Association between follicular fluid phthalate concentrations and extracellular vesicle microRNAs expression

Abstract

STUDY QUESTION

Are phthalate metabolite concentrations in follicular fluid (FF) associated with the expression of extracellular vesicle microRNAs (EV-miRNAs)?

SUMMARY ANSWER

Phthalate metabolite concentrations are associated with the expression of EV-miRNA and their associated pathways in FFs.

WHAT IS KNOWN ALREADY

Phthalate metabolites were recently detected in FF. Urinary phthalate metabolite concentrations alter the expression of EV-miRNAs in FF.

STUDY DESIGN, SIZE, DURATION

Prospective study including 105 women recruited between January 2014 and August 2016 in a tertiary university-affiliated hospital.

PARTICIPANTS/MATERIALS, SETTING, METHODS

We assessed FF concentrations of 12 phthalate metabolites. EV-miRNAs were isolated from aliquots of the same FF, and their expression profiles were measured using a human miRNA panel. Associations between EV-miRNAs that were present in >50% of the samples and phthalate metabolites that were measured in >74% of the FF samples were tested. Genes regulated by EV-miRNAs that were found to be significantly (false discovery rate q-value < 0.1) correlated with FF-phthalates were analyzed for pathways linked with female fertility using miRWalk2.0 Targetscan database, DAVID Bioinformatics Resources and Kyoto Encyclopedia of Genes and Genomes (KEGG).

MAIN RESULTS AND THE ROLE OF CHANCE

Of 12 phthalate metabolites, 11 were measured in at least one FF sample. Mono (6-COOH-2-methylheptyl) phthalate (MCOMHP), mono-2-ethyl-5-carboxypentyl phthalate (mECPP), mono-n-butyl phthalate (MnBP), monobenzyl phthalate (MBzP), mono-isobutyl phthalate (MiBP), monoethyl phthalate (MEP) and mono (7-COOH-2-methyloctyl) phthalate (MCOMOP) were detected in more than 74% of the samples. Of 754 EV-miRNAs tested, 39 were significantly associated either with MEP, MBzP, MCOMOP, MCOMHP and/or with mECPP, after adjusting for multiple testing (P < 0.05). KEGG-based pathway enrichment analysis of the genes regulated by these miRNAs showed that these EV-miRNAs may be involved in pathways related to ovary or oocyte development, maturation and fertilization.

LIMITATIONS, REASONS FOR CAUTION

The use of miRNA panel array limits the number of potential relevant miRNAs. Moreover, several of the phthalate metabolites examined may be biased due to internal (enzymatic activity) or external (contamination in medical interventions) causes.

WIDER IMPLICATIONS OF THE FINDINGS

Phthalate metabolites may alter follicular EV-miRNAs profile and thus impair pathways that are involved with oocyte development, maturation and fertilization. Our results contribute to understanding of possible mechanism(s) in which endocrine disruptor chemicals interfere with female fertility.

STUDY FUNDING/COMPETING INTERESTS

This work was supported by the National Institutes of Environmental Health Sciences [Grant R21-ES024236]; and Environmental Health Fund, Israel [Grant 1301], no competing interests.

TRIAL REGISTRATION NUMBER

N/A.

Introduction

Phthalates are man-made chemicals that are primarily added to plastics to increase their flexibility (Hauser and Calafat, 2005). They are commonly found in a variety of consumer products, including personal care products, plastic packaging, toys, vinyl flooring and wall covering, detergents, lubricating oils, food packaging and pharmaceuticals (Heudorf et al., 2007; Diamanti-Kandarakis et al., 2009; Koniecki et al., 2011). While exposure to phthalates is ubiquitous, urinary phthalate metabolite levels were found to be higher in women than in men (CDC, 2017). Phthalates are considered endocrine disruptors and are associated with a variety of adverse health outcomes, such as developmental anomalies (Hauser and Calafat, 2005; Meeker et al., 2009), cancer (López-Carrillo et al., 2010) and adverse female fertility outcomes (Toft et al., 2012; Meeker and Ferguson, 2014; Hauser et al., 2016; Jukic et al., 2016; Machtinger et al., 2018). In previous studies, we have shown negative associations between urinary phthalate metabolite levels and IVF outcomes with regard to total and mature numbers of oocytes retrieved, fertilization rates and Day 3 embryo quality (Hauser et al., 2016; Machtinger et al., 2018).

