Main

Both the medicinal and recreational use of marijuana (cannabis) are increasing worldwide1. It is one of the most popular illicit drugs used by pregnant women2, raising concern about marijuana's adverse effects during pregnancy. Maternal use of marijuana is associated with reduced birth weight as well as cognitive and memory deficits in offspring3,4. The mechanism by which marijuana exerts its effects was first discovered in the 1990s, when it was shown that active constituents of marijuana mediate most of their effects through two G-protein-coupled receptors, cannabinoid receptor 1 (CB1) and cannabinoid receptor 2 (CB2), encoded by the genes Cnr1 and Cnr2, respectively5,6. This led to the identification of two endogenous ligands (endocannabinoids), anandamide (N-arachidonoylethano-lamine)7 and 2-arachidonoyl glycerol8,9,10, for these receptors. These findings established that endocannabinoid signaling occurs in many central and peripheral systems11,12,13. With respect to reproduction, we first provided evidence in mice for the presence of CB1 in preimplantation embryos14 and anandamide in the oviduct and uterus15,16, suggesting that endocannabinoid signaling is operative during early pregnancy. This was further supported by our findings of biphasic effects of anandamide on embryo development and implantation17. For example, anandamide at a low concentration makes the blastocyst competent for implantation, whereas at a higher concentration it dampens this response17. Cannabinoids also influence reproductive functions in invertebrates18,19. Collectively, these observations led to studies in humans showing the presence of CB1 in the uterus20 and an association of spontaneous pregnancy loss with higher anandamide levels21.

Our results in mice, however, could not explain the underlying causes of pregnancy loss in a substantial number of Cnr1−/− mice that we observed during the course of our studies for many years. We show here that subfertility in Cnr1−/− mice is due to oviductal retention of embryos. Our results show that oviductal embryo transport during early pregnancy is coordinated by endocannabinoid signaling through oviductal CB1 in collaboration with adrenergic receptor signaling. This finding has important implications for ectopic pregnancy in women because one major cause of tubal pregnancy is embryo retention in the fallopian tube. Embryo retention has become an issue of concern in light of the recent evidence that marijuana use or dependence in adults aged 18 and over in the United States has increased by 22% since 1991 (ref. 22).

Results

CB1 deficiency causes early pregnancy loss

During the past several years, we have observed that ∼40% of Cnr1−/− mice show pregnancy loss (Fig. 1a). Ovulation, fertilization, embryo development, timely passage of embryos through the oviduct and implantation all contribute to the success of pregnancy. We showed that ovulation and fertilization, as assessed by the number of ovulated eggs and two-cell embryos, respectively, are normal in Cnr1−/− or Cnr1−/−/Cnr2−/− mice compared with wild-type females (Supplementary Fig. 1a online). Development of Cnr1−/− or Cnr1−/−/Cnr2−/− embryos was asynchronous when observed in the morning on day 4 of pregnancy (Supplementary Fig. 1b online). Normally, over 85% of the wild-type embryos recovered from the uterus at this time are blastocysts. Only 64% of Cnr1−/− and 58% of Cnr1−/−/Cnr2−/− embryos were at the blastocyst stage and a large number of embryos were at the morula stage. We also failed to recover embryos from uteri in a substantial number (∼35%) of vaginal plug-positive Cnr1−/− mice when examined on day 4 (Supplementary Fig. 1c online). This asynchronous development of Cnr1−/− embryos may account for the observation of underrepresented mendelian frequency of Cnr1−/− offspring from heterozygous matings23. We speculated that the loss of embryonic CB1 contributes to these defects in Cnr1−/− mice. This was addressed by mating Cnr1−/− females with wild-type males to generate all heterozygous embryos. We observed that development of heterozygous embryos is normal in mutant mothers (Supplementary Fig. 1b online), suggesting that embryonic CB1 is important for normal development. But a considerable number (36%) of mutant mothers still did not yield any embryos in the uterus (Supplementary Fig. 1c online).

