Cell-free mitochondrial DNA in human follicular fluid: a promising bio-marker of blastocyst developmental potential in women undergoing assisted reproductive technology

This research recruited 92 women enrolled in IVF (n = 53) or intracytoplasmic sperm injection (ICSI) (n = 39) cycles at the Center for Reproductive Medicine of Tongji Medical College at the Huazhong University of Science and Technology from October 2017 to July 2018. The women’s ages ranged from 21 to 45 years (mean ± SD: 32.13 ± 4.85 years) and their body mass index (BMI) ranged from 16.60 kg/m² to 33.90 kg/m² (mean ± SD: 22.38 ± 3.58 kg/m²). The duration of infertility was 4.20 ± 3.57 years, and 53% of couples had primary infertility. Female infertility had been identified in 60% couples, male factors in 23%, and mixed infertility in 17%. The number of days of stimulation ranged from 3 to 22 days (mean ± SD: 9.75 ± 2.74), and the total dose of gonadotropins received ranged from 225 to 6800 IU (mean ± SD: 2089.69 ± 835.12). Baseline hormonal levels (follicle-stimulating hormone [FSH], luteinizing hormone [LH], and 17β-estradiol [E₂]) and anti-Müllerian hormone (AMH) were assessed using the Beckman DXI800 chemiluminescence analyzer (Beckman Coulter Inc., Brea, CA) on the third day of the menstrual cycle for each patient. Levels of AMH < 1.1 ng/ml are considered to reflect a reduced ovarian reserve, and levels ≥1.1 ng/ml are normal ovarian reserve [26]. Levels of FSH ≥ 9 IU/L represent a reduced ovarian reserve.

The ovarian stimulation regimens used included long agonist protocols, ultra-long agonist protocols, antagonist protocols, and progestin-primed ovarian stimulation (PPOS). Oral progestin was first given to prevent a premature LH surge in PPOS. Pituitary inhibition was performed with a gonadotropin-releasing hormone agonist or antagonist for other protocols. FSH stimulation was monitored by measuring serum E₂ levels and follicular size. Human chorionic gonadotrophin (hCG) (Livzon, Zhuhai, China) was injected when at least three follicles reached a diameter of 18 mm or more by ultrasound inspection. Oocytes were retrieved by trans-vaginal ultrasound-guided puncture 36 h after hCG injection.

Cumulus cells were stripped to observe the extrusion of the first polar body of the oocyte prior to ICSI (Day 0), and oocytes extruding the first polar body were called mature oocytes. For IVF oocytes, oocyte maturity was confirmed when the cumulus cells were stripped 4–6 h after IVF. Zygotes with pronuclei present 18–20 h after microinjection or insemination were considered representative of fertilization. Oocytes that underwent cleavage on Day 2 without pronuclei on Day 1 were also considered representative of fertilization for IVF. Day 3 embryos were morphologically scored in accordance with the current consensus system [27]. High-quality embryos were defined as those with at least six blastomeres of a uniform shape on Day 3, and with fragments less than 25%. One or two high-quality embryos were transferred or frozen on Day 3, and the rest were cultured to the blastocyst stage to be transferred or frozen. Blastocysts were evaluated morphologically based on the expansion of the blastocoele (3–6 stages) and the number and cohesiveness of the inner cell mass and trophectoderm (Grade A, B, and C), according to Gardner’s criteria [28].

Two hundred and twenty-five follicular fluid samples without flushing media were collected from individual follicles of 92 patients, centrifuged at 3000×g for 15 min and 16,000×g for 10 min, and then immediately stored at − 80 °C until cf-nDNA and cf-mtDNA quantification. To prevent any blood pollution, only clear follicular fluid samples were involved, while bloodstained and cloudy follicular fluid samples were discarded.

cf-DNA was extracted and purified from follicular fluid by the BeaverBeads™ Circulating DNA kit (BEAVER, Suzhou, China) according to the manufacturer’s instructions. cf-nDNA and cf-mtDNA in individual follicular fluid samples were estimated by amplification using β-globin and ND1 primers using real-time quantitative polymerase chain reaction (qPCR). Primers were designed and synthesized by Sangon Biotechnology Co., Ltd. (Shanghai, China). Primers for ND1 were 5′-CCCTAAAACCCGCCACATCT-3′ (forward) and 5′-GAGCGATGGTGAGAGCTAAGGT-3′ (reverse), which amplify a 69 bp DNA fragment. β -globin was detected using the following primers: 5′-AAAGGTGCCCTTGAGGTTGTC-3′ (forward) and 5′-TGAAGGCTCATGGCAAGAAA-3′ (reverse), which amplify a 77 bp DNA fragment. The amplicons were detected using primer sequences and verified in the GenBank database (Additional file 1: Figure S1). Standard curves were made using purified plasmid DNA corresponding to ND1 and β-globin (Additional file 1: Figure S1). The relative content of cf-mtDNA in follicular fluid was expressed by the cf-ND1/cf-β-globin ratio. All reactions were performed in duplicate on the StepOne™ and StepOnePlus™ Software system (ThermoFisher Scientific, Waltham, MA, USA). Reactions were performed in a total volume of 20 μl containing 2 μl of sample template (elution product of the BeaverBeads™ Circulating DNA kit), 10 μM of forward and reverse primers, and 10 μl of Hieff™ qPCR SYBR® Green Master Mix (High Rox Plus; Yeasen, Shanghai, China). Cycling conditions were as follows: 95 °C for 30 s, then 40 cycles of 95 °C for 5 s and 60 °C for 30 s.