Phthalate metabolites have been also detected in follicular fluid (FF) (Krotz et al., 2012; Hannon et al., 2014; Du et al., 2019), and a recent study suggested that FF phthalate metabolite levels may alter FF hormone levels (Du et al., 2019). The mechanism(s) by which phthalates affect the human ovary remain largely unexplored.

MicroRNAs (miRNAs) are non-coding RNAs that regulate gene expression based on a partially complementary profile of the miRNA to its mRNA target, which may lead to mRNA degradation. MiRNAs are found in most tissues, bio-fluids and also in extracellular vesicles (EV), which are lipid bilayer-delimited particles that are released from various cells in both physiological and pathological conditions and serve as a vehicle to various molecules. Extracellular vesicle miRNAs (EV-miRNAs) are considered more resistant to degradation than free miRNAs (Bartel, 2009). While total miRNAs in biofluids may also be released from apoptotic cells or cell debris, EV-miRNAs are actively released by viable cells and therefore are expected to represent an active means of communication (Machtinger et al., 2016). We have previously shown that urinary concentrations of phthalate metabolites were correlated with altered EV-miRNAs expression in the FF (Martinez et al., 2019). Gene targets and pathways of these miRNAs were related to follicular development and oocyte competence.

Urine is the classical matrix used to assess an individual’s exposure to phthalates and has been used in many studies to assess their effects on human health. However, in some settings, FF may be considered a more appropriate matrix than urine because it provides the opportunity to measure exposure biomarkers (metabolite levels) at the target organ, the ovary. A recent study demonstrated weak to moderate associations between urinary and FF levels of monoethyl phthalate (MEP) and mono(2-ethyl-5-oxohexyl) phthalate (MEOHP) with no associations for other phthalate metabolites (Yao et al., 2020). The aim of the current study was to test associations between EV-miRNAs expression and levels of phthalate metabolites measured in the FF.

Materials and methods

Ethics

This study was approved by the Sheba Medical Center institutional review board in accordance with the Declaration of Helsinki. All methods were in accordance with the relevant guidelines and regulations. All participants provided written informed consent upon enrollment.

Study population and FF collection

One hundred and five women undergoing a fresh IVF cycle at a Tertiary University Affiliated Medical Center were enrolled as a part of a larger prospective cohort study from January 2014 to August 2016 (Machtinger et al., 2018), designed to determine associations between exposure to endocrine disruptor chemicals and IVF outcomes. Participants were enrolled during ovarian stimulation and followed through one fresh IVF cycle. Inclusion criteria were women ≤38 years old with adequate ovarian reserve, undergoing their first to fifth IVF treatment due to male factor or unexplained infertility, who were oocyte donors, or couples undergoing PGD for autosomal recessive diseases. All patients enrolled were treated with GnRH antagonist protocol.

FF (otherwise discarded material) was collected from a single follicle during oocyte retrieval from follicles >18 mm (see technical details in our previous study; Machtinger et al., 2018). Each woman contributed fluid from only one follicle for the analysis. All analyses were run on FF samples containing only mature oocytes that were released from metaphase II. Each oocyte, its corresponding embryo, and FF sample were tracked.

Normal fertilization was defined as the presence of two pronuclei 16–18 h after ICSI and standard insemination. Number of blastomeres was assessed on Day 3, 72 h after oocyte retrieval. Day 3 embryo quality was evaluated using standard criteria based on the number of blastomeres, symmetry and the extent of fragmentation (Alpha Scientists in Reproductive Medicine and ESHRE Special Interest Group of Embryology, 2011; Machtinger and Racowsky, 2013).

Exposure assessment

FF samples were shipped on dry ice to the National Science Foundation (NSF, Ann Arbor, MI, USA) for the quantification of concentrations of 12 phthalate metabolites: mono-benzyl phthalate (MBzP), mono (3-carboxypropyl) phthalate (MCPP), mono-(2-ethyl-5-carboxypentyl) phthalate (mECPP), di-2-ethylhexyl phthalate (DEHP (metabolites: mono ethyl hexyl phthalate (MEHP); mono (2-ethyl-5-hydroxyhexyl) phthalate (MEHHP) and MEOHP, MEP, mono-isobutyl phthalate (MiBP), mono-isononyl phthalate (MiNP), mono-n-butylphthalate (MnBP), mono (6-COOH-2-methylheptyl) phthalate (MCOMHP), mono (7-COOH-2-methyloctyl) phthalate (MCOMOP). The analytical approach, developed by NSF to replicate the Centers for Disease Control and Prevention (CDC), and based on solid-phase extraction coupled with high-performance liquid chromatography–isotope dilution tandem mass spectrometry, followed standard quality assurance/quality control procedures.

