Mucosal biomarkers for endometrial receptivity: A promising yet underexplored aspect of reproductive medicine

ABSTRACT

Annually, approximately 2 million assisted reproductive technology (ART) procedures are performed worldwide, of which, only ~25% lead to successful delivery. There are two major factors contributing to successful implantation: embryo quality and endometrial receptivity (ER). Although embryo quality might be assessed through morphological and genetic testing, no clinically approved techniques are available to evaluate ER. Mucus in different parts of the female reproductive tract contains many cytokines, chemokines, growth factors, and nucleic acids, which influence and reflect various implantation-related processes. Therefore, the aim of the present review was to summarize available data regarding noninvasively obtained mucosal biomarkers for ER and to investigate their ability to predict the outcome of ART procedures. A broad literature search was performed to define studies related to noninvasive ER assessments. More than 50 biomarkers detectable in endometrial fluid, embryo transfer cannula leftover cells and mucus, menstrual blood, cervicovaginal washings are discussed herein. The remarkable methodological heterogeneity of the reviewed studies complicates the comparison of their results. Nevertheless, certain promising analytical targets may already be identified, such as urocortin, activin A, IL-1β, TNF-α, IP-10, MCP-1, and several oxidative stress biomarkers. The present review contains a collection of currently available mucosal biomarker-related data, which may provide insights for future studies.

Abbreviations: ART: assisted reproductive technology; ER: endometrial receptivity; IVF: in vitro fertilization; ICSI: intracytoplasmic sperm injection; IUI: intrauterine insemination; MeSH: Medical Subject Headings; hDP 200: human decidua-associated protein 200; ET: embryo transfer; IL-18: Interleukin-18; LRG: leucine-rich α2-glycoprotein; ROC: receiver operating characteristic; AUC: area under the ROC-curve; LH: luteinizing hormone; LIF: leukemia inhibitory factor; TNF-α: tumor necrosis factor alpha; IFN-γ: interferon γ; MCP-1: monocyte chemoattractant protein-1; VEGF: vascular endothelial growth factor; SOD: superoxide dismutase; CAT: catalase; LPO: lipid peroxidation; TTG: total thiol groups; TAP: total antioxidant power; CE: chronic endometritis

KEYWORDS:

• Biomarkers
• endometrial receptivity
• endometrial fluid
• menstrual blood
• cervicovaginal washings
• assisted reproductive technologies

Introduction

Infertility is defined as the inability to conceive within one year of regular unprotected intercourse (or 6 months if a woman is ≥35 years old) (Practice Committee of the American Society for Reproductive Medicine 2013). Globally, this problem affects 15% of couples of reproductive age, with the male factor serving as a primary or contributing factor in approximately half of cases (Sun et al. 2019). Infertility is defined as a public health priority in many regions of the world (Duffy et al. 2020). Infertility and its treatment inevitably impact patients’ quality of life, causing economic strain, depression, social disruption, and family discord (Gdanska et al. 2017; Massarotti et al. 2019).

Currently, assisted reproductive technologies (ART), such as in vitro fertilization (IVF), intracytoplasmic sperm injection (ICSI), and intrauterine insemination (IUI), are available to treat infertility. Annually, approximately 2 million ART cycles are performed worldwide, of which, only ~25% lead to successful delivery (De Mouzon et al. 2020). Considering the high cost of ART procedures as well as possible adverse effects of an increased hormonal load during cycles, the development of new approaches to predict the outcome of ART treatment is of great value.

Embryo quality is known to be responsible for approximately 30% of the implantation failures (Oron et al. 2014; Zhu et al. 2014). Although the implementation of morphological grading and preimplantation genetic testing helped to moderately increase the likelihood of successful IVF and ICSI outcomes, the implantation rate is still suboptimal, thereby it is important to address the other major cause of implantation failures – impaired endometrial receptivity (ER) (Baltaci et al. 2006).

ER is defined as the ability of the endometrium to accept the embryo, enabling its development and securing optimal trophic and microenvironmental status (Lessey and Young 2019). Currently, there is no widely accepted approach for the quantitative evaluation of this parameter. A plethora of studies have been conducted to evaluate ER based on endometrial biopsy analysis, hysteroscopy, and ultrasound findings (Sakumoto et al. 1992; Jinno et al. 2001; Santi et al. 2012; Voa et al. 2015; Jin et al. 2017). Nevertheless, in a recent meta-analysis, it was concluded that these approaches are characterized by a poor ability to predict clinical pregnancy (Craciunas et al. 2019).

Despite the above-mentioned advances in endometrial biopsy analysis, the modern trends in precision medicine are gravitating toward noninvasive approaches. Fortunately, there are certain options to obtain biomaterial from the female reproductive tract without any traumatizing interventions, such as cervicovaginal/endometrial washing, direct endometrial fluid aspiration, and menstrual blood collection. ER studies involving these biological substrates are scarce and heterogeneous. Therefore, they are either completely ignored or simply mentioned in the currently published reviews and meta-analyses. The aim of the present review was to summarize the available knowledge regarding noninvasively obtained mucosal biomarkers for ER and to investigate their ability to predict the outcome of ART procedures.

