are not significantly different from the wild-type uterine levels.1 Nonetheless, the present genetic evidence suggests that cannabinoid receptors have some role in synchronizing embryo development in the reproductive tract in vivo.
Anandamide Levels Are Higher in the Nonreceptive Uterus
In the mouse, progesterone (P4) and estrogen sequentially program the uterus into prereceptive, receptive, and nonreceptive phases during pregnancy or pseudopregnancy (23, 25). Blastocysts implant only in the receptive uterus. The P4-primed uterus becomes receptive on day 4 when it is superimposed with preimplantation ovarian estrogen. Subsequently the receptive uterus proceeds to the nonreceptive phase, and blastocysts can no longer implant (21, 25). Similar uterine changes are experimentally produced during delayed implantation induced by ovariectomy before preimplantation ovarian estrogen secretion on day 4. Under this condition, blastocysts undergo dormancy and fail to implant. This condition is terminated by an injection of estrogen with activation of dormant blastocysts, attainment of the receptive uterus, and implantation (21, 23, 25). The mechanism by which estrogen initiates these events in the P4-primed uterus is not clearly understood.
The mouse uterus is receptive on day 4, whereas it was considered to be nonreceptive on day 5 (21). We previously observed that not only are uterine anandamide levels higher on day 5 of pseudopregnancy, the levels are lower at implantation sites than at interimplantation sites in day 5 pregnant mice (3). Thus, the regulated levels of anandamide correlate with uterine receptivity and implantation. However, we have recently discovered that the mouse uterus on day 5 of pseudopregnancy is still receptive to implantation. For example, most (86%) of the day 5 pseudopregnant recipients receiving transfer of day 4 blastocysts showed implantation. In contrast, transferred blastocysts completely failed to implant in day 6 pseudopregnant recipient uteri.3 To examine whether uterine anandamide levels correlate with gradual progression of a receptive uterus to a nonreceptive state, we measured uterine anandamide levels on days 4—6 of pseudopregnancy. The results show that anandamide levels are lower on day 4 but increase gradually reaching maximum levels on day 6 (Fig.2 A). These results show that uterine nonreceptivity results in higher anandamide levels, which could be a cause of implantation failure. In fact, experimentally induced, sustained, and higher levels of exogenous cannabinoids postpone implantation via activation of CB1 (4). Higher levels of anandamide during the nonreceptive phase could be due to lower levels of anandamide hydrolase activity in the uterus. Indeed the lower anandamide level at the implantation site is accompanied by higher anandamide hydrolase activity (18). A correlation between decreased anandamide hydrolase activity and increased pregnancy loss in women has recently been reported, suggesting an adverse effect of higher anandamide levels during pregnancy (26).
During delayed implantation blastocysts undergo dormancy, and implantation does not occur (21, 23, 25). We surmised that if the down-regulation of uterine anandamide levels is important for the receptive state, then its levels should be higher in the P4-treated delayed implant uterus and lower in the estrogen-induced receptive uterus. Indeed, uterine anandamide levels were considerably higher in the delayed implant uterus than in the estrogen-induced receptive uterus (Fig. 2 B). These results again suggest that uterine anandamide levels are critical to implantation.
Uterine Anandamide Levels Are Higher in Leukemia Inhibitory Factor-deficient Mice
In the mouse, LIF is expressed in a biphasic manner, first in endometrial glands on the morning of day 4 and then in stromal cells adjacent to blastocysts during the attachment reaction on the night of day 4 (20), suggesting its role in implantation. Indeed, mice deficient in LIF show implantation failure, and blastocysts undergo dormancy (27). This is similar to delayed implantation produced by ovariectomy during normal pregnancy. However,LIF −/− blastocysts implant successfully after transfer into wild-type uteri (27), establishing LIF as an essential maternal factor for uterine preparation prior to implantation. We speculated that uterine anandamide levels should remain high inLIF −/− uteri as compared with wild-type uteri on day 4 of pregnancy. Indeed, uterine levels inLIF −/− mice were significantly higher than those in wild-type mice (Fig. 2 C). Collectively, the above results suggest that regulated levels of anandamide are associated with various phases of uterine receptivity for implantation.
