Lipid metabolism and endometrial receptivity

Abstract

BACKGROUND

Obesity has now been recognized as a high-risk factor for reproductive health. Although remarkable advancements have been made in ART, a considerable number of infertile obese women still suffer from serial implantation failure, despite the high quality of embryos transferred. Although obesity has long been known to exert various deleterious effects on female fertility, the underlying mechanisms, especially the roles of lipid metabolism in endometrial receptivity, remain largely elusive.

OBJECTIVE AND RATIONALE

This review summarizes current evidence on the impacts of several major lipids and lipid-derived mediators on the embryonic implantation process. Emerging methods for evaluating endometrial receptivity, for example transcriptomic and lipidomic analysis, are also discussed.

SEARCH METHODS

The PubMed and Embase databases were searched using the following keywords: (lipid or fatty acid or prostaglandin or phospholipid or sphingolipid or endocannabinoid or lysophosphatidic acid or cholesterol or progesterone or estrogen or transcriptomic or lipidomic or obesity or dyslipidemia or polycystic ovary syndrome) AND (endometrial receptivity or uterine receptivity or embryo implantation or assisted reproductive technology or in vitro fertilization or embryo transfer). A comprehensive literature search was performed on the roles of lipid-related metabolic pathways in embryo implantation published between January 1970 and March 2022. Only studies with original data and reviews published in English were included in this review. Additional information was obtained from references cited in the articles resulting from the literature search.

OUTCOMES

Recent studies have shown that a fatty acids-related pro-inflammatory response in the embryo-endometrium boundary facilitates pregnancy via mediation of prostaglandin signaling. Phospholipid-derived mediators, for example endocannabinoids, lysophosphatidic acid and sphingosine-1-phosphate, are associated with endometrial receptivity, embryo spacing and decidualization based on evidence from both animal and human studies. Progesterone and estrogen are two cholesterol-derived steroid hormones that synergistically mediate the structural and functional alterations in the uterus ready for blastocyst implantation. Variations in serum cholesterol profiles throughout the menstrual cycle imply a demand for steroidogenesis at the time of window of implantation (WOI). Since 2002, endometrial transcriptomic analysis has been serving as a diagnostic tool for WOI dating. Numerous genes that govern lipid homeostasis have been identified and, based on specific alterations of lipidomic signatures differentially expressed in WOI, lipidomic analysis of endometrial fluid provides a possibility for non-invasive diagnosis of lipids alterations during the WOI.

WIDER IMPLICATIONS

Given that lipid metabolic dysregulation potentially plays a role in infertility, a better understanding of lipid metabolism could have significant clinical implications for the diagnosis and treatment of female reproductive disorders.

lipid metabolism, fatty acid, phospholipid, endocannabinoid, cholesterol, endometrial receptivity, embryo implantation

Introduction

Overnutrition and/or a sedentary lifestyle cause obesity. Owing to the economic boom and social transition in China, the mean BMI for adult women increased from 22.9 kg/m² in 2004 to 23.9 kg/m² in 2013–2014 (Pan et al., 2021). In the USA, 39.7% of women aged 20–39 years are obese (BMI ≥30.0 kg/m²), according to the current National Center for Health Statistics report (Hales et al., 2020). Obesity leads to either intracellular fat accumulation or elevated circulating lipid profiles. About 45% of reproductive-age women are affected by lipid abnormalities (Pugh et al., 2017). The growing obesity epidemic, along with its related dyslipidemia and other metabolic disorders have become a public health concern worldwide.

Obesity exerts various deleterious effects on human reproduction (Mintziori et al., 2020). Although overweight or obese women can conceive naturally, they have reduced fecundability (Gesink Law et al., 2007; Loy et al., 2018). The clinical impacts of obesity on female infertility have been well documented (Broughton and Moley, 2017; Gambineri et al., 2019). One mechanism is that obesity affects the regulation of the hypothalamic–pituitary–ovarian axis resulting in ovulatory function disturbances. Oligo- or anovulation is common in overweight or obese women, especially those with polycystic ovary syndrome (PCOS) (Silvestris et al., 2018). Even after the restoration of ovulation, pregnancy outcomes in these patients remained to be far from satisfactory (Alebic, 2022). This suggests that either oocyte-embryo quality or the endometrial milieu was affected by increased BMI.

Indeed, there is emerging evidence showing that obesity may impact the oocyte and embryo quality (Robker, 2008; Gonzalez et al., 2022). Lower fertilization rates of oocytes and decreased blastocyst development rates are observed in obese women compared to those of moderate BMI (van Swieten et al., 2005; García-Ferreyra et al., 2021). Obese women tend to have fewer meiotically mature oocytes during controlled ovarian stimulation (COS), and impaired embryo growth after IVF/ICSI (Gonzalez et al., 2022). Despite the achievements in improving embryo quality and embryo selection, progress to a better understanding of endometrial function has lagged behind the biology of the blastocyst (Demiral et al., 2015). In particular, the role of lipid metabolism in endometrial receptivity is poorly understood.

This review aims to provide insights into effects of lipids and lipid-derived mediators on the embryonic implantation process as well as endometrial receptivity. First, we reviewed reproductive outcomes among women with obesity, dyslipidemia and PCOS seeking ART. Further, we have discussed key lipid-related metabolic pathways and the underlying mechanisms that are potentially involved in regulating endometrial receptivity. Finally, we highlighted several promising methods for evaluating endometrial receptivity. We acknowledge that lipid metabolism is a broad field but this review only focuses on effects of major lipids and their derivatives on embryo implantation.

Methods

PubMed and Embase were searched, from January 1970 to March 2022, independently by two researchers (T.Y. and J.Z). The relevant reference lists were searched manually for studies that met the inclusion criteria. The following search strategy was adopted by using the keywords: (‘lipid’ OR ‘fatty acid’ OR ‘prostaglandin’ OR ‘phospholipid’ OR ‘sphingolipid’ OR ‘endocannabinoid’ OR ‘lysophosphatidic acid’ OR ‘cholesterol’ OR ‘progesterone’ OR ‘estrogen’ OR ‘transcriptomic’ OR ‘lipidomic’ OR ‘obesity’ OR ‘dyslipidemia’ OR ‘polycystic ovary syndrome’) AND (‘endometrial receptivity’ OR ‘uterine recepitivty’ OR ‘embryo implantation’ OR ‘assisted reproductive technology’ OR ‘in vitro fertilization’ OR ‘embryo transfer’). Only studies with original data and reviews published in English were included in this review. Additional information was also obtained from references cited in the articles resulting from the literature search. Studies were excluded if: they were irrelevant or less relevant to the main topic; and they were published as a conference abstract, case report, editorial text, letter or erratum.

Results

The literature search from January 1970 to March 2022 led to the identification of 14 026 relevant citations. An additional 237 publications were obtained from the cited references during the literature search. Owing to duplication, 5381 of these citations were excluded. Another 2544 articles were excluded after screening the titles and publication type, and 5972 were excluded after reviewing the abstracts and full texts. Finally, 366 articles were identified as eligible for inclusion. The flow-chart for study inclusion is shown in Fig. 1.

Figure 1.

Flow diagram of studies identified in the review of lipid metabolism and endometrial receptivity.

Clinical effects of obesity, dyslipidemia and PCOS on ART outcomes

Obesity

Obesity is defined as an abnormal or excessive accumulation of body fat that might impair health (World Health Organization, 2000). In women undergoing ART, obesity has been associated with a higher requirement for gonadotrophin (Gn), longer duration of Gn stimulation (Rittenberg, et al., 2011), fewer oocytes retrieved (Maheshwari et al., 2007), decreased blastocyst formation rate (Comstock et al., 2015) and increased cycle cancellation rate (Fedorcsák et al., 2004; Kudesia et al., 2018). The impact of obesity on the pregnancy outcomes following ART remains debatable. Recently, a meta-analysis based on 682 532 cycles demonstrated that live birth rates are significantly lower in overweight and obese women, compared to women with normal weight (risk ratio (RR): 0.94, CI 95%: 0.91–0.97; RR: 0.85, CI 95%: 0.84–0.87, respectively) (Sermondade et al., 2019). Decreased live birth rates have been found in women with excess weight undergoing their first IVF/ICSI and embryo transfer (IVF/ICSI-ET) cycles and subsequent frozen embryo transfer (FET) cycles (Chavarro et al., 2012; Ding et al., 2019; Xue et al., 2020). However, other studies report that neither clinical pregnancy rate nor live birth rate is significantly affected by BMI (Dechaud et al., 2006; Ozekinci et al., 2015; Insogna et al., 2017; Brunet et al., 2020; Prost et al., 2020). These controversial results may be attributed to a variety of factors. For example, the heterogeneity in terms of study populations, sample size, BMI group definitions, type of embryo transfer (fresh or frozen cycle), number of cycles (first or cumulative cycles), oocyte origin (donor or non-donor), ovarian stimulation protocols and inclusion and exclusion criteria (age and PCOS status).

Regardless of the advances in the use of sophisticated equipment and improvements in procedures, the pregnancy rate after ART remains low. In fact, a considerable number of obese infertile women still suffer from implantation failure even with high-quality embryos transferred. This can be partially attributed to embryos transferred into a non-receptive uterus, since displacement of the window of implantation (WOI) increases in a BMI-dependent manner (Bellver et al., 2021). Mounting evidence indicates that female obesity may interfere with endometrial receptivity, which is necessary for normal implantation potential (Dessolle et al., 2009; Bellver et al., 2010, 2013; Boots et al., 2014; Provost et al., 2016; Zhang et al., 2019), even though several studies cast doubt on this point (Wattanakumtornkul et al., 2003; Styne-Gross et al., 2005; Prost et al., 2020). In the fresh cycles, a supraphysiologic hormonal milieu generated from COS would bias the conclusion, since endometrial receptivity might have been impaired by high doses of Gn (Acharya et al., 2018). To rule out the effect of lipotoxicity on oocyte-embryo quality or the effect of COS on endometrium, the ovum donation cycles with oocytes obtained from normal weight donors were adopted (Bellver et al., 2013). The overweight and obese recipients exhibited lower embryo implantation, clinical pregnancy and live birth rates in comparison to their normal weight counterparts (Bellver et al., 2013). Currently, in China, legislation and regulation pertaining to oocyte donation are still stringent. Chinese scholars analyzed 22 043 first FET cycles with high-quality autologous embryo transfer to control for the embryo factor (Zhang et al., 2019). After adjusting for potential confounders, obesity was associated with decreased embryo implantation, clinical pregnancy and live birth rates (Zhang et al., 2019). Moreover, the miscarriage rates (both in the first and second trimesters) were significantly higher in the obesity group than that in the control group (Zhang et al., 2019). Obese women were prone to suffer from euploid miscarriage, highlighting the role of the endometrium in impaired reproductive performances among these individuals (Boots et al., 2014; Cozzolino et al., 2021).

Dyslipidemia

Dyslipidemia is often characterized by elevated circulating levels of low-density lipoprotein cholesterol (LDL-C) and/or triglyceride (TG), concurrently with decreased high-density lipoprotein cholesterol (HDL-C) levels (Liu and Li, 2015). As a major contributing factor of cardiovascular disease, atherogenic dyslipidemia is strongly associated with obesity (Bamba and Rader, 2007). Overweight and obese women are more likely to have higher total cholesterol (TC), LDL-C, TG concentrations and lower HDL-C level compared to their counterparts with BMI <25 kg/m² (Bever et al., 2020). Moreover, gestational lipid profile alterations are associated with pregnancy loss (Horn et al., 2019), spontaneous (Catov et al., 2010) or induced preterm birth (Vrijkotte et al., 2012), pregnancy-related complications, such as pre-eclampsia (Wiznitzer et al., 2009), gestational diabetes mellitus (Han et al., 2016) and fetal and neonatal morbidities (Adank et al., 2020; Bever et al., 2020). TG and remnant cholesterol are positively associated with blood pressure even 6 and 9 years after pregnancy (Adank et al., 2019), highlighting the long-term effects of lipid abnormalities on maternal health.