Search strategy

A literature search was performed using PubMed resources: MEDLINE, PubMed Central, and Bookshelf. To gain insight into the historical evolution of ER studies, no restrictions regarding publication date were applied. The search input consisted of a broad spectrum of terms and keywords related to the topic and Medical Subject Headings (MeSH), such as endometrium, uterus, ER, assisted reproductive technologies, embryo implantation, pregnancy, biomarkers, and endometrial fluid. The PubMed function ‘Similar articles’ were used as an auxiliary tool to reveal articles beyond the scope of the implemented search input.

Single endometrial fluid biomarker studies

A summary of studies that sought to analyze a single molecular biomarker in endometrial fluid is presented in Table 1. In all cases, endometrial fluid samples were obtained by flushing the uterine cavity with a physiological solution. The timing of sample collection varied significantly from study to study.

Table 1. Summary of studies that sought to analyze a single endometrial fluid biomarker. 

Study	ART	Sampling	Biomarker	Cutoff	Statistical outcome^a
Halperin et al. 1995 (n = 109)	IVF	Flushing (0.1 mL), on the day of ET	hDP 200	>1000 mU/mg	Sensitivity 57.14 (28.86–82.34) %; specificity 69.47 (59.18–78.51) %
Lédée-Bataille et al. 2004 (n = 133)	IVF	Flushing (1 mL), on the day of oocyte retrieval	IL-18	>0 pg/mL	Sensitivity 85.71 (71.46–94.57) %; specificity 35.16 (25.44–45.88) %
Gillott et al. 2008 (n = 31)	N/A	Flushing (6 mL), timing is N/A	LRG	N/A	LRG III and IV isoforms increased in spontaneously pregnant group vs. treatment dependent pregnant and non-pregnant groups, p < 0.01
Florio et al. 2008 (n = 71)	IUI	Flushing (2 mL), on the day of IUI	Urocortin	>0.321 pg/mL	Sensitivity 60.70 (40.60–78.50) %; specificity 97.7 (87.7–99.6) %; AUC 0.88 (0.78–0.94)
Florio et al. 2010 (n = 50)	IUI	Flushing (2 mL), on the day of hCG prescription	Activin A	>0.04 ng/mL	Sensitivity 76 (56.6–88.5) %; specificity 100 (86.7–100) %; AUC 0.93 (0.82–0.98)
Bentin-Ley et al. 2011 (n = 96)	IVF	Flushing (1 mL), at LH+1 and LH+7^b	Glycodelin	N/A	At LH+1: AUC 0.58 (0.44–0.72); At LH+7: N/S

^aFormat of the statistical outcome regarding clinical pregnancy is not unified due to heterogeneity and limitations in data presentation of the reviewed articles, only statistically significant results are included, data are presented with 95% confidence interval, ^b days 1 and 7 after LH peak. n: number of patients included in the study, ART: assisted reproductive technologies, IVF: in vitro fertilization, IUI: intrauterine insemination, ET: embryo transfer, N/A: not available, N/S: not significant, AUC: area under the receiver operating characteristic curve, hCG: human chorionic gonadotropin, LH: luteinizing hormone, hDP: human decidua-associated protein 200, IL: interleukin, LRG: leucine-rich α2-glycoprotein.

Human decidua-associated protein 200

The first attempt to evaluate the relationship between endometrial fluid biomarkers and IVF outcomes was conducted by Halperin et al. (1995). Human decidua-associated protein 200 (hDP 200), which was previously described and detected in endometrial secretion by the same research group, was selected as a biomarker (Halperin et al. 1990; Golan et al. 1994). The secretory glycoprotein, hDP 200, was hypothesized to be involved in the functional development of the endometrial tissue. hDP 200 was evaluated in endometrial washings obtained on the day of embryo transfer (ET). A significant positive correlation (p < 0.05) was found between the hDP 200 level and pregnancy rate. However, analytical parameters such as sensitivity and specificity were suboptimal (57.14% and 69.47%, respectively). No further attempts to evaluate its clinical value regarding ART procedures outcomes were performed. Three years later, this research group attempted to investigate the ability of hDP 200 to detect ectopic pregnancy; however, no statistical significance was found (Halperin et al. 1998).

