Integrated CNV-seq, karyotyping and SNP-array analyses for effective prenatal diagnosis of chromosomal mosaicism

Background Emerging studies suggest that low‐coverage massively parallel copy number variation sequencing (CNV-seq) more sensitive than chromosomal microarray analysis (CMA) for detecting low-level mosaicism. However, a retrospective back-to-back comparison evaluating accuracy, efficacy, and incremental yield of CNV-seq compared with CMA is warranted. Methods A total of 72 mosaicism cases identified by karyotyping or CMA were recruited to the study. There were 67 mosaic samples co-analysed by CMA and CNV-seq, comprising 40 with sex chromosome aneuploidy, 22 with autosomal aneuploidy and 5 with large cryptic genomic rearrangements. Results Of the 67 positive mosaic cases, the levels of mosaicism defined by CNV-seq ranged from 6 to 92% compared to the ratio from 3 to 90% by karyotyping and 20% to 72% by CMA. CNV-seq not only identified all 43 chromosomal aneuploidies or large cryptic genomic rearrangements detected by CMA, but also provided a 34.88% (15/43) increased yield compared with CMA. The improved yield of mosaicism detection by CNV-seq was largely due to the ability to detect low level mosaicism below 20%. Conclusion In the context of prenatal diagnosis, CNV-seq identified additional and clinically significant mosaicism with enhanced resolution and increased sensitivity. This study provides strong evidence for applying CNV-seq as an alternative to CMA for detection of aneuploidy and mosaic variants.


Background
Chromosomal mosaicism is de ned by the presence of two or more cell populations within the body and results from either gamete meiotic or mitotic cleavage-stage errors in the early preimplantation embryo [1].
Depending on the differentiation stages when the mosaicism occurred, the abnormal cells can reside only in extra-fetal tissues (e.g. the placenta), only in the fetus, or in both. Therefore, it has an important impact on the phenotype variability in rst generation carriers but also on the recurrence risk and thus prenatal counselling [2].
Karyotyping, with a maximum resolution of 5 Mb, is referred as golden standard for identifying chromosomal abnormality in prenatal diagnosis for more than 50 years, the lower mosaicism detection limit of karyotype analysis has been reported as 19% [3] (conventional G-banded karyotype; 15 cells examined) and the mosaic pattern maybe caused by culture artifact. Chromosomal microarray (CMA) on uncultured cells from chorionic villus sampling or amniocentesis has gradually replaced conventional karyotyping for all prenatal diagnosis indications owing to a higher diagnostic yield, quicker turnaround time and elimination of cultural artifacts (pseudo mosaicism) [4]. Although it has been demonstrated to be a powerful tool to detect mosaicism at levels as low as 5% [5], but it is di cult to detect mosaicism in clinical research when the ratio is below 20% due to platform differences and specimen quality. Besides, the e ciency to detect the mosaicism of submicroscopic copy number variations (CNVs) is de ned by the design and upgrade of probes which was based on the chromosomal abnormality and clinical information from public databases.
More recently, low-coverage massively parallel copy number variation sequencing (CNV-seq) is emerging as a higher-resolution and lower-costs technology in clinical research for detecting CNVs [6]. The CNV-seq is reported to detect structural abnormalities larger than 100 bp and aneuploid chimerism lower to 5%, which is more effective than CMA [7,8]. However, no prospective back-to-back comparison study evaluating accuracy and e cacy of CNV-seq compared with CMA has been reported in routine prenatal diagnosis. Herein, we conducted a study to evaluate the diagnostic outcome and technical limitations of CMA and CNV-seq in detecting mosaicism.

Karyotyping
The amniotic uid and fetal cord blood were obtained under sterile conditions. G-banded (320-400 bands) karyotyping analyses were performed on 20 independent metaphases cells according to standard protocols. Karyotyping of at least 50 independent metaphases cells diagnosed the sample as mosaicism.

CMA analysis
Genomic DNA was extracted from amniocytes or fetal cord blood by using DNA Extraction Kit (Tissue and cells) and QIAamp DNA Blood Mini Kit (QIAGEN, Hilden, Germany) separately. SNP array was performed using Affymetrix CytoScan®750 K Array (Affymetrix Inc, CA, USA), according to the manufacturers protocol. The theoretical values for the detection of a single copy gain or loss are described as previously reported [9,10].