Unlike obesity, little is known about the relation between serum lipid disturbances and human reproductive outcomes after the ART procedures. Most clinical studies in this field are limited to the association between lipid compositions in follicular fluid (FF) and oocyte-embryo qualities (Browne et al, 2008; Von Wald et al., 2010; Jungheim et al., 2011a; Niu et al., 2014; Kim et al, 2017). There is a downward trend for cumulus–oocyte complex (COC) morphology and the number of cleavage stage embryos in women with elevated follicular fatty acids (FAs) (Jungheim et al., 2011a). Correlations between embryo fragmentation or blastomere score and follicular-specific FAs levels are also measured (Niu et al., 2014). The embryo fragmentation score positively correlates with the oleic acid concentration, and the blastomere score negatively correlates with the stearic acid levels (Niu et al., 2014). Higher concentrations of HDL-C subfractions in the large and medium particle size ranges are associated with poorer embryo quality (Kim et al., 2017).

Dyslipidemia exerts negative effects on ART clinical outcomes, independent of obesity and PCOS status (Liu et al., 2021). During the fresh cycles, pregnancy rate was significantly higher in the normal lipid metabolism group than in the dyslipidemia group (82.7% versus 74.4%, P = 0.000) (Liu et al., 2020). Preconception serum lipid and lipophilic micronutrient levels were associated with live birth rate and, more specifically, women with higher TG levels (65.64 mg/dl) were less likely to have a live birth (RR: 0.54, CI 95%: 0.33–0.90) (Jamro et al., 2019). A newly established prognostic model has been built to estimate the probabilities of IVF/ICSI success in patients with PCOS (Gao et al., 2020). Obesity, and high serum TC and basal FSH levels are found to be associated with unfavorable live birth outcome among these populations (Gao et al., 2020). It seems that LDL-C had the strongest negative effect on pregnancy outcomes after IVF/ICSI-ET, including lower pregnancy rate, lower live birth rate and higher miscarriage rate (Cai et al., 2021). On the contrary, higher serum HDL-C was associated with more favorable ovarian stimulation and pregnancy outcomes (Cai et al., 2021). It remains to be established whether oocyte-embryo quality or the endometrial microenvironment is responsible for these findings.

Recently, we performed a retrospective real-world analysis involving 5030 infertile women to investigate the associations between dyslipidemia and pregnancy outcomes (Yang et al., 2021). Dyslipidemia, especially serum TC of ≥5.20 mmol/l, was negatively associated with the live birth rate in the first complete IVF/ICSI cycle. In addition, the prevalence of type C endometrium was higher in women with dyslipidemia than the controls. We failed to find significant between-group differences in the number of metaphase II (MII) oocytes and good-quality embryos. Taken together, it is plausible to assume that the elevated TC concentrations exert detrimental effects on endometrial receptivity, which would eventually affect the subsequent maintenance of pregnancy.

PCOS

PCOS, a common endocrine disorder among reproductive-aged women, is diagnosed based on the presence of two out of three criteria: oligoanovulatory ovarian dysfunction, polycystic ovarian morphology, clinical or biochemical hyperandrogenism (HA) (Rotterdam ESHRE/ASRM-Sponsored PCOS Consensus Workshop Group, 2004). PCOS is frequently associated with risk factors of cardiometabolic disease including obesity and dyslipidemia (Lim et al., 2011, 2012; Zhang et al., 2012). Besides, women with PCOS are subfertile secondary to ovulatory dysfunction, impaired oocyte quality and endometrial receptivity (Balen et al., 2016). IVF is only recommended for those who do not respond to the first or second-line ovulation induction therapies or those with other indications (Teede et al., 2018). In fact, women with PCOS are more likely to receive IVF treatment than women without PCOS (Hart and Doherty, 2015).

The three complete phenotypes of PCOS and related morbidities, such as obesity or insulin resistance (IR), are at high risk of reduced oocyte competence (Palomba et al., 2017). High BMI causes a reduction in average serum levels of Gn during COS. Despite requiring higher doses of Gn, overweight or obese women with PCOS have a reduced number of oocytes retrieved (Akpınar et al., 2014; Zhou et al., 2020). PCOS patients with metabolic syndrome exhibited pronounced circulating lipid disturbances, decreased oocyte utilization rate and less available embryos (He et al., 2019). Obesity adversely affected oocyte size independent of PCOS (Marquard et al., 2011), which might be an important marker of compromised oocyte competence. HA negatively impacts oocyte qualities in PCOS by inducing follicular atresia and preventing oocyte meiotic resumption (Pan et al., 2015). The obesity-induced oxidative stress and IR were also associated with decreased oocyte developmental potential (Cardozo et al., 2011). From the evidence of transcriptomic-based analysis in women with PCOS, ovarian granulosa-lutein cells expressed an enhanced inflammatory pattern, and there might be a local regulatory loop comprising intrafollicular androgens and cytokines that exacerbates this phenomenon (Adams et al., 2016).

Even though COS and IVF can help to improve the morphological quality of oocytes and the embryos transferred, the PCOS population exhibits poorer than expected reproductive outcomes. After controlling for confounding effects, the live birth rate is lower among PCOS women with a single ideal blastocyst transferred (Steiner et al., 2020). Meanwhile, despite comparable embryological parameters, lower implantation rate, clinical pregnancy rate and live birth rate were observed in the obese PCOS group than in the normal weight PCOS group (Bu et al., 2013; Bailey et al., 2014). Obesity altered a cascade of genes related to biological processes, the immune system as well as protein binding in the receptive endometrium (Bellver et al., 2011). In PCOS, obesity exacerbates low-grade chronic inflammation, IR and HA, rendering hundreds of dysregulated genes involved with suboptimal endometrial receptivity (Jiang and Li, 2021). The three endocrinologic and metabolic disarrangements promote fat accumulation and aggravate obesity, resulting in a vicious cycle among this population (Jiang and Li, 2021).

Lipid-based endocrine factors involved in endometrial receptivity

The human endometrium undergoes dynamic changes at multiple levels throughout the menstrual cycle in response to ovarian hormones and paracrine secretions. After ovulation, the endometrium transits from a proliferative to secretory morphology and culminates in a spatially and temporally restricted period; the WOI is defined as the self-limited period when the uterine milieu is favorable to blastocyst acceptance and implantation. It usually coincides with the 20th to 24th day of a regular menstrual cycle (i.e. between the 7th and 11th day after the LH surge: LH + 7–LH + 11) (Achache and Revel, 2006). However, the optimal duration of the WOI is restrictedly short, and only lasts about 24–48 h in human (Galliano et al., 2015). Endometrial receptivity enables the endometrium to provide an optimal environment for embryo development and placenta formation (Craciunas et al., 2019).

During this period, both the uterine epithelium and stroma experience alterations orchestrated by many different factors such as steroid hormones, cytokines and chemokines, adhesion molecules, immune cells, growth and transcription factors (Munro, 2019). The cholesterol-derived steroid hormone estrogen (E2) and progesterone (P4) are crucial for the establishment of endometrial receptivity. E2 priming-induced endometrial proliferation and P4 receptor recruitment prior to P4 exposure are essential steps in the achievement of endometrial receptivity (Kodaman and Taylor, 2004; Vasquez and DeMayo, 2013). P4 is produced in the corpus luteum (CL) and is regarded as a critical element for early pregnancy maintenance (Csapo and Pulkkinen, 1978; Lessey and Young, 2019). On the one hand, P4 directs the luminal epithelium (LE) to a state that will promote embryo adhesion and invasion by decreasing tight junctions (Grund and Grümmer, 2018). On the other hand, P4 acts on glandular epithelium to produce uterine fluid that contains factors important for the embryo survival and LE-embryo communications (Ye, 2020). Meanwhile, P4-P4 receptor signaling drives decidualization in endometrial stromal cells to sustain pregnancy (Large and DeMayo, 2012).

In addition to steroid hormones, other lipid mediators participate in the control of female procreation. Prostaglandins (PGs) derived from FAs participate in neoangiogenesis in an autocrine and/or paracrine manner (Kaczynski et al., 2016). Sufficient blood flow has to be maintained both for ovaries and endometirum, which synthesize and secrete various factors into the systemic system for embryo nourishment and early implantation. In addition, PGE2 maintains luteal function for blastocyst development and stimulates chemokine expression for trophoblasts adhesion to the decidua (Niringiyumukiza et al., 2018). Phospholipid-derived endocannabinoids (eCBs), along with their receptors and the set of enzymes responsible for their synthesis and degradation, exert several actions either in the central nervous system or in peripheral tissues (Battista et al., 2008). This system modulates reproductive neuroendocrine function through binding to the endogenous cannabinoid receptors presented in the hypothalamic–pituitary–gonadal axis (Meah et al., 2021). Recently, more direct effects of eCBs on a plethora of biological processes, including gametogenesis, fertilization, embryo implantation and decidualization, have been confirmed (Sun and Dey, 2012). For instance, Anandamide (AEA, a member of eCBs), transmits biosignals through its receptor CB1 to regulate embryonic synchronous development, oviduct transport of the embryo and endometrial receptivity (Ye et al., 2021). Within a very narrow range, AEA regulates blastocyst activation and implantation by differentially modulating mitogen-activated protein kinase/extracellular regulated kinase (MAPK/ERK) signaling and Ca²⁺ influx (Sun and Dey, 2008).

Fatty acids

FAs are essential energy sources for the human body and commonly present in organisms as esterified FAs (i.e. TGs, phospholipids and cholesterol esters), which are transported in the circulation by lipoproteins (Shetty and Kumari, 2021). The non-esterified FAs, also known as free FAs, are insoluble in water and always bound to serum albumin (Kimura et al, 2020). According to the degrees of saturation, FAs are classified as saturated FAs (SFAs, containing no double bonds) and unsaturated FAs (UFAs, more than one double bond). UFAs can be further grouped into monounsaturated FAs (MUFAs, containing one double bond) and polyunsaturated FAs (PUFAs, containing more than two double bonds) (Nikolaidis and Mougios, 2004). Depending on the position of the first double bond from the omega (ω)-carbon, MUFAs and PUFAs are classified into the ω-3 (n-3), ω-6 (n-6), ω-7 (n-7) and ω-9 (n-9) series.

SFAs and some UFAs, such as oleic acid (18:1 n-9), which can be synthesized autonomously by the body are thus considered nonessential. Essential FAs can only be obtained from diet sources, such as n-3 and n-6 PUFAs. Linoleic acid (LA, 18:2 n-6) and α-linolenic acid (ALA, 18:3 n-3) are the primary PUFAs in Western diets (Simopoulos, 2002). FA metabolism has been well reviewed elsewhere (Wakil and Abu-Elheiga, 2009). In addition to their function as an energy source, FAs are involved in critical functions such as cellular membrane formation, ion homeostasis, signal transduction, gene expression and synthesis of lipid bioregulators (Kogteva and Bezuglov, 1998). Recently, a wealth of emerging evidence has addressed the effect of FA compositions on female reproduction (Table I).

Table I

Summary of the major fatty acids, highlighting their biological function related to female reproduction.