Interleukin-18

Interleukin-18 (IL-18) plays an important role in the regulation of the uterine cytokine microenvironment (Ida et al. 2000; Yoshino et al. 2001; Yasuda et al. 2019). IL-18 modulates innate and acquired immune responses, controls the Th1/Th2 balance, promotes the expansion of natural killer cells and changes in their phenotype (Smeltz et al. 2001; Senju et al. 2018). Due to its ability to possibly reflect the whole status of cytokine balance in the uterus, Lédée-Bataille et al. (2004) selected IL-18 as a biomarker for ER. In their study (n = 133), IL-18 was analyzed in endometrial fluid collected prior to oocyte retrieval in patients undergoing IVF. In a group of patients with undetectable levels of IL-18, the rate of pregnancy was higher than that found in the group with detectable IL-18 levels (37.9% vs. 15%, respectively, p < 0.05). Despite the statistical significance of raw group comparisons, the prognostic value of IL-18 as a sole biomarker for ER remains low. Nonetheless, in contrast to hDP 200, IL-18 has been studied in future multiple biomarker studies, which will be discussed later.

Leucine-rich α2-glycoprotein

Leucine-rich α2-glycoprotein (LRG) plays a role in the regulation of immune cell infiltration, re-epithelialization, and angiogenesis (Liu et al. 2020). Knowledge regarding the role of LRG in ER is limited, although there is indirect evidence of its participation in implantation and decidualization (Tabibzadeh and Babaknia 1995). Such findings served as the basis of a small-scale study conducted by Gillott et al. (2008). The study included 31 patients who underwent unspecified infertility treatment in an ART clinic. A statistically significant abundance of LRG isoforms III and IV (p < 0.01) was observed in women with treatment-independent pregnancies (n = 3) compared to non-pregnant (n = 25) and pregnant women (n = 3). These results did not promote any future studies on ER using LRG as a biomarker.

Urocortin

Urocortin is a peptide produced by the luminal and glandular epitheliocytes of the human endometrium (Florio et al. 2002). This peptide is closely related to corticotropin-releasing hormone in terms of structure, and partially shares receptors with this hormone (Reyes et al. 2001). There is evidence that urocortin might play a role in decidualization and endometrial immune response, inducing the expression of various cytokines (Torricelli et al. 2007; Novembri et al. 2011). Florio et al. (2008) conducted a prospective controlled study (n = 71) to evaluate urocortin as a biomarker of successful IUI outcomes. Endometrial wash fluid was obtained prior to the IUI in the periovulatory phase. Urocortin levels were significantly higher in pregnant than non-pregnant study participants (0.38 vs. 0.13 ng/mL, respectively, p < 0.001). Based on receiver operating characteristic (ROC) analysis, the researchers revealed that at a cutoff value of 0.321 ng/mL, this biomarker achieved sensitivity of 60.7% and specificity of 97.7%, with an area under the ROC-curve (AUC) as high as 0.88. Surprisingly, despite the great results of this study, subsequent biomarker studies did not include urocortin.

Activin A

In contrast to previously discussed biomarkers, the involvement of activin A in essential endometrial processes has been supported by numerous in vitro and in vivo studies (Jones et al. 2000; Popovici et al. 2000; Florio et al. 2003). This glycoprotein belongs to the transformation factor β superfamily and is mainly expressed in reproductive organs. Its functions include stimulation of decidualization, trophoblast invasion, embryonic differentiation, and fetal brain protection under hypoxic conditions (Bloise et al. 2019). Due to its prominent role in the implantation process, Florio et al. (2010) examined the ability of activin A to reflect the ER. Endometrial washing samples were collected during the periovulatory phase on the day of human chorionic gonadotropin prescription. The levels of activin A in the pregnant group were statistically significantly increased relative to levels in the non-pregnant group (0.800 vs. 0.022 ng/mL, respectively, p < 0.001). According to the results of ROC analysis, activin A levels were characterized by a sensitivity of 76%, specificity of 100%, and AUC of 0.926 (at a cutoff value of 0.04). Thus, activin A demonstrated excellent predictive power for successful implantation in patients undergoing IUI. Nevertheless, there have been no attempts to investigate its predictive power in other cohorts of patients receiving ART treatment, which already appears to be a common trend in the field of endometrial fluid biomarker research.

Glycodelin

Glycodelin is a secretory glycoprotein, which is primarily expressed in different parts of the male and female reproductive system (Uchida et al. 2013). Glycodelin is frequently referred to as placental protein 14 because of its initial discovery in amniotic fluid (Krivonosov et al. 1979). In humans, glycodelin is present in four glycoforms (glycodelin-A, -S, -F, -C), which are characterized by different functions, including various interactions with spermatozoa, endometrial epithelial cell adhesion toward embryos, and suppression of natural killer cytotoxicity (Uchida et al. 2013). The expression of glycodelin in endometrial tissue was found to be significantly upregulated during the peak estradiol and progesterone levels timed to luteinizing hormone (LH) surge – this exact timeframe is considered to be ‘the window of implantation’ by many researchers (Kao et al. 2002a).