Instrumental reading values were used for concentrations below the limit of detection (LOD). Molar sums were calculated for metabolites of di 2-(ethyl hexyl) phthalate (ΣDEHP) by summing the concentration of the metabolite: ΣDEHP = [(mEHP) + (mEHHP) + (mEOHP) + (mECPP)].

Expression analysis of EV-miRNAs in FF

FF aliquots were shipped on dry ice to the University of Milan for the assessment of EV-miRNAs. Briefly, extraction of RNA from biological fluids was done by sample thawing in ice, followed by 15 min centrifugation at 1200g at room temperature and three centrifugations (1000, 2000 and 3000g for 15 min at 4°C). The cleared supernatant was further ultracentrifuged (Beckman Coulter Optima-Max-XP) at 110 000g for 75 min at 4°C to obtain a pellet enriched in EVs. The obtained pellets were kept at −80°C until use. EV-miRNAs were extracted from the pellet using the miRNAeasy Kit and RNeasy CleanUp Kit per the manufacturer (Qiagen, Valencia, CA, USA) (Martinez et al., 2019). The final purified EV-miRNA-enriched RNA was eluted into 20 μl of RNAse-free water and stored at −80°C until further use.

EV-miRNAs analysis was previously described in detail (Martinez et al., 2019). In short, EV-miRNAs were extracted from the subjects’ FFs. We screened for 754 microRNAs in our EV-miRNA aliquot using the TaqMan OpenArray^® system (Life Technologies, Foster City, CA, USA). Each RNA sample was prepared and then reverse-transcribed to cDNA and pre-amplified. Pre-amplified samples were mixed with the TaqMan OpenArray^® Real-Time PCR Master Mix and loaded onto a TaqMan™ OpenArray^® Human miRNA panel with the QuantStudio™ AccuFill System Robot (Life Technologies, Foster City, CA, USA). Real-time quantitative polymerase chain reaction was performed on the QuantStudio™ 12 K Flex Real-Time PCR System with the OpenArray^® Platform [QS12KFLEX] (Life Technologies, Carlsbad, CA, USA) according to the manufacturer’s instructions. For each FF sample, we quantified 754 unique miRNAs and four internal controls (ath-miR159a, RNU48, RNU44 and U6). We calculated the expression levels in relative cycle threshold values (Crt), estimating the amplification cycle at which the fluorescence levels for each of the analyzed EV-miRNAs exceed the background fluorescence threshold.

Normalization was performed by the global mean method, as previously described in Pergoli et al. (2017). Briefly, EV-miRNA expression was determined using the relative quantification $2^{-\mathrm{ }\mathrm{\Delta }\mathrm{crt}}$ (where $\mathrm{\Delta }\mathrm{Crt}$ = Crt_miRNA − Crt_{global mean}) to find the normalized expression of the miRNA’s data. For each sample, the global mean was calculated using all the EV-miRNAs within the subject and dividing them by the total miRNAs (N = 754). For each amplification curve, we also obtained an AmpScore value, a quality measurement that indicates the low signal in the amplification curve linear phase. MiRNAs with Crt value >28 or AmpScore <1.1 or missing were considered unamplified, and the Crt value was set to 29.

Only EV-miRNAs that were expressed in at least 50% of the patients were included in the statistical analysis (see Supplementary Table SI for the expressed miRNAs among the 105 samples).

Statistical analyses

Adjusted linear regression models were used to study associations between FF EV-miRNA expressions and phthalate concentrations. The models included only EV-miRNAs that were found in at least 50% of the samples and phthalate metabolites detected in at least 74% of the samples. All models were adjusted for four factors: age, BMI (calculated from patient height and weight), smoking status and pre-IVF fertility status (fertile vs infertile). While fertile women were those undergoing IVF due to PGD, oocyte donors or male factor, infertile women were those with unexplained infertility. Both biomarker concentrations and EV-miRNA outcomes were log10 transformed to ensure normality.

Benjamini–Hochberg false discovery rate (FDR) correction was applied on the resulting P-values (Benjamini and Hochberg, 1995) for the correction of multi-hypotheses. Of note, we chose a less conservative threshold (FDR q-value < 0.1) due to our sample size. Previous studies demonstrated that endocrine disrupting chemicals, including phthalates, can have non-linear dose–response (Lagarde et al., 2015; Heggeseth et al., 2019). Therefore, we used both Pearson and Spearman correlations to capture both linear and non-linear associations between the levels of FF phthalate metabolites and urinary phthalate metabolites that were measures in our previous study (Martinez et al., 2019).

