Array-comparative genomic hybridization results in clinically affected cases with apparently balanced chromosomal rearrangements

Balanced chromosomal rearrangements comprising translocations or inversions with an incidence 0.52% in the general population are structural chromosomal rearrangements without cytogenetically detected imbalances [1]. As conventional karyotyping might not be able to detect the smaller than 5-10 Mb imbalances, some submicroscopic imbalances can be overlooked [2]. These types of rearrangements are called as apparently balanced chromosomal rearrangements (ABCRs) and can be familial or de novo in origin. Warburton [3] determined the risk of phenotypical abnormalities as 6.1% for de novo ABCR translocations and inversions in a large prenatal series. Madan et al. [4] estimated the risk as 23.0% for complex chromosome rearrangements (CCRs), defining structural rearrangements involving more than two breakpoints. Especially, if the ABCR is de novo, it is important to search the submicroscopic genomic imbalances. Not only the cryptic imbalances could affect the phenotype but also other mechanisms, such as impairment of the expression of the gene in its new location (positional effects), disruption of genes at the breakpoints, formation of a new chimeric gene and disruption of the parental imprinting pattern might have a role [5].

The application of chromosomal microarray (CMA) has become an essential tool in routine diagnostics to detect the submicroscopic genomic imbalances for genome-wide screening at kilobase (kb) levels. Chromosomal microarray contributes a 15.0-20.0% detection rate of cryptic chromosomal imbalance in selected patients who have multiple congenital anomalies/intellectual disability (MCA/ID) with normal karyotypes [6,7].

We report here the CMA results of the 34 clinically affected patients (13 prenatal, 21 postnatal) carrying de novo or familial ABCRs of a single center in Turkey. The aim of the study was to add new cases to the literature, to present the rate of submicroscopic genomic imbalances according to different type of chromosomal abnormalities such as translocations, inversions, and CCRs, which are important in genetic counseling, and these findings were combined with published studies to provide the newest integrated insights.

The study contains the cytogenetic and array-comparative genomic hybridization (aCGH) results of 34 affected patients (13 prenatal, 21 postnatal) carrying ABCRs and 76 of family members (including parents and siblings) investigated between the years 2001-2017. All cases were examined, counseled and laboratory work was performed at the Department of Medical Genetics, Istanbul University Faculty of Medicine, Istanbul, Turkey. Fetal ultrasonography (USG) and invasive procedures [one chorionic villus sampling (CVS), four fetal blood sampling (FSB) and eight by amniocentesis (AC)] were performed at the Perinatology Division of the Obstetrics and Gynecology Department, Istanbul University Faculty of Medicine, Istanbul, Turkey.

Lymphocytes cultures of peripheral and cord blood samples, amniotic fluid and chorionic villus samples using short and long-term cell cultures were performed according to traditional techniques. Routine cytogenetic analyses were performed on metaphase chromosomes at 550-600 banding levels using Giemsa-Pancreatin-Leishman’s banding.

Chromosomal Microarray Analysis. Genomic DNA was extracted using High Pure polymerase chain reaction (PCR) Template Preparation kit following the manufacturer’s protocol (Roche Diagnostics, Indianapolis, IN, USA). Chromosomal microarray testing was performed in patients using NimbleGen 3 × 1.4 M Whole-Genome Tiling Array, which has 1,400,000 copy number variations (CNV) probes across the entire genome according to the manufacturer’s instructions (Roche NimbleGen, Madison, WI, USA).

The SurePrint G3+SNP (single nucleotide polymorphism) Human CGH Microarray (4 × 180 K) dual-color array containing more than 170,334 distinct biological probes with 13 kb genome-wide median probes spacing was carried out following the protocols provided by the manufacturer (Agilent Technologies Inc., Santa Clara, CA, USA) in parents to detect inheritance and for confirmation of detected CNVs that were larger than 1 Mb in index cases.