CNV-seq analysis
The DNA library was constructed as previously described [7]. Multiple libraries were indexed and pooled into a single lane and sequenced to 45 bp (with 8 bp being the index sequence) on the Nextseq CN500 instrument (Illumina, Inc.). Data analysis pipeline of sequencing reads and the percentage of chromosomal mosaicism were described as previously published [7,8].

Diagnostic concordance of CNV-seq and CMA vs karyotyping
Positive result of CMA was obtained on 63.08 % (41/65) samples which showed a mosaic pattern identi ed by karyotyping, of which, 37 samples showed a mosaic pattern with the level as low as 20% (Fig. 1). Besides, CMA detected additional mosaic trisomy 8 and mosaic partial trisomy 8 which was not detected by karyotyping in the cultured AF sample. In comparison, CNV-Seq not only identi ed all 43 chromosomal aneuploidies or submicroscopic CNVs detected by CMA, but also provided a 34.88% (15/43) increased yield compared with CMA( Figure 1). Besides, the level of mosaicism de ned by CNVseq range from 6% to 92%. The chromosomal map intervals, size, and copy number of the reportable mosaicisms detected by both DNA-based techniques were almost identical. In addition, there was a 100% negative concordance of CNV-Seq for the 9 remaining samples diagnosed as normal by CMA.

Chromosomal mosaicism for autosomal aneuploidy
As showed in Figure 1, a total of 22 subjects were enrolled in the group, which involving 21 cases identi ed by karyotyping and 1 case identi ed by CMA. High-risk noninvasive prenatal screening (16/22) were the most prevalent diagnosis indications ( Table 1). The relative frequencies of mosaic aneuploidies showed 54.54 % (12/22) for trisomy 21, 9.09 % (2/22) for trisomy 18, 9. 13 of 22 cases were con rmed by CMA with the level as low as 20%, while consistent CNV-seq and chromosome results were noted in 19 cases with the identi ed mosaicism at levels as low as 5 % (Fig. 2).
Of the 19 cases con rmed by CNV-seq, the percentages of cells for trisomy 21, 18 and 13 were in good agreement when CNV-seq and cytogenetics were compared. While in Case 16, Case 17, Case 18, Case 19, Case 22, the proportion of abnormal chromosome was much lower in culture samples compared with uncultured. Notably, in Case 18, the mosaic trisomy 8 was not detected in the cultured AF sample by metaphase analysis of 100 G-banded cells, while CMA and CNV-seq showed 24 % and 18 % trisomy 8 mosaicism, respectively (Supplemental Fig. S1). Of the remaining three cases (case 11, 20, 21), both CNV-seq and CMA showed a normal result in uncultured amniotic uid cells but karyotype showed a mosaic pattern of trisomy 21, trisomy 9 and trisomy 20 in cultured amniotic uid cells, respectively.
Based on the karyotyping data, the percentage of monosomic or trisomic cells varied from 3.8 % to 92 % (Table 1). Of the 40 sex chromosome aneuploidies identi ed by karyotyping, 25 cases including mosaic pattern in 23 cases were con rmed by CMA with the mosaicism level as low as 20 %, while consistent CNV-seq and chromosome results were noted in 34 cases with the level as low as 8 %. Namely, the incremental yield of mosaicism less than 20% achieved by CNV-Seq was 22.5% (9/40). There was a 100% positive concordance between CMA and CNV-Seq for 23 samples. It should be noted that the proportion of monosomy X or disomy Y was more than 30% differences in culture samples compared with uncultured samples in 5 samples (Case 38, Case 47, Case 50, Case 54 and Case 55).
In addition, for the 9 incremental cases of mosaicism identi ed by CNV-seq, CNV-Seq showed the level of mosaicism was range from 8 % to 23 %. Besides, for Case 36 and Case 41, which showed 1.92 and 1.90 haploid equivalents of chromosome X for the amniotic uid sample, CNV-seq of induced fetal placenta con rmed the placental mosaicism with chromosome X of 1.17-1.87 and 1.3-1.85 haploid equivalents (Supplemental Fig.S2).
In Case 23, 26, 27, 30, 32 and 61, both CNV-seq and CMA showed a normal result in uncultured amniotic uid cells but karyotype showed a mosaic pattern of monosomy X or disomy X in cultured amniotic uid cells. Among these cases, karyotype detected a mosaic pattern of monosomy X or disomy X less than 10% in 5 cases. The negative results of CMA and CNV-seq were most likely due to technical limitations.

