Extended application of PGT-M strategies for small pathogenic CNVs

Abstract

Purpose

The preimplantation genetic testing for aneuploidy (PGT-A) platform is not currently available for small copy-number variants (CNVs), especially those < 1 Mb. Through strategies used in PGT for monogenic disease (PGT-M), this study intended to perform PGT for families with small pathogenic CNVs.

Methods

Couples who carried small pathogenic CNVs and underwent PGT at the Reproductive and Genetic Hospital of CITIC-Xiangya (Hunan, China) between November 2019 and April 2023 were included in this study. Haplotype analysis was performed through two platforms (targeted sequencing and whole-genome arrays) to identify the unaffected embryos, which were subjected to transplantation. Prenatal diagnosis using amniotic fluid was performed during 18–20 weeks of pregnancy.

Results

PGT was successfully performed for 20 small CNVs (15 microdeletions and 5 microduplications) in 20 families. These CNVs distributed on chromosomes 1, 2, 6, 7, 13, 15, 16, and X with sizes ranging from 57 to 2120 kb. Three haplotyping-based PGT-M strategies were applied. A total of 89 embryos were identified in 25 PGT cycles for the 20 families. The diagnostic yield was 98.9% (88/89). Nineteen transfers were performed for 17 women, resulting in a 78.9% (15/19) clinical pregnancy rate after each transplantation. Of the nine women who had healthy babies, eight accepted prenatal diagnosis and the results showed no related pathogenic CNVs.

Conclusion

Our results show that the extended haplotyping-based PGT-M strategy application for small pathogenic CNVs compensated for the insufficient resolution of PGT-A. These three PGT-M strategies could be applied to couples with small pathogenic CNVs.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10815-024-03028-6.

Introduction

Chromosomal copy number variants (CNVs) are defined as a gain or loss of DNA compared with the reference human genome, ranging in size from a kilobase to several megabases or even an entire chromosome [1], whose clinical significance has been classified as benign, likely benign, likely pathogenic (LP), pathogenic (P), or uncertain (VUS) [2]. The Database of Genomic Variants showed that 983,845 CNVs have been discovered by February 2020, mostly ranging between 1 kb and 3 Mb with an average of 465 kb and accounting for at least 12% of the human genome. Although more than half of CNVs are referred to as genomic polymorphisms, a number of CNVs can result in complicated diseases (such as chromosomal microdeletion and microduplication syndromes [3–5]), and a few can even cause Mendelian diseases [6]. A study of 23,865 fetuses showed that 375 fetuses (1.6%) carried pathogenic CNVs in prenatal diagnosis and 44 (11.7%) contained two or more pathogenic CNVs [7].

Once an LP/P CNV is detected in a couple, preimplantation genetic testing (PGT) or prenatal diagnosis can be used to prevent the recurrence of the disorder in the next generation. PGT has been used to prevent birth defects, increase the chances of a successful pregnancy, and reduce the occurrence of spontaneous abortions since 1990s [8]. Currently, embryos are routinely generated by intracytoplasmic sperm injection (ICSI) in PGT, and a biopsy from each embryo is genetically tested for monogenic disease (PGT-M), structural rearrangements (PGT-SR), and/or genome-wide aneuploidy (PGT-A) [9]. Although the PGT-A platform has been widely implemented for CNV detection, its resolution is commonly between 5 and 10 Mb [10]. Whole-genome amplification (WGA) and next-generation sequencing (NGS) technology developments have only furthered this resolution to approximately 1 Mb [11]. The PGT-A platform is not currently available for small CNVs, especially those < 1 Mb.