Based on this evidence, Bentin-Ley et al. (2011) conducted a study (n = 96) to evaluate glycodelin as a biomarker for ER. Endometrial washing samples were obtained on days 1 and 7 after the LH surge (LH+1 and LH+7). At LH+7, no statistically significant difference was observed between the fertile and infertile groups, in contrast to LH+1, where glycodelin levels were found to be decreased in fertile patients. ROC analysis revealed poor accuracy of this biomarker at LH+1 to predict successful IVF outcomes (AUC of 0.58).

Multiple endometrial fluid biomarkers studies

Studies that sought to analyze multiple biomarkers for ER in endometrial secretion are summarized in Table 2. In these studies, uterine flushing was not the only option to acquire biomaterial owing to the emergence of a new approach – the direct aspiration of small volumes of endometrial fluid. All researchers in this section collected samples on the day of ET prior to the procedure. The development of bead-based multiplex immunoassays enabled the simultaneous quantification of up to 100 analytes in a single 50 µL sample (Houser 2012). This technology was expected to be prevalent in multiple endometrial fluid biomarkers studies, but was only implemented once (Boomsma et al. 2009).

Table 2. Summary of studies that sought to analyze multiple endometrial fluid biomarkers. 

Study	ART	Sampling	Biomarkers	Cutoff	Statistical outcome^a
Lédée-Bataille et al. 2002 (n = 33)	IVF or IUI	Flushing (1 mL), on the day of ET	LIF, TNF-α	N/A	LIF: Pregnant group (0 (0–177) pg/mL) vs. non-pregnant group (203 (0–1620) pg/mL)^b, p < 0.01
Boomsma et al. 2009b (n = 210)	IVF/ISCI	Direct aspiration (1–4 µL), on the day of ET	IL(−1β, −5, −6, −10, −12, −15, −17, −18) TNF-α, IFN-γ, MMIF, eotaxin, IP-10, MCP-1, DH-1, HBEGF, VEGF	N/A	IL-1β/TNF-α: AUC 0.61 (0.52–0.69)
Rahiminejad et al. 2015 (n = 50)	IVF	Direct aspiration (1–4 µL), on the day of ET	IL-1β, TNF-α, IP-10, MCP-1	N/A	TNF-α: 13.5 (10.3; 230.45) vs. 25.1 (17.9;32.7) pg/mL; MCP-1: 250.5(134.2; 372.5) vs. 1050.3 (106.3; 1995.6) pg/mL; IP-10: 298.2 (267.6; 407.4) vs. 634.7(290.7; 996.3) pg/mL^c; for pregnant and non-pregnant groups, respectively, p < 0.01
Rahiminejad et al. 2016 (n = 100)	IVF	Direct aspiration (1–4 µL), on the day of ET	SOD, CAT, LPO, TTG, TAP	N/A	CAT: AUC 0.77(0.66–0.87); TAP: AUC 0.79 (0.69–0.89); SOD: AUC 0.71 (0.60–0.82)
Khadem et al. 2019 (n = 76)	IVF	Flushing (2–3 mL), on the day of ET	IL-1β, TNF-α	N/A	N/S

^aFormat of the statistical outcome regarding clinical pregnancy is not unified due to heterogeneity and limitations in data presentation of the reviewed articles, only statistically significant results are included, data are presented with 95% confidence interval, ^b median (range), ^c median (interquartile range). n: number of patients included in the study, ART: assisted reproductive technologies, IVF: in vitro fertilization, IUI: intrauterine insemination, ICSI: intracytoplasmic sperm injection, ET: embryo transfer, N/A: not available, N/S: not significant, AUC: area under the receiver operating characteristic curve, IL: interleukin, TNF: tumor necrosis factor, IFV: interferon, MMIF: macrophage migration inhibitory factor, IP-10: IFN-γ-inducible 10 kDa protein, MCP-1: monocyte chemo-attractant protein-1, DH-1: Dickkopf homolog-1, HBEGF: heparin-binding epidermal growth factor, VEGF: vascular endothelial growth factor, SOD: superoxide dismutase, CAT: catalase, LPO: lipid peroxidation, TTG: total thiol groups, TAP: total antioxidant power.

Cytokines

In the past 20 years, a plethora of global gene expression studies have been conducted to describe the variability of the transcriptome at different points of the menstrual cycle (Kao et al. 2002a; Laird et al. 2006; Wang et al. 2020). For all the discussed cytokines, there is growing evidence of up-/downregulation either at the implantation window or in fertile/infertile patients. It is well known that cytokines guide all steps of implantation and embryo-endometrial crosstalk (Massimiani et al. 2020). Therefore, the reasons for the selection of various cytokines as ER biomarkers are clear.