Chromsomal mosaicism for submicroscopic CNVs
A total of 5 cases with a mosaic submicroscopic CNVs pattern were enrolled in the group. 4 out of 5 had a mosaic pattern with small supernumerary marker chromosome (sSMC) or unclari ed derived chromosome identi ed by karyotyping were clari ed the character, origin and pathogenicity of sSMC with the help of CMA (Number 64, 65, 66, 67)(Supplemental Fig. S3-S7). Details of the chromosomes involved and the clinical course of the 5 pregnancies are presented in Table 1.

Discussion
Current prenatal practice showed mosaicism can involve in any chromosome and be presented as many types of chromosome abnormalities including trisomy, monosomy, triploidy, deletions, duplications, rings and other types of structural rearrangements. A clinical cytogenetics laboratory performing prenatal diagnosis should understand the limitations of cell-based chromosome analyses and DNA-based CNVseq and CMA analysis on detecting mosaic aneuploidies and other submicroscopic CNVs.
To our knowledge, this is the rst prospective back-to-back study evaluating the e cacy of CNV-seq in detecting mosaicism by using CMA and karyotyping as a reference. In our study, 72 of 5367 cases showed a mosaic pattern in prenatal diagnosis with 1.39% (67/4825) detection rates among AF samples and 0.92% (5/542) detection rates among CB samples. This rate is higher than previously reported chromosomal mosaicism rates of 1%-2% in CVS [11,12] and 0.1%-0.5% in AF samples [13,14]. The high overall prevalence of 1.34%(72/5367) could be explained by the use of CMA which has a higher resolution than conventional karyotyping and therefore detects the additional mosaicism for CNVs. Besides, among the 72 mosaic pattern fetuses, high-risk of NIPS (53/72, 68.91%) are the most common prenatal diagnosis indications, NIPS provides an important clinically indications that an increased number of cells should be examined and these extra cells need to be available with analysis often targeted to the chromosome/region of interest, which is also the reason for what most of the mosaicism identi ed by karyotyping rather than CMA. Thus, doctors should be aware of the possibility of low levels of mosaicism or con ned placental mosaicism (CPM) when performing prenatal diagnosis to whom with a positive result of NIPS [15].
The current study demonstrated that CNV-seq provide an additionally diagnostic yield of 34.88% (15/43) compared to CMA, and can detect the level of mosaicism down to 5%. This is equal to the level reported by DNA models mimicking XXX and XO mosaicism [8] and supports the contention that CNV-seq is able to resolve lower levels of mosaicism than CMA. Although SNP arrays has been demonstrated to be a powerful tool to detect mosaicism at levels as low as 5% by using IlluminaQuad610 array [5], the detected rate is still variable among different CMA platform (9%-20% for array CGH [13,16] and 30%-70% for Affymetrix arrays [17,18]). Besides, it perform poorly on array CGH platforms with poor-quality, contaminated or fragmented DNA when compared with CNV-seq [19]. In addition, for submicroscopic CNVs mosaicism, the detected rate is not only due to size but also due to nonuniformly distributed probes of the CMA platform which was used [20]. Wang et al [21] had showed a variable probe density in the targeted region among different CMA platforms, and identi ed a pathogenic 298.7-kb deletion (affecting FBN2) by low-pass GS which was missed by CMA. This is indicating the advantages of applying lowpass GS for CNV analysis which is a technology relying on genome-wide uniformly distributed reads/windows. Variable proliferation of cells with different karyotype under in vitro cell culture may also contribute to the inconsistent results between CNV-seq/CMA (uncultured samples) and karyotyping (cultured samples). Cell culture may promote the in vivo selection of euploid over aneuploid cells, which has been reported to increase with age of the culture [22]. In our study, the percentages of cells for trisomy 21, 18 and 13 were in good agreement when DNA-based CNV-seq or CMA and cell-based chromosome analyses were compared. While mosaic trisomy 15 (Case 16 and Case 17), trisomy 8(Case 18), trisomy 2 (Case 22) and trisomy 22 (Case 19) were detected in the direct AF sample by CNV-Seq and CMA, however