Long-read genome sequencing resolves a de novo complex 18q12.1q21.2 triplication causing partial tetrasomy and reveals its underlying mechanism

Clinical description and post-mortem examination findings

A 29-year-old woman (first pregnancy) sought prenatal diagnosis due to increased fetal nuchal translucency (3.9 mm) at echographic examination along with a reversed ventricular septal defect and tricuspid valve insufficiency. The combined test (PAPP-A and free beta-HCG) resulted in a high risk for trisomy 21 (1:12). Therefore, a chorionic villus sampling was performed at 12 weeks of gestational age. Following the conventional and molecular karyotype results, the pregnancy was terminated voluntarily at 16 weeks and 4 days of gestation and the autopsy was performed. The measurements and the morphological appearance of the organs were consistent with those of a female fetus at gestational age of 15–17 weeks. External inspection revealed mild facial dysmorphisms, characterized by hypertelorism and low-set ears. The most significant finding was a left-sided Congenital Diaphragmatic Hernia (CDH), with intrathoracic herniation of the liver, spleen, and stomach. Examination of the brain revealed features suggestive of neuronal heterotopia involving the parietal lobe of the right hemisphere, and possible focal defects of corticogenesis.

Conventional and molecular karyotype

Cytogenetic analysis performed on both cytotrophoblasts and cultured mesenchymal cells showed a female karyotype with an abnormal chromosome 18, defined as a duplication of region q12.2q22. The duplication was observed in 11 metaphases from cytotrophoblast, whereas 3 cells showed a normal karyotype (46,XX,?dup(18)(q12.2q22)[11]/46,XX[3]). Given the extensive experience of our center and the rigorous cleaning protocol employed, MCC is considered statistically negligible (0.01%) in our direct preparations. Therefore, the 3/14 normal cells observed in the direct preparation most likely represent low-level confined placental mosaicism (CPM) or a monochorionic vanishing twin. Notably, short tandem repeat (STR) analysis on genomic DNA (gDNA) from cultured mesenchymal cells, where maternal cells would have had a greater opportunity to proliferate, conclusively excluded MCC. Interestingly, both conventional and molecular karyotypes performed on cultured mesenchymal cells failed to identify the normal cell line. Array-CGH analysis further validated the existence of the triplicated region:

arr[GRCh38] 18q12.1q21.2(34509207_52782563)x4dn,18q21.2q23(54447020_80221686)x2hmz. Since one chromosome 18 had a normal morphology and the other one was elongated, it is evident that three doses are located on the same chromosome 18 (Fig. 1A, B). Surprisingly, a 26-Mb region, extending from the distal breakpoint of the triplicated region to the 18q telomere, exhibited a loss of heterozygosity (LOH) (Fig. 1C).

Fig. 1Fig. 1

Cytogenetics and molecular karyotype characterization. A: QFQ banded foetal chromosomes 18; the arrow shows the abnormal one. B-C: CGH+SNPs-array results: blue area shows the tetrasomic region in 18q12.1q21.2 (from nt 34,509,207 to nt 52,782,563); aqua green area shows loss of heterozygosis (LOH) in 18q21.2q23 (from nt 54,447,020 to nt 80,221,686)

Molecular characterization of tetrasomy 18q12.1q21.2 by LRS

Genomic DNA (gDNA) purified from cultured mesenchymal cells, previously analyzed by array CGH, was subsequently analyzed by LRS to delineate the precise BP junctions and the orientation of the triplicated 18q12.1q21.2 region.

LRS confirmed the previously identified triplication of the 18q12.1q21.2 region. The results showed that the chromosome 18 CNV ranges approximately from 34,460,000 bp to 52,859,000 bp (GRCh38/hg38), for a total size of 18.4 megabases (Table S1). Moreover, two inversions were detected at the BP junctions’ regions (Table S1). Analysis of the alignment between the reads supporting the inversions and the reference genome revealed that the start and end points of both inversions turned out to be the BPs of the two junction fragments between the middle repeat of the 18 Mb CNV and the two flanking repeats (Fig. 2).

Fig. 2Fig. 2

Identification of the junction fragments of 18q12.1q21.2 triplicated region by LRS. The Integrative Genomics Viewer (IGV) visualization shows reads alignments around the 18q inversions’ BPs detected by LRS. The chimeric reads spanning BPs were highlighted by the presence of a misaligned colored end. A The read alignment revealed the proximal BP1 and BP2 of the junction fragment, mapping in the 18q21.2 cytogenetic band, between the middle and the last repeats of the CNV. B The read alignment identified the distal BP3 and BP4 of the junction fragment, between the first and the second CNV repeats in the 18q12.1

Specifically, the BPs of the first inversion were defined by chimeric reads mapping from position chr18:34,460,465 (on the forward or reverse DNA strand), to chr18:34,494,769 on the opposite DNA strand (Fig. 2A). These genomic coordinates represent the proximal BP1 (chr18:34,460,465) and BP2 (chr18:34,494,769), which are located at the junction between the second and third copies of the CNV. The BPs of the second 18q inversion were identified by chimeric reads spanning up to chr18:52,858,932, on one strand, and up to chr18:52,862,079, on the opposite strand (Fig. 2B). Their localization revealed the distal BP3 and BP4 involved in the formation of the junction fragment, between the first and second copy of the CNV. The double orientation of the chimeric reads, mapping on the junction fragments of the triplicated 18q12.1q21.2 region, suggested that the chromosome 18 carrying the CNV has three tandem repeats of 18 Mb, with the middle repeat inverted.