Cannabinoid Receptors Are Higher in Dormant Blastocysts
Because an intimate interaction between the blastocyst and the uterus is essential for implantation, we examined whether uterine anandamide levels at different phases of receptivity correlate with cannabinoid receptor expression in the blastocyst. As previously reported (1, 2), significant levels of [3H]anandamide binding were observed in normal blastocysts collected at 0900 h on day 4 of pregnancy (Fig.3, A, panel a, andB). This binding remarkably decreased in blastocysts recovered at 2000 h on day 4 just prior to the attachment reaction between the blastocyst and the uterine luminal epithelium (Fig. 3,A, panel c, and B). These results suggest that down-regulation of anandamide binding to the blastocyst is important for the initiation of implantation. If this speculation is correct, then anandamide binding should remain high in dormant blastocysts during delayed implantation but diminish with blastocyst activation. Indeed, dormant blastocysts exhibited an increased population of anandamide binding sites (Fig. 3, A, panel e, andB), which significantly diminished by 12 h after termination of the delay by an estrogen injection (Figs. 3, A, panel g, and B). The CB1 immunoreactive protein paralleled anandamide binding in dormant (delayed) and activated blastocysts (Fig. 3 C). As stated before, blastocysts fail to implant and undergo dormancy in LIF −/− mice (20, 27). We observed that the levels of anandamide binding to dormantLIF −/− blastocysts were also higher as compared with active wild-type blastocysts (data not shown). Collectively, these results demonstrate that the decreased levels of cannabinoid receptors in active blastocysts just prior to implantation correlate with decreased levels of anandamide in the receptive uterus.
Cannabinoid Receptor Mutant Mice Are Resistant to Cannabinoid-induced Implantation Failure
We speculated that if tight regulation of ligand-receptor signaling is important for implantation, maintaining a sustained level of exogenously administered cannabinoids should disrupt uterine receptivity for implantation in wild-type mice but not in mutant mice. An active natural cannabinoid, (−)-THC, or its inactive stereoisomer, (+)-THC, was delivered subcutaneously at a constant rate via miniosmotic pumps in wild-type orCB1 −/− /CB2 −/− double mutant mice from days 2—5 of pregnancy. To inhibit rapid systemic degradation of THC by cytochrome P450 enzymes, mice were injected twice daily with P450 inhibitors as previously described (4). The mouse uterus accumulates (−)-THC when its infusion accompanies P450 inhibitors; the levels are below the limit of detection when (−)-THC alone is infused (4). Mice were examined for implantation on day 5 by the blue dye method. As observed previously (4), (−)-THC in the presence of P450 inhibitors inhibited implantation in wild-type mice, and morphologically dormant-looking blastocysts were recovered from these mice. In contrast, similar treatments failed to inhibit implantation inCB1 −/− /CB2 −/− double mutant mice (Table III). Administration of (+)-THC plus the P450 inhibitors or the inhibitors alone had no adverse effects on implantation in wild-type or double mutant mice. These results, indeed, demonstrate that sustained and higher levels of cannabinoids are inhibitory to implantation and that this effect is mediated by cannabinoid receptors because the double mutant mice, but not the wild-type mice, are resistant to this effect.
DISCUSSION
The mechanism(s) by which the development of preimplantation embryos into active blastocysts is synchronized with uterine receptivity for implantation is not clearly understood. Our present observations of the asynchronous development of embryos deficient in cannabinoid receptors during the preimplantation period and of the coordinated down-regulation of both uterine anandamide levels and blastocyst cannabinoid receptors prior to implantation in wild-type mice suggest that ligand-receptor signaling with endocannabinoids locally helps in regulating the "window" of implantation. Although the Mendelian frequency of offspring resulting from heterozygous crossings of CB1-null mice is skewed resulting in a somewhat reduced number of homozygous offspring (19), and although the pregnancy rate in mutant mice resulting from homozygous mating is somewhat lower, the birth of viable CB1 −/−,CB2 −/−, orCB1 −/− /CB2 −/− double mutant offspring suggests that the absence of embryonic and/or uterine cannabinoid receptors is not indispensable for embryonic development or implantation. However, the observed increased mortality ofCB1 −/− offspring perhaps could be due to inferior fetal development resulting from the implantation of slowly developing embryos (19). The major phenotypes ofCB1 −/− orCB1 −/− /CB2 −/− double mutant mice are their resistance to exogenous cannabinoid exposure with respect to embryo development and implantation in vitro andin vivo. This suggests that embryonic development and implantation are likely to be affected by aberrant levels of exogenous or endogenous cannabinoids in the uterus and/or aberrant embryonic expression of cannabinoid receptors during early pregnancy.
