Copy number variation sequencing combined with quantitative fluorescence polymerase chain reaction in clinical application of pregnancy loss

Abstract

Purpose

In this study, we evaluated the feasibility of the combining CNV-seq and quantitative fluorescence polymerase chain reaction (QF-PCR) for miscarriage analysis in clinical practice.

Methods

Over a 35-month period, a total of 389 fetal specimens including 356 chorionic villi and 33 fetal muscle tissues were analyzed by CNV-seq and QF-PCR. Relationships between the risk factors (e.g., advanced maternal age, abnormal pregnancy history, and gestational age) and incidence of these chromosomal abnormalities were further analyzed by subgroup.

Results

Clinically significant chromosomal abnormalities were identified in 58.95% cases. Aneuploidy was the most common abnormality (46.84%), followed by polyploidy (8.16%) and structural chromosome anomalies (3.95%). In sub-group analysis, significant differences were found in the total frequency of chromosomal abnormalities between the early abortion and the late abortion group, as well as in the distribution of chromosomal abnormalities between the advanced and the younger maternal age group. Meanwhile, the results of the logistic regression analysis identified a trend suggesting that the percentage of fetal chromosomal abnormalities is significantly higher in advanced maternal age, lesser gestational age, and lesser number of prior miscarriages.

Conclusion

Our study suggests that CNV-seq and QF-PCR are efficient and reliable technologies in the fetal chromosome analysis of miscarriages and could be used as a routine selection method for the genetic analysis of spontaneous abortion.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10815-021-02243-9.

Introduction

Spontaneous abortion (SA), the spontaneous loss of a clinically established intra-uterine pregnancy before the fetus has reached viability, is the most common complication of early pregnancy. It is estimated that about 15–20% of clinically recognized pregnancies end in miscarriage, about 25% of all women experience at least one spontaneous abortion, and approximately 1% of couples experience recurrent (at least two) pregnancy losses [1–3]. Multifarious factors, both maternal and fetal, may cause miscarriage. It is generally accepted that pregnant women with genetic abnormalities, endocrine dysfunction, pro-thrombotic tendency, uterine abnormalities, or advanced age are at high risk of miscarriages [4]. In fact, the presence of fetal chromosomal abnormalities is the most frequent cause of pregnancy loss, accounting for more than 50% losses in early pregnancy and about 8–10% intra-uterine fetal demises and/or stillbirths in the second or third trimester [5–7]. As the underlying cause of pregnancy loss is complex, an etiologic analysis is essential and provides important information for medical management, reproductive counseling, and supportive patient care [1, 8].

At present, cytological analysis methods represented by karyotype analysis, some molecular cytogenetic methods such as fluorescence in situ hybridization (FISH), and molecular methods represented by chromosome microarray analysis (CMA) are commonly applied for genetic diagnosis of products of conception samples (POCs) [9–11]. As the gold standard for cytogenetic diagnosis, karyotype analysis has been used to evaluate POCs for many years. However, chromosome karyotype analysis has many limitations, such as the stringent requirement of sample conditions, complicated operation process, long cell culture cycle, high failure rate, and difficulty in identifying maternal cell contamination (MCC) [7, 12]. FISH can detect the duplication/deletion of a specific genomic region by using fluorescently labeled probes and is widely used in the clinical diagnosis of common chromosomal aneuploidies. Nevertheless, FISH is time-consuming as it involves cell culture and the sensitivity of the assay depends upon probe design, thus limiting its application [13]. In comparison, CMA has been considered to replace traditional karyotyping as the standard in genetic screening of chromosomal abnormalities due to its higher efficiency in detection of both numerical chromosomal abnormalities and sub-chromosomal copy number variations. Although the high resolution of CMA is remarkable, the technical demands and costs are very high, which restrict its use as a routine detection method for spontaneous abortion [7, 14].