Figure 1: Impaired oviductal embryo transport causes pregnancy loss in Cnr1−/− mice.
figure 1

(a) Pregnancy loss frequently occurs in Cnr1−/− mice. The litter size is also smaller in Cnr1−/− mice giving birth, suggesting loss of embryos during pregnancy (unpaired t-test, *P < 0.01). Numbers within bars indicate number of mice with failed pregnancy/number of vaginal plug–positive mice examined. (b) Number of mice with oviductal retention of embryos/total number of mice examined. (c) Percentage of embryos recovered from oviducts or uteri. (d) Morphology of embryos recovered from oviducts or uteri. Both morulae and blastocysts are recovered from oviducts or uteri of Cnr1−/− mice on day 4 midmornings. Bar, 50 μm. (e) Implantation in Cnr1−/− mice. Numbers within bars in e indicate number of mice without implantation sites/total number of mice examined. (f) Implantation failure was noted in many Cnr1−/− mice on day 5 of pregnancy with unimplanted embryos recovered from oviducts or uteri. Bar, 50 μm. (g) A representative histological section of a day 7 pregnant Cnr1−/− oviduct showing a trapped blastocyst (Bl, arrow) at the isthmus. Bar, 100 μm. (h) Trapped blastocysts recovered from Cnr1−/− oviducts on day 5 show implantation after transfer into a day 4 wild-type pseudopregnant recipient. Blue bands (arrows), sites of implantation. Arrowheads, sites of embryo transfer. WT, wild-type; Bl, blastocyst; Mus, muscularis; S, serosa; Mu, mucosa.

Full size image

CB1 deficiency derails oviductal embryo transport

To determine the underlying cause of this observation, we examined oviductal embryo transport in Cnr1−/− females. As expected, a considerable number of Cnr1−/− and Cnr1−/−/Cnr2−/− mice, but not Cnr2−/− or wild-type mice, did not yield any embryos when their uterine lumens were flushed at midmorning of day 4 of pregnancy (Fig. 1b). However, embryos at the morula and blastocyst stages were recovered from oviducts of these Cnr1−/− or Cnr1−/−/Cnr2−/− mice (Fig. 1c,d). Normally implantation is initiated at midnight of day 4 and implantation sites are detected as discrete blue bands along the uterus after an intravenous injection of a blue dye solution24. On day 5 of pregnancy, implantation sites (blue bands) become more distinct. A large number (∼35%) of Cnr1−/− mice did not show any sign of implantation when examined on this day (Fig. 1e). Unimplanted blastocysts were frequently recovered from uteri of these mutant females (Fig. 1f). This provided evidence that oviductal embryo transport was delayed, preventing timely implantation in the uterus. Furthermore, when examined on day 7 of pregnancy, 29 embryos were still retained in oviducts of 5 (28%) of 18 Cnr1−/− mice. In fact, zona-free blastocysts were found trapped within Cnr1−/− oviducts at the isthmus-uterine junction (Fig. 1g); no embryos were retained at the ampulla. These trapped blastocysts appeared morphologically normal and they implanted upon transfer to day 4 pseudopregnant receptive uteri (Fig. 1h), suggesting that they remained implantation-competent. These results indicate that many embryos remain trapped in the oviduct and some are trapped for an extended period, resulting in pregnancy failure and contributing to reduced fertility in Cnr1−/− mice.

CB1 and CB2 differentially contribute to pregnancy loss

The above results led us to speculate that the oviduct is a target for cannabinoid signaling. Thus, in wild-type oviducts we examined the expression of Mbldc1, which encodes N-acyl-phosphatidylethanolamine-hydrolyzing phospholipase D (NAPE-PLD), an enzyme considered specific for anandamide synthesis in vivo25, and of Cnr1 and Cnr2 using RT-PCR. We found that Mbldc1 and Cnr1, but not Cnr2, are expressed in oviducts on days 1–4 of pregnancy (Fig. 2a–c). Furthermore, analysis by reverse-phase positive-ion ESI-HPLC-MS-MS detected anandamide in day 3 oviducts (66.6 ± 2.6 pmol/g, n = 3; each assay was comprised of oviducts pooled from 13–15 mice). These results are consistent with a previous observation of the presence of anandamide in human oviductal secretions26, suggesting that the oviduct is indeed a target for endocannabinoid signaling to influence embryo transport. To obtain genetic evidence for the relative importance of maternal versus embryonic CB1, and of CB1 versus CB2 in oviductal embryo retention, we used reciprocal embryo transfers between Cnr mutant and wild-type mice. As shown in Table 1, whereas normal implantation occurred after transfer of wild-type embryos to wild-type or Cnr2−/− recipients, a considerable number (28%) of wild-type embryos were retained in Cnr1−/− recipient oviducts on day 5. In addition, another 12% of transferred embryos did not show timely implantation in Cnr1−/− uteri and unimplanted blastocysts were recovered upon uterine flushing, indicating delayed oviductal embryo transport. In contrast, Cnr1−/− or Cnr2−/− embryos implanted normally after transfer into wild-type recipients (Table 1). These results mimic the phenotypes of oviductal embryo retention in Cnr1−/− mice after natural mating, demonstrating that this effect is independent of embryonic CB1 or CB2, or both. Collectively, these results provide genetic evidence that although embryonic CB1 primarily contributes to normal embryo development, oviductal CB1 directs the timely oviductal transport of embryos.