On day 5 of pseudopregnancy when the uterus is still receptive,LIF expression persists in uterine glands. In contrast,LIF expression is undetectable or extremely low in the nonreceptive day 6-uterus.3 Moreover, estrogen is essential for the induction of uterine LIF in mice (20, 28). This is consistent with the absence of LIF expression in the P4-primed delayed implant mouse uterus and its rapid induction after an estrogen injection (20). Thus, the virtual absence of uterine LIF on day 6 and during delayed implantation correlates with higher uterine anandamide levels and implantation failure. However, we do not know whether the absence of LIF is the cause of higher uterine anandamide levels or whether the higher levels are the consequence of implantation failure in the absence of LIF.
The physiological significance of anandamide in the uterus and cannabinoid receptors in the blastocyst is still not fully understood. Although it is clear that higher uterine anandamide and blastocyst cannabinoid receptor levels are detrimental to the implantation process, uterine anandamide and blastocyst cannabinoid receptors still persist, albeit at lower levels, at the time of implantation. This suggests that lower levels of anandamide and cannabinoid receptors are beneficial to implantation. This suggestion is consistent with our recent observation that whereas higher anandamide levels are detrimental to blastocyst outgrowth in culture, lower levels stimulate this event (5). Similar biphasic (inhibitory and stimulatory) effects of anandamide at high and low concentrations are evident for other neural and behavioral functions (29), although the definitive cause of these biphasic effects of anandamide is not yet clearly understood (29). More recently, a bidirectional regulation of airway responsiveness by endogenous cannabinoids has been documented (30). Nonetheless, it is envisioned that a biphasic paracrine signaling via anandamide and cannabinoid receptors influences the fate of the embryo-uterine interactions during implantation and that aberrant levels of uterine anandamide and/or embryonic cannabinoid receptors are likely to adversely affect embryonic development and implantation. This could be a mechanism to prevent implantation of abnormal embryos resulting from exposure to aberrant levels of cannabinoids. In conclusion, the present study highlights the importance of the ligand-receptor signaling with cannabinoids in female fertility and places the embryo and/or the uterus as targets for this signaling.
Source, Graphs and Figures: Dysregulated Cannabinoid Signaling Disrupts Uterine Receptivity for Embryo Implantation
The mechanisms by which synchronized embryonic development to the blastocyst stage, preparation of the uterus for the receptive state, and reciprocal embryo-uterine interactions for implantation are coordinated are still unclear. We show in this study that preimplantation embryo development became asynchronous in mice that are deficient in brain-type (CB1) and/or spleen-type (CB2) cannabinoid receptor genes. Furthermore, whereas the levels of uterine anandamide (endocannabinoid) and blastocyst CB1 are coordinately down-regulated with the onset of uterine receptivity and blastocyst activation prior to implantation, these levels remained high in the nonreceptive uterus and in dormant blastocysts during delayed implantation and in pregnant, leukemia inhibitory factor (LIF)-deficient mice with implantation failure. These results suggest that a tight regulation of endocannabinoid signaling is important for synchronizing embryo development with uterine receptivity for implantation. Indeed this is consistent with our finding that while an experimentally induced, sustained level of an exogenously administered, natural cannabinoid inhibited implantation in wild-type mice, it failed to do so inCB1 −/− /CB2 −/−double mutant mice. The present study is clinically important because of the widely debated medicinal use of cannabinoids and their reported adverse effects on pregnancy.