Next-generation sequencing (NGS), a low-cost technique with a short turn-around time, unprecedented resolution, and reliable high-throughput and requirement for only small amounts of DNA, has been widely used in clinic [15]. Compared to CMA, NGS has significant advantages in terms of quality, speed, and affordability [16–18]. Copy number variation sequencing (CNV-seq), a NGS-based method, has been used in most pediatric and prenatal diagnostic applications as a viable alternative methodology to CMA owing to its ability to simultaneously detect aneuploidies and sub-microscopic chromosomal imbalances [18–20]. Nevertheless, CNV-seq fails at the detection of MCC and polyploidy, which limits its application in abortion detection. Quantitative fluorescence polymerase chain reaction (QF-PCR) is a rapid method of chromosome detection commonly used in clinic. It could identify MCC, some euploidies and some common aneuploidies by the amplification of selected short tandem repeats (STRs) sites and quantitative analysis of the allelic dosage ratios to evaluate the numbers of copies of specific chromosomes [21]. Therefore, we speculated that the combination of the CNV-seq and QF-PCR would be a reliable approach for chromosome detection of POCs, which has been confirmed in prenatal [20, 22].

In previous studies, genetic diagnosis of abortion samples was conducted using karyotype analysis, different molecular cytogenetic methods, or CMA. However, there is no report in which a combination of CNV-seq and QF-PCR was systematically used in the clinical setting. In this study, we aim to evaluate the combined application of CNV-seq and QF-PCR as a tool for the identification of chromosome abnormalities, investigate the frequency and type of chromosome aberrations in POCs of couples who had at least one miscarriage, and probe into the influencing factors of chromosomal abnormalities related to miscarriages.

Materials and methods

Subjects

A total of 389 fetal specimens including 356 chorionic villi and 33 fetal muscle tissues were obtained from female subjects who had undergone spontaneous abortion (gestational age ≤ 28 weeks) at the West China Second University Hospital of Sichuan University from February 2017 through December 2019. The study was approved by the Medical Ethics Committee of West China Second University Hospital of Sichuan University (medical research 2016-7). All the participants provided written informed consent for genetic investigation involving detection of fetal chromosomal anomalies using CNV-Seq combined with QF-PCR. The mean maternal age of the study population was 31.0 years (range: 20 to 46).

Sample preparation

Chorionic villi or fetal tissues were obtained via protocols approved by the institutional review board. Briefly, in aseptic condition, put the sample in aseptic vessel, take 10 mg of the abortion material, and wash it with aseptic PBS buffer. Genomic DNA was extracted from chorionic villi or fetal muscle tissues using the DNeasy Blood & Tissue Kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. The genomic DNA was measured using the Qubit quantitative kit (Thermo Fisher Scientific Inc., Rockford, IL, USA) with a Qubit 2.0 Fluorometer (Thermo Fisher Scientific, Waltham, MA, USA) strictly adhering to the protocols.

QF-PCR

QF-PCR was conducted using a set of 20 highly polymorphic STR markers including D21S1433, 21q11.2, D21S1411, D21S1414, D21S1412, and D21S1445 for chromosome 21, D18S1002, D18S391, D18S535, and D18S386 for chromosome 18, D13S628, D13S742, D13S634, and D13S305 for chromosome 13, and AMXY, DXS1187, DXS8377, SRY, DXS6809, and DXS981 for the sex chromosomes X and Y. PCR was performed using a Bio-Rad PTC-200 PCR system (Bio-Rad, Mexico City, Mexico). The PCR products were analyzed on the 3500 ABI Genetic Analyzer (Applied Biosystems, Waltham, MA). QF-PCR was used to measure the quality of the specimen DNA and assist in the determination of euploidy. Each DNA sample was subjected for CNV-seq after exclusion of MCC.

CNV-seq

CNV-seq was performed for all samples with a modified protocol [19, 22]. First, the specimen DNA was used for DNA library construction. Briefly, 50 ng of DNA was fragmented and DNA libraries were constructed by end filling, adapter ligation, and PCR amplification. Then, the DNA libraries were subjected to massive parallel sequencing on the NextSeq 500 platform (Illumina, San Diego, CA) to generate approximately 5 million raw sequencing reads with 36 bp long genomic DNA sequences. With the hg19 genomic sequence as reference, a total of 2.8–3.5 million reads were uniquely and precisely mapped using the Burrows-Wheeler algorithm [23]. Mapped reads were allocated progressively to 20 kilobase (kb) bin sizes from the p to q arms of the 24 chromosomes. Counts in each bin were then compared between all test samples run in the same flow cell to evaluate copy number changes using previously described algorithms [18, 22].