Wilcoxon nonparametric tests were used to compare the phthalate concentrations between normal and abnormal fertilization. In cases of normal fertilizations, we used adjusted linear regressions to study the associations between each of the examined phthalate and the number of cells on Day 3. These models were also adjusted for: age, BMI, smoking and pre-IVF fertility status (fertile vs infertile). We used Wilcoxon nonparametric tests to compare the phthalate concentrations between follicles that yielded a top quality embryo on Day 3 (i.e. 7–8 cells, <10% fragments) and non-top quality embryos.

Statistical analyses were conducted using MATLAB (version: 9.7.0.1261785 (R2019b) Update 3, November 2019).

Pathways analysis

EV-miRNAs that were found to be significantly (FDR q-value < 0.1) correlated with phthalates were further investigated to discern the genes that are regulated by these miRNAs. We studied the regulated genes and the pathways in which they are involved, focusing on pathways related to female fertility.

Pathways were analyzed using several web-based tools: using miRWalk2.0 (Sticht et al., 2018) (score = 1) with the Targetscan database (Dweep et al., 2011), we identified genes potentially inhibited by the statistically significant (corrected q-value < 0.1) EV-miRNAs. In order to find enrichments for pathways, we ran the DAVID Bioinformatics Resources 6.8 (Huang et al., 2007) with Kyoto Encyclopedia of Genes and Genomes (KEGG) (Altermann and Klaenhammer, 2005), using the genes we found as input (see also Fig. 1 for the flow of the analysis) and looked for statistically significant pathways (Benjamini–Hochberg FDR corrected q-value < 0.05).

Figure 1.

Flow chart of the study analysis. Analysis flow chart including numbers of phthalate metabolites, extracellular vesicle microRNAs (EV-miRNAs), genes and pathways in this study.

Results

Study population

Analyses were performed on FF collected from 105 participants. Mean patients’ age was 30.7 ± 3.7 years (range 19.3–38.8), BMI of 23.7 ± 4.6 kg/m² (range 16–37.3).

Most participants (65%) underwent their first IVF attempt (range 1–5). Patients’ characteristics are shown in Table 1.

Table I.

Demographic information of study participants (N = 105).

Age (mean ± SD)		30.7 ± 3.7
BMI, kg/m2 (mean ± SD), n = 104		23.7 ± 4.6
Number of oocytes retrieved (mean ± SD, range)		10 ± 6, (2–49)
Number of oocytes fertilized (mean ± SD, range)		5 ± 4, (0–29)
Infertility diagnosis, n (%)	Male	41 (38.0%)
	Unexplained	16 (15.2%)
	PGD	48 (45.8%)
	Egg donor	1 (1%)
Smokers	Yes	22 (21%)
Number of FSH IU (mean ± SD)		1800  ± 825

Phthalate concentrations in FF

Eleven of the 12 phthalate metabolites investigated were detected in the FF samples. MCOMHP, mECPP, MnBP and MBzP, MiBP, MEP and MCOMOP were detected in more than 74% of the samples. Five patients (4.7%) had 10 phthalate metabolites in their FF and 80 out of the 105 patients (76%) had at least seven phthalate metabolites detected in their FF. For two patients, no phthalate metabolites were detected at all. FF chemical data are shown in Table 2. Next, we compared the FF phthalate metabolite levels to those measured by the study team in urine in a previous study (Martinez et al., 2019). Medians of FF phthalate concentrations were lower by one to three orders of magnitude than the urinary phthalate concentrations. However, we found statistically significant correlations between the FF and the urinary phthalate metabolite levels for six of the 10 phthalate metabolites that were measured in both studies: mECPP, MEHHP, MEOHP, MEP and MiBP (P < 0.05, see Supplementary Fig. S1). Three out of the 10 metabolites compared showed significant linear and non-linear correlations between urine and FF levels (mECPP, MHEPP and MEP). For two phthalate metabolites (MEOHP and MiBP), we found statistically significant Spearman correlations (P < 0.05) but not Pearson correlations, implying that there are significant non-linear correlations between their urine and FF levels. For one phthalate metabolite (MnBP), we found statistically significant Pearson correlation (P < 0.05) but not Spearman, implying that there is a significant linear correlation between their urine and FF levels. There was no correlation between MBzP, MCPP, MEHP levels in FF and urine.

Table II.

