The role of abnormal hypermethylation of the HOXA10 and HOXA11 promoters in implantation failures in IVF programs

Abstract

Objective: The objective of this study is to investigate the association between methylation levels of HOXA10 and HOXA11 promoters, clinical parameters, and implantation outcomes after in vitro fertilization-embryo transfer (IVT-ET) cycles in women with repeated implantation failures and tubal infertility.

Methods: Endometrium samples were collected from 34 women during implantation window before IVF-ET cycle to assess methylation status of HOXA10 and HOXA11 promoters using bisulfite sequencing. All participants had a tubal factor of infertility and at least two implantation failures in the anamnesis. A logistic regression model was used to predict the implantation outcome depending on methylation status and clinical parameters.

Results: The methylation of CpG-islands of HOXA10 and HOXA11 promoters was identified in 76.5 and 100% of participants, respectively. The median methylation levels did not differ significantly between the groups with different implantation outcomes, but a logistic regression model based on HOXA10 and HOXA11 methylation and clinical parameters allowed to classify the implantation outcomes with the total percentage of correct predictions of 85.19%.

Conclusions: Abnormal methylation levels of the HOXA10 and HOXA11 promoters were found in the endometrium of women with tubal infertility and repeated implantation failures. The findings suggest that methylation status could be an important factor of implantation failure during IVF-ET cycles.

Keywords:

• In vitro fertilization
• implantation failure
• infertility
• receptivity
• methylation
• HOXA10
• HOXA11

Introduction

About 10–15% of couples of the reproductive age suffer from infertility. Causes of infertility vary and involve male, female, or combined factors. When initial treatments do not work, a couple may apply for assisted reproductive technology. Sometimes couples undergoing in vitro fertilization (IVF) treatment experience recurrent implantation failure (RIF). RIF has been defined as failure of implantation in at least three consecutive IVF attempts in which 1 or 2 embryos of high-grade quality were transferred in each cycle [1]. The implantation process consists of three main components: hormonal environment, a healthy embryo that should have the ability to implant, and a receptive endometrium. Any abnormality attributed to one of these components may result in unsuccessful IVF attempts [2–4]. It is known that the transfer of an embryo without genetic abnormalities after preimplantation genetic screening in many cases does not result a successful pregnancy, which may be due to alteration of endometrial receptivity (ER) or hormonal stimulation [5,6].

A functioning and receptive endometrium is crucial for embryo implantation [7]. The ‘window of implantation’ (WOI) is defined as a period when the uterus is receptive for implantation of embryo. WOI is characterized by histological and morphological changes of endometrium and ER is determined by several factors [8]. It should be recognized that at the moment there is no single and optimal marker of ER, because of lack of accurate and objective methods of ER assessment [9]. Therefore, it is important to find new informative biomarkers and noninvasive methods of ER assessing, which could predict the pregnancy rate in IVF programs.

The purpose of this work was to study the methylation status of the promoter regions of the HOXA10 and HOXA11 genes in endometrial biopsies in women with infertility caused by tubal factor and unsuccessful IVF attempts in history.

Material and methods

Study population

This study was conducted in the National Medical Research Center for Obstetrics, Gynecology and Perinatology named after Academician V.I. Kulakov. The inclusion criteria were as follows: infertility caused by tubal factor; at least two cases of implantation failure after IVF-ET; age from 18 to 40 years; normal ovarian reserve. The exclusion criteria were as follows: previous ovarian surgery; abnormal semen analysis of male partner; endometriosis; leiomyoma in or near the uterine cavity; anovulation. In 2017–2018, 34 patients who were undergoing IVF treatment were included. Patients were divided into two groups based on pregnancy outcomes. Study population characteristics are shown in Table 1. The study was approved by the Ethics Committee of the National Medical Research Center for Obstetrics, Gynecology and Perinatology named after Academician V. I. Kulakov. Written informed consent was provided by all participants.

Table 1. Comparison of demographic and patient characteristics between two groups.