Analysis of detection

For each informative STR marker, if fluorescent ratios of the fetally inherited maternal allele to the fetally inherited paternal allele was consistently >1.1, the sample was deemed to have maternal DNA contamination levels >10%. In these cases, DNA samples were considered unacceptable for a reliable chromosome test result. If all the allele ratio ranged from 1.8 to 2.4 or 0.45 to 0.65, the height ratio of three individual peaks was 1:1:1 or there was only a single peak for each marker, it was considered euploidy positive and karyotyping, and FISH or CMA were used as confirmatory tests.

For CNV-seq, chromosome profiles were plotted as copy number (Y-axis) vs. 20 kb count windows (X-axis). A blue line was used to indicate the mean copy number across each chromosome to identify the nature and map position of any deleted or duplicated regions. Detected CNVs were evaluated based on scientific literature review and public databases, including DECIPHER (http://decipher.sanger.ac.uk/), Database of Genomic Variants (DGV, http://dgv.tcag.ca/dgv/app/home/), ClinGen (http://www.ncbi.nlm.nih.gov/projects/dbvar/clingen/), GeneReviews (http://www.ncbi.nlm.nih.gov/books/NBK1116/), and Online Mendelian Inheritance in Man (OMIM, http://www.omim.org/). The pathogenicity of detected CNVs was assessed according to the guidelines outlined by the American College of Medical Genetics (ACMG) for interpretation of sequence variants. The chromosomal microdeletions/microduplications were classified as benign CNVs, variants of uncertain significance (VOUS), or pathogenic CNVs [24]. For clinical reporting of patient results, only pathogenic chromosome anomalies were considered in this study. All the abnormalities including aneuploidy or pathogenic CNVs were confirmed by repeating the CNV-Seq analysis in an independent laboratory. FISH was conducted to confirm the results when samples were indicated chimeras or euploidy.

Results

Overall results

From February 2017 through December 2019, 389 samples were received for CNV-seq and QF-PCR analysis. Of these, 9 cases (8 chorionic villi and 1 tissue) with serious MCC were removed from the primary study, leaving 380 cases available for further study. Normal results were obtained in 156 cases (41.05%) and abnormal results were obtained in the remaining 224 cases (58.95%). A total of 51.92% (81/156) of the cases with normal results were identified as female while 48.08% (75/156) were identified as male, with a female-to-male ratio of 1:08. Abnormal results consisted of 178 aneuploidies, 31 polyploids, and 15 pathogenic CNVs. The detailed results are summarized in Fig. 1.

Fig. 1.

Summary and characterization of the chromosomal analysis results conducted by CNV-seq and QF-PCR in 389 fetal specimens of spontaneous abortion

Spectrum of abnormalities

Aneuploidy was the most frequent abnormality; it was identified in approximately 79% of cases (178/224), including 162 single aneuploidies, 12 multiple aneuploidies, and 4 chimeras. Most of the single aneuploidies were trisomies (142/162, 87.65%), while the others were monosomies that were found only in chromosome X (20/162, 12.35%). Figure 2 presents the details of each chromosome aneuploidy in this study. Aneuploidies were identified in all the chromosomes, except chromosomes 1 and 19. Trisomy 16 was the most frequent trisomy (47/142, 33.10%), followed by trisomy 22 (17/142, 11.97%). Thirty-one cases with polyploidy were detected in this study, all of which were triploidy. Pathogenic CNVs were detected in 15 (3.95%) cases (Table S1).

Fig. 2.

Summary of each chromosome aneuploidy in fetal specimens of spontaneous abortion. a Represents the detected numbers of each chromosome aneuploidy in 380 spontaneous abortion samples. b Represents the percentage of each chromosome aneuploidy in all detected aneuploidies. Two values were ignored because aneuploidy was not identified in chromosomes 1 and 19. Abbreviations: Chr, chromosome

Maternal age and abnormal CNV-seq and QF-PCR

We classified the results according to maternal age. Comparisons of the distribution of chromosomes in the advanced maternal age group (≥35 years old) with the younger maternal age group is listed in Table 1. No statistical difference was found between these two groups in the frequency of total chromosomal abnormalities. However, the distribution of chromosomal abnormalities was different between the advanced maternal age group and the younger maternal age group. Compared with the younger maternal age group, higher frequency of aneuploidy and lower frequency of pCNV were identified in the advanced maternal age group.