Concentrations (µg/l) of follicular fluid phthalates.

Parent compound	Biomarker	LOD	% Detected > LOD	Median (ng/ml)	IQR (ng/ml)	Max (ng/ml)
Diethyl phthalate (DEP)	MEP	1	74.3	1.66	(0.99, 2.67)	28.2
Di-n-butyl phthalate (DBP)	MnBP	0.5	95.2	1.88	(1.17, 3.30)	224.91
Di-iso-butyl phthalate (DiBP)	MiBP	0.2	87.6	1	(0.55, 1.91)	55.42
Benzylbutyl phthalate (BBzP)	MBzP	0.2	91.4	0.52	(0.34, 0.84)	6.52
Di-n-octyl phthalate (DOP)	MCPP	0.2	2	0	(0, 0.03)	0.63
Di(2-ethylhexyl) phthalate (DEHP)	ΣDEHP			1.5	(0.91, 2.37)	6.96
	MEHP	1	39.1	0.86	(0.43, 1.21)	5.61
	MEHHP	0.1	37.1	0.08	(0.06, 0.12)	0.34
	MEOHP	0.1	15.2	0.04	(0.01, 0.08)	0.5
	mECPP	0.2	96.2	0.49	(0.38, 0.78)	2.59
Di-isononyl phthalate (DiNP)	MiNP	0.5	0	0.04	(0, 0.11)	0.37
	MCOMHP	0.2	98.1	0.55	(0.44, 0.78)	6.76
Di-isodecyl phthalate (DiDP)	MCOMOP	0.2	84.8	0.26	(0.22, 0.35)	0.97

LOD, limit of detection; IQR, inter quartile range; MEP, monoethyl phthalate; MnBP, mono-n-butyl phthalate; MiBP, mono-isobutyl phthalate; MBzP, monobenzyl phthalate; MCPP, mono-3-carboxypropyl phthalate; ΣDEHP, di 2-(ethyl hexyl) phthalate; MEHP, mono-2-ethylhexyl phthalate; MEHHP, mono-2-ethyl-5-hydroxyhexyl phthalate; MEOHP, mono-2-ethyl-5-oxohexyl phthalate; mECPP, mono-2-ethyl-5-carboxypentyl phthalate; MiNP, mono-isononyl phthalate; MCOMHP, mono (6-COOH-2-methylheptyl) phthalate; MCOMOP, mono (7-COOH-2-methyloctyl) phthalate.

Profile of EV-miRNAs in FF

A 754-miRNA panel was used to screen the EV-miRNAs. Three hundred and two EV-miRNAs were detected in at least one of the 105 FF samples. For the analysis, we restricted the EV-miRNAs to those that were detected in at least 50% of our samples, resulting in 131 EV-miRNAs. EV-miRNAs that contributed to the analysis are shown in Supplementary Table SI.

FF phthalate concentrations, fertilization and embryo development

As each oocyte, its corresponding embryo and FF sample were tracked, we were able to test associations between FF phthalate concentrations, fertilization and embryo quality.

We compared the phthalate concentrations between normal fertilized and abnormal/failed to fertilize oocytes. MEOHP levels were statistically significantly lower (mean of 0.05 ng/ml) in FF from normal fertilized oocytes compared with FF that contained oocytes that did not fertilized or abnormally fertilized (mean of 0.09 ng/ml) (P = 0.02). Adjusted linear regression showed that embryos derived from follicles with higher MEP levels had statistically significantly lower number of blastomeres on Day 3 (β = −0.13, P = 0.02). No association was found between any of the phthalates and top quality embryos on Day 3.

Associations between FF phthalates and EV-miRNA expression in FF

We examined the associations between the FF EV-miRNAs and the phthalate metabolites, which were detected (>LOD) in ≥ 74% of the FF samples. After adjusting for multiple testing, 39 EV-miRNAs were statistically significantly correlated either with MEP, MCOMOP, MCOMHP, MBzP and/or with mECPP. Twenty-eight EV-miRNAs were associated with one phthalate metabolite, 10 EV-miRNAs were significantly associated with two phthalate metabolites, and one EV-miRNA (has-miR-572) was significantly associated with three phthalate metabolites (MBzP, mECPP and MEP). Table 3 shows these associations and their statistical significance.

Table III.

Associations between extracellular vesicle microRNAs and phthalate metabolites in follicular fluids.

