LAUSR

To the Editor: Simpson-Golabi-Behmel syndrome (SGBS) is an X-linked multiple congenital anomalies (MCA) and overgrowth syndrome caused by a lossof-function of the glypican-3 gene (GPC3). Almost, a hundred distinct GPC3mutations have been identified, mostly large deletions, nonsense or frameshift mutations. Only 6 intragenic GPC3 duplications were described in SGBS patients, 2 of them detected by array-CGH in patients with MCA. Although their transcriptional effects were assessed in only 2 of them, all these duplications clearly segregated with the SGBS phenotype. Here, we describe the characterization of 2 novel intragenic GPC3 tandem duplications identified by array-CGH during prenatal diagnosis. In both cases increased maternal plasmatic alpha-fetoprotein and abnormal ultrasound findings (polyhydramnios, macrosomia and pyelectasy for the first patient; macrosomia, ventricular septal defect and large hyperechogenic kidneys for the second patient) raised the diagnosis of SGBS. Informed consent for genetic analysis was obtained from the parents according to the French law on Bioethics and following the Helsinki Declaration. In patient 1 (Figure 1A), array-Comparative Genomic Hybridization (array-CGH) revealed a 52-kb Xq26.2 duplication encompassing the exon 7 of GPC3 (arr[hg19] Xq26.2(132 717 085-132 769 148) x2mat). Multiplex ligation-dependent probe analysis (MLPA) confirmed this hemizygous duplication and cDNA sequencing demonstrated a tandem duplication of exon 7. The open reading frame was disrupted by the insertion of 10 amino acids followed by a premature termination codon. In patient 2 (Figure 1A), array-CGH revealed a 825.3-kb Xq26.2 duplication encompassing 3 RefSeq genes (USP26, TFDP3, GPC4) and exons 3 to 8 of GPC3 (arr[hg19] Xq26.2(132 161 510-132 986 815) x2mat). MLPA analysis revealed that only exons 3 to 6 of GPC3 were duplicated while all exons of GPC4 were duplicated. This duplication was composed of 2 adjacent duplications, one involving USP26, TFDP3 and GPC4 while the second involved only the exons 3 to 6 of GPC3. cDNA sequencing conducted after Reverse Transcription-Polymerase Chain Reaction (RT-PCR) showed a tandem orientation of the duplicated exons predicted to cause the addition of 24 amino acids followed by a premature termination codon. In both cases, the diagnosis of SGBS was confirmed postnatally. We reviewed Xq26 duplications registered in public databases (DECIPHER, ISCA and dbVar) (Figure 1B). We excluded the large duplications encompassing the whole GPC3 gene and many other genes in which a loss-of-function of GPC3 is very unlikely. Twelve duplications, partially overlapping GPC3 could potentially disrupt its transcription. Knowing that most interstitial duplications are tandem rearrangements, the 4 duplications totally included within GPC3 (nssv584468, nssv13650346, DECIPHER 258050, DECIPHER 326611) and the 2 duplications with their 30 breakpoint within GPC3 (nssv1415234, nssv13644225) could lead to a GPC3 loss-of-function by disrupting the reading frame. However, features consistent with SGBS were documented in only 2 of them (DECIPHER 258050, nssv13644225) and no functional molecular analysis was performed. The 6 remaining duplications (DECIPHER 253888, nssv1610317, DECIPHER 277238, nssv3397178, nssvs579235 and DECIPHER 278212) either including the whole GPC3 gene or with only a 50 breakpoint in GPC3 probably maintain the production of a full length transcript even if we cannot rule out the presence of a microrearrangement of the gene or a dosage alteration due to a “position effect.” This is in agreement with the clinical data available in the databases which were not suggestive of SGBS. On the other hand, assuming that GPC3 belongs to the class of dosage-sensitive genes, whole GPC3 duplications could give rise to a mirrored SGBS phenotype. However, neither DECIPHER 253888 nor nssv1610317 were documented with clinical features mirroring those of SGBS. In conclusion, this report expands the molecular spectrum of GPC3 mutations in SGBS and illustrates the questions raised in interpreting GPC3 duplications. Array-CGH is useful to determine the exact boundaries of the duplication when one of them involves the 50 or the 30 end of the gene. Nevertheless, as illustrated by our second case, the resolution of array-CGH may be insufficient to detect more complex or intragenic micro-rearrangements within the duplication and a more precise quantitative analysis, such as MLPA, is necessary to determine exactly which genes are involved in the duplication. As microarray analysis is now widely used in both preand postnatal diagnosis, GPC3 duplications might be detected in patients with unexplained developmental problems and/or MCA. If cytogenetics tools such as fluorescence in situ hybridization do not locate the duplicated fragment elsewhere in the genome, or in case of small GPC3 Received: 19 September 2017 Revised: 4 October 2017 Accepted: 5 October 2017