Previous investigation suggested that cannabinoid exposure has adverse effects on pregnancy outcome (reviewed in Refs. 1-5). Cannabinomimetic drugs interact with two types of cannabinoid receptors, brain-type (CB1) and spleen-type (CB2) (6-9). These receptor subtypes are negatively coupled to adenylate cyclase and to N-type and P/Q-type calcium channels and positively coupled to mitogen-activated protein kinase and to A-type potassium channels through Gi/o proteins (7, 10). CB1 is expressed in brain and other peripheral tissues (7-11). CB2 is expressed primarily in immune tissues including the spleen, leukocytes, and tonsils (12). We have previously shown that CB1 is expressed in the preimplantation mouse embryo at much higher levels than in the brain (1, 2). The discovery of cannabinoid receptors led to the identification of endogenous cannabinoid ligands, N-arachidonoylethanolamine (anandamide), and 2-arachidonoylglycerol (13-17). We observed that anandamide is synthesized in the pregnant mouse oviduct and uterus (18) and that its levels are far higher in the uterus than in any other mammalian tissue examined (3). However, anandamide levels are significantly lower at implantation sites than at interimplantation sites, suggesting endocannabinoid ligand-receptor signaling during implantation (3). Indeed anandamide at higher levels adversely affects embryo development and implantation, whereas at lower levels it stimulates embryo growth and differentiation via CB1 (1-5). Interestingly, 2-arachidonoylglycerol is present in the mouse uterus at amounts similar to or lower than the lowest anandamide levels (1—5 nmol/g of tissue), and its level does not vary significantly between implantation and interimplantation sites.1 Using cannabinoid receptor mutant mice (19) and physiological approaches, we further defined the role of ligand-receptor signaling with cannabinoids in preimplantation embryo development and in implantation.
EXPERIMENTAL PROCEDURES
Animals
Adult CD-1 (Charles River Breeding Laboratories),CB1, and CB2 mutant mice on a C57BL/6J background or (LIF)2 mutant mice on a mixed background were housed in the animal care facility at the University of Kansas Medical Center. Story Landis (NINDS, National Institutes of Health) kindly provided LIF mutant mice that were originally generated by Philippe Brulet (Pasteur Institute). We have characterized these mice previously (20). Heterozygous breeding was performed to generate littermate (+/+), (+/−), and (−/−) mice of specific mutation. Homozygous males and females with specific null mutations were crossed to generate homozygous embryos with respective mutations or double mutation. Females were mated with fertile or vasectomized males of the same strain to induce pregnancy or pseudopregnancy (day 1 = vaginal plug), respectively. To induce conditions of delayed implantation, mice were ovariectomized on the morning (0800—0900) of day 4 of pregnancy and maintained with a daily injection (subcutaneously) of progesterone (P4, 2 mg/mouse) from days 5—7. To activate dormant blastocysts and initiate implantation, P4-primed delayed implanting pregnant mice were injected with estradiol-17β (E2, 25 ng/mouse) (21).
Embryo Culture
Two-cell embryos recovered on day 2 were cultured in 25 μl of Whitten's medium in groups of 5—10 under silicon oil in an atmosphere of 5% CO2 in air for 72 h with or without anandamide (7.5 nM) (1, 2). The control cultures contained the same concentration of the vehicle. The embryos were observed every 24 h to monitor their development. At the termination of culture the number of embryos forming blastocysts was recorded.
Anandamide Binding
Autoradiographic ligand binding in blastocysts was performed as previously described (1, 2). Blastocysts were incubated with 4.5 nM [3H]anandamide (specific activity 221 Ci/mmol, PerkinElmer Life Sciences) in the presence or absence of a 500-fold molar excess of unlabeled anandamide. After incubation, blastocysts were washed, fixed in 2% paraformaldehyde, cytospun onto slides, air-dried, and subjected to autoradiography using a liquid emulsion. Blastocysts were poststained in hematoxylin. The autoradiographic signals were quantitated under dark field conditions using an OPTIMA II program with an image analysis system. Statistical analysis was performed using one-way analysis of variance followed by a Newman-Keuls test.
CB1 Immunostaining
Blastocysts were cytospun onto slides, fixed in 2% formalin for 15 min, washed in phosphate-buffered saline, and incubated in blocking solution for 10 min. They were then incubated with an affinity-purified rabbit antipeptide antibody specific to CB1 (1 μg/ml) overnight at 4 °C (2). Immunostaining was performed using a Zymed Laboratories Inc. Histostain-SP kit (2). Blastocysts were counterstained with hematoxylin. Red deposits indicated the sites of immunoreactive CB1 protein. Control blastocysts were incubated parallel with the antibody that was preneutralized with excess antigenic peptide.