Figure 2: Analysis of Mbldc1, Cnr1 and Cnr2 mRNAs in the oviduct.
figure 2

(a) Southern blot analysis of RT-PCR amplified products of Mbldc1 (449 bp). Lanes: 1, mouse brain (positive control); 2–5, mouse oviducts from days 1–4 of pregnancy, respectively; 6, brain RT negative control; 7, primer control. (b) Southern blot analysis of RT-PCR amplified products of Cnr1 (284 bp). Lanes: 1, mouse brain (positive control); 2–5, mouse oviducts from days 1–4 of pregnancy, respectively; 6, brain RT negative control; 7, primer control. (c) Southern blot analysis of RT-PCR amplified products of Cnr2 (182 bp). Lanes: 1, mouse spleen (positive control); 2–5, mouse oviducts from days 1–4 of pregnancy, respectively; 6, spleen RT negative control; 7, primer control.

Full size image
Table 1 Reciprocal embryo transfers between the Cnr mutant and wild-type mice
Full size table

Aberrant cannabinoid signaling impairs embryo transport

To further understand the physiological relevance of cannabinoid signaling in embryo development and oviductal transport, we examined the consequences of either silencing or enhancing cannabinoid signaling in wild-type mice by sustained infusion through miniosmotic pumps of selective cannabinoid antagonists or a cannabinoid agonist, respectively. For example, when examined at midmorning on day 4, 70% of wild-type mice treated with a CB1 selective antagonist (SR141716) showed oviductal retention of embryos with asynchronous development, as opposed to none treated with the vehicle or a CB2 selective antagonist (SR144528) (Fig. 3a–c). This high rate of oviductal embryo retention after pharmacological intervention of CB1 in wild-type mice is notable, as compared to the retention rate (30–40%) in Cnr1−/− mice. This effect was specific, because similar treatment of Cnr1−/− mice with SR141716 did not further increase the rate of oviductal embryo retention; only 3 of 9 (33%) treated mice showed embryo retention. Moreover, pretreatment with methanandamide substantially reversed the effects of SR141716 on oviductal embryo transport in wild-type mice in 1; only 1 of 8 (13%) treated mice showed embryo retention. These data reinforce the essential role of CB1 for normal journeying of embryos through the oviduct. We speculate that an acute inactivation of CB1 by SR141716 is more effective than the genetic ablation of the receptor, perhaps owing to a compensatory adaptation that occurs in Cnr1−/− mice with the loss of CB1 throughout their life span. Indeed, specific effects on ethanol consumption and dependence produced by an acute blockade of CB1 by SR141716 are not similar to those observed in Cnr1−/− mice27. There is also evidence that an acute silencing of a gene is more effective in presenting phenotypes than the chronic silencing of the same gene28,29.

Figure 3: Aberrant cannabinoid signaling causes impaired oviductal embryo transport in wild-type mice.
figure 3

(a) Number of mice with embryos retained in the oviduct/total number of mice examined. (b) Percentage of embryos recovered from oviducts or uteri of mice with different treatments. (c) Differential distribution of morulae and blastocysts among the drug-treated groups. Oviductal retention of embryos was evidenced in wild-type mice by either silencing of CB1 signaling by SR141716 or enhancement of cannabinoid signaling by sustained infusion of methanandamide. Statistical significance between treatment groups was evaluated using unpaired t-test (*P < 0.01). Meth-ANA, methanandamide.