Common name	Chemical designation	Main findings that related to female reproduction	References
Saturated fatty acids SFAs (no double bonds)
Lauric acid	C12:0	Expand the number of mammary gland alveoli and improve the mammary gland development;	Yang et al., 2020
		Increase the body weight of offspring and improve the lactation function of lactating mice.
Myristic acid	C14:0	Preconception levels positively correlate with improved live birth outcome.	Kim et al., 2019
Palmitic acid	C16:0	Induce apoptosis in ovarian granulosa cells;	Mu et al., 2001; Jungheim et al., 2011c; Wu et al., 2012; Rampersaud et al., 2020
		Restrict the embryo development and fetal growth;
		Induce inflammatory responses in the extravillous trophoblasts of placenta;
		Impair trophoblasts migration and early placental development.
Stearic acid	C18:0	Induce apoptosis in ovarian granulosa cells;	Mu et al., 2001; Valckx et al., 2014b
		Restrict the follicular growth and antrum formation.
Arachidic acid	C20:0	Preconception levels positively correlate with improved live birth outcome.	Kim et al., 2019
Monounsaturated fatty acids MUFAs (one double bond)
Omega-7 fatty acids
Palmitoleic acid	C16:1	Decreased syncytiotrophoblast synthesis may contribute to insulin resistance and low-grade inflammation in pregnancies complicated by obesity.	Ferchaud-Roucher et al., 2019
Omega-9 fatty acids
Oleic acid	C18:1	Support oocyte development and counteract detrimental effects by palmitic acid.	Fayezi et al., 2018; Yousif et al., 2020; Zarezadeh et al., 2021
Polyunsaturated fatty acids PUFAs (two or more double bonds)
Omega-6 fatty acids
Linoleic acid (LA)	C18:2	Restrict the oocyte and embryo development;	Marei et al., 2010; Shrestha et al., 2020
		Induce inflammatory responses in the placenta;
		Alter fatty acid compositions and genes responsible for metabolism, nutrient transport and angiogenesis in the placenta.
Dihomo-γ-linolenic acid (DGLA)	C20:3	Induce 1 series prostaglandin (PG) productions which modulate the establishment of pregnancy.	Wathes et al., 2007
Arachidonic acid (AA)	C20:4	Modulate oocyte maturation, embryonic development and the process of steroidogenesis;	Vilella et al., 2013b; Khajeh et al., 2017
		The derivatives two series PGs participate in the embryo implantation process.
Omega-3 fatty acids
α-Linolenic acid (ALA)	C18:3	Serum levels negatively correlate with embryo implantation and clinical pregnancy rates;	Wathes et al., 2007; Jungheim et al., 2011b; Angkasa et al., 2017
		Modulate oocyte growth and differentiation, offspring birth weight, and the process of steroidogenesis.
Eicosapentaenoic acid (EPA)	C20:5	Induce three series PGs productions;	Wathes et al., 2007; Jawerbaum and Capobianco, 2011; Chiu et al., 2018
		Serum levels positively correlate clinical pregnancy and live birth rates after ART treatments;
		Involve in embryonic and placental development by binding to peroxisome proliferator-activated receptors (PPARs).
Docosahexaenoic acid (DHA)	C22:6	Exert dual effects on oocyte quality;	Hammiche et al., 2011; Carlson et al., 2013; Oseikria et al., 2016; Basak et al., 2020
		Essential for improved embryo morphology, fetal-placental and fetal brain development;
		Prevent preterm birth and very low birth weight of the offspring.

Association between fatty acid composition and ART outcomes

Serum FAs levels, particularly n-3 and n-6 PUFAs, have been associated with pregnancy outcomes in ART scenarios (Jungheim et al., 2011b, 2013; Mirabi et al., 2017; Chiu et al., 2018). ALA levels on oocyte retrieval day were adversely associated with embryo implantation rates and occurrences of clinical pregnancy (Jungheim et al., 2011b). No association was found between circulating ALA levels and COC morphologies, suggesting a possible role for elevated ALA in disturbing embryo implantation (Jungheim et al., 2011b). They further determined that it was the LA/ALA (n-6 to n-3) ratio, not the specific PUFAs that mattered (Jungheim et al., 2013). Women with increased LA/ALA ratios had a higher chance of embryo implantation, clinical pregnancy and live birth compared to women with lower ratios (Jungheim et al., 2013). In line with a previous study (Jungheim et al., 2011b), Chiu et al. (2018) reported a negative association between serum ALA levels and clinical pregnancy (RR: 0.47, CI 95%: 0.24–0.92). However, long chain n-3 PUFAs were associated with pregancny outcomes in the different direction. Pregnant women had higher serum levels of eicosapentaenoic acid (EPA, 20:5 n-3) compared to the non-pregnant women (Mirabi et al., 2017). Higher concentrations of EPA were associated with higher probability of clinical pregnancy and live birth (Chiu et al., 2018).

Most serum and FF metabolite concentrations correlated significantly, indicating that serum metabolite concentrations are partly mirrored in the follicular microenvironment (Valckx et al, 2012). FF provides an important milieu for oocyte development, and its metabolic composition directly affects oocyte quality (Dumesic et al., 2015). There are growing numbers of studies assessing the effects of follicular FAs composition on oocyte or embryo quality during IVF or ICSI procedures (Shaaker et al., 2012; O'Gorman et al., 2013; Niu et al., 2014; Mirabi et al., 2017; Ruiz-Sanz et al., 2019; Zarezadeh et al., 2021). PUFAs may participate in steroidogenesis regulation by activating their transcriptional sensors peroxisome proliferator-activated receptors (PPARs) (Komar, 2005; la Cour Poulsen et al., 2012). Obesity significantly altered the FF metabolites related to FAs metabolism (Yang et al., 2012; Valckx et al, 2014a; Bou Nemer et al, 2019; Song et al, 2020). Mouse COCs exposed to lipid-rich human FF during their maturation had elevated levels of endoplasmic reticulum stress markers and decreased MII oocyte rates (Yang et al., 2012).

Perhaps the natural FET cycle is a better approach to understand the potential involvement of the endometrium than the IVF cycle, since the results are free from the impacts of exogenous hormones. Higher erythrocyte ratios of SFAs to UFAs on the day of FET and shorter average FAs chain lengths have been found to predict ongoing pregnancy (Onyiaodike et al., 2018). Unfortunately, only one article refers to this issue and more clinical evidence is required to clarify whether FAs impact pregnancy rates in FET cycle.

Fatty acid-related inflammation during embryo implantation

There are several ways in which FAs may contribute to successful embryo implantation. One mechanism involves a modified FAs-related pro-inflammatory response at the conceptus–maternal inferface. Embryos secret a plethoral of pro-inflammatory factors during the peri-implantation period, suggesting their role of provoking inflammatory pathways around the endometrial lining of uterus (Waclawik et al., 2017). Inflammation-prone locations are favorable sites of embryo implantation. A local injury-induced endometrial inflammatory reaction appears to optimize the uterine receptivity and facilitates the subsequent decidualizaiton process in women undergoing IVF with previous histories of implantation failure (Barash et al., 2003; Gnainsky et al., 2010). This is consistent with several clinical findings that local endometrial biopsy or scratching prior to the IVF/ICSI-ET treatments improved pregnancy outcomes (Raziel et al, 2007; Karimzadeh et al., 2009; Narvekar et al., 2010; Shohayeb and El-Khayat, 2012; Gibreel et al., 2013). The embryo attachment-associated inflammation is a balanced and delicately controlled process. It ensures trophoblast invasion, but free from the neutrophil-induced acute inflammatory response, which may destroy the conceptus at the same time (Chavan et al, 2017).

LA is converted to the longer chain n-6 PUFA arachidonic acid (AA, 20:4 n-6), while ALA can be converted to EPA and further docosahexaenoic acid (DHA, 22:6 n-3) through multistep desaturation and elongation (Decsi and Kennedy, 2011). The eicosanoids are bioactive mediators and regulators of inflammation and are synthesized from the 20 carbon PUFAs, i.e. AA (n-6) and EPA (n-3). In general, most eicosanoids derived from n-6 PUFAs, such as 2-series PGs (e.g. PGD2, PGE2, PGF2α and PGI2), 2-series thromboxanes (e.g. TXA2 and TXB2) and 4-series leukotrienes (e.g. LTB4), are pro-inflammatory and pro-aggregatory (Hata and Breyer, 2004; Schmitz and Ecker, 2008). Both LTB4 and PGE2 increase vascular permeability and recruit inflammatory cytokines, chemokines and immune cells (Calder, 2009; Nakanishi and Rosenberg, 2013; He et al., 2020), which ensures proper blastocyst implantation. PGE2 has both pro-inflammatory and anti-inflammatory roles. The initial pro-inflammatory effect of PGE2 was followed by the effect of inflammation resolution, since there was a shift from predominantly LTB4 biosynthesis to lipoxin A4 (Levy et al., 2001).

Other eicosanoids formed from n-3 PUFAs, such as 3-series prostaglandins (e.g. PGD3, PGE3, PGF3 and PGI3), 3-series thromboxanes (e.g. TXA3 and TXB3) and 5-series leukotrienes (e.g. LTB5), are less potent as pro-inflammatory mediators (Calder, 2011). For example, LTB5 was 10- to 30-fold less potent as a neutrophil chemoattractant than LTB4 (Goldman et al., 1983). PGE3 was less efficient in inducing cyclooxygenases-2 (COX-2) gene expression and interleukin-6 synthesis compared with PGE2 (Bagga et al., 2003). One possible mechanism for this is that eicosanoid receptors have a lower affinity for the n-3 PUFAs-derived mediator than for the n-6 PUFAs-derived one (Wada et al., 2007). Resolvins and protectins generated from EPA and DHA are also reported to exert anti-inflammatory actions (Schwab et al., 2007). Nuclear factor-kappaB (NF-κB) is one of the most prominent transcription factors that plays a crucial role in various inflammatory signaling pathways. Pretreatment of n-3 FAs inhibited lipopolysaccharide-stimulated macrophage tumor necrosis factor-α production by decreasing IκB phosphorylation (Novak et al., 2003), implying an inhibitory role of n-3 FAs in NF-κB activity.

The conversion of LA to AA (n-6) is in competition with the conversion of ALA to EPA (n-3), since the same elongase and desaturase enzymes are used. However, LA is much more prevalent in most human diets than ALA, and so the conversion to AA is quantitatively more active than that of EPA (Calder, 2013). Therefore, it is plausible that an increase in the n-6 to n-3 PUFAs ratio may enhance endometrial inflammation and ultimately facilitate a successful pregnancy. Moreover, the anti-inflammatory effects of n-3 PUFAs protect the conceptus from further exacerbated inflammation (Fig. 2). In line with this, a higher rate of embryo implantation occurred in mice fed with n-6 PUFAs supplementation compared to those given n-3 PUFAs (Fattahi et al., 2018). Levels of total n-6 PUFAs, AA and LA in murine uterine tissue were positively correlated with embryo implantation rate (Fattahi et al., 2018).

Figure 2.

Fatty acid metabolism and possible mechanisms of embryo implantation. Fatty acids are converted to acyl-CoA by ACS and converted into acetyl-CoA via β-oxidation in the mitochondria. Acetyl-CoA enters the TCA cycle, yielding energy and citrate. In the cytoplasm, citrate is cleaved into acetyl-CoA, which is used for de novo FAs synthesis. Eicosanoids derived from n-3 PUFAs in membrane phospholipids are low pro-inflammatory or anti-inflammatory, while eicosanoids generated from n-6 are pro-inflammatory. An increase in n-6 to n-3 ratio may promote endometrial inflammation in support of pregnancy. Series 2 PGs derived from n-6 participate in embryo implantation via several approaches including promotion of embryonic adhesiveness to the endometrium and localized vascular permeability. AA, arachidonic acid; ACC1, acetyl-CoA carboxylase 1; ACC2, acetyl-CoA carboxylase 2; ACLY, ATP citrate lyase; ACS, acyl-CoA synthase; CPT-1, carnitine palmitoyltransferase 1; DHA, docosahexaenoic acid; EPA, eicosapentaenoic acid; FAS, fatty acid synthase; FAs, fatty acids; n-3, omega-3 fatty acids; n-6, omega-6 fatty acids; PGs, prostaglandins; PUFAs, polyunsaturated fatty acids; TCA: tricarboxylic acid. This image contains some elements adapted from the Servier Medical Art Image Bank (Servier Laboratories (Aust) Pty Ltd) licensed under a Creative Commons Attribution 3.0 Unported Licence.