	Group A (n = 10)	Group B (n = 24)	p
Average age (year)^b	33.0(32.0–34.0)	36.0(32.0–39.0)	.053
BMI (kg/m^2)^b	25.9(23.4–26.9)	23.5(20.4–25.6)	.354
Gynecologic and obstetric history
Age of menarche (year)^b	13.0(13.0–14.0)	12.0(13.0–14.0)	.678
Length of period^b	29.0(27.0–30.0)	28.0(27.5–28.0)	.511
Sexually transmitted diseases^a	80.0%(8/10)	50.0%(12/24)	>.05
Endometriosis^a	20.0%(2/10)	12.5%(3/24)	>.05
Endometrial polyps and/or endometrial hyperplasia^a	30.0%(3/10)	37.5%(9/24)	>.05
Infertility history
Duration of infertility (year)^b	5.5(5.0–10.0)	5.5(2.5-9.5)	.834
Etiology of infertility^a			>.05
Tubal factor only	50.0%(5/10)	50.0%(12/24)	–
Tubal factor + male factor	50.0%(5/10)	50.0%(12/24)	–
Type of infertility^a			>.05
Primary infertility	60.0%(6/10)	54.0%(13/24)	–
Secondary infertility	40.0%(4/10)	46.0%(11/24)	–
Gravidity^b	1(0–1)	0(0–2)	1.000
Gravidity after IVF^b	0(0–1)	0(0–0)	.266
Parity^b	0(0–0)	0(0–0)	.623
Spontaneous abortion^b	0(0–1)	0(0–0)	.658
Serum AMH (ng/mL)^b	2.31(1.18–5.47)	1.25(1.01–2.02)	.967

^a Data were presented as relative (%) and absolute frequencies.

^b Data were presented by median (interquartile range).

Group A: women who became pregnant following fresh embryo transfer. Group B: women who did not become pregnant following fresh embryo transfer.

Ovarian stimulation

All patients underwent controlled ovarian hyperstimulation with gonadotropins using a GnRH antagonist protocol. The starting dose of recombinant FSH (Gonal-F, Serono, Geneva, Switzerland) was 75–225 IU on days 3–5 of the menstrual cycle. Follicular development was assessed by transvaginal ultrasonography. A dose 0.25 mg of GnRH antagonist (Cetrotide, Serono, Geneva, Switzerland) was administered when the leading follicle is 14–15 mm and then administered daily by subcutaneous injection, until human chorionic gonadotropin (hCG) administration. A dose of 10,000 IU of hCG (Pregnyl, Organon, Oss, The Netherlands) was administered when the leading follicle reached 18 mm in diameter. Transvaginal oocyte retrieval, along with IVF and embryo culture, was performed 36 h after hCG injection. All patients underwent ET with 1 good-quality embryo on day 5 after oocyte retrieval and fertilization. The luteal phase was supported using 600 mg of progesterone vaginal capsules (Utrogestan, Besins Healthcare, London, UK) three times per day starting from the oocyte retrieval day. Clinical pregnancy was defined as the presence of a gestational sac on transvaginal ultrasound approximately 5 weeks after ET.

Endometrium sampling

Endometrial specimens were collected using the Pipelle Endometrial Suction Curette during the implantation window, between the seventh and ninth day after ovulation, determined by ultrasound and the Clearblue ovulation test (SPD Swiss Precision Diagnostics GmbH, Geneva, Switzerland) . Endometrial specimens were washed three times with a sterile PBS solution and frozen at −20 °C until further analysis.

DNA extraction and bisulfite conversion

Before DNA extraction, endometrial specimens were thawed and homogenized using the FastPrep-24 Classic Grinder with the Lysing Matrix D (MP Biomedicals, Solon, OH). DNA extraction was performed using the innuCONVERT Bisulfite All-In-One Kit (Analytik Jena AG, Jena, Germany) from no more than 1 mg of homogenized fresh tissue. DNA concentration was determined using the Qubit 2.0 Fluorometer with the Qubit dsDNA HS Assay Kit (Life Technologies, Carlsbad, CA). For bisulfite conversion using the innuCONVERT Bisulfite All-In-One Kit, 150 ng of isolated DNA were used. The DNA concentration after bisulfite conversion was determined using the ND-1000 Spectrophotometer (Thermo Fisher Scientific, Waltham, MA), as described previously [10,11].