PGT-M presents an alternative method for solving these issues and has been traditionally used for nuclear DNA pathogenic variant(s) causing monogenic disorders, mitochondrial DNA pathogenic variant(s), and human leukocyte antigen typing with or without concurrent testing for monogenic disorders [12]. Recently, PGT-M was performed for a chr14q32 microdeletion (108 kb) in a family with Kagami–Ogata and Temple syndromes, indicating that the PGT-M strategy could be used to block rare genomic disease transmission [13]. PGT has also been applied after IVF for VUS CNVs, demonstrating the feasibility of PGT on VUS CNV as the clinical interpretation of VUS may change over time, and suggesting the necessity of re-evaluation for these patients before each pregnancy [14]. However, no recommendations or guidelines have been provided for PGT-M involving small CNVs. Moreover, determining whether to perform PGT for small CNVs clinically is difficult, as the clinical phenotypes of individuals with small CNVs are highly heterogeneous [15] due to incomplete penetrance, variable expressivity, or delayed dominance [14]. Therefore, regulation and standardization of CNV testing in PGT, including patient background evaluation, understanding decision-making factors regarding seeking PGT treatment, and developing the clinical analysis strategy and workflow, is crucial [16].

In this study, we focused on patients with small pathogenic CNVs and aimed to summarize the PGT-M strategies used for evaluating these small CNVs. Moreover, we investigate patient backgrounds seeking PGT-M in our hospital, the characteristics of these CNVs, and report the clinical outcomes. Our study contributes to the fast-evolving PGT-M strategy practices for future genomic medicines.

Materials and methods

Patient selection

The selection criteria included couples who carried small pathogenic CNVs and underwent PGT at the Reproductive and Genetic Hospital of CITIC-Xiangya, Hunan, China, between November 2019 and April 2023. CNVs were identified using chromosomal microarray (CMA) or low-pass genome sequencing (WGS, 0.5× in depth, 100 kb in resolution) and classified according to American College of Medical Genetics and Genomics and ClinGen guidelines [2]. Only CNVs of < 4 Mb and classified as LP or P were included. In most cases, the origin of the CNVs was evaluated and identified as inherited or de novo. Couples in the following situations were admitted undergoing PGT-M: (i) the asymptomatic couple had at least one affected child/fetus that carried the LP/P CNVs or (ii) one of the partners was diagnosed with chromosomal microdeletion or microduplication syndromes. Couples with other genetic defects, including chromosomal segmental abnormalities (> 4 Mb), balanced chromosomal rearrangements, and gene mutations/small CNVs resulting in monogenic diseases, were excluded. This study was approved by the Institutional Ethics Committee of the Reproductive Genetic Hospital of CITIC-Xiangya (LL-SC-XJS-2017-001), and the principles of the Declaration of Helsinki were followed. All patients underwent genetic counseling and signed consent forms at the beginning of their treatment.

Preliminary PGT-M process

Preliminary experiments on familial samples were conducted to assess the feasibility of PGT-M by establishing high-risk haplotypes using informative single-nucleotide polymorphism (SNP) markers [12]. An SNP was considered informative when a genotype could be assigned to one of the chromosomes inherited from the mother or father. Preliminary experiments were performed as previously described [17, 18]. Genomic DNA was extracted from available samples using a QIAamp DNA Blood Midi kit (QIAGEN, Hilden, Germany). Targeted NGS sequencing (NEXTSeq 500; Illumina, San Diego, CA, USA) or SNP arrays (Illumina Infinium Asian Screening Array-24 v1.0) were used to investigate the presence of informative SNPs inside the microdeletion or 1 Mb upstream to 1 Mb downstream of the microdeletions/microduplications. Data were then mapped to the human reference genome (hg19, GRCh37) using the Burrows–Wheeler Alignment Tool (http://bio-bwa.sourceforge.net/). Genome Analysis Toolkit (http://www.broadinstitute.org/gatk/) and Bluefuse Multi Software (Illumina, San Diego, CA, USA) were used for SNP-based analysis [19, 20].

ICSI and embryo biopsy

For PGT, ovarian stimulation was performed based on the patient’s status [21]. ICSI was performed on all metaphase II oocytes, and fertilization was determined by visualizing two pronuclei and two polar bodies 16–18 h post-injection. Embryos were cultured in sequential media (G1 and G2; Vitrolife, Gothenburg, Sweden) until the blastocyst stage under 6% carbon dioxide, 5% oxygen, and 89% nitrogen. Laser-mediated trophectoderm (TE) biopsy was performed on day 5 or 6 after zona pellucida dissection; 3–8 cells were collected in sterile RNase- and DNase-free PCR tubes for subsequent genetic analyses. The blastocysts were cryopreserved for further embryo transfer.