Analysis of Anandamide and Other N-Acylethanolamines
Extraction, isolation, and analysis ofN-acylethanolamine (NAE) were done essentially as previously described (3). Uteri were homogenized in chloroform/methanol (2:1), and the homogenate was partitioned against 2.5% aqueous NaCl. Internal standards (N-acyl-1,1,2,2-2H4-ethanolamines with chain lengths of 16:0, 17:0, 18:0, 18:1n-9, 18:2n-6, and 20:4n-6; d 4 NAEs, 0.1 μg each) were added to the extract. The NAEs were isolated using solid phase silica extraction columns (Alltech Associates). Neutral lipids were eluted with 4 ml of chloroform followed by elution of the NAEs with 4 ml of chloroform:methanol (98:2). The NAE fractions were taken to dryness and treated with 50 μl oftert-butyldimethylchlorosilyl/imidazole reagent (Alltech) at 80 °C for 1 h. Thetert-butyldimethylchlorosilyl/imidazole derivatives were extracted into 50 μl of hexane for analysis by gas chromatography-mass spectrophotometry in the selected ion monitoring mode. Endogenous NAEs were quantitated by comparing the peak size of their (M − 57) ions to those of the corresponding stable isotope-labeled internal standards (22). Anandamide accounted for over 90% of total NAEs.
Administration of Drugs
(−)-Δ9-Tetrahydrocannabinol ((−)-THC) or its inactive stereoisomer (+)-THC was delivered at a constant rate via miniosmotic pumps (Alzet Corp., Palo Alto, CA) to study their effects on implantation in wild-type andCB1 −/− /CB2 −/−pregnant mice. Miniosmotic pumps (mean pumping rate of 0.52 μl/h and fill volume of 96 μl) containing either (−)THC or (+)THC were placed under the back skin of mice on day 2 (1000 h) of pregnancy and continued through day 5. These pumps released THC at a rate of 20 μg/h. Cytochrome P450 inhibitors, metyrapone and clotrimazole (50 mg/kg body weight), were first injected intraperitoneally 2 h before the installation of pumps and injected twice daily until day 4 of pregnancy (4). On day 5 (1000 h), implantation sites were determined by the blue dye method (21). If implantation sites were absent, uterine horns were flushed with saline to recover unimplanted blastocysts. Mice without implantation sites or blastocysts were excluded from the experiments.
RESULTS
Embryo Development Is Asynchronous in Cannabinoid Receptor-deficient Mice
Because synchronized embryo development and uterine preparation are important for successful implantation, we first compared the in vivo developmental potential of cannabinoid receptor mutant embryos with wild-type embryos. We observed that the development of CB1 −/−,CB2 −/−, orCB1 −/−/CB2 −/− double mutant embryos recovered from the oviduct on day 3 (TableI) and from the uterus on day 4 (TableII) of pregnancy was asynchronous. On the morning of day 3, about 84% of the wild-type embryos recovered were at the 8-cell stage. In contrast, only about 24% ofCB1 −/−, 42% ofCB2 −/−, and 50% of theCB1 −/−/CB2 −/− double mutant embryos were at the 8-cell stage, resulting in an increased population of 2-, 4-, and 6-cell embryos (Table I). On the morning of day 4, about 98% of the wild-type embryos were blastocysts. However, only about 62% of CB1 −/−, 71% ofCB2 −/−, or 61% ofCB1 −/−/CB2 −/− embryos were at the blastocyst stage; a considerable number of embryos were at the morula stage. However, these morulae were apparently viable and normal because they rapidly developed into blastocysts in culture. We speculated that developmentally retarded embryos eventually form blastocysts and implant. Thus, we next compared the status of implantation in the wild-type and mutant mice on day 5 of pregnancy. Increased localized endometrial vascular permeability at the site of the blastocyst is one of the early markers of implantation (23). This can be monitored by an intravenous injection of a blue dye (Chicago blue B) solution which specifies implantation sites as distinct blue bands along the uterus (4, 21, 23). We observed that whereas all of the vaginal plug-positive wild-type or CB2 −/− mice (n = 7) showed an average of 9 or 7 implantation sites, respectively, 6 of 10 (60%) CB1 −/− and 14 of 18 (78%)CB1 −/− /CB2 −/−plug-positive mice showed an average of 7 and 9 implantation sites per mouse, respectively. These results suggest that the loss of CB1 had modest, if any, adverse effects on implantation. This is perhaps because of the implantation of blastocysts that eventually formed from the slowly growing embryos. Our next objective was to examine whether the mutant embryos are responsive to an endocannabinoid exposurein vitro.
