Improved clinical outcomes of preimplantation genetic testing for aneuploidy using MALBAC-NGS compared with MDA-SNP array

This study was approved by the Ethics Committee of The First Affiliated Hospital of Zhengzhou University. Between January 2013 and December 2017, a total of 1418 couples received PGT-A treatment after being evaluated by both a geneticist and an infertility specialist in the reproductive center of The First Affiliated Hospital of Zhengzhou University. Factors leading to PGT-A treatment included abnormal karyotypes, advanced maternal age (AMA≥35 years), repetitive implantation failures (RIF ≥ 3times) and/or recurrent miscarriage (RM ≥ 2times). To avoid DNA contamination from the paternal genome, all PGT-A patients underwent intracytoplasmic sperm injection (ICSI) treatment irrespective of their sperm quality.

The patients were divided into two main groups: MALBAC-NGS-PGT-A group and MDA-SNP-PGT-A group. Each group was subsequently classified into five subgroups according to the treatment indications of couples: Robertsonian translocation, reciprocal translocation, inversion, sex chromosome abnormality karyotype and undergoing PGS (patients with AMA or RIF or RM) treatment.

In the MALBAC-NGS-PGT-A group, a total of 805 couples were treated between January 2016 and December 2017: 157 with a Robertsonian translocation, 405 with a reciprocal translocation, 40 with an inversion, 92 with a sex chromosome abnormality karyotype and 111 with PGS treatment indications. Among the couples undergoing PGS treatment,54 were of advanced maternal age, 47 had experienced recurrent miscarriages and 10 were patients with repeated implantation failures.

In the MDA-SNP-PGT-A group, 613 couples were treated from January 2013 to December 2015. Five hundred nineteen couples were found to be chromosomally abnormal. Among them, the number of couples with a Robertsonian translocation, reciprocal translocation, inversion and sex chromosome abnormality karyotype were 105, 342, 14 and 58 respectively. The remaining 94 couples underwent PGS treatment due to advanced maternal age (39 couples), recurrent miscarriages (46 couples) and repeated implantation failures (9 couples).

Controlled ovarian stimulation (COS) was performed using the standard long protocol. Gonadotropin-releasing hormone agonist (GnRH-a), Decapeptyl (Germany) or Diphereline (France) were used to induce complete down-regulation (follicle stimulating hormone FSH ≤ 5 IU/L, luteinizing hormone LH ≤ 5 IU/L, estrogen E2 ≤ 50 pg/ml, diameter of the largest follicle≤10 mm and endometrial thickness ≤ 5 mm). Gonal-F (Gn, Switzerland) was subsequently administered. The starting dose of Gn was determined based on body mass index, age, basal FSH level and antral follicle count. Where the diameter of the dominant follicle was more than 14 mm, fasting blood was collected every day to determine the levels of LH, E2 and P present. The regimen of Gn was adjusted based on the hormone levels and follicle size. When the amount of follicles with a diameter ≥ 16 mm was more than two-thirds, and the diameter of the dominant follicle was greater than 20 mm, human chorionic gonadotropin was intramuscularly injected. Thirty-seven hours later, oocyte retrieval was performed using ultrasonic guidance. Fertilization took place 3–4 h after oocyte retrieval. Embryo cleavage was recorded every 24 h. Blastocysts were graded on the fifth or sixth day after oocyte retrieval using the Gardner criteria. Trophectoderm (TE) biopsy was performed on selected blastocysts higher than 3BB, and 5 to 10 TE cells were isolated and cut using a laser. The biopsied cells were then immediately placed in the RNAse and DNAse-free polymerase chain reaction tube and transported to our PGT-A center. After testing karyotypes of the biopsied blastocysts, a single thawed euploid embryo was transferred for each transfer cycle.

MALBAC was used for WGA in NGS testing, following the manufacturer’s protocol (XK028,Yikon Genomics, China). Amplified DNA was then diluted, fragmented, adaptor ligated, subjected to PCR reaction and purified. After measuring the concentration, libraries were processed using rapid single-end 50 cycle mode sequencing using a Hiseq 2500 sequencer (Illumina, USA) in our center.

The sequenced raw data were demultiplexed and converted to the FASTQ format using CASAVA 1.8.4(Illumina, USA).MALBAC primers, adaptors and low-quality bases were removed from the FASTQ data using Trimmomatic [16],generating an average of ~ 2 million filtered reads per sample, with an average ~ 0.03 × sequencing depth.