PCR and sequencing

For PCR, 20 ng of DNA after bisulfite conversion were used. Amplification was performed using EpiMark Hot Start Taq DNA Polymerase (New England Biolabs, Ipswich, MA) in two parallel reactions with one of two primer pairs:

Table

HOXA10F	TATGATATTGTTGTGGGATAATTTG
HOXA10R	CATAAACACCCCACTTATTAACTAC
HOXA11F	ATTAGTATTTTTTTGTTAAGGATGGG
HOXA11R	CCTACAAACAATCTCTATACAC

Amplification products were separated by electrophoresis in 2.5% agarose gel with the addition of ethidium bromide and 100 bp Molecular Ruler (Bio-Rad, Hercules, CA). Target amplification products were isolated from the gel using the GeneJET Gel Extraction Kit (Thermo Fisher Scientific, Waltham, MA). Bisulfite sequencing was performed using 16 ng of amplification products using the 48-capillary 3730 DNA Analyzer with the Platinum Taq DNA Polymerase and the BigDye Terminator version 3.1 Cycle Sequencing Kit (Thermo Fisher Scientific, Waltham, MA) based on the Engelhardt Institute of Molecular Biology of Russian Academy of Sciences. The sequencing results were analyzed using the DNA Sequencing Analysis Software version 5.1 (Thermo Fisher Scientific, Waltham, MA) and the Quantification Tool for Methylation Analysis online resource (quma.cdb.riken.jp). The degree of promoter methylation of the HOXA10 and HOXA11 genes was determined as the proportion of methylated CpG-islands.

Statistical analysis

All data were analyzed using IBM SPSS Statistics version 23.0.0.0 (IBM, Armonk, NY) and Statistica version 12.0 (StatSoft, Tulsa OK). The data are presented as the medians (quartile range) for quantitative indicators and the percentages (absolute number) for qualitative indicators. Univariate analyses of categorical data were performed using a Fisher’s exact test, and quantitative indicators were compared using a Mann–Whitney U-test. p < .05 was considered significant. A logistic regression model was used to predict the pregnancy outcome for an individual woman. The factors included were binary variables (the presence of artificial and/or spontaneous abortion in medical history; the presence of inflammatory diseases in medical history) and quantitative indicators (age; body mass index (BMI); the number of births; AMH level; duration of infertility; the level of methylation of HOXA10 and HOXA11 genes).

Results

No significant differences were detected in age, BMI, gynecologic, obstetrics, and infertility history among the two groups (Table 1). While analyzing the sequencing result, it was found that methylation of promoter region of HOXA10 was observed in 76.5% (26/34) patients and HOXA11 – in 100.0% (34/34) patients. There were no statistically significant differences between the studied groups in the methylation levels of HOXA10 and HOXA11 genes (Table 2).

Table 2. Comparison of methylation levels of HOXA10 and HOXA11 genes between two groups.

	Group A (n = 10)	Group B (n = 24)	p
Methylation of HOXA10 promoter, %	12.07 (0.00–32.00)
	12.07 (0.00–21.14)	11.90 (0.00–33.24)	.658
Methylation of HOXA11 promoter, %	72.56 (55.56–84.21)
	80.65 (57.89–94.74)	68.42 (51.47–78.71)	.150

Data were presented by median (interquartile range).

Group A: women who became pregnant following fresh embryo transfer. Group B: women who did not become pregnant following fresh embryo transfer.

A multivariate logistic regression model was developed to determine key factors associated with implantation (and correspondingly pregnancy). Significant interactions were not observed between predictors. The best predictors for implantation were age, level of AMH duration of infertility, sexually transmitted diseases in history, level of methylation of HOXA10 and HOXA11 genes (Table 3).

Table 3. Pregnancy prediction model for the logistic regression analysis.