Mutant Embryos Are Resistant to Anandamide
Two-cell wild-type or mutant embryos were cultured in the presence or absence of anandamide as previously described (1, 2). As observed previously (1,2), most (>79%) of the 2-cell wild-type embryos failed to develop to the blastocyst stage in the presence of anandamide. In contrast, more than 80% of the CB1 −/− orCB1 −/− /CB2 −/− double mutant embryos developed into blastocysts in the presence of anandamide (Fig. 1). However, in vitrodevelopment of CB2 −/− embryos, like wild-type embryos, was severely compromised in the presence of anandamide. These results provided genetic confirmation to our previous findings that CB1, but not CB2, in wild-type blastocysts responds to cannabinoid agonists in a stage- and dose-dependent manner in vitro (1-5) although these embryos express CB2mRNA (1). Whereas the developmental response ofCB2 −/− embryos to anandamide is similar to that of wild-type embryos in vitro (1, 2), early asynchronous development of CB2 −/− embryosin vivo, like CB1 −/− orCB1 −/− /CB2 −/−embryos, is surprising. It is possible that the reproductive tract environment differentially modulates the development of mutant embryos in response to endocannabinoids in vivo but not during their development in vitro. In this respect, the mouse uterus expresses the CB1 but not the CB2 gene (24), and thus uterine CB1 may influence embryonic development in vivoin the absence of embryonic CB2. Another possibility could be that uterine anandamide levels in mutant mice are different from those of the wild-type mice. However, uterine anandamide levels inCB1 −/− /CB2 −/− double mutant mice are not significantly different from the wild-type uterine levels.1 Nonetheless, the present genetic evidence suggests that cannabinoid receptors have some role in synchronizing embryo development in the reproductive tract in vivo.
Anandamide Levels Are Higher in the Nonreceptive Uterus
In the mouse, progesterone (P4) and estrogen sequentially program the uterus into prereceptive, receptive, and nonreceptive phases during pregnancy or pseudopregnancy (23, 25). Blastocysts implant only in the receptive uterus. The P4-primed uterus becomes receptive on day 4 when it is superimposed with preimplantation ovarian estrogen. Subsequently the receptive uterus proceeds to the nonreceptive phase, and blastocysts can no longer implant (21, 25). Similar uterine changes are experimentally produced during delayed implantation induced by ovariectomy before preimplantation ovarian estrogen secretion on day 4. Under this condition, blastocysts undergo dormancy and fail to implant. This condition is terminated by an injection of estrogen with activation of dormant blastocysts, attainment of the receptive uterus, and implantation (21, 23, 25). The mechanism by which estrogen initiates these events in the P4-primed uterus is not clearly understood.
The mouse uterus is receptive on day 4, whereas it was considered to be nonreceptive on day 5 (21). We previously observed that not only are uterine anandamide levels higher on day 5 of pseudopregnancy, the levels are lower at implantation sites than at interimplantation sites in day 5 pregnant mice (3). Thus, the regulated levels of anandamide correlate with uterine receptivity and implantation. However, we have recently discovered that the mouse uterus on day 5 of pseudopregnancy is still receptive to implantation. For example, most (86%) of the day 5 pseudopregnant recipients receiving transfer of day 4 blastocysts showed implantation. In contrast, transferred blastocysts completely failed to implant in day 6 pseudopregnant recipient uteri.3 To examine whether uterine anandamide levels correlate with gradual progression of a receptive uterus to a nonreceptive state, we measured uterine anandamide levels on days 4—6 of pseudopregnancy. The results show that anandamide levels are lower on day 4 but increase gradually reaching maximum levels on day 6 (Fig.2 A). These results show that uterine nonreceptivity results in higher anandamide levels, which could be a cause of implantation failure. In fact, experimentally induced, sustained, and higher levels of exogenous cannabinoids postpone implantation via activation of CB1 (4). Higher levels of anandamide during the nonreceptive phase could be due to lower levels of anandamide hydrolase activity in the uterus. Indeed the lower anandamide level at the implantation site is accompanied by higher anandamide hydrolase activity (18). A correlation between decreased anandamide hydrolase activity and increased pregnancy loss in women has recently been reported, suggesting an adverse effect of higher anandamide levels during pregnancy (26).