Phthalate metabolite	miRNA name	Beta coefficient	q-value
MBzP	hsa-miR-125b	−0.3189	0.0216
MBzP	hsa-miR-126	0.5235	0.03
MBzP	hsa-miR-127	0.1598	0.0405
MBzP	hsa-miR-1274b	0.0733	0.0455
MBzP	hsa-miR-143	0.4628	0.0431
MBzP	hsa-miR-150	−0.2545	0.0405
MBzP	hsa-miR-193a-5p	0.5219	0.0405
MBzP	hsa-miR-221	0.6156	0.0361
MBzP	hsa-miR-25	−0.2351	0.0318
MBzP	hsa-miR-320	0.089	0.0431
MBzP	hsa-miR-320b	0.5803	0.03
MBzP	hsa-miR-324-3p	0.452	0.0475
MBzP	hsa-miR-328	0.1077	0.0282
MBzP	hsa-miR-345	0.5732	0.03
MBzP	hsa-miR-376a	−0.2136	0.0446
MBzP	hsa-miR-483-5p	0.0864	0.0441
MBzP	hsa-miR-532-3p	0.6952	0.0216
MBzP	hsa-miR-572	−0.2933	0.0018
MBzP	hsa-miR-574-3p	−0.2753	0.0441
MBzP	hsa-miR-886-3p	−0.3398	0.0003
MBzP	hsa-miR-92a	0.1231	0.0216
MBzP	hsa-miR-99b	0.0934	0.0348
MBzP	hsa-miR-451	0.1839	0.0455
MCOMHP	hsa-miR-192	−0.3438	0.0411
MCOMHP	hsa-miR-335	−0.455	0.03
MCOMHP	hsa-miR-509-5p	−0.5689	0.0475
MCOMOP	hsa-miR-203	0.5765	0.0411
mECPP	hsa-miR-1274b	0.1883	0.0254
mECPP	hsa-miR-142-3p	0.8283	0.0361
mECPP	hsa-miR-15a	0.9494	0.0411
mECPP	hsa-miR-26a	0.1682	0.0441
mECPP	hsa-miR-26b	0.2624	0.0475
mECPP	hsa-miR-29c	0.8565	0.0361
mECPP	hsa-miR-345	1.0769	0.03
mECPP	hsa-miR-572	−0.3371	0.0361
mECPP	hsa-miR-638	−1.2655	0.0216
mECPP	hsa-miR-942	0.7409	0.0455
mECPP	hsa-miR-99b	0.1524	0.0452
mECPP	hsa-miR-451	0.3677	0.0431
MEP	hsa-miR-125b	−0.0574	0.0231
MEP	hsa-miR-15a	−0.0966	0.0411
MEP	hsa-miR-192	−0.0544	0.0446
MEP	hsa-miR-210	−0.1593	0.03
MEP	hsa-miR-222	−0.1028	0.0216
MEP	hsa-miR-25	−0.0494	0.0282
MEP	hsa-miR-34a	−0.1203	0.0361
MEP	hsa-miR-376a	−0.0597	0.0216
MEP	hsa-miR-376c	−0.0942	0.03
MEP	hsa-miR-572	−0.0364	0.03
MEP	hsa-miR-590-5p	−0.0642	0.0282
MEP	hsa-miR-886-3p	−0.0562	0.0021

KEGG pathway analyses

We further investigated the 39 EV-miRNAs that were found to be statistically significantly correlated (after FDR correction) with the aforementioned phthalates. Using miRWalk (Sticht et al., 2018) and TargetScan database, we identified 2444 genes that were putative targets of these EV-miRNAs. Next, we tested the full list of 2444 genes for pathway enrichments (KEGG pathway annotation) and identified 77 statistically significant pathways (see Supplementary Table SII). Twenty out of these 77 pathways are linked to the ovary or oocyte development, oocyte maturation and fertilization: estrogen signaling pathway, cAMP signaling pathway, PI3K-Akt signaling pathway, FoxO signaling pathway, oocyte meiosis, insulin resistance, prolactin signaling pathway, regulation of actin cytoskeleton, cell cycle, TNF signaling pathway, MAPK signaling pathway, focal adhesion, ErbB signaling pathway, Wnt signaling pathway, TGF-beta signaling pathway, mTOR signaling pathway, VEGF signaling pathway, GnRH signaling pathway, progesterone-mediated oocyte maturation, and ubiquitin-mediated proteolysis (see Fig. 2 and Supplementary Table SII).

Figure 2.