PGT for small CNVs and aneuploidy

Multiple displacement amplification (MDA)–based WGA was performed on biopsied cells using the REPLI-g Single Cell Kit (150345, QIAGEN). All MDA products and genomic DNA from the family members or affected offspring were used for library construction and subsequent sequence analysis was performed using targeted NGS (NEXTSeq 500; Illumina, San Diego, CA, USA) or SNP array (Illumina Infinium Asian Screening Array-24 v1.0) platforms. The mean sequencing depths for the SNPs should be more than 30× when using targeted NGS platforms. The mean coverage of SNP arrays is more than 90%. Informative SNPs were selected as described in the preliminary process and SNP-based haplotypes of the embryos were used to determine whether the embryos were affected. For couples who matched the indications for PGT-A, NGS-based PGT-A analysis was applied [11]. Only unaffected embryos were selected for transfer.

Pregnancy confirmation, prenatal genetic diagnosis, and follow-up

A positive serum β-human chorionic gonadotropin (hCG) level of >5 mIU/mL was used to confirm gestation 12 days after transfer. The presence of a gestational sac detected by ultrasound examination was used to define clinical pregnancy. Prenatal molecular analysis through amniocentesis was recommended to confirm the PGT-M results in the second trimester (usually 18-20 weeks of pregnancy). Karyotype analysis using the G-banding and CNV analysis using CMA or low-pass WGS (1.5× in depth, 100 kb in resolution) was performed during prenatal diagnosis [22]. A follow-up of the pregnancy and newborn of each family was conducted until February 2023. The clinical pregnancy rate after transplantation (clinical pregnancy numbers/total transplantation numbers) was analyzed.

Results

Clinical characteristics of the families undergoing PGT-M

A total of 20 families were included in this study, of which 15 were affected by chromosomal microdeletions and 5 were affected by chromosomal microduplications (Table 1 and Supplemental Table 1). Though PGT for monogenic or chromosomal disorders had definite indications, the clinical features of couples applied for PGT due to small CNVs seemed to be complex (Fig. 1). Of 17 couples that had adverse pregnancy histories (including early pregnancy loss, fetal malformation, or affected offspring) (Supplemental Table 2), 8 experienced at least two adverse pregnancies. Moreover, in three families (F6, F13, and F16), the couples requested PGT-M because one partner had a phenotype related to microdeletion/microduplication syndrome (Supplemental Table 2).

Table 1.

