Metabolomics and its applications in assisted reproductive technology

Abstract

Metabolomics, an emerging omics technology developed in the post‐gene age, is an important part of systems biology. It interprets the pathophysiological state of the subject by quantitatively describing the dynamic changes of metabolites through analytical methods, mainly mass spectrometry (MS) and nuclear magnetic resonance (NMR). Assisted reproductive technology (ART) is a method used to manipulate sperm, oocytes, and embryos to achieve conception. Recently, several studies have reported that metabolomics methods can be used to measure metabolites in ART samples; these metabolites can be used to evaluate the quality of gametes and embryos. This article reviews the progress of research on metabolomics and the application of this technology in the field of ART, thus providing a reference for research and development directions in the future.

1. INTRODUCTION

The study of metabolomics emerged from the 1970s to the 1990s after the development of genomics, transcriptomics, and proteomics [1]. The discipline is an important part of systems biology [2] that interprets the pathophysiological state of the research subject by quantitatively describing the dynamic changes of various metabolites in organisms. In systems biology, genomic studies based on proteomics and transcriptomics provide insight into the regulatory level, and metabolomics is the first choice to explore the metabolic phenotypes and phenotypic disorders of diseases affected by dynamic environmental changes [3, 4]. In short, genomics, transcriptomics, and proteomics are the starting points of biological events, and metabolomics is the endpoint. Metabolomics has several advantages. First, metabolites can directly reflect the changes in the body; changes in genes and proteins can be amplified by changes in metabolites, which are easy to observe. Second, the total number of metabolites is limited, eliminating the need to establish a huge database. Third, metabolomics data are easy to analyse. Fourth, sample collection and pretreatment are convenient and can be performed with general technology [5, 6]. Because of these advantages, metabolomics technology has been applied in the field of assisted reproductive technology (ART).

ART is a method used to manipulate sperm, oocytes, and embryos to achieve conception. ART techniques, such as artificial insemination (AI) and in vitro fertilisation and embryo transfer (IVF–ET) and its derivatives, are commonly known as ‘test‐tube baby’ technologies [7, 8, 9]. However, in recent years, the incidence of infertility caused by age and other factors has gradually increased; thus, ART has become a key treatment for infertility. With the continuous development and improvement of science and technology, the clinical pregnancy rate of IVF–ET technology is steadily increasing. The continuous research and optimisation of new methods and technologies enable more patients with infertility to realize their dream of having children [10].

However, as a new medical technology, ART and its related technologies are accompanied by many issues, the most prominent of which are a low implantation rate and a high multiple pregnancy rate. Morphological evaluation is still used to evaluate and select gametes and embryos, but this technique cannot fully reflect the quality of gametes and embryos or reveal genetic or epigenetic defects in embryos with normal morphology [11]. New methods are needed to evaluate the potential of gametes and embryos in clinical settings to improve implantation rates and avoid multiple pregnancies.

Recently, several studies have successfully used metabolomics to measure metabolites such as amino acids, proteins, reactive oxygen species and specific small molecules in the semen, follicular fluid, and embryo culture media [12, 13, 14, 15, 16]. Although no evidence has shown that metabolomics can improve ART outcomes (such as clinical pregnancy and live birth rates), many studies have used metabolomics to analyse the fluid components in the female reproductive tract, which could lead to new technologies and methods to treat infertility. [17] Metabolites associated with oocyte quality could serve as biomarkers for oocyte maturation and ART outcomes [18]. These small molecules can be used to evaluate the quality of gametes and embryos to improve pregnancy rates and reduce the risk of multiple pregnancies. However, the clinical benefits and safety of all types of ART must be considered.

2. TECHNICAL PLATFORM FOR METABOLOMICS

To conduct metabolomics research, samples are collected according to the established metabolomics research methods, and metabolites are obtained by technical means to assess statistical differences in metabolites between groups and conduct further analyses. No analytical methods have been developed to quantify all metabolites in human samples. Researchers have used a variety of analytical techniques to conduct complementary coverage analysis of metabolites; the most widely used technology platforms in metabolomics are nuclear magnetic resonance (NMR) and mass spectrometry (MS) [19]. The following sections provide an introduction to these technical platforms, and Table 1 compares their sensitivity, specificity, and other characteristics [20, 21, 22, 23, 24, 25].