The first study that evaluated the ability of more than one biomarker to reflect the ER was conducted by Lédée-Bataille et al. (2002) (n = 33). Leukemia inhibitory factor (LIF) and tumor necrosis factor alpha (TNF-α) were the analytical targets of choice. Their levels were evaluated in endometrial washings of patients undergoing either IVF or IUI. Statistically significant differences were only found for LIF concentrations, which differed between pregnant (0 [range: 0–177] pg/mL) and non-pregnant (203 [range: 0–1620] pg/mL) patients (p < 0.01). These results were surprising as the researchers initially expected contrary outcomes due to the results obtained by Laird et al. (1997), who demonstrated that LIF levels were higher in fertile patients than in infertile patients. Unfortunately, no data regarding the prognostic value of LIF as a biomarker of successful IVF or IUI outcomes are available.

The next study, carried out by Boomsma Carolien et al. (2009) (n = 210), is remarkable because it is the only study that utilized the above-mentioned bead-based multiplex immunoassay, which allowed the simultaneous determination of 17 cytokines. Endometrial fluid was directly aspirated in a volume of 1–4 µL from patients receiving IVF/ISCI treatment. The multiplex immunoassay included: IL-1β, IL-5 (excluded from the analysis due to low detection rate), IL-6, IL-10, IL-12, IL-15, IL-17, IL-18, TNF-α, interferon γ (IFN-γ, was undetectable in all samples), macrophage migration inhibitory factor, eotaxin, IFN-γ-inducible 10 kDa protein (IP-10), monocyte chemoattractant protein-1 (MCP-1), Dickkopf homolog 1, heparin-binding epidermal growth factor (excluded from the analysis due to low detection rate), and vascular endothelial growth factor (VEGF). The results of multivariable logistic regression revealed statistically significant associations between MCP-1 (p < 0.01) and IP-10 (p < 0.05) with implantation, and associations between IL-1β (p < 0.05) and TNF-α (p < 0.05) with clinical pregnancy. The AUC for IL-1β and TNF-α as biomarkers for clinical pregnancy was suboptimal (0.61), and inclusion of embryo quality slightly improved the model (AUC of 0.67).

Two small-scale studies that aimed to prove the results obtained by Boomsma Carolien et al. (2009) were conducted by Rahiminejad et al. (2015) and Khadem et al. (2019). Both studies included patients undergoing IVF, although Khadem et al. (2019) (n = 76) analyzed IL-1β and TNF-α in endometrial washings, while Rahiminejad et al. (2015) (n = 50) used the direct aspiration method to assess the levels of IL-1β, TNF-α, IP-10, and MCP-1. In contrast to Boomsma Carolien et al. (2009), who used a multivariable approach, in both studies, the predictive power of the biomarkers was evaluated individually in a direct comparison of the pregnant and non-pregnant groups. Thus, for further evaluation of the data, it is important to mention that Boomsma Carolien et al. (2009) did not observe any statistically significant differences for individual biomarkers, but found statistically significant differences for certain combinations.

Rahiminejad et al. (2015) demonstrated that TNF-α levels were decreased in the pregnancy group (p < 0.01), which contradicts the results of Khadem et al. (2019). Neither study demonstrated the ability of IL-1β in endometrial fluid to predict clinical pregnancy. Finally, IP-10 and MCP-1, which were the subjects of only one study, were indicative of clinical pregnancy (p < 0.01).

This discrepancy in the results of the three multiple biomarkers studies originates from methodological heterogeneity. Currently, it is difficult to evaluate the real prognostic value of TNF-α, IL-1β, IP-10, and MCP-1 in endometrial fluid, and an additional investigation is needed.

Biomarkers of oxidative stress

Oxidative stress is defined as an imbalance between free radicals and antioxidants in a certain environment (Jones 2008). The role of oxidative stress in various pathological conditions is mostly clear, but is still extensively investigated (Sies 2015). Endometrial cells are also subject to reactive oxygen species-mediated damage, which may be part of the ER reduction process (Wojsiat et al. 2017). Rahiminejad et al. (2016) (n = 100) conducted a study to evaluate common oxidative stress biomarkers in endometrial fluid by means of direct aspiration on the day of ET in patients undergoing IVF/ICSI. The activities of superoxide dismutase (SOD), catalase (CAT), lipid peroxidation (LPO), total thiol groups (TTG), and total antioxidant power (TAP) were analyzed. Statistically significant differences were found in the levels of SOD, LPO, CAT, and TAP in pregnant and non-pregnant patients (p < 0.01). ROC analysis revealed descent predictive power for CAT, TAP, and SOD individual levels (AUC of 0.766, 0.788, and 0.711, respectively). Unfortunately, the data are limited, and a multifactor model simultaneously including CAT, TAP, and SOD levels was not developed. Nevertheless, this study clearly demonstrated that oxidative stress biomarkers have great potential for the prediction of clinical pregnancy. Their inclusion in future multiple biomarkers studies is desirable.