Full size image

Anandamide shows biphasic effects in that it accelerates blastocyst competency to implantation at a low concentration, but inhibits it at a higher concentration17. Thus, we asked whether exaggerated cannabinoid signaling would interfere with oviductal embryo transport. We observed that a large number (56%) of wild-type pregnant mice exposed to sustained infusion of methanandamide, a stable anandamide analog, showed oviductal retention of embryos (Fig. 3a–c). These findings suggest that during normal pregnancy oviductal CB1 interacting with locally produced endocannabinoids elicits an 'endocannabinoid tone' conducive to oviductal embryo transport, but an aberrant cannabinoid signaling, either blockade or enhancement, disrupts this tone, leading to oviductal embryo retention.

Cannabinoid-adrenergic signaling guides embryo transport

In rodents, embryos transit rapidly through the oviduct ampulla by the forward-moving beatings of the cilia present on the epithelial cell surface. Transport of embryos is slowed once they reach the ampulla-isthmus junction, and they reside at the isthmus for approximately 3 d. Embryos journey from this region into the uterus aided by a wave of regulated contraction and relaxation of the oviduct muscularis. Thus, a combination of ciliary and muscular activity contributes to the overall success of oviductal embryo transport30. Because embryos are not trapped at the oviduct ampullary region in Cnr1−/− females, we believe that the impaired oviductal embryo transport does not result from defective ciliary function. It is thought that the sympathetic neuronal circuitry, under the direction of ovarian hormones, coordinates the 'closing and opening' of the sphincter at the isthmus-uterine junction, thereby regulating the timely passage of embryos from the oviduct into the uterus31,32. During pregnancy, rising progesterone levels from the newly formed corpus lutea decrease the levels of noradrenaline, a ligand with higher affinity for the α-adrenergic receptor (AR) than the β-AR at the adrenergic nerve endings, and its turnover rates33. In contrast, the sensitivity of the β-AR is increased in the circular muscle of the oviduct isthmus under progesterone dominance, causing muscle relaxation and facilitating embryo transport through the oviduct31.

We found that oviductal smooth muscle in both wild-type and Cnr1−/− mice contains adrenergic innervation (data not shown). More importantly, CB1 is colocalized in the oviductal muscularis at the isthmus region with β2-AR (Fig. 4a), the predominant subtype in the rodent oviduct34, suggesting that endocannabinoid signaling is functionally coupled to adrenergic signaling to regulate oviductal motility conducive to embryo transport. We speculated that oviductal embryo retention in Cnr1−/− mice is due to increased release of noradrenaline at the oviductal nerve terminals, leading to heightened smooth muscle contractility through α-AR and impairing embryo transport through the isthmus-uterine junction. Indeed, noradrenaline overflow from the superfused Cnr1−/− but not Cnr2−/− oviducts, preloaded with [3H]noradrenaline, exceeded that of wild-type oviducts by 33% (P < 0.05) after challenged with increasing concentration of potassium chloride (KCl) (4.8–50 mM), although preloaded tissue [3H]noradrenaline content of oviduct preparations from wild-type, Cnr1−/− or Cnr2−/− mice was similar after 30 min of preincu-bation with [3H]noradrenaline (Fig. 4b, c). Furthermore, although KCl-evoked [3H]noradrenaline overflow was inhibited by methanandamide over 40% (P < 0.01) in wild-type or Cnr2−/− oviduct preparations, this agonist did not influence the overflow in Cnr1−/− oviducts (Fig. 4d). In contrast, SR141716, but not SR144528, facilitated KCl-evoked [3H]noradrenaline overflow by 37% (P < 0.05) in wild-type oviducts (Fig. 4d). These results suggest that increased noradrenaline release from oviductal adrenergic nerve terminals in the absence of CB1 maintains a smooth muscle contractile tone through α-AR, thereby impeding oviductal embryo transport. In contrast, impaired oviductal embryo transport in wild-type mice with sustained infusion of methanandamide suggests that the relaxation phase of the oviductal muscularis dominates as a result of attenuated noradrenaline release, impairing embryo transport. Collectively, these observations reinforce the idea that an aberrant cannabinoid signaling, arising from either silencing or amplification, impedes the highly coordinated oviductal smooth muscle contraction and relaxation that are critical to normal embryo transport during early pregnancy. We speculated that balancing the signaling by α- and β-AR would help in normalizing oviductal embryo transport in Cnr1−/− mice. Indeed, administration of isoproterenol, a β-AR agonist, restores normal embryo transport in Cnr1−/− mice (Fig. 5a,b). However, similar treatment with a α1-AR agonist phenylephrine, alone or in combination with a β2-AR antagonist butoxamine, in wild-type mice showed oviductal retention of embryos, and as we had speculated, the response was more pronounced after the combined treatment (Supplementary Table 1 online).