During delayed implantation blastocysts undergo dormancy, and implantation does not occur (21, 23, 25). We surmised that if the down-regulation of uterine anandamide levels is important for the receptive state, then its levels should be higher in the P4-treated delayed implant uterus and lower in the estrogen-induced receptive uterus. Indeed, uterine anandamide levels were considerably higher in the delayed implant uterus than in the estrogen-induced receptive uterus (Fig. 2 B). These results again suggest that uterine anandamide levels are critical to implantation.
Uterine Anandamide Levels Are Higher in Leukemia Inhibitory Factor-deficient Mice
In the mouse, LIF is expressed in a biphasic manner, first in endometrial glands on the morning of day 4 and then in stromal cells adjacent to blastocysts during the attachment reaction on the night of day 4 (20), suggesting its role in implantation. Indeed, mice deficient in LIF show implantation failure, and blastocysts undergo dormancy (27). This is similar to delayed implantation produced by ovariectomy during normal pregnancy. However,LIF −/− blastocysts implant successfully after transfer into wild-type uteri (27), establishing LIF as an essential maternal factor for uterine preparation prior to implantation. We speculated that uterine anandamide levels should remain high inLIF −/− uteri as compared with wild-type uteri on day 4 of pregnancy. Indeed, uterine levels inLIF −/− mice were significantly higher than those in wild-type mice (Fig. 2 C). Collectively, the above results suggest that regulated levels of anandamide are associated with various phases of uterine receptivity for implantation.
Cannabinoid Receptors Are Higher in Dormant Blastocysts
Because an intimate interaction between the blastocyst and the uterus is essential for implantation, we examined whether uterine anandamide levels at different phases of receptivity correlate with cannabinoid receptor expression in the blastocyst. As previously reported (1, 2), significant levels of [3H]anandamide binding were observed in normal blastocysts collected at 0900 h on day 4 of pregnancy (Fig.3, A, panel a, andB). This binding remarkably decreased in blastocysts recovered at 2000 h on day 4 just prior to the attachment reaction between the blastocyst and the uterine luminal epithelium (Fig. 3,A, panel c, and B). These results suggest that down-regulation of anandamide binding to the blastocyst is important for the initiation of implantation. If this speculation is correct, then anandamide binding should remain high in dormant blastocysts during delayed implantation but diminish with blastocyst activation. Indeed, dormant blastocysts exhibited an increased population of anandamide binding sites (Fig. 3, A, panel e, andB), which significantly diminished by 12 h after termination of the delay by an estrogen injection (Figs. 3, A, panel g, and B). The CB1 immunoreactive protein paralleled anandamide binding in dormant (delayed) and activated blastocysts (Fig. 3 C). As stated before, blastocysts fail to implant and undergo dormancy in LIF −/− mice (20, 27). We observed that the levels of anandamide binding to dormantLIF −/− blastocysts were also higher as compared with active wild-type blastocysts (data not shown). Collectively, these results demonstrate that the decreased levels of cannabinoid receptors in active blastocysts just prior to implantation correlate with decreased levels of anandamide in the receptive uterus.
Cannabinoid Receptor Mutant Mice Are Resistant to Cannabinoid-induced Implantation Failure
We speculated that if tight regulation of ligand-receptor signaling is important for implantation, maintaining a sustained level of exogenously administered cannabinoids should disrupt uterine receptivity for implantation in wild-type mice but not in mutant mice. An active natural cannabinoid, (−)-THC, or its inactive stereoisomer, (+)-THC, was delivered subcutaneously at a constant rate via miniosmotic pumps in wild-type orCB1 −/− /CB2 −/− double mutant mice from days 2—5 of pregnancy. To inhibit rapid systemic degradation of THC by cytochrome P450 enzymes, mice were injected twice daily with P450 inhibitors as previously described (4). The mouse uterus accumulates (−)-THC when its infusion accompanies P450 inhibitors; the levels are below the limit of detection when (−)-THC alone is infused (4). Mice were examined for implantation on day 5 by the blue dye method. As observed previously (4), (−)-THC in the presence of P450 inhibitors inhibited implantation in wild-type mice, and morphologically dormant-looking blastocysts were recovered from these mice. In contrast, similar treatments failed to inhibit implantation inCB1 −/− /CB2 −/− double mutant mice (Table III). Administration of (+)-THC plus the P450 inhibitors or the inhibitors alone had no adverse effects on implantation in wild-type or double mutant mice. These results, indeed, demonstrate that sustained and higher levels of cannabinoids are inhibitory to implantation and that this effect is mediated by cannabinoid receptors because the double mutant mice, but not the wild-type mice, are resistant to this effect.