Endometrial fluid sampling

According to the methodologies of the discussed studies, there are two main approaches to endometrial fluid collection: flushing and direct aspiration. Cheong et al. (2013) argued that washings are inferior to direct aspirates due to increased dilution and leakage of flushing solution through fallopian tubes, which makes it difficult to compare the extent of dilution in different samples. The statement regarding increased sample dilution is undeniable, as the mean volume of saline water, which was used to obtain endometrial washings in the previously discussed articles, was 2 mL. However, for direct aspiration, the samples were diluted with only 70 µL of physiological solution. Nevertheless, the extent of dilution is difficult to compare in both sampling methods, as the direct aspirates, according to Boomsma Carolien et al. (2009), are characterized by inconsistent volumes of 1–4 µL. Thus, samples required normalization by the total protein level, which was successfully implemented by the researchers. Of note, normalization was not used in any of the studies that utilized endometrial flushing.

Indeed, extensive dilution may force the researchers to work with low target concentrations, neighboring the immunoassay limit of detection, which was the case in all the discussed studies. Information about the linearity of the immunoassays used in the lowest concentration range was questionable and has not been presented in most studies.

Thereby, it is interesting to compare the results of IL-18 quantification in endometrial washings and direct aspirates. Lédée-Bataille et al. (2004) reported that only 38/133 patients had detectable levels of this cytokine in endometrial washings and that the difference in pregnancy rates between the IL-18 positive and negative groups was statistically significant (p < 0.05). In contrast, Boomsma Carolien et al. (2009) demonstrated the absence of statistical significance, while IL-18 was detectable in all 210 patients. However, in the latter, an analysis technique with a 10-fold lower limit of detection was used; this discrepancy was, by a large margin, a result of an excessive dilution, which was more than 10-fold higher in cases of uterine flushing.

Other sources of mucosal ER biomarkers

Endometrial fluid is the most obvious source of ER biomarkers; however, it has certain drawbacks, such as low sampling volume and relative invasiveness, as access to the uterine cavity is required. Secretion in other parts of the female reproductive tract may also reflect endometrial status. The available sources of mucosal ER biomarkers are summarized in Figure 1. Although follicular fluid has been extensively studied regarding ART procedures outcomes, it was not considered to be the subject of the present review. This is because follicular fluid biomarkers predict the outcome of ART procedures via reflection of certain oocyte related characteristics, such as oocyte quantity and quality, fertilization rate, and embryo fragmentation rate, rather than ER (Ciepiela et al. 2018; Nagy et al. 2019; Yang et al. 2020).

Figure 1. Sources of mucosal biomarkers for endometrial receptivity. ET: embryo transfer, IVF: in vitro fertilization.

Cervicovaginal washings

Cervicovaginal secretion has a clear advantage over endometrial fluid because its sampling is incomparably easier and less invasive. Notably, cervical secretion has a distinctive cytokine profile compared to endometrial fluid, as shown by Boomsma Carolien et al. (2009) in a within-patient comparison (n = 22). However, these data do not exactly mean that cytokine levels in cervicovaginal washings are not suitable for the analysis of ER. In fact, Zanotta et al. (2019) proved the opposite in a study with 155 women undergoing IVF cycles. A panel of 48 cytokines, chemokines, and growth factors was analyzed using a multiplex bead-based immunoassay in cervicovaginal washings obtained prior to oocyte retrieval. Logistic regression revealed a pattern of four pro-inflammatory immune factors that were significantly increased in patients who achieved clinical pregnancy (IL-12p40, IFN-α, macrophage migration inhibitory factor, and MCP-3, p < 0.001). Data regarding predictive power for successful IVF outcomes of these biomarkers are not available. Nevertheless, these results are of great value as they clearly demonstrate the potential of cervicovaginal secretion as a source of biomarkers for ER. Further large-scale studies of this biomaterial with the inclusion of additional biomarkers are needed.

Embryo transfer cannula leftover cells and mucus

Endometrial epithelial cells at the time of ET might serve as the best sources of information regarding ER in the current IVF cycle. However, it is difficult to imagine an ethics committee that allows traumatizing endometrial biopsy immediately after (or prior to) ET. A sophisticated approach to solve this issue was proposed by Camargo-Díaz et al. (2017) (n = 76), who used the ET cannula leftover cells and mucus as a biomaterial for ER assessment. Hematoxylin-eosin staining with microscopic evaluation confirmed that cells that were attached to the ET cannula were mostly of endometrial origin. Total RNA was isolated from a mixture of endometrial epithelial cells and mucus. The levels of mucin 1 (MUC-1), homeobox A10 (HOXA-10), LIF, and colony stimulating factor-1 (CSF-1) RNA molecules were selected as ER biomarkers because of their altered expression profiles in infertile patients (Othman et al. 2012). LIF and CSF-1 were significantly upregulated in pregnant patients (p < 0.05). The increased LIF expression contradicts the previously discussed results of Lédée-Bataille et al. (2002), who observed a significant decrease in LIF levels in the endometrial fluid of pregnant patients and confirmed the results of Laird et al. (1997), who compared fertile and infertile women. Thus, the ability of LIF to reflect the ER remains unclear and requires further investigation.