Table 1.

Comparison of distribution of chromosomes identified by CNV-seq and QF-PCR according to maternal age

Groups	Total number (n)	Normal (n) (frequency (%))	Chromosomal abnormalities (n) (frequency (%))
			Aneuploidy	Polyploidy	pCNV	Total
Age<35	302	131 (43.38%)	133 (44.04%)	23 (7.62%)	15 (4.97%)	171 (56.62%)
Age≥35	78	25 (32.05%)	45 (57.69%)	8 (10.26%)	0	53 (67.95%)
P value	-	0.070	0.031	0.448	0.048	0.070

Note: (a) Data are presented as number and percentage for every group. (b) Significant p-values (<0.05) are highlighted in bold. Abbreviation: pCNV, pathogenic copy number variation

Miscarriage frequency and abnormal CNV-seq and QF-PCR

We also compared the results between the first spontaneous abortion (FSA) and recurrent miscarriage (RM) groups. Here, RM was defined as two or more pregnancy losses. The detection rates in the FSA and RM groups were compared in Table 2. No conspicuous difference was observed.

Table 2.

Comparison of distribution of chromosomes identified by CNV-seq and QF-PCR according to the number of miscarriages

Groups	Total number (n)	Normal (n) (frequency (%))	Chromosomal abnormalities (n) (frequency (%))
			Aneuploidy	Polyploidy	pCNV	Total
FSA	155	62 (40.00%)	75 (48.39%)	12 (7.74%)	6 (3.87%)	93 (60.00%)
RM	225	94 (41.78%)	103 (45.78%)	19 (8.44%)	9 (4.00%)	131 (58.22%)
P value	-	0.729	0.616	0.806	0.949	0.729

Note: Data are presented as number and percentage for every group. Abbreviation: FSA, first spontaneous abortion; RM, recurrent miscarriage; pCNV, pathogenic copy number variation

Gestational age and abnormal CNV-seq and QF-PCR

The results were divided into early abortion and late abortion groups. Early abortion refers to spontaneous abortion with gestational age less than 12 weeks, while late abortion refers to spontaneous abortion between 12 and 28 weeks. The detection rates in the two groups were compared in Table 3. Distribution of chromosomes was found to be statistical different between the early abortion group and late abortion group. The frequency of chromosomal abnormalities in the early abortion group was significantly higher than that in the late abortion group, mainly reflected in the aneuploidy cases.

Table 3.

Comparison of distribution of chromosomes identified by CNV-seq and QF-PCR according to the gestational age

Groups	Total number (n)	Normal (n) (frequency (%))	Chromosomal abnormalities (n) (frequency (%))
			Aneuploidy	Polyploidy	pCNV	Total
Early abortion	317	120 (37.85%)	158 (49.84%)	27 (8.52%)	12 (3.79%)	197 (62.15%)
Late abortion	63	36 (57.14%)	20 (31.75%)	4 (6.35%)	3 (4.76%)	27 (42.86%)
P value	-	0.004	0.009	0.566	0.723	0.004

Note: (a) Data are presented as number and percentage for every group. (b) Significant p-values (<0.05) are highlighted in bold. Abbreviation: pCNV, pathogenic copy number variation

Logistic regression

Logistic regression models were used for calculating odds ratios (95% confidence interval) and corresponding P-values for association of clinical characteristics with the occurrence of chromosomal abnormalities in spontaneous abortion specimens using SPSS. Upon carrying out single factor regression first, the P-value for the continuous variables—maternal age, gestational age and number of miscarriages (the number of previous miscarriages)—were all less than 0.05. This was followed by multi-factor regression. Maternal age, gestational age, and number of miscarriages were found to be significantly associated with the occurrence of chromosomal abnormalities in abortion specimens (Table 4). Women with advanced age, lesser gestational age, and lesser miscarriage times were more likely to find chromosomal abnormalities in abortion specimens.

Table 4.