Figure 4: Lack of CB1 increases oviductal nor-adrenaline release.
figure 4

(a) Colocalization of CB1 and β2-AR in the oviduct. Red, TRITC-labeled CB1; green, FITC-labeled β2-AR; yellow, merged images CB1 and β2-AR are colocalized in oviductal mus-cularis at the isthmus with undetectable localiza-tion in the ampulla. Bar, 100 μm. (b) Differences in oviductal [3H]noradrenaline contents are insig-nificant among wild-type (WT), Cnr1−/− or Cnr2−/− mice after preincubation with [3H]noradrenaline. (c) Cnr1−/− oviducts release higher amounts of [3H]noradrenaline after stimulation with increasing concentration of KCl (4.8–50μM) (*P < 0.05, unpaired t-test). (d) Effects of cannabinoid agonist or antagonists on KCl-evoked [3H]noradrenaline release. Inhibition of [3H]noradrenaline overflow by methanandamide is noted in oviduct preparations of WT and Cnr2−/− mice, but not those from Cnr1−/− mice (**P < 0.01). In contrast, SR141716 enhances [3H]noradrenaline overflow in WT oviducts (*P < 0.05). No significant changes in [3H]noradrenaline overflow are noted in Cnr1−/− oviducts after exposure to vehicle or Meth-ANA. The values (mean ± s.e.m.) represent the ratio of [3H]noradrenaline overflow evoked by the second (S2) over the first (S1) KCl stimulation. Numbers within bars indicate the number of oviducts exam-ined. Mus muscularis; S, serosa; Mu, mucosa; WT, wild-type; Meth-ANA, methanandamide.

Full size image
Figure 5: Normal embryo development and transport in Cnr1−/− mice are restored by prostaglandins and β-AR signaling, respectively.
figure 5

(a) Number of mice with embryos retained in oviducts/total number of mice examined. (b) Percentage of embryos recovered from oviducts or uteri. (c) Differential distribution of morulae and blastocysts for different treatments. Normal Cnr1−/− embryo development is noted after supplementation with PGE2+cPGI, whereas isoproterenol treatment restores normal embryo transport in Cnr1−/− mice (*P < 0.01).

Full size image

In addition to signaling through adrenergic nerves33, the embryo's journey through the oviduct during pregnancy is influenced by other factors including ovarian progesterone, estrogen35 and prostaglandins36. Thus, we examined whether supplementation with steroids or prostaglandins would restore oviductal embryo transport in Cnr1−/− mice. Progesterone or estrogen individually (Supplementary Table 2 online) or in combination did not rescue either the asynchronous embryo development or impaired oviductal transport in Cnr1−/− mice (Fig. 5a,b), suggesting that these defects are not the result of reduced levels of ovarian steroids. Prostacyclin (PGI2) and prostaglandin E2 (PGE2) are major prostaglandins produced in oviducts and have beneficial effects on embryo development37. We found that sustained infusion of cPGI (a stable PGI2 analog) and PGE2 restores normal development of Cnr1−/− embryos, but the same treatment does not restore normal oviductal embryo transport (Fig. 5a–c). Cannabinoid signaling influences embryo development and function in a biphasic and stage-dependent manner through embryonic CB1 involving cAMP, MAPK and calcium signaling pathways15,17. It is possible that these signaling pathways become aberrant in the absence of CB1, leading to asynchronous embryo development. Whether endocannabinoid signaling collaborates with prostaglandin signaling or works independently in embryo development is not known. Nonetheless, the present observation of normal development of Cnr1−/− embryos, but not the defective oviductal embryo transport, by prostaglandins suggests that oviductal embryo retention in the absence of CB1 is independent of embryonic developmental stages. These data support our new concept that CB1 maintains an endocannabinoid tone in the oviduct that appropriately regulates the activity of neurotransmitters, ensuring normal passage of embryos through the oviduct for their timely homing into the uterus for implantation.