Aberration calls using Nexus Copy Number (Bio-Discovery, El Segundo, CA, USA) was implemented following thresholds for copy numbers; 100 kb gains, 50 kb losses across the genome with a minimum of 25 markers to make the call. For regions of known significance, the minimal setting was 25 kb and 25 markers for gains and losses. Detected variations were classified according to the American College of Medical Genetics and Genomics (ACMG) guidelines [8,9]. Only likely pathogenic or pathogenic variations were reported and uncertain CNVs were evaluated by taking into consideration of ACMG classification scores [8,9].

This study was reviewed and approved by the Ethics Committee of the Istanbul Medical Faculty [N. 08/13]. Written informed consent was obtained from all subjects or their legal guardians included in this study.

Postnatal Cases. Postnatal cases (n = 21) consisted of four inversions, eight CCRs and nine translocations. Familial inheritance was found in seven cases (33.3%) (three inversions, four translocations) and 14 cases were de novo in origin (66.6%) (one inversion, five translocations and eight CCRs). Clinical findings, initial karyotypes, and CMA results of all cases are shown in Tables 1 and 2.

No cryptic genomic imbalance by CMA was observed in familial cases (n = 7). Of 14 de novo cases, seven presented with at least one imbalance (50.0%). These CNVs were 12 deletions and one duplication, of which the size ranged from 1.1 Mb to 10.6 Mb. Five imbalances were at breakpoints or near to breakpoints of the rearrangements (38.4%), and eight were at different regions (61.5%).

Of five de novo translocations, only one had a 4.8 Mb duplication (20.0%) at the breakpoint of the translocation (case 1) (Table 1). Chromosomal microarray revealed a deletion near to the breakpoint of a de novo inversion (case 10) (Table 1).

All CCRs (n = 8) were de novo and CMA revealed 11 deletions in five of them (62.5%). Of 11 deletions, two were at breakpoints (cases 10 and 21), one was near to breakpoints (case 14) and eight were at different regions (cases 15, 16, 20, 21) (Table 1).

Prenatal Cases. Of 13 prenatal cases, nine were translocations (two familial and seven de novo in origin), three were inversions (all familial), and one was de novo CCR. An imbalance has been found in two of 13 cases (15.4%) (cases 23 and 26) (Table 2); both were de novo translocations. The imbalance rate for de novo cases was 25.0%. One of the CNVs was a deletion 1.2 Mb in size at the breakpoint of rearrangement. Another one was a duplication 5.5 Mb in size on the same chromosome, but at different bands. This CNV was inherited from a healthy mother.

Molecular karyotyping allowed us to detect genomewide chromosomal imbalances even in size of kbs using DNA copy number and/or SNP variation probes [10]. Recent reports have suggested that ABCRs in patients having abnormal phenotypes can be more complex at the molecular level than suspected by karyotyping. Therefore, ABCRs either de novo or familial, should be investigated at the molecular level, if the phenotype is affected [5,11, 12, 13, 14, 15].

Molecular karyotyping reveals the breakpoints of the rearrangements, especially for inversions and insertions, more precisely than the karyotyping seen in our cases 10, 14, and 21. As CMA is a genome-wide technique, it allowed us to detect the imbalances elsewhere in the genome such as in our cases 15, 16, 20, and 21. Eight CNVs at different locations, apart from the suspected breakpoints, indicates the complexity of the CCRs and effectiveness of CMA [16].

No molecular imbalances were observed in our familial ABCRs (five prenatal, seven postnatal). However, Gijsbers et al. [16] reported a deletion at the unrelated chromosome in one case of four familial translocation cases (25.0%). Sismani et al. [5] determined two CNVs at the translocation breakpoints in one out of the six familial translocations (16.6%). In another study from Schluth-Bolard et al. [14] containing 14 familial cases (seven translocations, five inversions, two CCRs), imbalance rates were 14.3% for translocations, 40.0% for inversions and 50.0% for CCR and one inversion case had two different CNVs, one at the breakpoint and the other one at an independent region. Tabet et al. [15] reported four familial translocation and three inversion cases, of which one had a deletion at the related region of the inversion (14.3%). Sezin et al. [17] reported four familial cases with no cryptic imbalance. To sum up, the CNV rate in familial cases was 14.9% in a total of 47 cases. Six of eight imbalanced cases were at the breakpoints (75.0%), whereas the remaining two were at unrelated locations (25.0%). These results show that the risk for imbalances is higher for inversions and CCRs than translocations, and familial ABCRs can also have imbalances at the breakpoints or somewhere in the genome, coincidentally or not. If aCGH reveals no imbalances in cases presenting with phenotypical findings, they should be further investigated for monogenic disorders [18]. In our postnatal series, six of the seven familial ABCR carriers had a consanguineous marriage, which enhances the possibility of monogenic disorders.