Logistic regression of clinical characteristics with chromosomal abnormalities occurrence

Characteristics	P value	OR	95%CI
Maternal age	<0.001	1.036	1.028–1.044
Gestational age	<0.001	0.989	0.988–0.991
Number of miscarriages	0.047	0.968	0.937–1.000

Abbreviation: OR, odds ratios; 95%CI, 95% confidence interval

Discussion

Among the many factors that contribute to clinical pregnancy loss, fetal chromosomal abnormalities is the most common etiology and increasing maternal age and previous losses are frequently cited associations. Some studies have showed that chromosomal factors in the parents, such as largely balanced chromosomal rearrangements, account for 2–5% couples experiencing recurrent pregnancy loss. Meanwhile, up to 50–60% early spontaneous abortions were caused by fetal chromosomal abnormalities [1, 7, 25, 26]. Considering the etiological explanations of such magnitude, fetal chromosome analysis should be conducted routinely in all pregnancy losses. By doing so, the couple can benefit from specific recommendations. It can also help to decide whether additional etiologies need to be investigated and whether or not prenatal and/or pre-implantation genetic testing needs to be conducted [2, 27]. However, there are technical challenges and limitations for routine evaluating chromosomes of POCs in both traditional cytogenetic techniques and new molecular genetics technology [7, 12–14]. In this study, we evaluate the feasibility of CNV-seq and QF-PCR for chromosomal abnormalities analysis of spontaneous abortion specimens in clinical practice. Our results confirmed that chromosomal abnormalities are the most common cause of pregnancy loss and that maternal age, gestational age, and number of miscarriages were associated with fetal chromosome aberrations. Meanwhile, our study shows that a combination of CNV-seq and QF-PCR testing is efficient in the fetal chromosome analysis of miscarriages and could be used as a routine selection method for genetic analysis of abortion.

In this study, 389 specimens were successfully investigated, including 356 chorionic villi and 33 fetal muscle tissues. Of these, 2.25% (8/356) of chorionic villi and 3.03% (1/33) of tissues were removed from the primary analysis because of significant MCC. The overall detection frequency of clinically significant chromosomal abnormalities was 58.95% (224/380). Aneuploidy was the most common abnormality (178/380, 46.84%), followed by polyploidy (31/380, 8.16%) and structural chromosome anomalies (15/380, 3.95%) including partial aneuploidy and pathogenic microdeletions/microduplications. The rate and distribution of chromosomal abnormalities in our study are similar to the frequencies obtained in previous studies [7, 28–30]. Among the cases with multiple aneuploidy, double trisomy was observed in 9 cases (2.37%), which accorded with previous study [31]. Besides, three novel cases were identified in our study, including 1 autosomal trisomy with chimera of sex chromosome and 2 triploidy with autosomal tetralogy. The rate of chromosome structural anomalies (3.95%) was slightly higher than the rate observed in a previous large-scale study (2.4%) [7]. The possible reason for this discordance could be the low average maternal age as well as the use of CNV-seq and QF-PCR detection strategy in our study.

Sub-microscopic genomic imbalances or CNVs have been shown to play an important role in prenatal ultrasound anomalies and neuro-developmental disorders such as intellectual disability, autism, and epilepsy [32, 33]. Attempts are being made to identify lethal human CNVs all the time. Nevertheless, little is known about the association between specific CNVs and spontaneous abortion. The frequency of pathogenic CNV reported by previous studies was inconsistent, ranging from 0 to 11% [28, 30, 34, 35]. In our study, we detected sub-microscopic pathogenic genomic imbalances in 3.95% (15/380) of the cases. Among these cases, deletions, including 22q11.2 microdeletion, 1p36 microdeletion, 5p deletion, 8p23.1 deletion, and 5q35 microdeletion, and duplications, including 1q21.1 microduplication, 17p11.2 microduplication, and 3q29 microduplication, were found, some of which were also reported in other studies concerning miscarriage [28, 30, 36, 37]. However, it remains to be determined whether these deletions and duplications contribute to miscarriage. Although some studies considered that these microdeletions/microduplications might be related to pregnancy loss by comparing the CNVs prevalence in miscarriage products and the general population, there is still no definite conclusion due to the lack of more powerful evidence. More large-scale studies are required to confirm whether these CNVs are causative of miscarriage.