Discussion

Using multiple approaches, we show here that aberrant cannabinoid signaling prevents oviductal embryo transport, ultimately leading to pregnancy failures in a large number of mice. Although there is no evidence of ectopic pregnancy in organisms other than humans, these results have high clinical importance because embryo retention in the fallopian tube as a result of dysfunctional muscular contraction is one cause of ectopic pregnancy in women, the incidence of which has significantly increased during the past decade38. The ability of trapped blastocysts within Cnr1−/− oviducts to implant after transfer into wild-type pseudopregnant recipient uteri suggests that similar embryo retention in the fallopian tube would result in ectopic pregnancy in women. Although it is a major health issue, functional studies underlying the mechanism of ectopic pregnancy are lacking. Our report describes a genetic model to understand the molecular mechanism of oviductal embryonic transport. Our results show that genetic or pharmacological intervention of evolutionarily conserved G-protein-coupled CB1 enhances the oviductal noradrenaline release, thereby impeding embryo transport. There is evidence that presynaptic noradrenaline release in the vas deferens is subject to oscillation by an endocannabinoid tone through CB139. Anandamide also produces a nitric oxide–mediated inhibition of KCl-stimulated noradrenaline release from renal arterial sympathetic nerves40. Our present observations suggest that one function of oviductal CB1 is to regulate presynaptic release of noradrenaline that works through α- or β-AR, or both, in coordinating smooth muscle contraction and relaxation waves for normal embryo transport. We propose that CB1 serves as a presynaptic receptor and maintains an endogenous tone, regulating neurotransmitter release that coordinates oviductal muscle contraction and relaxation to facilitate embryo transport. In humans, cannabinoid signaling through CB1 exerts a direct relaxant effect on the pregnant myometrium20. It is interesting to note that cigarette smoking is thought to increase the risk of ectopic pregnancy41, which is consistent with impaired oviductal transport of rat embryos by nicotine42. There is evidence that chronic exposure of rats to nicotine increases or decreases brain endocannabinoid levels in a region-specific manner43. Whether there is any crosstalk between the G-protein-coupled nicotinic receptors with CB1 in female reproductive functions is not known.

Our work shows that under normal physiological conditions, endocannabinoid signaling through CB1 is crucial to various female reproductive functions that include development of embryos, their oviductal transport and ultimately their homing and implantation in the receptive uterus. Conversely, an aberration in endocannabinoid signaling, either silenced or enhanced, derails these processes. Our study adds a new dimension to the concern that the adverse effects of maternal use of cannabinoids on offspring may be seeded very early in pregnancy. There is now evidence that defective implantation creates adverse effects during the subsequent course of pregnancy both in humans and mice44,45. Collectively, our results raise caution for women of reproductive ages regarding chronic marijuana consumption for recreation or pain alleviation.

There is evidence that polymorphism of the CNR1 gene occurs in humans46 and is associated with variable drug dependency47. It remains to be seen whether the incidence of ectopic pregnancy is associated with Cnr1 gene polymorphism, mutation or chronic use of cannabinoids or anticannabinoids. Because genetic background–dependent phenotype variations are observed for Cnr1−/− mice23,48, our studies on Cnr1 mutant mice on a C57BL/6J background need to be expanded on other genetic backgrounds to ascertain whether impaired oviductal embryo transport is a function of genetic makeup.

Methods

Mice.

Cnr-deficient mice on a C57BL/6J genetic background were generated as previously described23,49. Adult wild-type, Cnr1−/−, Cnr2−/− and Cnr1−/−/Cnr2−/− mice were housed in the Institutional Animal Care Facility according to National Institutes of Health and institutional guidelines for laboratory animals. Females were mated with fertile or vasectomized males of the same strain to induce pregnancy or pseudopregnancy, respectively. The morning (09:00 h) on which a vaginal plug was found was considered day 1 of pregnancy or pseudopregnancy.

In vivo delivery of drugs.

SR141716, SR144528 (National Institute on Drug Abuse), methanandamide, or PGE2+cPGI (Cayman Chemical) was dissolved in propylene glycol to desired concentrations: 20, 20, 40 or 5+5 mg/ml, respectively. Miniosmotic pumps (mean fill volume 96 μl and pumping rate 0.52 μl/h) loaded with each drug were placed subcutaneously on day 1 (1000 h) of pregnancy and continued through day 4. In another set of experiments, day 1 pregnant wild-type mice were first placed on methanandamide-filled pumps for 1 h, and then coinfused with pumps containing SR141716 to examine the specificity of this CB1 antagonist. To examine the effects of ovarian steroids on oviductal embryo transport, pregnant Cnr1−/− mice were injected subcutaneously with estradiol-17β (3 ng/mouse) on days 1–2 and progesterone (1 mg/mouse) on days 2–3. To examine the role of adrenergic signaling on oviductal embryo transport in Cnr1−/− mice, pregnant females were intraperitoneally injected twice with isoproterenol (β-adrenergic receptor agonist, 200 mg/mouse) on day 3 of pregnancy. Mice were killed midmorning on day 4 and oviducts and uterine horns were separately flushed with Whitten's medium to recover embryos. Mice without fertilized ova were excluded from the experiments.