PGT cycles for 20 families affected by small LP/P CNVs

No.	Female ages (years)	CNVs (hg19)	Inherited or de novo	PGT cycle	PGT-M for small CNVs				PGT-A: Normal/total	Prenatal diagnosis	Clinical outcomes
					Platform	Analysis strategies	Normal/total detected embryos	Embryos with recombination*
Microdeletions (F1-F15)
F1	35	1q21.1	Unknown	1	SNP array	Strategies 1 and 3	1/1	0	1/1	NA	No pregnancy
				2	SNP array		3/6	0	1/3	NA	Early abortion
F2	27	1q21.1q21.2	Inherited	1	Targeted NGS	Strategies 1 and 3	2/2	0	NA	NA	No pregnancy
				2	Targeted NGS		1/2	0	NA	NA	No pregnancy
F3	32	1q31.3	Unknown	1	Targeted NGS	Strategy 1	0/2	0	NA	NA	No unaffected embryos
F4	36	2p16.3	Unknown	1	SNP array	Strategies 1 and 3	0/1	0	NA	NA	No unaffected embryos
				2	SNP array		0/2	0	NA	NA	No unaffected embryos
F5	27	6q27	De novo	1	SNP array	Strategies 2 and 3	3/4	0	3/3	NA	In pregnancy (8 weeks)
F6	21	7q11.23	De novo	1	Targeted NGS	Strategies 2 and 3	2/4	0	2/2	Normal	One healthy boy
F7	30	13q24.21	Inherited	1	Targeted NGS	Strategies 1 and 3	3/6	0	2/3	Not performed	One healthy girl
F8	28	15q11.2	Inherited	1	Targeted NGS	Strategy 3	4/6	2	3/4	Normal	One healthy boy
F9	30	15q11.2	Inherited	1	Targeted NGS	Strategies 1 and 3	4/6	1	3/4	NA	Early abortion
F10	26	15q11.2	Inherited	1	Targeted NGS	Strategies 1 and 3	1/3	0	1/1	NA	Not transplanted
				2	Targeted NGS		2/5	1	1/2	NA	In pregnancy (9 weeks)
F11	36	16p11.2	Unknown	1	Targeted NGS	Strategies 1 and 3	2/4	0	2/2	Normal	One healthy girl
F12	24	16p11.2	De novo	1	Targeted NGS	Strategies 1 and 3	2/4	0	1/2	Normal	One healthy girl
F13	38	16p11.2	De novo	1	Targeted NGS	Strategies 2 and 3	0/1	0	NA	NA	No unaffected embryos
F14	36	16p11.2	De novo	1	SNP array	Strategies 1, 2, and 3	3/4	0	2/3	NA	Early abortion
F15	30	Xq28	De novo	1	Targeted NGS	Strategies 1 and 3	2/3	0	2/2	Normal	One healthy boy
Microduplications (F16-F20)
F16	29	7q36.3	Inherited	1	SNP array	Strategy 1	2/4	1	NA	NA	In pregnancy (20 weeks)
F17	35	16p11.2	Unknown	1	Targeted NGS	Strategy 1	3/4	0	3/3	Normal	One healthy boy
F18	28	16p13.11	Inherited	1	SNP array	Strategy 1	0/2	0	NA	NA	No unaffected embryos
				2	SNP array	Strategy 1	5/5	0	2/2	NA	No pregnancy
F19	28	Xq22.2	Inherited	1	Targeted NGS	Strategy 1	2/5	1	1/2	Normal	One healthy girl
F20	30	Xq27.1	Unknown	1	Targeted NGS	Strategy 1	0/3	0	1/1	Normal	One healthy girl

PGT-M preimplantation genetic testing for monogenic diseases, LP/P likely pathogenic/pathogenic, CNVs copy number variants, PGT-A preimplantation genetic testing for aneuploidy, NSG next-generation sequencing, SNP single-nucleotide polymorphism, del microdeletion, dup microduplication, NA not applicable

*Recombination events within 1Mb upstream and downstream of the CNVs were counted

Fig. 1.

Preimplantation genetic testing for monogenic diseases (PGT-M) for 20 small copy number variants (CNVs) in 20 families. a Sizes and chromosomal location of 20 involved small CNVs. b Statistics of the strategies used in PGT-M for 20 small CNVs. c Clinical outcomes of 19 transfer cycles for 17 families

All 20 identified CNVs were classified into LP/P with sizes ranging from 57 to 2120 kb, with an average size of 777.85 kb (Supplemental Table 1). Sixteen of 20 (80.0%) CNVs were < 1 Mb (Fig. 1a), and the minimum size of CNVs involved in this study was 57kb. The 20 involved small CNVs were distributed on chromosomes 1, 2, 6, 7, 13, 15, 16, and X (Supplemental Fig. 1a). The most common hotspot for these small CNVs was 16p11.2, which was implemented for PGT-M in five families (Fig. 1a). Notably, three of the CNVs classified as LP/P were not documented in OMIM, among which two were reported in other studies and one (13q24.21 microdeletion in F7) was reported for the first time in this study (Supplemental Table 1). Among the 14 CNVs whose origins could be determined (14/20, 70.0%), inherited and de novo CNVs in the affected partner accounted for 57.1% (8/14) and 42.9% (6/14), respectively (Table 1 and Supplemental Fig. 1b). Overall, these couples carrying small pathogenic CNVs were considered to have high-risk pregnancies and had indications for PGT-M due to varied penetrance rates of 10.4–100% (Supplemental Table 1) [23–25].