The rate of de novo imbalances detected by CMA was higher in postnatal than prenatal cases (50.0 vs. 25.0%) in this study, because postnatal cases were preselected due to their distinct abnormal phenotype. In our center, CMA was offered in the presence of pathological fetal ultrasound findings after normal fetal karyotypes, but also in cases with de novo ABCRs, even if ultrasound findings were normal. A prenatal case presented with polyhydramnios and abdominal cysts had a 1.2 Mb deletion containing GRID1, Mir_544, AK097624, LOC100507470, AX746544, 7SK genes at the 10q23.1 band. Van Bon et al. [19] reported 12 patients having a deletion on 10q22.3-q23.3. One patient had an atrioventricular septal defect (AVSD), dysmorphic features, diaphragmatic eventration and undescended testes, and CMA revealed a small interstitial GRID1 deletion comprising exons 5 to 8, 0.2 Mb in size. The GRID1 gene has been also reported to be associated with schizophrenia [20,21]. Because the ultrasound findings of our case was not compatible with the cases of Van Bon et al. [19], and deletions of this region have not yet been associated with a particular phenotype, this deletion was interpreted as variations of unknown significance (VUS). In another case, presenting with only choroid plexus cysts, a 5.5 Mb duplication was inherited from the healthy mother. Surprisingly, prenatal cases presenting with major structural anomalies (cases 31, 32, 33 and 34) had no imbalances by CMA, and further studies to search for monogenic disorders were planned. De Gregori et al. [11] also reported 14 de novo prenatal translocations including one presenting with abnormal USG findings, one had normal USG but delayed psychomotor development at 6 months of age and none of them had an imbalance by CMA. Evangelidou et al. [22] reported one familial and four de novo translocation cases with abnormal USG findings; all cases were also found to be normal by CMA. Due to the limited number of published cases, experiences with prenatal cases are not sufficient to estimate a risk figure for prenatal ABCRs with abnormal USG findings. When de novo ABCRs detected by fetal karyotyping, independent from the phenotypical findings, molecular karyotyping should be applied, because clinical findings are limited with only ultrasound examination.

Variations of unknown significance detected in prenatal CMA studies are confusing and genetic counseling is difficult. Therefore, to decrease the VUS possibility, the standards and guidelines of 2013 [9] recommends reporting the variations that are >400 kb for both deletions and duplications in prenatal and postnatal studies by whole genome array platforms. Different countries have their approach for prenatal cases, such as including genes number [>18 genes (Belgium), size of imbalances (>500 kb for deletions and >1.0 Mb for duplications (Canada)], choosing array platforms reducing densities of the probes [23,24, 25]. In this study, we initially used higher CMA resolution (1.4 Mb) for both postnatal and prenatal cases. Detected imbalances in prenatal cases were over 1.0 Mb, which could be confirmed by 180 K resolution. In some cases, a lower resolution can cause difficulties to search the whole genome or the size of the CNV may differ from the actual size due to the limited number of probes [26].