TABLE 1.

Comparison of mass spectrometry (MS) and nuclear magnetic resonance (NRM) technologies [20, 21, 22, 23, 24, 25].

	MS	NMR
Sensitivity	High, detection limit up to femto mole	Low, but can be improved by high field intensity, low temperature, and microprobe and dynamic nuclear polarization
Selectivity	Can be used for selective and nonselective (targeted and nontargeted) analysis	Although several selective experiments have been conducted, such as experiments using selective total correlation spectroscopy (TOCSY), it is usually used for nonselective analysis
Measurement of samples	Different chromatographic techniques are typically required for different types of metabolites	All metabolites with NMR concentration levels can be detected with a single measurement
Sample recovery rate	Destructive technology, but requires a small number of samples	Nondestructive; samples can be recovered and stored for a long time, and the same sample can be analysed multiple times
Repeatability	Moderate	Very high
Sample preparation	Relatively demanding; different columns and optimised ionisation conditions are required	Minimal sample preparation is required
Targeting analysis	Superior targeting analysis	Not related to targeted analysis

2.1. Mass spectrometry

2.1.1. Gas Chromatography–Mass spectrometry

Gas chromatography–mass spectrometry (GC–MS) is an analytical method that measures the ion–charge mass ratio (charge–mass ratio). Depending on the ionisation source, GC–MS can be divided into two types—electron bombardment ionisation mass spectrometry and chemical ionisation mass spectrometry—both of which are used in metabolomics studies. However, the application of GC–MS has been limited to some extent in recent years because the processed samples must meet certain requirements (they should be volatile and stable after volatilisation) [23]. The recently developed full two‐dimensional gas chromatography (GC×GC) is considered to be more suitable for the analysis of complex samples in metabolomics because of its high resolution, short analysis time, and other characteristics, which make up for the shortcomings of GC–MS.

2.1.2. Liquid Chromatography–Mass spectrometry

Compared with GC–MS, liquid chromatography–mass spectrometry (LC–MS) has a higher sensitivity and separating degree, requires a shorter separation time, and uses a lower dosage of reagents; therefore, it has been widely used in the study of metabolomics in recent years [24]. Depending on the column efficiency of front‐end liquid chromatography, LC–MS can be divided into high‐performance liquid chromatography (HPLC) and ultra‐high‐performance liquid chromatography (UHPLC), each of which meets different analytical needs. Furthermore, depending on the ionisation mode of back‐end mass spectrometry, LC–MS methods can be divided into electrospray ion source and atmospheric pressure chemical ionisation source techniques, allowing the comprehensive analysis of metabolites. LC–MS allows a variety of measurement methods through different front‐end and back‐end combinations. The many combinations of LC–MS, along with ultra‐performance liquid chromatography in tandem with time‐of‐flight mass spectrometry (UPLC–TOF), enable the analysis of metabolites and the discovery process of potential biomarkers to be performed quickly and accurately.

2.2. Nuclear magnetic resonance

NMR technology uses the nuclear magnetic resonance spectroscopy of biological body fluids to provide rich information about all small molecule metabolites in an organism. Multivariate statistical analysis and pattern recognition processing of this information have elucidated the status and dynamic changes in related organisms in functional genomics, pathophysiology, and other aspects and revealed their biological significance at the molecular level. NMR causes no damage; its test conditions can be selected within a range of temperatures and buffers; and it enables the study of chemical exchange, diffusion, and internal motion. In addition, a variety of editing techniques with flexible and diverse test methods can be designed. NMR technology allows relatively simple sample analysis; the pretreatment step can be omitted, and the sample can be injected directly. Moreover, it is not destructive to the original sample, enabling superior analysis. Despite these advantages, the sensitivity of NMR is lower than that of mass spectrometry, and it is difficult to guarantee the accuracy of quantitative analysis, so its application is somewhat limited [23]. However, the development of high‐resolution nuclear magnetic resonance (HR‐NMR) technology has greatly improved the sensitivity of this platform [25].

3. SAMPLE COLLECTION AND PRETREATMENT IN METABOLOMICS

Traditional omics cannot fully explain complex physiological and pathological responses, whereas metabolomics attempts to describe all the metabolites within cells or biological systems; thus, metabolomics is used to study the metabolites of biological samples [22].