Three strategies for PGT of small pathogenic CNVs

During PGT-M treatment for the 20 families in this study, three SNP analysis strategies were summarized to establish high-risk haplotypes, or directly distinguish embryos (Fig. 2). When there are samples from another family member carrying the small CNVs (affected offspring/fetus, or his/her parent) besides the affected partner, the high-risk haplotype could be distinguished by linkage analysis using informative SNPs 1 Mb upstream and downstream of the CNVs (strategy 1, Fig. 2). Specifically for microdeletions, another two strategies could also be used because SNPs within the deletion regions could be reliable. In strategy 2, when there are samples from the affected partners’ parents not carrying the CNVs, informative SNPs within the microdeletion region (see SNP combinations in Supplemental Table 3) are used to first determine the origin of the high-risk chromosome, which is subsequently used to establish the high-risk haplotype (strategy 2, Fig. 2). For example, the origin of the high-risk chromosome could be determined according to the missing allele of the affected partner in strategy 2 (Supplemental Table 3) when his/her unaffected parents carry homozygous wild alleles and homozygous mutant alleles, respectively. Alternatively, the embryos could be directly judged as unaffected if there were heterozygous SNPs within the microdeletion region from the embryos (strategy 3, Fig. 2; see SNP combinations in Supplemental Table 4).

Fig. 2.

A working-flow of PGT for small CNVs in this study. The whole process includes clinical diagnosis, molecular diagnosis, genetic counseling, preliminary analysis, and PGT analysis. During preliminary/PGT analysis, three single-nucleotide polymorphism (SNP)–based strategies were summarized to establish high-risk haplotype and determine affected embryo. In each strategy, representative pedigree is depicted. A dot in the middle of the symbol indicates CNV carriers. Black dotted lines indicate informative SNPs upstream or downstream the microdeletion (del) or microduplication (dup), while red dotted lines represent informative SNPs within the microdeletion. Strategies 1 and 2 can be done in preliminary analysis, while strategy 3 is only available during PGT analysis. E1-E4 indicate embryos

Specifically, in 17 families, samples were collected from another family member carrying the small CNVs, and strategy 1 was effective for 16 families (80%, 16/20) (Fig. 1b and Supplemental Fig. 2). Only in one family (F8) strategy 1 was not effective because there was no upstream SNPs. Strategies 2 and 3 could not be used for families with microduplications because no informative SNPs could be provided within the duplication region using the routine sequencing platforms. Specifically, for 15 families with microdeletions, the efficiency rate for strategy 3 was 93.3% (14/15) (Supplemental Fig. 2). Only in one family (F3) it was invalid to use strategy 3 because the deletion region was too small (57kb) to have informative SNPs. In three families (F5, F6, and F1), samples from parents of the affected partner were available and strategy 2 was applied.

Overall, 13 out of 20 families used more than one strategy, including ten (10/20, 50%) with strategy 1 plus strategy 3, and three (3/20, 15%) with strategy 2 plus strategy 3 (Fig. 1b and Supplemental Fig. 2). Six families used only strategy 1 (6/20, 30%), and one family used only strategy 3 (1/20, 5%). Our results show that these three strategies can be used either alone or in combination to produce consistent results for judgment. To the best of our knowledge, the 20 included cases carrying small CNVs represented most clinical scenarios, indicating that our three strategies could be adapted to most PGT for small CNVs.