High-quality reads were mapped to a hg19 reference genome using the Burrows-Wheeler Aligner software package [17]. Unique mapped reads were extracted from the alignment reads (bam file). The reference genome was then divided into non-overlapping observation windows (bins) with a size of 1000Kb. The number of reads and GC-content were calculated in each bin and GC bias correction was performed for every 1% GC-content [18]. The R programming language was then used to graph the final relative number of reads for each bin to visualize copy number variations.

MDA was performed using REPLI-g Single Cell Kit (150,345, QIAGEN, Germany) in SNP array according to the authors’ previous report [19]. Briefly, biopsied TE cells (5 to 10 cells) were lysed in DTT and DLB buffer. Whole genome DNA was amplified in an amplification buffer created by mixing REPLI-g single-cell DNA polymerase and the reaction buffer. The MDA product was then hybridized on the Illumina HumanCyto12 microarray by following the user manual. Data were analyzed using GenomeStudio Software (2011, Illumina, USA). The genotype was evaluated by using the B allele frequency and log₂R ratio (log₂R = Log₂ normalized R value/expected normalized R value).

NGS testing was successfully performed on 805 couples and 3321 biopsied blastocysts. As is shown in Tables 1, 766 (23.1%) euploid/normal and 2555(76.9%)chromosomally abnormal samples were identified. Of these abnormalities, 760(22.9%) had numerical abnormalities, 404(12.2%)contained structural abnormalities (deletions and duplications), while 1391(41.8%) had complex abnormality (defined as two or more duplications/deletions involving the same, or two or more chromosomes) (Fig. 1).

The most common chromosome abnormality type in each subgroup was also determined, as shown in Fig. 2. Numerical abnormality was most frequently found in blastocysts from couples with Robertsonian translocation (43.8%, 264/603), sex chromosome abnormality (27.5%, 95/345) and PGS couples (33.2%, 131/395), while structural abnormality showed a relatively balanced distribution among the five subgroups. Meanwhile, complex abnormality was more common than the other kinds of abnormalities for couples with reciprocal translocation (56.6%, 1034/1825) and inversion (31.4%, 48/153) (Table 1).

Among the 2450 blastocysts tested by SNP array from 613 couples, 1095 (44.7%) were euploid/normal and 1355(55.3%) were abnormal. Among all of the detected abnormalities, 654 (26.7%) had a numerical abnormality, 393 (16%) had a structural abnormality (duplication or deletion), 299 (12.2%) contained complex abnormalities, and the remaining 9 (0.4%) showed uniparental disomy (UPD) (Fig. 1).

With regard to the distribution of chromosomal aberration types among each subgroup, numerical abnormality was most commonly found in samples from couples with Robertsonian translocation, inversion, PGS and sex chromosome abnormality, accounting for 38.7% (177/457), 32.9% (27/82), 28.3% (86/303) and 34.9%(75/215) for each of the above, respectively. Among couples with reciprocal translocation, structural abnormality (24.3%, 338/1393) was identified to be the most frequent (Table 2 and Fig. 3).

Among the 805 couples in the MALBAC-NGS-PGT-A group, 270 had no euploid/normal embryos to transfer, 522 underwent embryo transfer, and 13 are still waiting for transfer. Of the 522 couples with embryo transfer, 675 euploid/normal embryos were transferred in 675 cycles. Clinical pregnancy was identified in 341 (50.5%) of the transferred cycles. Healthy baby rate was 39.6% (267/675,including two onset of twins). Of the clinical pregnancy cycles, 15.5%(54/341) of the cycles ended with miscarriage. In the SNP array-PGT-A group, 145 couples were found with no euploid embryos to transfer, while 468 couples had euploid/normal embryos. To date, 452 couples have undergone embryo transfer while 12 couples were excluded in this study because they transfered two embryos in once transfer cycle, and 16 couples are still waiting for transfer. Of the 440 couples with embryo transfer, 547 euploid/normal embryos were transferred in 547 cycles. Of the 547 transferred cycles, 41.7% (228/547) of the cycles resulted in clinical pregnancy, healthy baby rate was 31.4% (172/547,including two onset of twins). Of the clinical pregnancy cycles, 22.8% (52/228) of these miscarried.

There were no significant differences in terms of maternal age, BMI, basal FSH level and endometrial thickness on the day of embryo transfer between the MALBAC-NGS-PGT-A and MDA-SNP-PGT-A groups (P > 0.05). However, a significantly higher clinical pregnancy rate (50.5% vs41.7%, P = 0.002), healthy baby rate (39.6% vs31.4%, P = 0.003) and lower miscarriage rate (15.5% vs22.8%, P = 0.036) were observed in the MALBAC-NGS-PGT-A group when compared with the MDA-SNP-PGT-A group (Table 3).