Involvement of prostaglandin signaling

PUFAs are the direct precuros of PGs (Wathes et al., 2007). Dietary changes in PUFAs could affect endometrial PGs signaling in animal models and this was considered as another possible mechanism by which FAs contribute to implantation (Cheng et al., 2005; Coyne et al., 2008; Shahnazi et al, 2018). PG compositions within the uterine environment are thought to be critical for successful implantation (Salleh, 2014), although there are conflicting reports regarding the precise roles of each PG.

During the receptive stage, PGE2 and PGF2α concentrations, as well as enzymes related to PGs synthesis, were accumulated in the human uterine lumen (Vilella et al., 2013b). Patients with recurrent implantation failure (RIF) exhibited aberrant endometrial PGs synthesis characterized by significantly reduced levels of COX-2 and cytosolic phospholipase 2α (cPLA2α) (Achache et al, 2010). PG synthesis is disrupted in women with repeated failure of IVF-ET, suggesting that reduced PG synthesis in the human endometrium leads to compromised endometrial receptivity. The deferred implantation occurred beyond the normal WOI (i.e. shifting from Day 4 to Day 6 of pregnancy) in cPLA2α^−/− mice, and exogenous supplementations of PGE2 and carbaprostacyclin (cPGI) restored this fertility defect (Song et al., 2002).

Besides the pro-inflammatory effect on the endometrium mentioned above, mounting evidence has suggested that PGs also play a direct and crucial role in embryo implantation. The involvement of PGs in the regulation of CL function has been established. Particularly, PGF2α is the major driver for luteolysis in ruminants and other species with an estrous cycle (Mondal et al., 2011). PGE2 released from the uterus and conceptus are involved in the control of porcine CL lifespan by antagonizing PGF2α in a luteotropic or antiluteolytic manner (Christenson et al., 1994; Waclawik and Ziecik, 2007). In cows, luteoprotective PGE2 induced the mRNA expression of LH receptors and prevented the decrease in unoccupied or occupied luteal receptors for LH (Weems et al., 2011).

Compared to non-primate species, far less is understood about the roles of PGs in the regulation of the human luteal function. Intraluteal injection of PGF2α gave rise to suppression of serum P4 levels and a shortening of the luteal phase in healthy women (Bennegård et al., 1991). Genes responsible for PGF production, AKR1B1 and AKR1C1-3, are increased in the regressing CL, suggesting a luteolytic effect of PGF in human (Nio-Kobayashi et al., 2017). PGE exerts hCG-like luteotropic effects on human luteinized granulosa cells by inducing steroidogenic enzyme transcripts (e.g. STAR, CYP11A1 and HSD3B1) (Nio-Kobayashi et al., 2017). One mechanism by which PGE2 stimulates luteal P4 secrection is by accumulating cAMP via functional receptors EP1 and EP2 (Harris et al., 2001).

The role of PGs in facilitating trophoblast invasion was largely unknown until recently. During implantation, PGE2 secreted from the conceptus and endometrium stimulated adhesion of trophoblast cells to the extracellular matrix (ECM) via EP2 (Waclawik et al., 2013). This effect was mediated via the MAPK1/3 signaling pathway as well as up-regulated expression of cell adhesion proteins such as focal adhesion kinase (FAK) and intracellular adhesion molecular-1 (Waclawik et al., 2013). Meanwhile, exogenous administration of PGE2 up-regulated integrin expression in the endometrial epithelial cell line RL95-2 (Huang et al, 2017). Similar to PGE2, PGF2α promotes adhesion of porcine trophoblast cells and the human trophoblast cell line HTR-8/SVneo to ECM by activating the FAK and MAPK signaling pathways (Kaczynski et al., 2018; Baryla et al, 2019).

Increased endometrial vascular permeability at the implantation site is another crucial event during embryo implantation. Prostacyclin (PGI2), a potent vasodilator, has been reported to improve uterine vascular permeability and decidualization by activation of its nuclear receptor PPARδ (Lim et al., 1999). Exogenous administration of a PGI2 agonist, namely cPGI, significantly improves implantation rate and incidence of decidual response in COX-2 deficient mice (Lim et al., 1999). So far, there is no direct evidence showing that activation of the transmembrane G protein-coupled receptor (PTGIR) is associated with endometrial vascular permeability. Increased expression of PTGIR was observed in the porcine endometrium during the peri-implantation period (Blitek et al., 2015). Incubation of stromal cells with iloprost (PGI2 analogue) activated PTGIR, resulting in an intracellular cAMP rise and increased expression of angiogenic genes, including fibroblast growth factor 2 (FGF2) and vascular endothelial growth factor (VEGF) (Blitek et al., 2015). In an endometrial epithelial cell line, iloprost-PTGIR interaction induced gene expression of basic FGF, angiopoietin-1 (Ang-1) and Ang-2 via cross talk with the epidermal growth factor receptor (Smith et al., 2006).

PGE2 was found to be more effective in inducing endometrial vascular permeability and decidualization than PGI2 and agonists of PPARδ and its obligate partner retinoid x receptor α (Gillio-Meina et al., 2009). Via its receptor, EP4, PGE2 was shown to induce vasodilation through endothelial nitric oxide synthase (eNOS) activation (Hristovska et al., 2007), suggesting an involvement of PGE2 in regulation of the utero-ovarian blood circulation. An augmentation of VEGF164 mRNA expression and VEGF secretion by PGE2 was detected in the porcine endometrial stromal cells (Kaczmarek et al., 2008). Similar results were observed in the PGF2α-stimulated porcine endometrial explants (Kaczynski et al., 2016). In an autocrine/paracrine manner, PGF2α was involved in uterine vascular remodeling and angiogenesis at the embryo–maternal interface (Kaczynski et al., 2016). As in porcine endometrium (Kaczynski et al., 2016) or in human endometrial adenocarcinoma cells (Sales et al., 2005), PGF2α-PTGFR interaction enhanced synthesis and secretion of VEGF via activation of the MAPK signaling pathway (Fig. 2).

Phospholipids

Phospholipids are amphiphilic molecules with one hydrophilic head of a phosphate moiety and two hydrophobic tails of fatty acyl chains esterified to the alcohol or glycerol molecule (Li et al., 2015). Based on the alcohol present, phospholipids are categorized into two classes: glycerophospholipids, in which the alcohol is glycerol, and sphingolipids, in which the alcohol is sphingosine. Phospholipids act as structural constituents of cellular membranes and are implicated in signal transduction as the precursor of various messenger molecules, such as diacylglycerol, inositol trisphosphate and sphingosine-1-phosphate (S1P) (Sunshine and Iruela-Arispe, 2017; Fakhr et al., 2021).

Profiles of uterine phospholipids associated with endometrial receptivity

Correlations between the implantation event and metabolites derived from phospholipid metabolism, e.g. eCBs and lysophosphatidic acid (LPA), have been largely studied. However, there is limited information regarding uterine phospholipid profiles during WOI. By employing molecular specificity of mass spectrometry (MS), Burnum et al. (2009) first globally mapped spatiotemporal phospholipid distributions in peri-implantation mouse uterus. Most phospholipids substantially increased in the stroma immediately surrounding the implantation sites, uncovering the complexity of biological processes involved in embryo implantation and postimplantation growth (Burnum et al., 2009). A few years later, the same team replicated the experiments by using another spectrometry technique (i.e. liquid chromatography-ion mobility spectrometry-MS) to quantify lipid distributions. In line with previous results, phospholipids that predominately localize to the mesometrial (top) pole are phosphatidylcholines (PCs) and phosphatidylethanolamines (PEs) containing 18:1; PEs and phosphatidylinositols (PIs) with 20:4 as their PUFAs (Burnum-Johnson et al., 2017). At the antimesometrial (bottom) site, PCs and PIs with 18:2; PC with 20:4; PCs and PIs with 22:6 show higher intensities (Burnum-Johnson et al., 2017).

A defined uterine phospholipid profile might be modulated by different periovulatory endocrine milieus. Distinct phospholipid profiles characterized by the abundance of ceramide (Cer) (42:1) and three species of PC (31:0, 32:1 and 34:4) were observed in bovine uterus associated with greater embryo acceptance (i.e. large follicle-large CL, LF-LCL group) at D4 (Belaz et al., 2016). In contrast, six ions associated with higher-molecular weight PC were more abundant in the uterus with lower receptivity phenotypes (i.e. small follicle-small CL, SF-SCL group) (Belaz et al., 2016).

Growing demands for phospholipid substrates are essential for conceptus nourishment. Total PCs, PIs, phosphatidylglycerol and diacylglycerols were more abundant in the uterine lumen fluid of pregnant versus cyclic sheep (O'Neil and Spencer, 2021). A recent lipidomic analysis revealed that glycerophospholipids and PUFAs were mostly altered in the endometrial fluid of implantative versus non-implantative IVF cycles, suggesting their importance to uterine receptivity (Matorras et al., 2020). Given that glycerophospholipids also serve as precursors of eicosanoids, lysophospholipids and eCBs, glycerophospholipid metabolism disorders may impact pregnancy success rates via the aforementioned pathways.

The sphingolipid metabolic pathway

During early gestation in mice, the sphingolipid metabolic pathway, in particular Cer and the enzymes catalyzing the production of sphingosine and S1P, were highly activated (Kaneko-Tarui et al., 2007). By acting on its G protein-coupled receptors (S1P1-5), S1P signaling was important in uterine decidualization (Mizugishi et al., 2007). Interestingly, the mRNA expression of S1P receptors was down-regulated in human endometrial stromal cells (Brünnert et al., 2014). Under regulation of P4, activated de novo synthesis of sphinoglipids was observed in mouse uterine stromal cells during the decidualization process (Ding et al., 2018). Blockage of this synthetic pathway resulted in decreased number of implantation sites and a defective decidualization process (Ding et al., 2018).

Uterine angiogenesis co-ordination has been implicated as one potential mechanism by which S1P signaling facilitates decidualization. Pioneering studies have discovered that S1P induced COX-2 expression in predecidualized stromal cells, suggesting a link between the sphingolipid and PG signaling pathways in early gestation (Skaznik-Wikiel et al., 2006). Disruption of the sphingosine kinase genes (Sphk, enzyme that catalyze the formation of S1P) caused sphingoid base accumulation and reduced PE levels, followed by uterine vascular bed instability and severely impaired decidualization (Mizugishi et al., 2007). However, COX-2 and PGE2 levels were not affected by Sphk deficiency, indicating that PG biosynthesis was unlikely to be involved in the pathogenesis of Sphk1–/–Sphk2+/– uterus (Mizugishi et al., 2007). Massive infiltration of neutrophils and an exaggerated maternal immune response would be another proposed mechanism in the Sphk deficiency-mediated pregnancy loss (Mizugishi et al., 2015). Neutrophils recruited to the fetomaternal interface and predisposed neutrophil extracellular traps (i.e. the web-like structures composed of DNA fibers, histones and antimicrobial proteins released from neutrophils in response to inflammation and autoimmune stimuli) were recently demonstrated to be involved in early pregnancy loss (Mizugishi and Yamashita, 2017). Nevertheless, knowledge on the functional roles of phospholipids in the implantation process remains limited and more detailed cellular studies are required.

Endocannabinoid system

The endocannabinoid system (ECS) is a lipid signaling system that consists of eCBs, cannabinoid receptors (CB1 and CB2), and the enzymes responsible for eCBs synthesis, degradation and membrane transporters (Walker et al., 2019). The N-arachidonoyl ethanolamine (AEA) and 2-arachidonoylglycerol (2-AG) have been highlighted as two principal eCBs that direct the timely homing of embryos into the uterus (Wang et al., 2007). Besides CB1 and CB2, AEA and 2-AG also bind to the G-protein coupled receptor (GPR) 55, GPR119, PPARs and transient receptor potential vanilloid 1 (Brown, 2007). The biosynthesis and degradation of AEA and 2-AG were thoroughly illustrated in previous studies (Wang and Ueda, 2009; Karasu et al., 2011) (Fig. 3). The eCBs were once postulated to be synthesized and released on demand from phospholipid precursors residing in the plasma membrane (Maccarrone, 2009). However, later studies have challenged this by demonstrating that AEA could be accumulated in lipid droplets (adiposomes), which may improve the half-life of AEA and allow it to trigger nuclear receptors such as PPARs (Maccarrone et al., 2010).