To validate the LRS results, the 18q CNV junction fragments were amplified by PCR assays and sequenced by Sanger Sequencing. The proximal breakpoints, BP1 and BP2, mapped within a 2-bp microhomology sequence (chr18:34,460,465-34,460,466 and chr18:34,494,753-34,494,754, on opposite DNA strands) (Fig. S1A). Sanger Sequencing confirmed BP1 as identified by LRS, whereas BP2 showed a 16-bp shift compared to the LRS-defined breakpoint, at chr18:34,494,753. Visualization of the BP2 region via Integrative Genomics Viewer (IGV) revealed that while the majority of reads indicated a start position at chr18:34,494,769, a smaller subset aligned with the coordinates obtained by Sanger sequencing (Fig. 2A). BP1 is located in an intergenic region between NOL4 and DTNA, whereas BP2 is situated within intron 1 of the DTNA gene.

The distal BP3 and BP4 shared a 3-bp microhomology sequence (chr18:52,858,930-52,858,932 and chr18:52,866,374-52,866,376, on opposite strands) (Fig. S1B). While the BP3 position identified by LRS was confirmed, Sanger Sequencing revealed a different position for BP4 compared to the one initially identified by LRS. Specifically, the same sequence identified by LRS was found 80 bp downstream of BP4 (chr18:52,866,294-52,866,373), indicating the presence of a 4.2-kb deletion. Genomic annotation showed that this deletion, along with BP3 and BP4 is located within intron 2 of the DCC gene. This intron contains LTR49-int and L1PA7 repetitive elements, as well as several deletions reported in the Database of Genomic Variants (Fig. S2).

Based on the results, we hypothesized that the sequences between BP1 and BP2 and between BP3 and BP4 would be present in three copies, consistent with their lower LRS coverage compared to that detected for the region mapping from BP2 and BP3, which is present in four copies (Fig. S3).

Potential molecular mechanism generating the 18q12.1q21.2 tetrasomy and the 18q21.2q23 LOH

The generation of the 18q12.1q21.2 triplicated segment requires a three-chromatid exchange, mediated by two events. LRS and Sanger Sequencing allowed us to identify the sequences involved in the two events, which showed 50% homology if aligned in an inverted orientation (Fig. S4), suggesting the potential underlying mechanisms.

Our data supported the hypothesis that the fusion of two sister chromatids occurred through a U-type exchange mechanism, likely mediated by microhomology-based mechanisms such as replication fork stalling and template switching and microhomology-mediated break-induced replication (FoSTeS/MMBIR), or microhomology-mediated end-joining (MMEJ), during S-phase. This event, involving the distal BP3 and BP4 sequences, led to the formation of a transient dicentric chromosome and the loss of an acentric derivative (Fig. 3). Following the U-type exchange event, the dicentric chromosome had a mitotic break at either BP1 or BP2 during anaphase, producing two unequal chromatids: one with a significant terminal deletion, which is probably unstable, and another with the duplicated inverted 18q12.1q21.2 segment (Fig. 3). The chromatid containing the duplication was sustained by subsequent telomere capture, resulting in a chromosome 18 with a triplicated 18q12.1q21.2 region organized in a direct–inverted–direct orientation (Fig. 3). Obviously, in addition to the rearranged chromosome, a normal chromosome 18 was also expected.

Fig. 3Fig. 3

Schematic representation of the proposed mechanism explaining the 18q12.1q21.2 tetrasomy identified in the fetus. Regions including the BPs involved in a three-chromatid exchange mechanism are represented by colored rectangles. A transitory dicentric chromosome 18 and an acentric derivative, which is assumed to be lost, may result from a U-type exchange between sister chromatids (1), involving the distal BP3 and BP4. The dicentric chromosome may have broken at BP1 or BP2 (2) during mitosis, producing a derivative chromosome with a terminal deletion, del(18)(q12.1–qter), which is most likely lost, and a rearranged chromatid bearing the duplicated inverted 18q12.1q21.2 segment. A subsequent telomere capture event (3) would stabilize the broken chromatid with the duplication, resulting in a chromosome harboring three copies of the 18q12.1q21.2 region in a direct-inverted-direct orientation, as indicated by arrows. Specifically, the rearranged chromosome has three copies of the BP2–BP3 interval and two copies of the BP1–BP2 and BP3–BP4 intervals, respectively. However, as the four alternative final conformations show, it is impossible to identify which of the three repeated segments has lost one copy of the BP1–BP2 and BP3–BP4 intervals. A normal chromosome 18 is also expected, in addition to the rearranged chromosome

The modified chromosome exhibited three copies of the ~ 18 Mb region contained between BP2 and BP3 and two copies of the shorter portions BP1–BP2 and BP3–BP4, although it is impossible to determine which of the three repeated segments lost these sequences. In fact, four different conformations for the rearranged chromosome are consistent with the suggested mechanism and the sequencing findings of the junction fragments (Fig. 3).

Lastly, in line with the telomere capture event, both array-CGH (Fig. 1) and LRS (Fig. S5) results revealed a ~ 26 Mb LOH region, distal to the triplicated region, up to the telomere of chromosomes 18.

Study of 18q12.1q21.2 gene content

Using an in-silico approach, the potential contribution of the genes included in the 18q12.1q21.2 triplicated region to the phenotype of the fetus was examined. The pTriplo score that predicts the probability of triplosensitivity for the protein-coding genes was chosen as predictor of potential intolerance to increased DNA dosage of the 18q12.1q21.2 genes. In addition, possible genotype-phenotype correlations were evaluated using the Online Mendelian Inheritance in Man (OMIM) database. The 18q12.1q21.2 triplicated region involved 58 protein-coding genes. Among them, nine genes (MAPRE2, CELF4, SETBP1, SMAD2, ACAA2, MBD1, CXXC1, SMAD4, MEX3C) showed a pTriplo score ranging from 0.9 to 1, the maximum value associated with the increased dosage intolerance (Fig. S6).

Comments (0)

No login
gif