Of 225 individual follicular fluids, a total of 190 had mature oocytes (84%), 17 had immature oocytes (8%), 16 had no oocytes (7%), one had a degenerated oocyte, and one was naked without a zona pellucida. The relative cf-mtDNA content (cf-ND1/cf-β-globin ratio) was significantly higher in follicular fluids with than without oocytes (P < 0.01; Fig. 1a). Because the criteria for oocyte maturation and fertilization differ between IVF and ICSI, we carried out statistical analysis for IVF and ICSI oocytes separately (Table 1). The relative cf-mtDNA content did not differ between mature and immature oocytes, or between fertilized and non-fertilized oocytes for IVF or ICSI (P > 0.05).

According to cleavage embryo evaluation criteria, 89 (89/142, 63%) embryos were of high quality and 53 (53/142, 37%) were poor quality. There was no significant difference in the relative cf-mtDNA content between high-quality and poor-quality embryos (P > 0.05). In our samples, 43 embryos were cultured until Day 6. Twenty-two (51%) embryos developed into blastocysts, and 21 (49%) embryos arrested or failed to develop on Day 5 or Day 6. The relative cf-mtDNA content in the blastocystgroup was significantly lower than in the non-blastocyst group (2.84 versus 7.80; P = 0.030; Fig. 1c and Table 2).

The relative cf-mtDNA content in follicular fluid correlated significantly positively with age (β ± SE: 0.22 ± 0.085, P = 0.009) (Table 3; Fig. 2) but not with AMH, basal FSH level, or antral follicle count (AFC).

Patients were divided into four groups according to age and AMH level: age ≥ 38 years old and AMH > 1.1 ng/ml, age ≥ 38 years old and AMH ≤1.1 ng/ml, age < 38 years old and AMH > 1.1 ng/ml, and age < 38 years old and AMH ≤1.1 ng/ml. The relative cf-mtDNA content in the two groups of age ≥ 38 years old was significantly higher than that in both age < 38 years old groups (P < 0.05; Fig. 3a). Moreover, for women aged ≥38 years old, the relative cf-mtDNA content when AMH ≤1.1 ng/ml was significantly higher than when AMH > 1.1 ng/ml (P < 0.05; Fig. 3a). Similar results were observed for patient age and FSH levels (Fig. 3b).

The relative cf-mtDNA content in follicular fluid was not associated with BMI, the ovarian stimulation regimen, number of days of stimulation, or the total dose of gonadotropins (Table 3).

The median quantity and the IQR of cf-nDNA levels measured by β -globin qPCR in 225 human follicular fluid samples from individual follicles was 0.23 ng/ml (IQR: 0.12–0.46 ng/ml). cf-nDNA levels did not differ in follicular fluids with and without oocytes (P > 0.05; Fig. 1b). cf-nDNA levels in follicular fluids were not associated with oocyte maturation, fertilization, or Day 3 embryo morphological scoring. cf-nDNA levels were higher in follicular fluids where the oocytes were fertilized and reached the blastocyst stage than in those with no blastocyst development (0.63 ng/ml versus 0.14 ng/ml), but the difference was not significant after multivariate logistic regression (P = 0.071; Fig. 1d; Table 2).

There was no significant correlation between cf-nDNA levels in individual follicular fluid samples and patient age, BMI, AMH, basal hormone (basal FSH, LH, and E₂), or the ovarian stimulation regimen (P > 0.05). The AFC was significantly positively correlated with the amount of cf-nDNA in follicular fluid (β ± SE: 0.026 ± 0.017; P = 0.012; Table 3). Moreover, both the number of days and total dose of gonadotropin administration significantly and negatively affected cf-nDNA levels in follicular fluid (β ± SE: − 0.019 ± 0.05 and − 0.027 ± 0.0002, respectively; P = 0.039 and P = 0.015, respectively; Table 3).