Metabolomics analysis can be conducted with various samples, including blood, urine, cerebrospinal fluid, saliva, and biopsy tissue or cell extracts. Sample sources are chosen according to the research objectives, sample availability, and analysis platform used. Blood and urine are the most commonly used samples for human metabolomics studies, but researchers can also detect metabolites in other tissue samples, such as saliva, tonsil tissue, adipose tissue, and exhaled breath condensate. In addition, specimens related to systemic diseases can be selected for metabolite analysis to improve the sensitivity and specificity of detection. For example, exhaled breath condensates can be selected for respiratory diseases, follicular fluid can be selected for reproductive system diseases, faeces and gastric juice can be selected for digestive system diseases, saliva can be selected for oral diseases, and joint fluid can be selected for bone and joint diseases.

In the process of sample collection, appropriate collection tubes containing anticoagulants are required for blood samples. Before analysis, the serum or plasma should be separated and the proteins should be removed at 4°C; this step may be one of the main sources of pre‐analysis error in blood metabolomics studies [26]. It is generally believed that the time from blood sample collection to cell separation should not exceed 35 min because a longer interval could allow glucose metabolism in the blood cells and increase the lactic acid level. In addition, repeated freeze–thaw steps should be avoided in experiments [27]. Urine samples have a relatively simple biological composition and low protein content compared with blood samples and typically do not require additional metabolite extraction steps; however, high urea levels in urine samples may damage the analytical instruments. Traditional tissue analysis also requires extraction of a large amount of metabolites, but NMR technology allows high‐resolution magic angle spinning (HRMAS)–NMR to be used to directly analyse tissue samples without the need for pretreatment.

However, metabolomics samples are very vulnerable to circadian rhythms, diet, sex, age, and weight, and these variables are difficult to control. This creates ethical and economic limitations for metabolomics in studies on human samples [28].

4. DATA PROCESSING IN METABOLOMICS

Pattern recognition techniques, such as principal component analysis (PCA); partial least squares (PLS); orthogonal signal correction (OSC); and orthogonal partial least squares (OPLS), which combined OSC and PLS, are commonly used for data analysis in metabolomics.

PCA is the simplest method and requires the least supervision; it reflects the original status of the data and can effectively find and eliminate abnormal samples. This algorithm reduces the dimensions of high‐dimensional datasets to account for as much data variation as possible. The accuracy of PCA can be improved through supervised methods; thus, PCA is often only used as a first step to help develop models with superior supervised classification methods [28]. When differences between groups are small and differences within groups are large, it is difficult to draw accurate conclusions. Once potential biomarkers are identified, supervised methods (such as PLS) can be used to maximise regional classification and identify the most reliable biomarkers. Artificially adding group variables can make up for the shortcomings of PCA and strengthen the significance of the differences between groups. The OSC method is used to filter out noise that is unrelated to the research object and strengthen the variables that are meaningful to the group. OSC is commonly used to improve the integrity of data sets, and in many cases, it can improve the accuracy of the model [29, 30, 31]. OPLS is an extension of PLS that integrates the OSC method and improves the PLS method, making data analysis more accurate and targeted. Therefore, the OPLS method is becoming more widely used in metabolomics [32].

5. PROGRESS IN THE APPLICATION OF METABOLOMICS IN ASSISTED REPRODUCTIVE TECHNOLOGY

Infertility is defined as the inability of a couple of reproductive age to conceive spontaneously after 1 year of regular intercourse without contraception. Studies show that approximately 15% of couples are affected by reduced fertility; the underlying causes may be attributable to the female or the male, or they may be unexplained. ART involves the processing of human germ cells outside the uterus; this processing is consistent with exogenous ovarian stimulation and endometrial preparation and aims to achieve a live birth. Unfortunately, the effect of ART is limited because only 10%–30% of embryos can be successfully implanted—only 30% of embryos produced by ART are transferred to the uterus, and most do not progress to live births. The success rate of ART varies from 4.5% to 40.1% and is closely related to the age and ovarian function of the woman [33, 34, 35, 36].