PGT for couples with small CNVs

During the study period, 25 PGT-M cycles were performed for 20 families (five of them had two PGT-M cycles). The amplification efficiency of MDA was 100% in 89 embryo biopsies. A total of 47 embryos were determined as unaffected, 36 as affected, three as X-linked carriers, two with other abnormalities (likely monosomy), and one as indeterminate due to chromosome recombination (Table 1 and Supplemental Fig. 3). The overall diagnostic yield was 98.9% (88/89). Recombination events within 1Mb upstream and downstream of the CNVs were identified in six embryos (6/89, 6.7%) (Table 1). Moreover, 17 of 25 (68.0%) PGT-M cycles proceeded to PGT-A analysis, and 31 of 40 (77.5%) embryos were normal after PGT-A (Table 1).

PGT was performed for one 1.77 Mb paternal deletion in the 15q11.2 region near the centromere in F8 (Fig. 3). The couple (F8: II1 and II2) was informed of fetal malformations of the right kidney and left ventricle (F8: III1) during the 6-month ultrasound examination during their first pregnancy (Fig. 3a). NGS-based WGS (100kb resolution) for the fetus and related family members revealed a 685.4-kb heterozygous microdeletion of 15q11.2 transmitted from the grandfather (F8:I1). The deletion region contained the chromosome 15q11.2 deletion syndrome (OMIM:615656)-related region, with an approximately 10.4% penetrance. As the transmission of this deletion to their offspring in subsequent pregnancies was 50%, the couple requested PGT following genetic counseling. During the preliminary analysis, targeted NGS was performed for I1, I2 (affected), and II1 (affected). However, no SNPs was detected upstream the deletion region because this deletion was near the centromere (Fig. 3b–c). Using only downstream SNPs for haplotype establishment was risky, because recombination may happen and led to wrong judgment for embryos. Therefore, we used strategy 3 in this family. Specifically, seven preferred SNPs were identified in the preliminary experiments (i.e., 1 and 3) (Fig. 2 and Fig. 3b), which were informative for determining genotypes of all embryos (Fig. 3c). Moreover, a total of 13 usable SNPs (i.e., 2 and 4) were also identified in the preliminary experiments (Fig. 2 and Fig. 3b), but they were not always informative for determining the embryo genotypes (Fig. 3c). Linked informative SNPs of the embryos could be combinedly used to deduce or strengthen the determination for embryos. Using this strategy, we could distinguish four low-risk embryos (E1, E3, E5, and E6) and two high-risk embryos (E2 and E4). Maternal chromosome recombination in this region (the breakpoint near the downstream boundary of the 15q11.2 microdeletion) was observed in E4 (Fig. 3c), implying there was a chance of recombination within 1Mb downstream of the microdeletion. Three of the four embryos were suitable for transfer after PGT-A testing (E1, E5, and E6). Finally, the E6 euploid-unaffected embryo was transplanted, and a healthy boy was born.

Fig. 3.

PGT-M strategy for one paternal 15q11.2 microdeletion in F8. a The pedigree of a consanguineous family (F8) affected by the 15q11.2 microdeletion. The dots indicate members carrying the microdeletion. The black solid triangle shows one malformed fetus. b-c Representative SNPs in the family members (b) and the embryos (c). SNPs within the deletion region and upstream/downstream are depicted in red and black, respectively. The asterisk indicates the centromere region lacking SNPs. The informative SNPs within the del region in preliminary process were classified into two groups (b) showing in the table where A and B represent two different nucleotides (A/T/C/G). For SNPs of the embryos, the gray-marked ones within the del region were not informative, and the informative SNPs downstream used for analysis were depicted by black doted boxes (c)

PGT-M clinical outcomes for small CNVs

A total of 25 PGT-M cycles were performed for the 20 families with small pathogenic CNVs. In four cycles from four families (F3, F4, F13, and F18), no unaffected embryos were suitable for transfer. Overall, 19 transfers were performed for 17 families that had suitable embryos during the study period (Table 1 and Fig. 1c). Four transfers did not result in pregnancy (F1, F2, and F18), leading to a final clinical pregnancy rate of 78.9% (15/19), which is equivalent to that of regular assisted reproductive technology. Among 15 transfers that had clinical pregnancies, three ended up with early abortion (F1, F9, and F14), three resulted in ongoing pregnancy (F5, F10, and F16), and nine contributed to healthy babies (Table 1). Eight families underwent prenatal diagnosis through amniocentesis, and all had normal results (Table 1). All offspring born in this study through PGT were healthy to date (no preterm birth, low birth weight, macrosomia, or birth defects). Collectively, these results indicate that our PGT strategies for small CNVs were safe and effective.