Figure 3.

The biosynthesis and degradation of endocannabinoids. AEA is mainly generated from NAPE which is catalyzed by NAPE-PLD. AEA can be metabolized to AA and ethanolamine by FAAH or by NAAA. DAG is the precursor of 2-AG and is PA or PI catalyzed by PA-phosphohydrolase or by specific PLC. 2-AG can be degraded by MAGL or by FAAH to generate AA and glycerol. 12-HAEA, 12-hydroxy-AEA; 12-HETE-G, 12-hydroxy-eicosatetraenoic-glycerol; 12-LOX, 12-lipoxygenase; 2-AG, 2-arachidonoylglycerol; 5,6-EET-EA, 5,6-epoxyeicosatrienoic acid ethanolamide; AA, arachidonic acid; Abh4, alpha-/beta hydrolase 4; AEA, anandamide; COX-2, cyclooxygenase-2; CYP450, cytochrome P450; DAG, diacylglycerol; DAGL, sn-1-diacylglycerol lipase; FAAH, fatty acid amide hydrolases; GP-AEA, glycerophospho-N-arachidonoylethanolamine; lyso-PI, lyso-phosphatidylinositol; lyso-PLD, lyso-phospholipase D; MAGL, monoacylglycerol lipase; NAAA, N-acylethanolamine-hydrolyzing acid amidase; NAPE, N-arachidonoylphosphatidylethanolamine; NAPE-PLD, N-acyl-phosphatidylethanolamines-specific phospholipase D; NAT, N-acyltransferase; PA, phosphatidic acid; p-AEA, phospho-AEA; PC, phosphatidylcholine; PE, phosphatidylethanolamine; PG-EAs, prostaglandin-ethanolamides; PG-GEs, prostaglandin-glycerol esters; PI, phosphatidylinositol; PLA1, phospholipase A1; PLC, phospholipase C; sPLA2, secretory phospholipase A2. This image contains some elements adapted from the Servier Medical Art Image Bank (Servier Laboratories (Aust) Pty Ltd) licensed under a Creative Commons Attribution 3.0 Unported Licence.

The ECS is modulated by steroid hormones. Peripheral AEA level peaks at the time of ovulation, and then decreases, paralleled by the growing concentrations of its degrading enzyme, fatty acid amide hydrolase (FAAH), in the period that temporally coincides with the putative WOI (Habayeb et al., 2004; Lazzarin et al., 2004; El-Talatini et al., 2010; Cui et al., 2017). Plasma AEA levels positively correlate with serum FSH, LH and E2 levels (El-Talatini et al., 2010; Cui et al., 2017). In vitro, P4 has been shown to stimulate human lymphocyte FAAH activity, and this may account for the decrease in peripheral AEA levels in the secretary phase of menstrual cycle (Maccarrone et al., 2001). The ECS has also been characterized in the endometrium through the menstrual cycle, and a fine regulation of AEA levels in the endometrial environment is necessary for early pregnancy (Taylor et al., 2010; Scotchie et al., 2015).

The effects of maternal steroid hormone on rodent uterine AEA production depend on its reproductive state and the activation of blastocysts (Ribeiro et al., 2009). Earlier evidence in mice indicated that FAAH activity was suppressed by sex hormones (Maccarrone et al., 2000). Considering the low abundance of uterine FAAH in mice (Paria et al., 1999), it is suggested that uterine AEA levels are primarily modulated by the AEA synthetic enzyme, i.e. N-acylphosphatidylethanolamine-hydrolyzing phospholipase D (NAPE-PLD) (Guo et al., 2005). Specifically, E2 and P4 down-regulated uterine NAPE-PLD expression via their nuclear receptors (Guo et al., 2005). However, a recent study revealed that the cannabinoid receptors, alongside NAPE-PLD and FAAH, had been up-regulated by E2 administration, and this effect could be blocked by previous tamoxifen (Maia et al., 2017). In addition, no changes in uterine AEA turnover, but instead an increase of plasma levels, were induced by E2 (Maia et al., 2017). A suggested peripheral contributor to this may be the endothelial cells, since E2 activates NAPE-PLD and inhibits FAAH and, as a consequence, AEA is released into the circulation (Maccarrone et al., 2002a).

There is evidence that AEA exerts a dual function depending on its local concentration, and thereby regulates the WOI by synchronizing trophoblast differentiation with preparation of the uterus for the receptive state. For example, trophoblast differentiation and outgrowth were accelerated by lower levels of cannabinoids, as opposed to their inhibitory effects at higher concentrations (Wang et al., 1999). It seems that enhanced AEA signaling is detrimental in preparing the embryo for implantation (Schmid et al., 1997; Paria et al, 2001). Levels of AEA in the implantation sites were lower than those at the inter-implantation sites (Schmid et al., 1997). The expression of uterine AEA and blastocyst CB1 is co-ordinately down-regulated with the onset of endometrial receptivity and blastocyst activation (Paria et al, 2001). Higher levels of uterine AEA and blastocyst CB1 are observed in the non-receptive uterus and dormant blastocyst, respectively (Paria et al, 2001). Subcutaneous administration of an active natural cannabinoid, Δ9-tetrahydrocannabinol, resulted in implantation failure in wild-type mice (Paria et al, 2001). By acting preferentially on CB1, a narrow range of AEA regulates blastocyst implantation by differentially modulating MAPK signaling and Ca²⁺ channel activity (Wang et al, 2003). Specifically, a low concentration of AEA (7 nM) stimulated blastocyst competency for implantation in the receptive uterus by MAPK/ERK signaling activation. Higher AEA concentrations (28 nM) dampened this response by inhibition of Ca²⁺ influx without affecting MAPK signaling (Wang et al, 2003). In keeping with the results obtained from rodents, low AEA levels have been detected during the WOI in women with a successful pregnancy after IVF-ET procedures (El-Talatini et al., 2009). In contrast, higher AEA levels were noted in IVF-ET patients who failed to achieve ongoing pregnancy (Maccarrone et al., 2002b). It is noteworthy that silencing of eCB signaling by double deletion of CB1 and CB2 is also detrimental to early pregnancy events, including increased risk of midgestational resorption and compromised formation of decidua (Li et al., 2019b, 2020).

In the post-implantation period, the decidua undergoes regression associated with apoptosis and secondary necrosis after reaching its maximum development (Correia-da-Silva et al., 2004). AEA signaling exerts dual effects on cell death, acting as a modulator of decidual regression. Pharmacological activation of CB1 with the synthetic CB1 agonist WIN 55,212-2 inhibited human decidualization and induced apoptosis with the involvement of a cAMP-dependent mechanism (Moghadam et al., 2005). In a rat primary decidual cell model, low levels of AEA included apoptosis mediated by CB1 in a caspase-3/7-dependent manner (Fonseca et al., 2009). This apoptotic signaling pathway depends on AEA-triggered Cer de novo synthesis and p38 MAPK phosphorylation, followed by a decrease in mitochondrial potential and release of reactive oxygen species (Fonseca et al., 2013). Higher concentrations of AEA provoked dramatic alterations in cell viability and morphology, and also a release of lactate dehydrogenase (LDH), probably through the cholesterol-rich lipid rafts (Fonseca et al., 2009). Similarly, AEA interfered with stromal cell differentiation in an immortalized human endometrial stromal cell line (St-T1b) and human decidual fibroblasts from term placenta (Almada et al., 2016a). It has been recently demonstrated that this compromised decidualization could be ascribed to an AEA-induced anti-aromatase effect through down-regulation of CYP19A1 mRNA as well as E2 secretion (Almada et al., 2019).

Unlike AEA, the role of 2-AG in embryo implantation is less well known. Uterine 2-AG concentration is approximately 200-fold higher than AEA, with patterns showing higher levels at the inter-implantation sites and lower levels at the implantation sites (Wang et al., 2007). The spatiotemporal expression of the biosynthetic enzyme diacylglycerol lipase α (DAGLα) coincides with 2-AG, which is up-regulated in the luminal epithelium of the inter-implantation site with the initiation of implantation (Wang et al., 2007). Conversely, the hydrolytic enzyme monoacylglycerol lipase (MAGL) is restricted to the luminal and glandular epithelium of the inter-implantation site, whereas high levels of the enzyme are induced at the site of blastocyst attachment (Wang et al., 2007). This dynamic expression of metabolic enzymes provides evidence that DAGLα and MAGL co-ordinately regulate uterine levels of 2-AG during normal pregnancy, protecting embryos from excessive exposure to 2-AG. Similar to AEA, high levels of 2-AG (50 μM) exert detrimental effects on rat primary decidual cells, leading to a dramatic loss of cell viability and increased LDH release via CB1 activation (Fonseca et al., 2010). In BeWo cell cultures, 2-AG exposure decreased cell viability and proliferation in both a time- and a concentration-dependent manner (Costa et al., 2014).

AEA and 2-AG are also metabolized by COX-2 to produce PG-ethanolamides and PG glycerol esters, respectively, highlighting the crosstalk between the eCBs and PG metabolic pathways (Maia et al., 2020). AEA levels were lower in decidualizing cells compared to non-decidualized ones, and exogenous AEA administration down-regulated COX-2 and PGE2 expression by activating CB1 (Almada et al., 2016b). These results are consistent with the finding that AEA disrupts the decidualizaiton process of rat uterine stromal cells (Fonseca et al., 2015). Two decidual markers, α2-macroglobulin and alkaline phosphatase, were markedly diminished, and accompanied by decreased levels of COX-2 and VEGF (Fonseca et al., 2015).

The LPA-LPA receptor (LPA3) signaling

LPA is a small glycerophospholipid that exerts numerous biological functions, such as promoting cell growth, differentiation, migration and platelet activation and aggregation (Hama and Aoki, 2010). LPA3, a G-protein-coupled receptor, is the major subtype of LPA receptor involved in embryo implantation. It is expressed in the uterine luminal epithelium, with peak expression at 3.5 days post-coitus (D3.5) in mice, when embryo spacing occurs (Ye et al., 2005). LPA3 signaling is highly dependent on the ovarian steroid hormones. P4 administration is able to restore low levels of LPA3 expression in ovariectomized mice while E2 counteracts this effect (Hama et al., 2006). Patients with RIF had decreased levels of LPA3 in the endometrium, indicating a positive effect of LPA-LPA3 signaling on embryo implantation (Achache et al., 2010).

By acting on its cognate receptor LPA3, LPA plays a key role in regulating endometrial receptivity and a local balance of steroid hormone signaling in the uterus (Ye, 2020). LPA3-deficient female mice displayed delayed embryonic implantation, altered embryo spacing and a reduced number of implantation sites (Ye et al., 2005). No obvious defects were observed in ovulation, ovum transportation and blastocyst development (Ye et al., 2005). The delayed implantation process could be ascribed to the persistent epithelial P4 receptor (PGR) signaling after LPA3 deletion (Diao et al., 2015). Loss of PGR expression in the uterine luminal epithelium is essential for the establishment of mammalian uterine receptivity (Bazer et al., 2010). On day 4.5, PGR still remained strongly expressed in the uterine luminal epithelium of LPA3^−/− female mice (Diao et al., 2015). Up-regulation of its target genes, such as Indian hedgehog (Ihh) and alcohol dehydrogenase 5 (ADH5), was also observed at the transcriptional level. Suppression of the overexpressed PGR signaling by RU486 restored appropriately timed implantation in LPA3^−/− mice.

