Molecular mechanisms generating and stabilizing terminal 22q13 deletions in 44 subjects with Phelan/McDermid syndrome

Abstract

In this study, we used deletions at 22q13, which represent a substantial source of human pathology (Phelan/McDermid syndrome), as a model for investigating the molecular mechanisms of terminal deletions that are currently poorly understood. We characterized at the molecular level the genomic rearrangement in 44 unrelated patients with 22q13 monosomy resulting from simple terminal deletions (72%), ring chromosomes (14%), and unbalanced translocations (7%). We also discovered interstitial deletions between 17-74 kb in 9% of the patients. Haploinsufficiency of the SHANK3 gene, confirmed in all rearrangements, is very likely the cause of the major neurological features associated with PMS. SHANK3 mutations can also result in language and/or social interaction disabilities. We determined the breakpoint junctions in 29 cases, providing a realistic snapshot of the variety of mechanisms driving non-recurrent deletion and repair at chromosome ends. De novo telomere synthesis and telomere capture are used to repair terminal deletions; non-homologous end-joining or microhomology-mediated break-induced replication is probably involved in ring 22 formation and translocations; non-homologous end-joining and fork stalling and template switching prevail in cases with interstitial 22q13.3. For the first time, we also demonstrated that distinct stabilizing events of the same terminal deletion can occur in different early embryonic cells, proving that terminal deletions can be repaired by multistep healing events and supporting the recent hypothesis that rare pathogenic germline rearrangements may have mitotic origin. Finally, the progressive clinical deterioration observed throughout the longitudinal medical history of three subjects over forty years supports the hypothesis of a role for SHANK3 haploinsufficiency in neurological deterioration, in addition to its involvement in the neurobehavioral phenotype of PMS.

Figure 1. Schematic representation of the 22q13 rearrangements.

An ideogram of chromosome 22 is shown at the top with genomic coordinates of the boxed terminal region of interest shown at 1 Mb intervals. The location of the SHANK3 gene is marked in red. Each patient is represented by a horizontal line corresponding to the size of his deletion as determined by aCGH analysis. Each patient's code number is shown on the right side of the lines; asterisks (*) indicate previously published cases. Double asterisks (**) indicate mosaic deletions. The lines' colors correspond to 22q13 rearrangement categories: simple deletions are depicted in black, derivative chromosomes 22 in green, rings 22 in pink, and interstitial deletions in brown. Forty-four patients are represented; the breakpoint interval (represented in grey) in subject P35 was narrowed down to ∼400 kb by FISH analysis with BAC clones RP11-194L8 (chr22:44,951,438–45,122,714, still present) and RP11-266G21 (chr22:45,543,178–45,711,912, deleted).

Figure 2. Molecular characterization of the 22q13.2 terminal deletion in subject P20.

A, Magnified view of the aligned breakpoint boundaries detected by array-CGH analysis using an oligonucleotide-based custom 22q13 microarray (top) and a 180k Agilent kit (bottom); the deleted regions are shaded in blue. Arrowheads delimit two mosaic-deleted regions: the BP1–BP2 deletion region (from 42406240 to 42603381 bp) has an average log ratio of −0.3; the BP2–BP3 deletion region (from 42603381 to 42726895 bp) has an average log ratio of −0.5; the deleted region between BP3 and the telomere (from 42726895 to the end of chromosome 22) has an average log ratio of −0.8. The aligned UCSC map (hg18) is depicted at the bottom. The red bar indicates the map position of the RP11-141N8 BAC clone we used to confirm by FISH the mosaicism of the BP1–BP2 region. All genes (blue bars) mapping within the BP1–BP3 regions are shown. B, FISH analysis using the RP11-141N8 clone confirms a mosaic deletion of the BP1–BP2 region revealing: (top) the presence of hybridization signals (green signal) on only one chromosome 22 (arrowhead) in 30% of the metaphases analyzed; (bottom) the presence of hybridization signals (green signals) on both chromosome 22 homologues in the remaining 70% of the metaphases analyzed (bottom). C, Tel-ACP amplification and direct sequencing of the amplified fragments revealed the breakpoint junctions at BP1, BP2 and BP3. A telomere repeat is present at all three breakpoints.

Figure 3. 22q13.3 interstitial microdeletion detected by array-CGH analysis.

A, aligned aCGH profile (P37–38, P43–44: 180k Agilent kit; P42: 244k Agilent kit) details of all interstitial deletions; the deleted regions are shaded. B, map of the distal 22q13.3 region; the deletions are represented by black bars; the region overlapping the SHANK3 gene is shaded in light blue. All genes mapping in the region are shown. C, sequence alignment of the breakpoint junctions of subject P42 showing the homology with three genomic regions. The proximal breakpoint sequence is shown in red, the middle 24 bases in inverted orientation are blue, the distal breakpoint sequence in green; microhomologies between sequences at the breakpoints are depicted in bold. D, cartoon showing the respective position and orientation of the breakpoint sequences in P42 as arrows, colored as in C.