5.1. Factors influencing pregnancy rate of assisted reproductive technology‐assisted pregnancy

Three critical factors influence the success of ART methods, including intracytoplasmic sperm injection (ICSI) and in vitro fertilisation (IVF): gamete quality, embryo quality, and endometrial receptivity. In clinical practice, gamete quality is assessed by the morphological score. A low abundance of oocytes requires a more advanced evaluation system than morphology alone. In clinical practice, oocytes with a poor morphological grade may result in embryo development without a live birth. High‐quality oocytes may not even be fertilised by ICSI, or they may produce poor‐quality embryos or embryos lacking the potential for endometrial implantation [37, 38]. To assess endometrial receptivity, the endometrium is typically evaluated by ultrasound measurement of the endometrial thickness. Current data suggest that this practice has a limited ability to identify women with a low chance of conceiving through ART treatment.

One way to improve the current pregnancy rate is to transfer multiple embryos; however, this increases the incidence of multiple pregnancies and the associated risks to the mother and foetus during pregnancy. This method also places an economic burden on national health systems because it may lead to life‐threatening complications in newborns and long‐term health effects [39, 40, 41, 42]. In addition, researchers have determined a large number of morphological parameters, such as pronuclear morphology; polar body structure and location; cytoplasm appearance; early cleavage; and blastomere number, size, symmetry, fragmentation, compaction, and swelling, that are associated with higher survival rates of embryos at different stages. [43] Researchers have sought noninvasive methods to assess embryo quality for at least 10 years, [44] leading to the development of many embryo grading systems and specially designed incubators to monitor embryo development and record morphodynamic parameters during cell division. These incubators assess the time required to complete a cell division and reach a continuous cell cycle in a division synchronisation framework [45].

5.2. Application of metabolomics in assisted reproductive technology

Traditional morphological evaluation is the most common method for screening embryos, but it cannot accurately reflect the quality and developmental potential of gametes or embryos. A noninvasive, reliable, and accurate embryo selection scheme for selective single embryo transfer would improve the outcome of assisted reproduction. Low‐molecular‐weight metabolites are the final products of various physiological regulatory processes that occur during oocyte and embryo development. The development of oocytes and embryos can be evaluated by measuring metabolites in the follicular fluid or an embryo culture medium, allowing embryo optimisation. Metabolomics, the qualitative and quantitative analysis of small molecule metabolites in living organisms, reveals the overall change in the metabolic level of the organism. Recently, more attention has focussed on the application prospects of metabolomics in ART, including the detection of semen, follicular fluid, embryo culture fluid, and other specimens.

Ashish et al. measured the levels of lactate, choline, citric acid, alanine, glycerophosphocholine (GPC), glutamine, and other components in semen by NMR and found that GPC, citric acid, tyrosine, and phenylalanine could be used to evaluate sperm quality. This metabolomics method is a noninvasive and rapid approach to probing infertility. [46] Wallace et al. [47] extracted the oocytes and corresponding follicular fluid from 58 patients undergoing IVF and used 1H‐NMR analysis technology to analyse follicular fluid that could be fertilised but not cleaved as well as follicular fluid that could be both fertilised and cleaved. They found significant differences between the metabolite profiles of the two categories of follicular fluid; for example, the glucose, lactate, protein, and choline/choline phosphate profiles differed. Thus, metabolomic profiling of follicular fluid could be used in gamete and embryo selection.

Yun et al. [18] studied the oocyte metabolites of 30 women who received ART treatment. The women were divided into an advanced maternal age group and a young control group. The authors assessed 311 metabolites in the follicular fluid, and 70 metabolites showed significant differences between the groups. Eight metabolites were significantly and positively associated with maternal age. Among them, three metabolites were negatively correlated with the number of oocytes retrieved, and five metabolites were negatively correlated with cleaved embryo numbers. The authors summarised these metabolomics analyses in a schematic representation of potentially altered metabolic pathways in the follicular fluid as age advances (Figure 1).

FIGURE 1.

Schematic representation of the metabolic pathways that are putatively altered in follicular fluids as age advances [18].