In addition, LPA-LPA3 signaling participates in vascular remodeling at the maternal-fetal interface and guaranteed adequate blood flow in response to the increasing metabolic demand of the embryos (Sordelli et al., 2017). The intrauterine administration of LPA3 antagonist to pregnant rats on Day 5 of gestation decreased the cross-sectional length of the uterus and arcuate arteries. Defects in the uterine vasculature were paralleled by histological anomalies in decidua, and accompanied by decreased mRNA expression of vascularization markers at the resorbed implantation sites. Activation of LPA-LPA3 also increased the activity of inducible nitric oxide synthase (iNOS) (Beltrame et al., 2013), and iNOS-derived nitric oxide, known to be mediators of decidualization and vascularization in embryo implantation (Khorram, 2002). The autotaxin (ATX)-LPA-LPA3 signaling was demonstrated to induce decidualization via canonical heparin-binding epidermal growth factor (HB-EGF) and COX-2 pathways at the embryo-epithelial boundary (Aikawa et al., 2017).

Mounting evidence suggested an interaction between LPA-LPA3 signaling and PG biosynthesis. The implantation phenotypes of LPA3^−/− mice were similar to rodents deficient in cPLA2α or treated with the PG synthesis inhibitor, indomethacin (Shah and Catt, 2005). Incubation of uterine strips obtained from pregnant rats on Day 5 of gestation with LPA increased the production of COX-2 derived PGE2 (Sordelli et al., 2012). In LPA3-deficient female mice, uterine expression of COX-2, PGE2 and PGI2 were suppressed (Ye et al., 2005). However, exogenous administration of PGE2 and cPGI only corrected the delayed implantation defects, but failed to rescue the uneven embryo spacing and limited implantation sites (Ye et al., 2005). Embryo spacing could be regulated separately from implantation timing downstream of LPA-LPA3 signaling. Uterine myometrial contraction mediated by LPA3 signaling would be one possible mechanism controlling this process (Hama et al., 2007). A TXA2 receptor agonist (11-deoxy PGF2a) alleviated embryo crowding by inducing myometrial contraction (Ye et al., 2012).

Cholesterols

Cholesterols, which can be derived either from the diet or synthesized endogenously (the main source in humans), are essential for membrane biogenesis and modulation of membrane fluidity. They occupy the spaces between the polar headgroups of the phospholipid molecular bilayer and reduce the fluidity of the lipid bilayers. Apart from being the precursors of bile acids, vitamin D and steroid hormones, cholesterols also play a key role in membrane trafficking, transmembrane signaling processes and cell proliferation (Fernández et al., 2005). In order to provide a reservoir of FAs for fetal growth and steroid hormone biosynthesis in support of pregnancy, cholesterols and TGs are physically increased during the first and second trimesters (Nasioudis et al., 2019). However, hypercholesterolemia and overweight are risk factors for the live birth rate after IVF/ICSI treatments (Gao et al., 2020). Unfavorable lipid profiles characterized by elevated TC, LDL and TG were observed in women with endometriosis or unexplained infertility, either of which was associated with potentially compromised endometrial receptivity (Melo et al., 2010; Verit et al., 2017).

Serum cholesterol profiles during the menstrual cycle

Several studies have revealed that serum cholesterol levels remain relatively stable during the menstrual cycle (Lebech et al., 1990; Moller et al., 1996; Elhadd et al., 2003; Magkos et al., 2006; Codner et al., 2016; Tzeravini et al., 2021). However, mounting evidence indicated that blood cholesterol profiles fluctuated in response to female reproductive hormones (Lussier-Cacan et al., 1991; Azogui et al., 1992; De Leòn et al., 1992; Karpanou et al., 1992; Nduka and Agbedana, 1993; Schijf et al., 1993; Cullinane et al., 1995; Tonolo et al., 1995; Muesing et al., 1996; Pahwa et al., 1998; Barnett et al., 2004; Saleh et al., 2009; Mumford et al., 2010; Devi et al., 2012; Saxena et al., 2012; Alves et al., 2014; Vashishta et al., 2017; Draper et al., 2018). Here we have summarized publications from 1991 to the present with respect to cyclical changes in serum cholesterol among healthy women (Table II).

Table II

Summary of selected findings (from 1991 to present) evaluating cyclical changes in serum cholesterol levels throughout the menstrual cycle among healthy women.

Study (year)	N	Age	BMI	Cycle	Sample measurements	Changes from follicular to luteal phase	Statistical test
		(years)	(kg/m2)	(days)
Draper et al. (2018)	34	26.6 ± 5.9	22.9 ± 3.5	30.4 ± 3.3	One cycle, 5 samples: menses, follicular, periovular, luteal and premenstrual phase.	TC and HDL-C: decreased in luteal phase (P < 0.05).	Student’s t-test
Vashishta et al. (2017)	111	15–45	<18.5 or >24.9 were exclude	24–38	One cycle, 2 samples: follicular and luteal phase.	TC and LDL-C: decreased in luteal phase (P = 0.006, P = 0.004);	Paired t-test
						TC/HDL-C and LDL-C/HDL-C: decreased in luteal phase (P = 0.006, P = 0.01);
						VLDL-C and HDL-C: NS.
Alves et al. (2014)	10	Mean age 23	Not mentioned	Regular	One cycle, 2 samples: follicular and luteal phase.	LDL-C: decreased in luteal phase (P = 0.03)	Student’s t-test
						TC, HDL-C and VLDL-C: NS.
Saxena et al. (2012)	20	25.3 ± 0.9	23.5 ± 0.6	26–32	One cycle, 2 samples: early follicular and early luteal phase.	TC and HDL-C: increased in early luteal phase (P = 0.003, P < 0.0001);	Paired t-test
						LDL-C and LDL-C/HDL-C: NS.
Devi et al. (2012)	29	19.3 ± 1.2	22.2 ± 0.5	Regular, an average of 28	One cycle, 3 samples: mid-follicular, ovulatory and mid-luteal phase.	LDL-C: decreased in mid-luteal phase (P = 0.01);	Paired t-test
						TC/HDL-C and LDL-C/HDL-C: decreased in mid-luteal phase (P = 0.006, P = 0.001);
						TC and HDL-C: NS.
Mumford et al. (2010)	259	27.3 ± 8.2	24.1 ± 3.9	28.4 ± 2.2	Two cycles, 8 samples: menses, mid-follicular, late follicular, LH/FSH surge, ovulation, early luteal, mid-luteal and late luteal phase.	TC and LDL-C: decreased in luteal phase (P < 0.001);	Linear mixed model
						HDL-C: highest around ovulation (P < 0.001), but did not change across other phases.
Saleh et al. (2009)	19	20–24	21.5 ± 0.74	28.7 ± 1.4	One cycle, 5 samples: early follicular, mid-follicular, late follicular, ovulatory and mid-luteal phase.	LDL-C: decreased in luteal phase (P < 0.02);	Paired t-test
						VLDL-C: lowest around ovulation (P < 0.05);
						TC and HDL-C: NS.
Barnett et al. (2004)	44	26.8 ± 4.1	21.8 ± 1.8	28.5 ± 2.1	One cycle, 2 samples: follicular and luteal phase.	LDL-C: decreased in luteal phase (P = 0.015);	Paired t-test
						TC/HDL-C and LDL-C/HDL-C: decreased in luteal phase (P = 0.0006, P = 0.002);
						TC, VLDL-C, HDL-C: NS.
Pahwa et al. (1998)	30	20.6 ± 2.7	Not mentioned	Regular	One cycle, 3 samples: menses, follicular and luteal phase.	TC and LDL-C: decreased in luteal phase (P = 0.02, P = 0.05);	Paired t-test
						HDL-C and VLDL-C: NS.
Muesing et al. (1996)	12	27.0 ± 3.0	21.0 ± 2.0	27.0 ± 2.0	One cycle, 4 samples: menses, early follicular, late follicular and mid-luteal phase.	LDL-C: decreased in mid-luteal phase (P < 0.01);	ANOVA
						LDL-C/HDL-C: decreased in mid-luteal phase (P < 0.05);
						HDL-C: higher in the late follicular phase (P < 0.05).
Cullinane et al. (1995)	18	33.0 ± 5.7	Not mentioned	Regular	One cycle, 7 or 8 samples:	TC and LDL-C: higher on Days 5 and 10 (follicular phase) after the menses (P < 0.05);	Student’s t-test
					Days 1–2, Day 5 and every 5 days thereafter through Days 1–2 of the next cycle.	HDL-C: highest around days 10 and 15 (late follicular and ovulatory phases) (P < 0.05).
Tonolo et al. (1995)	15	30.0 ± 1.0	20.2–26.1	29.0 ± 1.0	One cycle: Day 1, Day 4 and thereafter daily until the Day 1 of the next cycle.	TC and LDL-C: higher in the preovulatory phase until the day after LH surge compared with menses (P < 0.05);	Paired t-test
						HDL-C: highest around LH surge (P < 0.05).
Nduka and Agbedana (1993)	14	23.0 ± 2.7	<25	Regular	One cycle, 3 samples: follicular, ovulatory and luteal phase.	TC and HDL-C: higher during the ovulatory phase (P < 0.05).	Multivariate Repeated Measures analysis
Schijf et al. (1993)	56	28.0 ± 4.7	22.6 ± 2.4	25–31	One cycle, 2 samples: follicular and luteal phase.	TC and LDL-C: decreased in luteal phase (P = 0.01, P = 0.002);	Wilcoxon’s test
						TC/HDL-C and LDL-C/HDL-C: decreased in luteal phase (P = 0.007, P = 0.004);
						HDL-C: NS.
De Leòn et al. (1992)	29	22.6 ± 1.6	Not mentioned	29.0 ± 1.6	One cycle, 2 samples: follicular and luteal phase.	TC: decreased in luteal phase (P < 0.05);	Student’s t-test
						HDL-C: NS.
Karpanou et al. (1992)	30	39.9 ± 4.3	24.6 ± 2.8	27.8 ± 1.0	One cycle, 3 samples: follicular, ovulatory and luteal phases.	HDL-C: decreased in luteal phase (P = 0.0009);	Paired t-test
						LDL-C: lowest in ovulatory phase (P = 0.01);
						TC and TC/HDL-C: NS.
Azogui et al. (1992)	18	19–44	18–25	Regular, 28–30	One cycle, 3 samples: early follicular, preovulatory and mid-luteal phases.	HDL-C: higher in the preovulatory phase (P < 0.01);	Paired t-test
						LDL-C/HDL-C: lower in the preovulatory phase compared to early follicular phase (P ≤ 0.05);
						TC, LDL-C and VLDL-C: NS.
Lussier-Cacan et al. (1991)	20	33.0 ± 5.0	Not mentioned	Regular	One cycle, 4 samples: menses, follicular, ovulatory and luteal phases.	TC: increased during the follicular and ovulatory phase (P < 0.01, P < 0.05), decreased in luteal phase (NS);	ANOVA

HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; NS, not significant; TC, total cholesterol; VLDL-C, very low-density lipoprotein cholesterol.

These studies showed that TC and LDL-C concentrations increased rapidly after menses, peaking during the follicular phase prior to the rise of E2, and then declining throughout the luteal phase. On the other hand, HDL-C levels were higher during the late follicular and ovulatory phases, but did not change across the other phases. One study reported that the very low-density lipoprotein cholesterol (VLDL-C) levels were lowest around ovulation (Saleh et al., 2009). However, most other studies found that VLDL-C levels did not differ significantly during the ovarian cycle (Azogui et al., 1992; Pahwa et al., 1998; Barnett et al., 2004; Alves et al., 2014; Vashishta et al., 2017).