KEGG pathway enrichment. Enriched KEGG (Kyoto Encyclopedia of Genes and Genomes) pathways involved in ovary and follicle development, maturation and fertilization associated with the 19 false discovery rate (FDR)-significant EV-miRNAs. Red dashed line represents the statistically significant FDR threshold (q < 0.05). Small numbers (in italic) within each bar indicate the number of predicted genes associated with each KEGG pathway.

These 20 pathways that are related to female fertility, involved 304 genes (out of the 2444 genes), whose transduction is potentially inhibited by the aforementioned EV-miRNAs. One hundred thirty one of these genes are linked to more than one pathway. Fifty one of these 131 genes consist of protein kinase or protein kinase-like domains (Fig. 1). Two of the 304 genes are regulated each by more than four EV-miRNAs that were statistically significantly correlated with the phthalate metabolites: Neuregulin 1 (NRG1) is regulated by hsa-miR-143-3p, hsa-miR-193a-5p, hsa-miR-221-3p, hsa-miR-222-3p, hsa-miR-25-3p, hsa-miR-92a-3p; and Abl interactor 2 (ABI2) is regulated by hsa-miR-15a-5p, hsa-miR-25-3p, hsa-miR-376a-3p, hsa-miR-532-3p, hsa-miR-92a-3p (Fig. 1 and Supplementary Table SIII). Of note, nine genes, MAPK1, MAP2K1, AKT3, PIK3CA, PIK3CB, PIK3R1, PIK3R3, MAPK10 and MAPK8, were involved in more than 10 fertility-related pathways (Supplementary Table SII). Many of the genes found in this analysis that were involved in more than one fertility-related pathway (Supplementary Table SII) were kinases. Indeed, oocyte fertilization, development and maturation processes are all dependent on signal transduction pathways driven by protein kinases (McGinnis et al., 2011; Richard et al., 2017).

Discussion

We detected 11 of the 12 phthalate metabolites investigated in the FF of women undergoing IVF. Seven phthalate metabolites—MCOMHP, MBzP, MCOMHP, mECPP, MiBP, MEP and MnBP—were detected in the FF of over 74% of our study population. Although the levels of the phthalate metabolites in the FF were relatively low, we found significant correlations between FF levels of MEP, MBzP, MCOMOP, MCOMHP, and mECPP and FF EV-miRNA profiles. These EV-miRNAs alter pathways that are linked to oocyte development, maturation and fertilization.

We ran KEGG analyses for the 39 FDR-adjusted significant EV-miRNAs and identified 20 pathways (involving 304 genes) that were associated with female fertility. Two of these genes, NRG1 and ABI2, were regulated each by at least five EV-miRNAs that were associated with phthalate metabolites. Neuregulin 1 (NRG1) is a member of the EGF-like factor family. This gene is induced by LH in granulosa cells and might play a role in ovulation and/or oocyte maturation. NRG1 seems to play a key role in oocyte maturation via cumulus cell- and mural granulosa cell–mediated mechanisms that comprise the regulation of protein kinase C activity and intracellular calcium (Noma et al., 2011; Kawashima et al., 2014). We did not find information regarding the association between ABI2 and female fertility.

FF phthalate metabolite concentrations in our study were of the same order of magnitude as previously reported by Krotz et al. (2012) and Du et al. (2019). However, in contrast with Du et al. (Du et al., 2019) who found MEOHP metabolites in 99.5% of the samples, this metabolite was identified in only 15.2% of our FF samples. This difference may be attributed to various populations and diverse exposures.

We identified positive associations between 39 EV-miRNA isolated from FF and follicular phthalate concentrations of MEP, MBzP, MCOMOP, MCOMHP and mECPP. In a previous study (Martinez et al., 2019), we showed both positive and negative associations between FF EV-miRNAs and urinary phthalate metabolites. Ten phthalate metabolites, MEHP, mECPP, mEHHP, MEOHP, MBzP, MCPP, MEP, MiNP, MnBP and MiBP, were measured in both studies, while other metabolites tested did not overlap. Of note, there were statistically significant positive correlations between urinary and FF levels for six out of the ten phthalates measured in both studies. We found that mECPP levels, either urinary or in FF, were associated with altered EV-miRNA expression. While urinary levels of mECPP were associated with hsa-let-7c, follicular levels of mECPP were associated with hsa-miR-1274b, hsa-miR-142-3p, hsa-miR-15a, hsa-miR-26a, hsa-miR-26b, hsa-miR-29c, hsa-miR-345, hsa-miR-572, hsa-miR-638, haa-miR-942, hsa-miR-99b and hsa-miR-451. Interestingly, urinary MEHP altered hsa-miR-125 whereas the same EV-miRNA was altered by follicular MBzP and MEP levels. These two studies are not comparable since urinary levels of phthalates in our previous study were drawn from pooled urine samples collected from women both during ovarian stimulation (i.e. not fasting) and the day of oocyte retrieval (fasting); whereas in the present study, FF was collected only from fasting women on the day of oocyte retrieval. Given that diet is a source of some of the phthalates, a transient fasting condition prior to the procedure will result in lower metabolite levels.