Discussion

We performed an integrated analysis of 20 families affected by LP/P CNVs. PGT analyses were carried out for 89 embryos. These results demonstrated the effectiveness and feasibility of SNP-based haplotype analysis for judging the affected embryo carrying small CNVs during PGT.

The last two decades have witnessed rapid advancements in sequencing technology and bioinformatics, resulting in an explosion of small CNVs associated with human diseases [15, 26]. Clinical concerns and needs have also been growing regarding PGT with the LP/P CNVs that could result in chromosomal microdeletion/microduplication syndromes or adverse pregnancies. PGT for CNVs is usually achieved by PGT-A, where WGA is used as a DNA enrichment step; however, amplification bias of WGA can artificially introduce regions of the genome being over or less, resulting in a limited chromosome resolution of approximately 5 Mb [27, 28]. Although we established a 1-Mb resolution platform for PGT-A [11], further validation using large sample sizes is still required. Small LP/P CNVs, especially those under 1 Mb, which could not be clearly determined using PGT-A platforms, were detected using the PGT-M strategy. PGT-M analysis relies on informative SNPs by resolving the sequence differences of parents (e.g., heterozygous alleles) within or near variant sites. Therefore, PGT-M analysis is particularly personalized and typically performed on a case-by-case basis.

In this study, a workflow of PGT for small CNVs was depicted and three SNP-based analysis strategies were summarized (Fig. 2). Linkage analysis based on upstream and downstream SNPs was still a basic strategy (strategy 1); as for microdeletions, another two strategies taking advantage of the informative SNPs within the deletion region were also available (strategies 2 and 3). Importantly, when using linkage analysis of strategy 1, both upstream and downstream SNP analyses are required. Unilateral SNPs would be at high risk of misjudgment due to recombination events (Supplemental Fig. 3).While non-recurrent CNVs (microdeletions/microduplications) are mediated by mechanisms such as non-homologous end joining or replication-based mechanisms (fork stalling and template switching), the major mechanism responsible for recurrent CNVs is non-allelic homologous recombination (NAHR) occurring between two paralogous low-copy repeats or segmental duplications [29]. Therefore, CNV sites are also potentially susceptible to NAHR-mediated recombination, underlining the importance of tracking these events in future PGT-M strategies.

In our study, only one strategy (strategy 1) could be applicable in PGT for microduplications. It should also be noted that only when the duplications are arranged in tandem arrays or very close to the original locations can the SNP-based linkage analysis (strategy 1) be used for PGT towards microduplications. Careful consideration should be conducted when the CNVs are located in special regions (such as telomeric or centromeric regions and regions with highly repetitive sequences), as limited SNPs could be used. In other situations, if samples from other family members (apart from the couple) are unavailable, phasing can only be worked out during PGT cycle, subject to the number of embryos available. In such a case, couples should be counseled for the possible outcomes. In addition, these three strategies used in our current study still had a few limitations. For example, PGT-A cannot be performed simultaneously with PGT-M through targeted sequencing platform, so most of these patients had to proceed with further PGT-A analysis, which is often time-consuming and less cost-effective. Currently, we are attempting to resolve these issues by improving the segment resolution of PGT-A or using comprehensive genomic analysis for PGT [11, 20]. Moreover, with technology development regarding direct small CNV determinations [29, 30], we believe additional effective strategies will become available to assist clinical decision-making.