	Model 1	Model 1b	Model 1c	Model 1d
Correctly predicts pregnancy (%)	80.00	70.00	70.00	80.00
Correctly predicts absence pregnancy (%)	88.24	82.35	70.59	76.47
Overall percentage (%)	85.19	77.78	70.37	77.78
p Value	.009	.016	.010	.009

Model 1: the final model which includes: age, level of AMH, duration of infertility, sexually transmitted diseases in history, level of methylation of HOXA10 and HOXA11 genes.

Model 1b: The level of methylation of HOXA10 gene was excluded from Model 1.

Model 1c: The level of methylation of HOXA11 gene was excluded from Model 1.

Model 1d: The levels of methylation of HOXA10 and HOXA11 genes were excluded from Model 1.

Discussion

Our study showed that women with a history of repeated implantation failures show abnormal methylation of the HOXA10 and HOXA11 promoters, the multivariate model for predicting pregnancy in this group has a higher predictive ability when taking into account methylation status. Regulated HOXA10 and HOXA11 expression is necessary for ER; promoter hypermethylation and decreased expression of HOXA10 and HOXA11 lead to decreased implantation rates associated with polycystic ovarian syndrome, endometriosis, leiomyoma, endometrial polyps, and hydrosalpinx [12,13]. However, the role of HOXA10 and HOXA11 expression in tubal infertility remains controversial. One study showed that HOXA11 expression was significantly impaired in women with unexplained infertility during implantation window, while in women with tubal infertility factor, the expression of HOXA10, and HOXA11 was not significantly different from fertile women [14]. In another study, the endometrial HOXA10 expression was decreased in women with RIF and recurrent miscarriage comparing to fertile women [15]. None of these studies considered the methylation level of the promoters.

The promoter methylation level may change under the influence of various substances and factors, which may become the basis for the development of new approaches to the management of infertility. In vitro, in Ishikawa cells, the proportion of methylated CpG islets of the HOXA10 gene was 72%; however, it was decreased to 38 and 35% when exposed to 5-aza-2’-deoxycytidine (AZA), DNA-methyltransferase inhibitor, at a concentration of 1 and 10 μM, respectively. AZA strongly induced the mRNA and protein expression of HOXA10 and its targets ITGB3 and IGFBP1. Knockdown of HOXA10 led to decreased expression of HOXA10, ITGB3, and IGFBP1, with and without AZA treatment. The attachment rate in a Jeg-3 spheroid-endometrial cell attachment assay was 82% in control experiment and increased to 95 and 96% after 1 and 10 μM AZA treatment, respectively [16]. To date, the only approved DNA-methyltransferase inhibitors for therapy belong to the nucleoside-based family of drugs, but they display relevant side effects making them unacceptable in pregnancy planning. Increasing literature evidence is highlighting that natural sources could help to regulate gene methylation level, because several polyphenols, flavonoids, anthraquinones, and others are described able to inhibit DNA-methyltransferase inhibitors activity and/or expression, thus decreasing the methylation/silencing of different genes [17].

Gonadotropin releasing hormone (GnRH) analog mouse models were treated with either human menopausal gonadotropin (HMG) and a GnRH agonist or HMG and a GnRH antagonist. Uterus samples were collected 48 h after GnRH analog treatment or ovulation for determination of HOXA10 methylation. A total of 37 (17.6%) methylated CpG sites were observed in the samples from the GnRH antagonist group, compared with 14 (6.7%) methylated sites in the samples from the GnRH agonist group and four (1.9%) methylated sites in the samples from the natural cycle control group; all differences are significant [18]. It was also shown that chronic exposure to polychlorinated biphenyls produced an increased number of implantation failures in association with a defective uterine morphology during the implantation period in mice, and alterations in methylation of HOXA10 could explain, at least in part, the mechanism of these effects [19]. These results may indicate the need to control factors potentially influencing the HOXA10 and HOXA11 methylation when planning pregnancy in women with repeated implantation failures.

In conclusion, we found abnormal methylation levels of the HOXA10 and HOXA11 promoters in the endometrium of women with tubal infertility and repeated implantation failures. Our data showed that methylation status could be an important factor of implantation failure during IVF-EC cycles.

Disclosure statement

The authors report no conflicts of interest.