The CL is the principal source of steroid hormone production (including P4 and E2) during the post-ovulatory phase and early pregnancy in humans (Jensen et al., 2017). Dietary cholesterol reaches the gonadal cells through blood circulation in the form of lipoprotein cholesterol, and serves as the primary substrate for steroidogenesis (Ye et al., 2021). There are several mechanisms involved in this process: membrane scavenger receptor B1 (SR-B1)-mediated selective cholesterol ester (CE) uptake; LDL receptor (LDLR)-mediated endocytosis of LDL-C; intracellular de novo biosynthesis; utilization of plasma membrane cholesterol and stored intracellular lipid droplet CE (Fujimoto et al., 2010). The LH surge is found to up-regulate LDLR expression transcriptionally via a cAMP-dependent mechanism (Murphy and Silavin, 1989). After ovulation, CL vascularization facilitates LDL-C uptake via endocytosis mediated by the LDLR and utilizes cholesterol for steroidogenesis. This may explain the decrease of TC and LDL-C during the mid-luteal phase of menstrual cycle. The altered circulating cholesterol profile mirrors a dramatic demand for steroidogenesis in order to establish a receptive endometrium.

The initial and rate-limiting steps of P4 synthesis include the steroidogenic acute regulatory protein (STAR)-mediated entry of cholesterol into the mitochondria, followed by CYP11A1-catalyzed cholesterol side-chain cleavage (DeWitt et al., 2020). CYP11A1 is responsible for converting cholesterol to pregnenolone, which is subsequently converted to P4 by 3β-hydroxysteroid dehydrogenase, type 2 (3βHSD2; encoded by HSD3B2) in the adrenal glands and gonads (Miller and Auchus, 2011). E2 is derived from either estrone or testosterone, which can be produced from androstenedione in ovarian granulosa cells (Miller, 2017).

CYP11A1 transgenic mice show aberrant implantation, anomalous placentation, and delayed parturition as a result of decreased cholesterol availability and dysregulated P4 production (Chien et al., 2013). Decreased transcript abundance associated with cholesterol transport and P4 biosynthesis, such as STAR, LDLR, SR-B1, CYP11A1 and HSD3B2, is observed in the CL of primates with increased weight (Kuokkanen et al., 2016). This finding suggests that obesity may exert detrimental effects on the endometrium by altering the transcription of genes related to cholesterol metabolism and steroidogenesis.

Steroid hormone participation in establishing endometrial receptivity

P4 and E2 are two major ovarian-derived steroid hormones that direct endometrial receptivity in the lipid metabolism system (Dey et al., 2004). The role of P4 in the maintenance of gestation was first demonstrated from the study of ovariectomized rabbits (Corner and Allen, 1970). New evidence shows that P4, via binding to its cognate receptors PGR-A and PGR-B, is important not only for the maintenance but also the initiation of the pregnancy (Wetendorf and DeMayo, 2014). Insufficient P4 production from the CL or poor endometrial response to P4 may result in compromised implantation (Orazov et al., 2017). Thus, P4 supplementation is routinely recommended for patients with suspected luteal phase deficiency to support successful pregnancies (Liang et al., 2018). P4 has been found to extend endometrial receptivity through day 6 of pseudo-pregnancy in mice (Song et al., 2007). On the contrary, engineered mice in which the function of PGR is abrogated in the endometrial eptithelium exhibited abnormal receptivity phenotypes including embryo attachment failure and defective decidualization (Franco et al., 2012). Consistent with these findings, the PGR antagonist RU486 precluded human embryo implantation in vitro (Boggavarapu et al., 2016). In addition, a disrupted P4 biosynthesis pathway impaired implantation without affecting ovulation, while P4 administration improved pregnancy rates (Chien et al., 2013).

Endometrial receptivity establishment cannot be fully attributed to P4 alone, whose function is well known to be modulated by the counteracting, and sometimes co-operating, action of E2. E2, by binding to its nuclear receptor, such as estrogen receptor-α and -β (ERα and ERβ) in cells, acts together with P4 to regulate a cascade of gene transcription associated with embryo implantation (Ozturk and Demir, 2010). The timing of endometrial receptivity coincides with the down-regulation of epithelial ERα in mid-secretory endometrium (Fox et al., 2016). ERα persistence in the glandular epithelium is associated with infertility and suspected implantation defects (Lessey et al., 2006). In addition, low E2 concentrations have been demonstrated to maintain the WOI, while higher concentrations rapidly closed this period (Ma et al., 2003). In the pre-receptive phase, E2 stimulates proliferation of both endometrial luminal and glandular epithelial cells via stromal ERα (Winuthayanon et al., 2017). Subsequently, increased P4 levels counteract the E2-induced proliferation through PGR in the epithelium and stroma (Li et al., 2011). At the WOI, proliferation in the epithelial compartment ceases and the stromal compartment undergoes the P4-dependent processes of proliferation and differentiation, also termed decidualization (Vasquez and DeMayo, 2013).

IHH signaling is a pivotal P4-PGR mediator in the endometrium, contributing to the crosstalk required for epithelial–stromal interaction and facilitation of a successful pregnancy (DeMayo and Lydon, 2020). Epithelial IHH is controlled by P4 via the PGR as well as GATA binding protein 2 (GATA2), a PGR target gene modifier (Rubel et al., 2016). Remarkable overlap among the sex determining region Y box 17 (SOX17), PGR and GATA2-bound regions are observed, and SOX17 acts on an enhancer region 19 kb distal to the IHH promoter (Wang et al., 2018). As a mediator between the epithelium and stroma, the epithelium-derived-IHH activates its stromal receptor patched-1 (Ptch1) in a paracrine manner. Then, the intracellular transducer smoothened (SMO) and transcription factor glioma-associated oncogene homolog (GLI) are sequentially stimulated, thereby activating the stromal orphan nuclear receptor: chicken ovalbumin upstream promoter transcription factor II (COUP-TFII) (DeMayo and Lydon, 2020). FGF production induces ERK-ERα signaling and maintains epitheial cell proliferation (Chung et al., 2015). COUP-TFII terminates this E2-induced epithelial proliferation by stimulating the stromal heart and neural crest derivatives expressed 2 (HAND2), a repressor of FGF expressed in the stroma (DeMayo and Lydon, 2020). On the other hand, COUP-TFII promotes signaling cascades involving the bone morphogenetic protein 2 (BMP2) and wingless-related MMTV integration site 4 (Wnt4) to initiate stromal decidualization (Wu et al., 2018). Wnt4 is a downstream effector of BMP2 and is positively regulated by P4-PGR signaling (Lee et al., 2007) (Fig. 4).

Figure 4.

Schematic mechanisms of progesterone-progesterone receptor signaling in the regulation of endometrial receptivity. P4-PGR signaling induces the epithelial IHH expression, which subsequently activates stromal Ptch1-SMO/GLI-COUP-TFII signaling, promoting the growth of stromal cells and initiating epithelial differentiation. Stromal P4-PGR signaling promotes the HAND2 expression, subsequently interfering with the FGF/FGFR-ERK-ERα signaling and leading to the cessation of epithelial proliferation. Stromal COUP-TFII promotes activation of Wnt4, a downstream effector of BMP2, to initiate decidualization. BMP2, bone morphogenetic protein 2; COUP-TFII, chicken ovalbumin upstream transcription factor II; E2, estrogen; ER, estrogen receptor; ERK, extracellular signal-regulated kinases; FGF, fibroblast growth factor; FGFR, fibroblast growth factor receptor; GATA2, GATA binding protein 2; GLI, glioma-associated oncogene homolog; HAND2, heart and neural crest derivatives expressed 2; IHH, Indian hedgehog; P4, progesterone; PGR, progesterone receptor; Ptch-1, patched-1; SMO, smoothened; SOX17, sex determining region Y box 17; Wnt4, wingless-related MMTV integration site 4. This image contains some elements adapted from the Servier Medical Art Image Bank (Servier Laboratories (Aust) Pty Ltd) licensed under a Creative Commons Attribution 3.0 Unported Licence.

By abnormally accelerating endometrial maturation, premature P4 elevation during IVF cycles promotes detrimental effects on the endometrium, leading to asynchrony between embryonic development and endometrial receptivity (Drakopoulos et al., 2019). A premature P4 rise above 1.50 ng/ml on the hCG day alters the endometrial genetic pattern on the day of oocyte retrieval or WOI (Labarta et al., 2011; Van Vaerenbergh et al., 2011). The most relevant functions of these differentially expressed genes (DEGs) are related to cellular development, cell adhesion, immune response and intercellular signaling, and the Wnt/β-catenin signaling is the most relevant canonical pathway in regulating these processes (Labarta et al., 2011; Van Vaerenbergh et al., 2011).

Genomic clues hidden in transcriptomic analysis

Transcriptomics, also known as functional genomics, refers to the analysis of the composition, variability and abundance of RNA transcripts that is generated within the cells under particular conditions (Bellver et al., 2012). Global transcriptomic profiling of human endometrium has provided important insights into the underlying mechanisms in response to steroid hormone fluctuation across the menstrual cycle (Teh et al., 2016). The first evidence from microarray analysis concerning lipid participation in uterine receptivity could go back to 20 years ago (Kao et al., 2002). Genes associated with cholesterol trafficking and transport (apolipoprotein E and D, ApoE and ApoD) as well as PG biosynthesis and action (PLA2 and PGE2 receptor) are markedly up-regulated during the WOI (Kao et al., 2002). Up-regulated expression of VLDL receptor (VLDLR), ApoD, ApoL, EP1 and PPARδ were reported almost at the same time (Carson et al., 2002). The elevated genomic abundances of ApoD and PLA2 in the receptive phase was also determined in a later study (Riesewijk et al., 2003),

Transcriptomics based on gene expression microarrays is now the most mature and stable technology available for endometrial dating. A new technology called RNA-sequencing (RNA-seq), which is capable of true genome-wide analysis based on next-generation sequencing, is also emerging. So far, by using microarrays or RNA-seq, studies have identified many DEGs associated with lipid homeostasis between non-receptive and receptive endometrium (Borthwick et al., 2003; Talbi et al., 2006; Haouzi et al., 2009; Tseng et al., 2010; Díaz-Gimeno et al., 2011; Hu et al., 2014; Sigurgeirsson et al., 2017; Suhorutshenko et al., 2018; Yu et al., 2021) (Table III). Furthermore, in women with unexplained infertility endometrial gene expression at the receptive phase was markedly different from that in fertile women (Altmäe et al., 2010). Several aspects of lipid metabolism, such as LA metabolism, AA metabolism and glycosphingolipid biosynthesis, were affected by the dysregulated genes (Altmäe et al., 2010).

Table III

Examples of genes responsible for lipid metabolism that correlate with endometrial receptivity—evidence from transcriptomic analysis.