In addition to the analysis strategy involved in PGT for small CNVs, another long-standing, unanswered issue that should be mentioned is necessity [14]. In this study, the prerequisite for introducing PGT-M to affected couples was the pathogenicity (LP/P) of these CNVs. However, decision-making for performing PGT for these CNVs remains complicated by variable phenotypes among people carrying the same CNVs. Although there is consensus that PGT should be applied to lethal or severe childhood diseases, there is little agreement on the use of PGT for variable expression conditions [31]. Importantly, all 20 families affected by LP/P CNVs wanted to take advantage of PGT to have healthy babies, regardless of phenotypic heterogeneity. In this study, only three probands were the male/female partner themselves, who presented with CNV-related diseases. However, there were 14 families in which some members that inherited small CNVs were symptomless, and the phenotypes of the affected offspring were highly related to the small CNVs. The remaining three families, in which some carriers were also symptomless, had an uncertain genotype–phenotype association due to unexplained early pregnancy loss or recurrent abortion. The mechanisms underlying phenotypic heterogeneity are extremely intricate [32, 33]. The preferential inactivation of the defective X chromosome due to a skewed inactivation mechanism [34] was one explanation for the symptomless carriers of X-linked small CNVs. However, the male offspring had a 50% risk of being affected. Moreover, because X chromosomal inactivation is reprogrammed during early development, female offspring are not guaranteed against being affected. PGT is highly effective in preventing related diseases in these scenarios. For small pathogenic CNVs in autosomes, phenotypic heterogeneity is usually explained by incomplete penetrance and variable expressivity. Overall, reproductive counseling in the context of PGT-M for CNVs with phenotypic heterogeneity should be performed carefully until more clinical evidence is available. Those with a sick child or previous termination could preferentially consider the use of PGT, which is also in agreement with the motivation of most patients [16, 31].

Genetic counseling is critical when patients decide to accept PGT to thoroughly inform them of the probability of success, the need for more than one cycle, and the potential of no suitable embryos to transfer. Five families received more than one PGT cycle, and four families had no suitable embryos for transfer after PGT-M for small CNVs (all four families have had malformed fetuses or affected offspring before). These results indicate that these types of couples should be recommended to explore reproductive therapy at an earlier age. Moreover, some couples suffering from CNV-related diseases (e.g., mental disorders in F6 and F13) needed to take medications, which may have had reproductive toxicity. Therefore, interdisciplinary research should also be considered to explore reproductive protection in of CNV-related disease treatment [35].

We are currently standing in a challenging era of genetics flooded by unprecedented progress. By taking advantage of this progress, the goal of clinical PGT-M implementation is to accurately discriminate affected embryos and select the most cost-effective and the least time-consuming method for patients. In this study, we demonstrated the general PGT-M analysis strategy for small CNVs and discussed the effectiveness of PGT-M in the clinical context, which may shed light on clinical practice and further improvements in this field.

Author contribution

Conceptualization: Ju.D., G.L., G.X.L., Y.-Q.T.; investigation: X.H., W.L.W., Ji.D.; funding acquisition: Ju.D., Y.-Q.T., W.L.W.; methodology: X.H., Y.Z., W.B.H.; clinical resources: F.G., K.L.L., S.P.Z., L.L.Y; project administration and supervision: G.L., Y.-Q.T., Q.J.Z., W.L.; validation: Ji.D., Z.X.W., Y.Z., Q.T.; writing—original draft: W.L.W., X.H.; writing—review and editing: Ju.D., G.L.

Funding

This work was supported by the National Key Research and Developmental Program of China (2022YFC2702604 to Y.-Q.T.), the National Natural Science Foundation of China (81971447 and 82171608 to Y.-Q.T.,), the Hunan Provincial Natural Science Foundation of China (2023JJ40459 to W.L.W., 2022JJ30772 to Ju.D.), China Postdoctoral Science Foundation (2022M721124 to W.L.W.), and Research Grant from Reproductive and Genetic Hospital of CITIC-Xiangya (YNXM-202305 to W.L.W., YNXM-202002 to Ju.D.).

Data availability

The targeted sequencing or SNP array data in this study have not been deposited in public repository because of privacy and ethical restrictions but are available from the corresponding author on request.

Declarations

Conflict of interest

The authors declare no competing interests.