Embryo transfer, oviductal transport and implantation.

To explore the roles of maternal versus embryonic CB1 in embryo development, oviductal embryo transport and implantation, we used reciprocal embryo transfer experiments between Cnr mutant and wild-type mice. Two-cell embryos collected on day 2 of pregnancy were transferred into recipient oviducts on day 1 of pseudopregnancy. Implantation sites were examined midmorning on day 5, after intravenous injection of a Chicago Blue solution24. Oviducts and uteri in mice devoid of signs of implantation were flushed to recover any unimplanted blastocysts. To test that the oviduct-trapped blastocysts are implantation-competent, they were recovered from oviducts of Cnr1−/− mice not showing any sign of blue reaction on day 5 and transferred into day 4 wild-type pseudopregnant recipients. Implantation was examined 24 h after embryo transfer by the blue dye method.

Reverse transcription (RT)-PCR analysis of Mbldc1 and Cnr mRNA.

Mbldc1, Cnr1 and Cnr2 mRNAs were analyzed by RT-PCR followed by Southern hybridization as described previously15.

Analysis of anandamide.

Oviducts (50–60 mg) from 13–15 day 3 pregnant mice in each group (n = 3) were assayed for anandamide as previously described50. Briefly, the preweighed samples were homogenized in ethyl acetate with 0.5% acetic acid. 2H8-labeled anandamide was added to a mortar immediately prior to homogenization. The homogenate was centrifuged and the supernatant was dried, reconstituted in chloroform and purified on a silica solid-phase extraction cartridge. The eluent was dried, reconstituted in 1:8 of aqueous silver acetate to methanolic silver acetate and analyzed by reverse-phase positive-ion ESI-HPLC-MS-MS. Quantification was performed by stable isotope dilution against the octadeuterated internal standard.

Double immunofluorescence.

Frozen sections of oviducts (10 μm in thickness) were fixed in cold acetone (−20 °C) for 10 min followed by blocking in PBS containing 3% BSA, 10% normal donkey serum and 0.4% Triton X-100 for 1 h at room temperature. The sections were incubated overnight in a mixture of two primary antibodies: CB1-specific antibody (goat, 1:200) and β2-adrenergic receptor–specific antibody (rabbit, 1:200) (Santa Cruz Biotechnology). After repeated washings in PBS containing 0.05% Tween 20, the sections were incubated for 1 h in a mixture of two appropriate secondary antisera: rabbit antibodies to IgG conjugated to FITC and goat antibodies to IgG conjugated to TRITC (Jackson ImmunoResearch Laboratories). The sections were then rinsed in PBS and mounted with fluorescent mounting solution.

Analysis of noradrenaline release.

Effects of cannabinoid agonists or antagonists on noradrenaline release from mouse oviducts were assessed as previously described39,40. Briefly, day 3 oviducts were incubated for 30 min at 37 °C in a physiological salt solution (PSS) containing 0.1% ascorbic acid and [3H]-noradrenaline (8 × 105 c.p.m.; 56.6 Ci/mmol, PerkinElmer Life Sciences). After incubation, oviducts were washed in 1 ml of PSS three times, and then each oviduct was superfused at 1 ml/5 min with PSS containing 1 μM desipramine (an inhibitor of noradrenaline transporter) and rauwolscine (an inhibitor of presynaptic α2-adrenoceptor) and the superfusate was collected every 5 min. [3H]noradrenaline overflow was evoked by increasing concentration of KCl (4.8–50 mM) for two 5-min periods after 20 (S1) and 60 min (S2) superfusions. To examine the effects of cannabinoid agonists or antagonists on noradrenaline release, drugs were added 30 min before the second KCl stimulation (S2).

Note: Supplementary information is available on the Nature Medicine website.