Gene name	Gene symbol	Functions	Regulation during WOI	Reproductive phenotype in gene knockout or deficient females	References
Apolipoprotein D	APOD	Cholesterol transportation	Up	Not mentioned.	Kao et al., 2002; Carson et al., 2002; Riesewijk et al., 2003; Borthwick et al., 2003; Talbi et al., 2006; Haouzi et al., 2009; Tseng et al., 2010; Hu et al., 2014; Sigurgeirsson et al., 2017; Suhorutshenko et al., 2018; Yu et al., 2021
Apolipoprotein L1	APOL1	Cholesterol transportation	Up	Not mentioned.	Carson et al., 2002; Haouzi et al., 2009; Hu et al., 2014; Sigurgeirsson et al., 2017; Suhorutshenko et al., 2018
Apolipoprotein E	APOE	Cholesterol transportation	Up	Decreased frequency and duration of estrus;	Kao et al., 2002; Talbi et al., 2006; Yu et al., 2021
				Decreased serum levels of estrogen and progesterone;
				Increased numbers of ovarian follicles at birth, but decrease with increasing age. (Zhang et al., 2013)
Fatty acid amide hydrolase	FAAH	Degrading enzyme of anandamide	Up	Defective oviductal embryo transport;	Hu et al., 2014; Sigurgeirsson et al., 2017; Suhorutshenko et al., 2018
				Impaired embryo development;
				Deferred on-time implantation. (Wang et al., 2006)
HMG CoA synthase2	HMGCS2	Regulation in ketogenesis	Down/Up	Not mentioned.	Talbi et al., 2006/Sigurgeirsson et al., 2017
17βHSD type2	HSD17B2	Interconversion of testosterone and androstenedione, as well as estradiol and estrone	Down/Up	Embryonic lethality and growth retardation;	Talbi et al., 2006; Hu et al., 2014; Sigurgeirsson et al., 2017
				Abnormal placenta. (Rantakari et al., 2008)
HMG CoA reductase	HMGCR	Cholesterol biosynthesis	Up	Embryonic lethality. (Ohashi et al., 2003)	Talbi et al., 2006; Hu et al., 2014; Sigurgeirsson et al., 2017
Lysophosphatidic acid receptor 3	LPA3	Receptor for LPA	Down	Altered embryo spacing;	Suhorutshenko et al., 2018; Yu et al., 2021
				Delayed embryonic implantations;
				Reduced number of implantation sites;
				Impaired decidualization. (Ye et al., 2005; Aikawa et al., 2017)
Phospholipase A2, group II A	PLA2G2A	Arachidonic acid biosynthesis	Up	Not mentioned.	Kao et al., 2002; Riesewijk et al., 2003; Talbi et al., 2006; Hu et al., 2014; Sigurgeirsson et al., 2017
Prostaglandin E receptor 2	PTGER2	Receptor for prostaglandin E	Up	Infertility;	Kao et al., 2002; Díaz-Gimeno et al., 2011; Hu et al., 2014; Suhorutshenko et al., 2018; Yu et al., 2021
				Reduced litter size;
				Impaired ovulation and fertilization. (Hizaki et al., 1999; Tilley et al., 1999)
Peroxisome proliferator-activated receptor delta	PPARδ	Ligand-activated transcription factor	Up	Subfertility;	Carson et al., 2002
				Embryonic lethality and growth retardation;
				Placental defects. (Nadra et al., 2006)
Prostaglandin D2 synthase	PTGDS	Conversion of PGH2 to PGD2	Up	Not mentioned.	Borthwick et al., 2003; Hu et al., 2014; Sigurgeirsson et al., 2017
Steroidogenic acute regulatory protein	STAR	Cholesterol translocation to the inner mitochondrial membrane for steroidogenesis	Up	Impaired follicular maturation;	Talbi et al., 2006; Díaz-Gimeno et al., 2011; Hu et al., 2014; Sigurgeirsson et al., 2017; Suhorutshenko et al., 2018; Yu et al., 2021
				Sexually immature uterus and oviducts;
				Absent corpus luteum. (Hasegawa et al., 2000)
Sphingosine kinase 1 and sphingosine kinase 2	SPHK1 and SPHK2	Catalyze the formation of sphingosine-1-phosphate	Up	Uterine vascular bed instability and severely impaired decidualization;	Hu et al., 2014; Yu et al., 2021
				Excessive neutrophil extracellular traps formation at the feto-maternal interface;
				Early embryonic lethality. (Mizugishi et al., 2007; Mizugishi and Yamashita, 2017)
Scavenger receptor class B member 1	SCARB1	Mediation of selective cholesteryl ester uptake	Down	Infertility;	Hu et al., 2014; Sigurgeirsson et al., 2017
				Impaired oocyte and embryonic development. (Trigatti et al., 1999; Miettinen et al., 2001)

17βHSD type 2, 17beta hydroxysteroid dehydrogenase type 2; HMG CoA, 3-hydroxy-3-methylglutaryl coenzyme A; LPA, lysophosphatidic acid; PGD2, prostaglandin D2; PGH2, prostaglandin H2; WOI, window of implantation.

Dysregulated endometrial transcriptomic profiles are found in obese women, particularly those with infertility or PCOS, with the affected genes being related to development, morphogenesis, immune response and molecular functions (Bellver et al., 2011). By comparing the transcriptomics of the receptive endometrium between well-defined obese and non-obese populations, Comstock et al. (2017) identified nine genes that could serve as obesity biomarkers. Aside from immune response, chemotaxis and ECM organization, these genes are also associated with MAPK-ERK signaling, which plays a role in the cellular proliferation, differentiation and invasion of trophoblasts into the endometrium (Comstock et al., 2017).

Even though transcriptomic approaches offer a holistic view of processes involved in endometrial receptivity owing to the simultaneous identification of thousands of DEGs in mid-secretory phase (Hernández-Vargas et al., 2020), there are several limitations to deciphering the complexity of the implantation process. The main obstacle for routine clinical practice is the invasiveness of endometrial biopsy, which delays embryo transfer. While previous transcriptomic studies yield a wealth of functional genes when the pre-receptive endometrium transitioned to the receptive phase, the overlap among different studies is relatively small (Altmäe et al., 2017). These disparities have been attributed to differences in study design, experimental populations, the timing and conditions of endometrial sampling, transcriptome array/sequencing platforms, data processing and statistical analysis methods. Still, it is noteworthy that alterations in gene expression do not always result in changes in protein expression. During embryo implantation, one of the most sophisticated biological processes, many cellular and physiological events can alter the relation between mRNA abundance and protein production. In this case, it is difficult to interpret-specific genetic functions and connect them with underlying mechanisms of implantation failure, thereby leading to a focus on other ‘omics’ method.

Lipidomic analysis: a new approach for evaluating endometrial receptivity

Following the advent of the genomics, transcriptomics and proteomics approaches, lipidomics, which is defined as the large-scale study of lipid metabolites and their related networks, has now emerged as a powerful tool for the study of endometrial receptivity (Gross, 2017). However, compared to other “omics”, lipidomics is probably the least used method for endometrial research (Wenk, 2005).

Endometrial fluid aspiration does not affect pregnancy rates during the same IVF cycle (Boomsma et al., 2009). Thereby, lipidomic analysis of uterine fluid seems to be a promising diagnostic tool for the non-invasive identification of lipids alterations at the WOI. Before MS was utilized to analyze lipid species, the classic methods of study were radioactivity assays or fluorescent probe-labeled lipids (Gandhi and Ross, 1989; Wolf et al., 1992). Modern approaches for analyzing lipid compositions and their structures include electrospray ionization techniques, MALDI, tandem mass spectrometry (MS-MS), gas chromatography, thin layer chromatography, high-performance liquid chromatography and nuclear magnetic resonance (Vilella et al., 2013a).

Studies have focused on lipid molecules that are differentially expressed at the time of the WOI, in order to understand the acquisition of a receptive endometrium, or to determine whether this receptive status of endometrium could be characterized by a specific pattern of lipids. Nine specific lipids have been identified in 51 endometrial fluid samples (i.e. AEA, N-palmitoyl ethanolamine, N-oleoyl ethanolamine, N-stearoyl ethanolamine, N-linoleoyl ethanolamine, 2-AG, PGF1α, PGE2 and PGF2α), with PGE2 and PGF2α being the most relevant lipids in implantation (Vilella et al., 2013b). The sphingolipid monohexosylceramides and Cers were found to be down-regulated in patients with ovarian endometriosis (Domínguez et al., 2017), in which impaired endometrial receptivity is detected (Lessey and Kim, 2017). Given that sphingolipids and Cers are associated with various cellular processes, such as proliferation, migration and apoptosis, decreased levels of these compounds could thus affect embryo implantation (Chrobak et al., 2009; Lee et al., 2014).

Braga et al. (2019) employed MALDI time-of-flight (MALDI-TOF) MS to compare the lipid signature of endometrial fluid samples between positive and negative implantation groups, and built up a prediction model by bioinformatics analysis to predict pregnancy success in FET cycles. In their model, an increased ratio of TG and phospholipids in endometrial fluid was found to associate with WOI displacement by interfering with steroid synthesis and release (Braga et al., 2019). Another lipidomic analysis revealed that eight altered metabolites, including seven glycerophospholipids and an n-6 PUFA, differed between implantation and non-implantation IVF cycles (Matorras et al., 2020). Two infertility-related lipid prostanoids are of significant importance: a higher TXB2 level in recurrent miscarriage and RIF, and high levels of 13, 14-dihydro-15-keto PGF2α in recurrent miscarriage (Demiral Keleş et al., 2020).

Supraphysiological hormone levels induced by FSH in ART had been found to exert adverse effects on endometrial receptivity (Shapiro et al., 2011). Substantial circulating P4 levels during the late follicular phase of COS may be attributed to an amplified response of the granulosa cells of multiple follicles to endogenous LH (Adda-Herzog et al., 2018). In a non-luteinized situation, noticeable P4 production was shown in the granulosa cells after incubation with FSH, primarily by stimulating the expression and enzymatic activity of HSD3B2 (Oktem et al., 2017). Instead of imparing oocyte or embryo quality (Hofmann et al., 1993; Fanchin et al., 1996), the possible detrimental effects of a premature P4 rise on IVF-ET outcome are probably mainly targeted on endometrial receptivity (Fanchin et al., 1993; Huang et al., 2012). Under high levels of P4, specific lipid profiles were identified in the endometrium during WOI, which might result in endometrial advancement and subsequently lead to asynchrony between embryonic development and endometrial maturation (Li et al., 2019a).

Conclusion

Evidence accumulated over the past 20 years highlights a principal role of lipid metabolism in endometrial receptivity, and metabolic disturbances would potentially diminish the endometrial capacity for implantation. Specific lipidomic alterations often occur before the appearance of clinical manifestations. Several lipid molecules have been shown to play crucial roles in modulating embryo implantation. Specifically, FAs could facilitate pregnancy by inducing a modified endometrial inflammatory response via mediation of PG production. PGs and eCBs are two widely studied lipid-derived molecules that participate in establishment of endometrial receptivity in both animal and human studies. LPA-LPA3 signaling serves a positive factor for implantation, embryo spacing and decidualization. Cyclical variation in serum cholesterol profiles throughout the menstrual cycle in healthy women mirrors a demand for steroid hormone synthesis at the time of WOI, during which P4 and E2 synergistically mediate structural and functional alterations in the uterus, ready for the blastocyst. Thus, the identification of lipid metabolites, which are capable of differentiating a receptive from a non-receptive endometrium, helps us not only to develop new biomarkers to predict endometrial receptivity, but also shed light on the molecular mechanisms governing this process.

The implementation of transcriptomics has opended the door to technological advances that will provide a better understanding of reproductive processes. With the help of transcriptomic analysis, accumulated DEGs responsible for lipid metabolism have been identified throughout the menstrual cycle or between infertile and fertile women. However, the specific mechanisms through which these genes modulate the implantation process remain to be further elucidated. Lipidomics research of endometrial fluid seems to be an ideal approach to more directly investigate the relation between lipid metabolism and endometrial receptivity. Current studies have been focusing on untargeted lipidomics, aiming to identify lipid biomarkers for endometrial dating or predicting pregnancy outcome in patients seeking IVF treatments. This is the first step in the long process of biomarker discovery, and further studies for molecular elucidation should be performed. However, compared to tissues, the volume of endometrial fluid available for sampling is relatively small, which may impact the results of lipidomic analysis. By integrating with the investigations from other ‘omics’ analyses, lipidomics will provide important new information on how lipid molecules work together with other secreted compounds to affect endometrial receptivity. A better understanding of the roles and mechanisms of action of lipid molecules in the regulation of reproductive processes would provide novel information for developing more effective therapeutic interventions for implantation failure.

Data Availability

The data underlying this article will be shared on reasonable request to the corresponding author.

Acknowledgements

Some parts of the images used in this article were adapted from the Servier Medical Art Image Bank (Servier Laboratories (Aust) Pty Ltd).

Authors’ roles

T.Y. designed the figures and wrote the manuscript. T.Y. and J.Z. performed literature search and selected the articles. J.Z. designed the tables. F.L. and Y.L. critically revised the manuscript. All the authors have seen and approved the final version.

Funding

This study was supported by the National Key Research and Development Program of China (2021YFC2700404) and the Key Project of Research and Development Plan in Hunan Province (2021SK2028).

Conflict of interest

The authors declare that there is no conflict of interest.