As we mentioned previously, increasing maternal age and previous losses were two frequently cited correlations. Since fetal chromosomal abnormality is the most common etiology for pregnancy loss, particularly prior to 20 weeks of gestation, gestational age has been considered an associated factors in some studies [3, 38]. In the subgroup analysis, we classified the results according to maternal age, number of miscarriages, and gestational age. For the frequency of total chromosomal abnormalities, significant difference was found only between the early abortion group and the late abortion group. The frequency of chromosomal abnormalities in the early abortion group was significantly higher than that in the late abortion group, and this was mainly reflected in the aneuploidy results. We speculate that different types of chromosomal abnormalities lead to SAs in different trimesters. It is notable that the distribution of chromosomal abnormalities was different between the advanced maternal age group and the younger maternal age group, though no statistical difference was found between these two groups in the frequency of total chromosomal abnormalities. As extensive studies about the relationship between chromosomal abnormalities and maternal age have been made, the discovery that aneuploidy increases with maternal age in cleavage stage embryos was widely known [39, 40]. In recent years, some studies have proposed that the incidence of post-meiotic abnormalities such as mosaicism, polyploidy, and structural abnormalities is not directly related to the maternal age [41]. In our study, higher frequency with aneuploidy and lower frequency with pCNV were identified in the advanced maternal age group. We provide more support for the theory that the incidence of embryonic aneuploidy increased with maternal age, while the incidences of chromosomal structural abnormalities and polyploidy seemed irrelevant to maternal age.

In order to further explore the association of maternal age, gestational age, and miscarriage times with chromosomal abnormalities of POCs, we used logistic regression models to analyze the influence of these factors on the incidence of chromosomal abnormalities. All these three factors were found to be significantly associated with the occurrence of chromosomal abnormalities. According to the results, women with advanced age, lesser gestational days, and lesser number of miscarriages were more likely to find chromosomal abnormalities in POCs. This is consistent with the suggestion that aneuploidy rates increase with maternal age and decrease with gestational age and number of prior miscarriages [42, 43].

For women, pregnancy loss is an unanticipated, physically and emotionally traumatic experience. A sense of self-accusation and guilt is often prominent because a precise medical cause is not found for the miscarriage. As an economic, robust, rapid, and high-resolution method for the genetic diagnosis in clinic, CNV-seq and QF-PCR provide a good choice for the chromosome analysis, which is the most important factor related to miscarriage. Information about the genetic etiology could avert the focus from unrelated factors, decrease the economic and psychological burden, avoid unnecessary treatment, and provide reference for specific recommendations.

Conclusion

Taken together, our results confirm that CNV-seq and QF-PCR are practical and valuable methods to determine the genetic etiology of miscarriage. We also identified a trend suggesting that the percentage of fetal chromosomal abnormalities was significantly higher in advanced maternal age, lesser gestational age, and lesser number of miscarriages than that in younger maternal age, advanced gestational age, and higher number of miscarriages. The detection rate of chromosomal abnormalities in POCs from SA can be increased by CNV-seq and QF-PCR, which is beneficial for couples with spontaneous abortion and offers better genetic counseling in the clinical setting.

Author contribution

Chen L., Wang L., and Wang J. designed the study; Tang F., Zeng Y., Yin D.S., Zhu H.M., Li L.P., and Zhang L.L. performed experiments; Chen L., Wang L., Zhou C., and Wang J. statistical analyses and interpreted results; Chen L and Wang J. wrote the manuscript. All authors reviewed the manuscript.

Funding

This work was supported by grants from the Program of Science and Technology Department of Sichuan Province (No.2020YFS0095). The funder supported our work including study design, data collection, decision to publish and preparation of the manuscript.

Data availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval

The study was approved by the Medical Ethics Committee of West China Second University Hospital of Sichuan University (medical research 2016-7).

Consent to participate

All the participants provided written informed consent for genetic investigation involving detection of fetal chromosomal anomalies using CNV-seq combined with QF-PCR.

Consent for publication

All authors of this research paper have directly participated in the analysis of the study. We all have read the final version, agreed to publish this article and confirmed that the contents of this manuscript have not been copyrighted or published previously.

Conflict of interest

The authors declare no competing interests.