DISCUSSION
The mechanism(s) by which the development of preimplantation embryos into active blastocysts is synchronized with uterine receptivity for implantation is not clearly understood. Our present observations of the asynchronous development of embryos deficient in cannabinoid receptors during the preimplantation period and of the coordinated down-regulation of both uterine anandamide levels and blastocyst cannabinoid receptors prior to implantation in wild-type mice suggest that ligand-receptor signaling with endocannabinoids locally helps in regulating the "window" of implantation. Although the Mendelian frequency of offspring resulting from heterozygous crossings of CB1-null mice is skewed resulting in a somewhat reduced number of homozygous offspring (19), and although the pregnancy rate in mutant mice resulting from homozygous mating is somewhat lower, the birth of viable CB1 −/−,CB2 −/−, orCB1 −/− /CB2 −/− double mutant offspring suggests that the absence of embryonic and/or uterine cannabinoid receptors is not indispensable for embryonic development or implantation. However, the observed increased mortality ofCB1 −/− offspring perhaps could be due to inferior fetal development resulting from the implantation of slowly developing embryos (19). The major phenotypes ofCB1 −/− orCB1 −/− /CB2 −/− double mutant mice are their resistance to exogenous cannabinoid exposure with respect to embryo development and implantation in vitro andin vivo. This suggests that embryonic development and implantation are likely to be affected by aberrant levels of exogenous or endogenous cannabinoids in the uterus and/or aberrant embryonic expression of cannabinoid receptors during early pregnancy.
On day 5 of pseudopregnancy when the uterus is still receptive,LIF expression persists in uterine glands. In contrast,LIF expression is undetectable or extremely low in the nonreceptive day 6-uterus.3 Moreover, estrogen is essential for the induction of uterine LIF in mice (20, 28). This is consistent with the absence of LIF expression in the P4-primed delayed implant mouse uterus and its rapid induction after an estrogen injection (20). Thus, the virtual absence of uterine LIF on day 6 and during delayed implantation correlates with higher uterine anandamide levels and implantation failure. However, we do not know whether the absence of LIF is the cause of higher uterine anandamide levels or whether the higher levels are the consequence of implantation failure in the absence of LIF.
The physiological significance of anandamide in the uterus and cannabinoid receptors in the blastocyst is still not fully understood. Although it is clear that higher uterine anandamide and blastocyst cannabinoid receptor levels are detrimental to the implantation process, uterine anandamide and blastocyst cannabinoid receptors still persist, albeit at lower levels, at the time of implantation. This suggests that lower levels of anandamide and cannabinoid receptors are beneficial to implantation. This suggestion is consistent with our recent observation that whereas higher anandamide levels are detrimental to blastocyst outgrowth in culture, lower levels stimulate this event (5). Similar biphasic (inhibitory and stimulatory) effects of anandamide at high and low concentrations are evident for other neural and behavioral functions (29), although the definitive cause of these biphasic effects of anandamide is not yet clearly understood (29). More recently, a bidirectional regulation of airway responsiveness by endogenous cannabinoids has been documented (30). Nonetheless, it is envisioned that a biphasic paracrine signaling via anandamide and cannabinoid receptors influences the fate of the embryo-uterine interactions during implantation and that aberrant levels of uterine anandamide and/or embryonic cannabinoid receptors are likely to adversely affect embryonic development and implantation. This could be a mechanism to prevent implantation of abnormal embryos resulting from exposure to aberrant levels of cannabinoids. In conclusion, the present study highlights the importance of the ligand-receptor signaling with cannabinoids in female fertility and places the embryo and/or the uterus as targets for this signaling.
Source, Graphs and Figures: Dysregulated Cannabinoid Signaling Disrupts Uterine Receptivity for Embryo Implantation