Parental array studies showed that the detected duplication at 16p11 in case 26 was inherited maternally, which, coincidentally, was on one of the ABCR-involved chromosomes. Imbalances (size range 579 kb to 4.6 Mb) including deletions and duplications at 16p11 showing incomplete penetrance/variable expressivity, can be associated with global development delay, behavioral problems, epilepsy, autism [27]. The critical region comprising breakpoints four and five (BP4-BP5) (600 kb, chr16; 29.6-30.2 mb-HG19) called 16p11.2 microdeletion/microduplication syndrome is reported in about three in 10,000 [28]. Most of the 16p11.2 microduplications (70.0%) are familial, and the clinical findings are variable from severe to mild [29]. The size of microduplication in our case was larger than the critical region, the healthy carrier mother decided to continue the pregnancy. Clinical evaluations of the newborn revealed normal results, clinical follow-up controls of the baby were planned. In general, the detection of duplications is troublesome, either cytogenetically or clinically due to the nonspecific and variable phenotypes. Based on these experiences, it was expected that the frequency of the duplications is higher by CMA than in karyotyping in patients with behavioral problems or autism without distinct dysmorphic features [30]. Therefore, genetic counseling for de novo duplications detected prenatally is still challenging.

When the data of published postnatal de novo series are combined, 35.9% of the apparently balanced de novo translocations were, in fact, unbalanced. This rate was 20.0% in our series (Table 3). Imbalance rate in inversions was as much as in translocations (37.0%). There was only one inversion patient in our series and she had a deletion. As expected, the highest submicroscopic imbalance rate (74.4%) was observed in CCRs. In our series, it was 62.5%. The difference (~12.0%) could be explained by the presence of two Xp21-autosome translocation cases in our series. Both cases manifested female carriers of Duchenne muscular dystrophy (DMD) had also been investigated by multiplex ligation-dependent probe amplification (MLPA) (MRC-Holland, Amsterdam, The Netherlands) and by next generation sequencing (NGS) techniques, and no mutations were detected. The causing factor was not imbalances at the molecular level, but the possibly skewed X inactivation, where the X chromosome carrying the normal dystrophin gene is preferentially inactivated to save the translocated autosomal segment on the derivative X chromosome [31]. If these two cases were excluded, the imbalance rate would be 83.3% in our series.

The imbalances at unrelated regions of the break-points/chromosomes observed in five CCR cases (cases 10, 15, 16, 20, 21) demonstrate the advantage of genome-wide array studies. Multiple breaks and consecutive micro-deletions in these cases show the complexity of the CCRs, and support that de novo CCRs occur due to the multiple breaks in the genome and increased genomic instability. The term ‘chromothripsis’ is used to describe ‘chromosome shattering,’ which means chromosomes are first fragmented into many pieces and then the fragments stick back together randomly due to the DNA repair processes [32]. Analysis of CCRs in patients with congenital disorders showed that chromothripsis is not applicable for all complex germline rearrangements. The term ‘chromosoanasynthesis,’ which means repeated chromosome synthesis was suggested by Liu et al. [33] for describing multiple template switch events that may occur during the germline CCRs formation process. Many more CCRs should be investigated by whole-genome analysis to increase the knowledge and understanding the underlying occurrence mechanism of the CCRs [34, 35, 36].

Altogether, the imbalance rate of de novo ABCRs is 24.6% at the breakpoints and 19.2% at different regions apart from the breakpoints. These rates can be used in genetic counseling related to de novo ABCRs.

In conclusion: 1) the current study supports previous studies, that the cryptic genomic imbalances are high (43.8%) in patients of de novo ABCRs with abnormal phenotype. 2) Patients presenting with multiple congenital anomalies, intellectual disability, and carrying de novo or familial ABCRs, whatever the type of rearrangement (translocation, inversion, or CCR), should be studied systematically by CMA. 3) Chromosomal microarray investigation is more important in cases with de novo ABCRs detected prenatally, as the results are essential for genetic counseling and decision-making process of the parents. Studies focused on prenatal ABCR carriers with abnormal ultrasound findings are limited and no submicroscopic imbalance was reported. Abnormality rate detected by CMA was 25.0% in our de novo prenatal cohort. 4) Chromosomal microarray promises to detect genome-wide imbalances at the kb level. However, further studies are needed for other mechanisms such as gene fusions or gene disruptions that might explain the phenotype of affected ABCR carriers. All unexplained patients should be examined for single gene disorders.