Menstrual blood

Menstrual blood has been extensively studied as a source of stem cells, which has been applied in regenerative medicine, although the use of this biological fluid as a substrate for diagnostic manipulations is limited (Bozorgmehr et al. 2020). Menstrual blood has certain advantages over endometrial fluid. This biomaterial can be easily and painlessly collected directly from the uterine cavity as the cervix is slightly dilated during menstruation. Its main drawback is obviously the fact that it can be sampled only in a certain short window of the menstrual cycle, which is quite delayed from the day of implantation.

Currently, menstrual blood is not used for ER evaluation; however, Tortorella et al. (2014) successfully attempted to detect three common pro-inflammatory cytokines (IL-1β, IL-6, and TNF-α) in this biomaterial of 64 infertile patients to develop a new noninvasive approach to chronic endometritis (CE) diagnosis (Tortorella et al. 2014). It is known that CE is prevalent in patients with infertility, repeated implantation failure, and recurrent miscarriage, although the clinical significance of this condition is still debated (Cicinelli et al. 2015). A recent meta-analysis evaluating the impact of antibiotic treatment of CE on IVF outcomes concluded that histological confirmation of CE cure leads to increased pregnancy and live-birth rates compared to cases of persistent CE (Vitagliano et al. 2018).

Tortorella et al. (2014) demonstrated that all three pro-inflammatory cytokines were significantly increased in women with histologically proven CE (p < 0.001). ROC analysis revealed that the IL-6/TNF-α-based model is an efficient predictor of CE with a sensitivity of 91.7% and specificity of 96.4% (AUC of 0.989). According to these results, menstrual blood is a promising biological substrate for the analysis of various biomarkers, although its suitability for the evaluation of ER is a subject for future studies.

Safety of endometrial sampling

Mainly among clinicians, the endometrium of women undergoing the IVF cycle is considered an extremely fragile system; thus, any attempt to obtain its biomaterial (especially close to the day of implantation) will be declared undesirable. Although the previously discussed endometrial fluid aspiration is hardly an invasive procedure, its non-disruptiveness for implantation must be demonstrated. van der Gaast et al. (2003) and Boomsma Carolien et al. (2009) conducted studies to prove the safety of direct endometrial fluid aspiration prior to ET using the same method. They compared 276 women undergoing the above-mentioned sampling with 276 matching controls; there was no statistically significant difference in pregnancy rates between the groups. Therefore, endometrial fluid has certain advantages over endometrial biopsy, namely it is less invasive and safe to collect even on the day of ET. However, it is clearly less suitable for genetic and histological evaluation because of the low content of the desired targets.

Evidently, no studies that performed the risk assessment of endometrial biopsy on the day of ET are available; however, certain studies have evaluated the safety of this procedure in the preceding ovarian stimulation cycle and found that endometrial injury, as a result of biopsy, might be beneficial for further implantation (Barash et al. 2003). Further small-scale studies and reviews indicated that endometrial scratching indeed improves the outcomes of ART procedures, and clinicians began to implement this intervention with quite limited scientific data on the underlying mechanism and real clinical benefit, which might be obtained only as a result of large randomized controlled trials (Lensen et al. 2016). The biological mechanism remained purely hypothetical, and it was thought that endometrial injury induces certain regenerative processes that lead to positive changes in cytokine balance and alterations in tissue thickness (Li and Hao 2009).

Following the completion of proper randomized controlled trials and respective meta-analyses years later, it was concluded that overall endometrial scratching had either no effect on clinical pregnancy and live-birth rates, or a slightly positive effect in the case of a certain subgroup analysis; thus, it is not advised for practical use until further evidence is available (van Hoogenhuijze et al. 2019; Sar-Shalom Nahshon et al. 2019). For ER biomarkers studies, these data indicate that endometrial biopsy in the cycle preceding IVF is a rather safe intervention. However, the main disadvantage is that this process is spaced in time from the moment of actual implantation.

Discussion

Healthy embryo-endometrial crosstalk is vital for a successful pregnancy. It is mediated by various complex molecular interactions that occur throughout gestation. However, this begins with implantation, and the goal of ART specialists is to guide a patient to this stage. All other steps of embryo-endometrial crosstalk are left to further echelons of obstetric care. Successful implantation requires proper ER and endometrial selectivity. Endometrial selectivity is the ability of the endometrium to distinguish low quality embryos (Macklon and Brosens 2014). In clinical practice, certain tools are available to filter embryos and meet the criteria raised by the endometrium, such as the previously mentioned morphological and genetic preimplantation testing. Currently, there is no option not only to reliably increase ER, but also to measure it.