In addition, the FF phthalate median levels were lower by one to three orders of magnitude than the urinary phthalate concentrations. Median MiBP levels in FF were 1 ng/ml compared with urinary median levels of 24.1 ng/ml and median mECPP levels in FF were 0.49 ng/ml compared with 20.16 ng/ml in the urine.

Several of the KEGG enrichment pathways that are involved in ovary and follicle development, maturation and fertilization and that were altered by both urinary and FF EV-miRNAs overlapped. Some examples include cAMP signaling pathway, oocyte meiosis, MAPK signaling pathway, progesterone mediated oocyte maturation, FoxO signaling pathway, ubiquitin-medicated proteolysis, TGF-beta signaling pathway and cell cycle. These findings may hint for the mechanism in which phthalates, as subgroup of endocrine disruptor chemicals, affect female fertility. Such results are in line with our previous clinical discoveries showing an inverse association between urinary levels of phthalates and IVF outcomes (Machtinger et al., 2018).

This study has several limitations: First, we used a 754 EV-miRNA panel array, while there may be other EV-miRNAs that are expressed in the FF, relevant for fertility and are associated with phthalate exposure but were not measure/analyzed. On the other hand, OpenArray is a method characterized by a very high sensitivity and reproducibility, as compared to other methods currently available (Mestdagh et al., 2014). Second, there is a concern among several researchers for the validity of some of the FF phthalate metabolites examined in this study: enzymes present in FF can break down ubiquitous diesters to their hydrolytic monoesters, so the hydrolytic monoester quantification in FFs may be biased. Six of the phthalate metabolites examined in this study are hydrolytic monoesters: MBzP, MEHP, MEP, MIBP, MiNP and MnBP (Calafat et al., 2015). Third, this study only included FF from one follicle containing a mature oocyte. Whether this could represent miRNAs and metabolite levels in a woman’s cohort of follicles is doubtful. Of note, for 20 of the patients we measured the phthalate concentrations in two follicles and found highly similar phthalate concentrations in the two follicles from the same patient (18 patients with correlation coefficient >0.75, 10 with correlation coefficient >0.95). Forth, the KEGG pathway analysis is based on the genes found by the miRWalk; gene expression analysis and/or gene-miRNA laboratory analyses were beyond the scope of this study. Our study has significant strengths: both the exposure (phthalate metabolites) and the outcome (EV-miRNAs) were taken from the same biological sample—FF, thereby most accurately reflecting the exposure of the oocyte to the chemicals. In addition, we limited our analysis only to FF that contained mature oocytes from young patients, treated by the same stimulation protocol, in order to reduce possible confounders.

Conclusions

In this study we showed for the first time, a statistically significant correlation between low levels of phthalates metabolites measured in FF and EV-miRNAs in the same matrix. Some of these miRNA target genes are associated with oocyte development, maturation and fertilization. These results may explain the mechanism(s) by which exposure to endocrine disruptor chemicals as phthalates alters female fertility.

Data availability

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

Authors’ roles

Z.B.I.: conception and design, drafting the article, revising the article critically for important intellectual content, final approval of the version to be published. S.K.: analysis and interpretation of data, drafting the article, final approval of the version to be published. C.A.: analysis and interpretation of data, drafting the article, final approval of the version to be published. C.R.: conception and design, revising the article critically for important intellectual content, final approval of the version to be published. R.H.: acquisition of data, revising the article critically for important intellectual content, final approval of the version to be published. V.B.: acquisition of data, revising the article critically for important intellectual content, final approval of the version to be published. A.A.B.: conception and design, revising the article critically for important intellectual content, final approval of the version to be published. R.M.: conception and design, acquisition of data, drafting the article, revising the article critically for important intellectual content, final approval of the version to be published.

Funding

This work was supported by the National Institutes of Environmental Health Sciences (R21-ES024236); Israeli Science Foundation (1936/12); the Environmental Health Fund, Israel (1301).

Conflict of interest

The authors have no conflict of interest.