Verloes et al. detected the presence and absence of human leucocyte Ag‐G (HLA‐G) in the follicular fluid and an in vitro culture medium at the cleavage and blastocyst stages and found that oocytes, embryos at the cleavage stage, and embryos at the blastocyst stage were secreted. [48] Another study also found that HLA‐G was positively correlated with embryo implantation potential and pregnancy rate. [49] Compared with healthy women of childbearing age, the metabolites of cholesterol, α‐tocopherol, and the high‐density lipoprotein phosphorylcholine were downregulated in patients with polycystic ovary syndrome (PCOS), whereas the metabolites of linoleic acid, fatty egg white, palmitic acid, unsaturated fatty acids, and low‐density fatty egg white were upregulated. The content of the main membrane lipids in plasma samples from patients with PCOS was significantly increased, whereas lysophosphatidylcholine and phosphatidylthanolamine were yellow, and the levels of control samples collected during the body stage were reduced. [50, 51, 52, 53, 54, 55, 56, 57] A preliminary metabolomics‐based study on the metabolic profiles of D3 (cleavage stage) embryo culture media from patients receiving IVF suggested that nontargeted omics analysis of this culture medium could be used to establish an ideal database and draw metabolic profiles. This technique is noninvasive and causes almost no damage to researchers or embryos; thus, it could assist in the screening of high‐quality embryos. The screened differential metabolites (such as L‐glutamine, which is involved in glucose metabolism and protein metabolism pathways; nicotinamide, which is involved in the anaerobic glycolysis of lipid metabolism and glucose metabolism in the body; and linoleic acid, arachidonic acid, and palmitoleic acid, which are involved in monounsaturated fatty acid metabolism) revealed that the fatty acid biosynthesis pathway and central carbon metabolism pathway may be the key molecular events affecting embryo quality. [58] A retrospective analysis used GC–MS to evaluate the metabolomics of the follicular fluid of 13 patients with repeated IVF failure and 15 patients with a successful first‐time IVF pregnancy in the same period. The profiles of the metabolites related to ovum development potential were as follows: increased valine, threonine, isoleucine, cysteine, serine, proline, alanine, phenylalanine, lysine, methionine, and ornithine; decreased dicarboxylic acid and cholesterol. These may be metabolic markers for predicting ovum development potential [59].

Timothy et al. included and compared 21 metabolomic studies of various biofluids in the female reproductive tract. Weak evidence has shown that embryo culture media might be utilised to predict the viability of individual embryos and the implantation rate better than standard embryo morphology. However, some of these studies are not supported by randomized controlled trials; inconsistencies exist between different studies, and no evidence has shown that metabolomics can improve ART outcomes. Nevertheless, metabolomics could provide a more comprehensive understanding of developing oocytes and embryos [17].

6. SUMMARY AND OUTLOOK

Metabolomics technology has very broad application prospects in the field of disease research. Compared with traditional clinical chemistry and other research methods, metabolomics is a ‘holistic view’ research concept that reflects the metabolic characteristics of organisms systematically and comprehensively. Through the analysis of high‐throughput, high‐resolution technology combined with pattern recognition analysis methods, researchers can explore the characteristics and regularity of life activities at the metabolic level and establish a direct relationship between the content of metabolites and biological phenotype changes. These techniques are fast, efficient, noninvasive, highly sensitive, and specific. With the rapid development of metabolomics technologies, they also provide new ideas for solving the bottleneck problems of ART, but more research is needed to further improve these solutions. Metabolomics will likely play a broader role in the future with the continuous progress of analytical and data analysis methods.

AUTHOR CONTRIBUTIONS

Jingying Gao: Conceptualisation; Data curation; Investigation; Project administration; Writing – original draft. Yan Xiao: Conceptualization; Supervision; Writing – review & editing.

CONFLICT OF INTEREST STATEMENT

All the authors do not have a conflict of interest to disclose.

PERMISSION TO REPRODUCE MATERIALS FROM OTHER SOURCES

Figure 1 in this manuscript was taken from other published articles and we have checked the original article and found that it was published with open access under the Creative Commons CC‐BY licence. We also contacted the publisher and the corresponding author by email. The publisher replied that we only had to correctly quote the original article and indicate it in the graph annotation (More details: https://www.frontiersin.org/guidelines/policies‐and‐publication‐ethics).

Gao, J. , Xiao, Y. : Metabolomics and its applications in assisted reproductive technology. IET Nanobiotechnol. 17(5), 399–405 (2023). 10.1049/nbt2.12141

DATA AVAILABILITY STATEMENT

Research data are not shared.