Thus, the main issue in the field of ER research is the lack of a gold standard for its evaluation. Currently, the clinical pregnancy rate is the only suitable option to confirm the presence of receptive endometrium. Unfortunately, the outcome of implantation is influenced by several other factors, including embryo quality, uterine structural defects, maternal systemic conditions, and various environmental factors. Therefore, the use of clinical pregnancy as a reference to the presence of a receptive endometrium will never allow mucosal biomarker-based tests to achieve a sensitivity of 100%. Although it is theoretically possible to build a comprehensive prognostic model that includes not only biomarker-related data, but all possible embryo characteristics, results of hysteroscopy, and demographic/health-related data, it is still impossible to account for all other factors. The cumulative pregnancy rate calculated in a series of ART cycles might reduce their influence. However, this approach complicates the study design, and patient adherence will become a major concern.

Although a great variety of mucosal biomarkers for ER has been studied, data regarding their clinical potential are limited. Owing to remarkable heterogeneity in the methodology and the selected biomarkers, it is impossible to conduct a meta-analysis.

The levels of various molecular agents and oxidative stress influence the ER; however, their interactions are complex. Owing to the simplicity of statistical analysis in most discussed studies, certain promising biomarkers were compared individually among the studied groups and no attempts were made to develop a complex multifactor model.

A multivariable approach, which might reflect the relationship between certain analytical targets, rather than individual comparisons, is the ideal path for ER biomarkers studies. Modern laboratory analysis markets provide various options for biomarker screening. In addition, premade 48-target multiplex immunoassays are commercially available. However, these immunoassays were underused in the discussed studies.

Among the mentioned molecular biomarkers, urocortin and activin A demonstrated the best prognostic value, with an overall accuracy of 79%–88% (Florio et al. 2008, 2010). Interestingly, both molecules were completely abandoned in subsequent studies. It is advisable to incorporate these molecules into future multiple biomarkers panels to further validate the promising results obtained previously.

Therefore, the design of a future ideal ER study must include simultaneous analysis of multiple biomarkers, preferably with an assessment of oxidative stress level (which also seems promising), multivariable approach to the statistical analysis, and an attempt to create a complex prognostic model. A topic beyond the scope of the present review is the influence of endometrial microbiota and immunological balance on ER. There is growing evidence that uterine bacteria play an important role in the modulation of immune cell subsets in the endometrium (Benner et al. 2018). Recently, it was reported that endometrium and ovulatory cervical mucus of patients with recurrent pregnancy loss are characterized by an increased density of natural killer cells CD56+ and CD16+, indicating a possible autoimmune component of infertility (Ulčová-Gallová et al. 2019). Menstrual blood might be also used as a source of endometrial lymphocytes, as it is known to have a clearly distinct immune microenvironment compared to peripheral blood (van der Molen et al. 2014). Thus, the results of microbiological and immunological assessment of endometrium-related biomaterial may reveal potential prognostic factors.

Various sample collection timings were employed in these studies. Intuitively, the patient’s biomaterial was the most predictive of ART treatment outcome on the day of ET. Accordingly, this day was the sampling timing of choice in most cases. However, from a purely practical point of view, this timing is useless for predicting the implantation outcome, as all clinical manipulations would have been carried out by the time the results arrive from the laboratory. Currently, there are insufficient data regarding the consistency of ER biomarkers levels in the following menstrual cycles. From a clinical perspective, the optimal implantation outcome prediction timing should be either at the beginning of the current cycle or in the previous cycle, assuming it is proven that a certain biomarker is consistent across menstrual cycles. Therefore, studies evaluating the consistency of ER biomarkers levels across different days of the cycle as well as in consecutive cycles are needed.

Conclusions

A receptive endometrium is crucial for successful implantation. The discovery of a reliable approach to measure ER may substantially promote an investigation of the influence of various internal and external factors on the likelihood of pregnancy, in addition to its prognostic value in ART procedures. Currently, our knowledge about the molecular mechanisms underlying impaired ER is limited. The studies reviewed had methodological heterogeneity and inconsistent results. Nevertheless, based on the data discussed, mucosal biomarkers can reflect the state of the endometrium and the noninvasiveness of their analysis makes them a remarkable alternative to endometrial biopsy. The challenge of studying ER is that it is influenced by various cross-disciplinary factors. Comprehensive research requires cooperation from specialists in reproductive medicine, endocrinology, immunology, and possibly microbiology. Therefore, future large-scale studies of mucosal ER biomarkers in endometrial fluid, menstrual blood, ET cannula leftover cells, and cervicovaginal washings are anticipated.

Ethics approval

Not applicable.

Authors’ contributions

Conceptualization: MJ, OP; methodology: EM, MJ; software: MJ; validation: OP; formal analysis: OP; investigation: EM, MJ; resources: LS; data curation: LS, OP; writing-original draft preparation: MJ; writing-review and editing: EM, OP; visualization: MJ; supervision: LS; project administration: LS; funding acquisition: LS. All authors have read and agreed to the published version of the manuscript.

Disclosure statement

The authors report no conflict of interest.

Additional information

Funding

This review was prepared as part of the Lomonosov Moscow University State Assignment [0908-2021].