obviously lower levels of mosaicism was detected by karyotyping on cultured amniocytes which is consistent with the research reported previously [23][24][25][26]. It is feasible that the normal cells may have had a growth advantage in culture or the abnormal cell line may have a culturing disadvantage [13]. While increased proportions of monosomy X in karyotyping compared to CNV-seq, it is seems that 7 of 30 the monosomy X cell line have growth advantage than the normal cells or this deviations is caused by arti cial counts. This highlights the advantage of using direct uncultured samples where a quicker result is possible and which can avoid artifact of culture and promote the accuracy of fetal outcome.
There are nine inconsistent results between CNV-seq (uncultured sample) and karyotype. In Case 27, the mosaic nding of monosomy X was con rmed by two independent prenatal clinics with the karyotype of 45,X[1]/46,XY[49] and 45,X [4]/46,XY[69], respectively. Thus, the negative results of CMA and CNV-seq were most likely due to technical limitations. Besides, a healthy 2870-g female baby was delivered with no phenotypic abnormality at 39 weeks of gestation. The postnatal blood karyotype was 46, XX. The abnormal chromosome con rmed by karyotyping in 5 cases (case 11,23,26,30,61) is consistent with results of NIPS. In case 20, 21 and 32, the levels of mosaicism detected by karyotyping were higher than 10%, which were less likely to be caused by culture artifact. Besides, studies [27,28] has showed the discrepancy in the trisomy mosaicism level between cultured amniocytes and uncultured amniocytes in prenatally detected mosaic trisomy 20 and trisomy 9. Comprehensive evaluation of prevalent diagnosis indications and karyotype, it implies that the negative results of CNV-seq is more likely limited by technology with the chimerism level less than 5% in uncultured samples. In Case 23, the mosaic pattern of monosomy X detected in cultured amniocytes would not rule out a pseudo-mosaicism caused by culture artifact.
CNV-Seq is a low-read depth platform and unable to detect balanced translocations and inversions. Clarifying mosaic pattern result from structural chromosomal rearrangements should combine with chromosome analysis. A clinical cytogenetics laboratory performing prenatal diagnosis should understand the technical limitations of cell-based chromosome analyses and DNA-based CNV-seq and CMA analysis on detecting mosaic aneuploidies and other submicroscopic CNVs. Thus, the analysis and the interpretation of a mosaic pattern should be cautious in the prenatal diagnosis. Collectively, these validation studies con rm that CNV-seq is a quantitative NGS protocol with high sensitivity and reproducibility for also detecting chromosome mosaicism in prenatal diagnosis.

Conclusions
This study evaluated the effectiveness of CNV-Seq for detecting low-level mosaicism in prenatal diagnosis. The retrospective analysis found that CNV-seq identi ed additional and clinically signi cant information with enhanced resolution and increased sensitivity of detecting mosaicism (34.88%) as compared with the CMA platform we used. So it provides strong evidence for applying CNV-seq as an alternative prenatal diagnostic test for detecting mosaicism. Whilst the diagnosis of mosaicism in a prenatal setting remains challenging for scientists, we believe that combined use of cell-based chromosome analyses and DNA-based CNV-seq and CMA analysis would provide help in prenatal genetic diagnosis and counseling for mosaicism.

Declarations
The datasets used or analysed during the current study are available from the corresponding author on reasonable request.

Ethics approval and consent to participate
Written informed consent for prenatal genetic investigation was obtained from each study participant. For this retrospective study, there were no pre-study requirements on the patient's specimens and clinical indications and there were no post-study interaction and intervention with the patients. This project was categorized as a chart review retrospective study and approved by the Ethics Committee of Hunan Provincial Maternal and Child Health Care Hospital (approved number 20180920-5).

Consent for publication
Written informed consent for publication of clinical and genetic data was obtained from all participants.