The recurrent SETBP1 c.2608G > A, p.(Gly870Ser) variant in a patient with Schinzel-Giedion syndrome: an illustrative case of the utility of whole exome sequencing in a critically ill neonate

Schinzel-Giedion syndrome (SGS, OMIM 269150) is an ultrarare autosomal dominant disorder characterized by neurodevelopmental disorder, characteristic facial features with midface hypoplasia, cardiac, skeletal and genitourinary malformations, and an increased pediatric cancer risk [1]. The exact frequency of SGS is unknown and only ~ 50 affected individuals have been reported to date. The molecular basis of SGS was resolved in 2010, with the identification of de novo heterozygous variants in the SETBP1 gene (OMIM 611060) [2]. The SETBP1 gene is located on chromosome region 18q21.1 and encodes for an oncogene-binding protein which binds to SET domains that are involved in the methylation of lysine residues on histone tails. The multisystemic involvement of SGS is recapitulated by the ubiquitous expression of SETBP1, whose biological functions remain not yet fully elucidated [3].

To the best of our knowledge, nine different SETBP1 deleterious variants have been reported in 25 individuals, and all of these variants cluster in exon 4, within a hotspot of 12 base pairs corresponding to aminoacids 867 to 871 of the encoded protein [1, 4‐13]. The affected protein region is highly conserved and is a peptide sequence crucial for the degradation of the protein itself. The SGS-specific variants increase protein stability by interfering with ubiquitination and lead to the accumulation of the protein in the cell [14]. Interestingly, some of the SETBP1 variants which cause SGS when inherited, have been repeatedly registered as somatic events in several types of myeloid malignancies [15‐18].

In addition to point mutations, two additional patients with a proximal interstitial 18q microdeletion selectively involving SETBP1, presenting with milder phenotypes apparently distinct from the full-blown SGS [19] have been described. Finally, also a subject with the SETBP1 frameshift p.(Asn395Lysfs*2) variant and dysmorphic features, coordination deficits in fine motor skills and behavioral problems has been described. Neurological evaluation showed no hypotonia; no abnormal findings were observed on genital examination. Hearing and vision were normal. Urinary tract ultrasound and brain magnetic resonance imaging were normal [20]. These cases testify that SETBP1 null alleles cause a less severe phenotype and that the point variants more typical for SGS are likely to act with a dominant negative effect.

Here, we report the case of a critically ill neonate presenting to the Neonatal Intensive Care Unit with multiple malformations and subsequently diagnosed with SGS. The diagnosis was reached by performing the Exome Sequencing, which revealed the presence of the SETBP1 recurrent c.2608G > A, p.(Gly870Ser) variant. We also compared the phenotypic data of our patient with the phenotypic data reported so far in the literature, which were collected from other patients carrying the same variant in the SETBP1 gene. Finally, we discuss the utility of WES in establishing an etiological basis in clinically and genetically heterogeneous conditions, which are often found in affected patients hospitalized in Neonatal Intensive Care Units (NICU) and/or Pediatric Intensive Care Units (PICU).

The proband is a Caucasian 4-day-old neonate. A clinical genetics assessment for a multisystemic disease -which was not recognized antenatally- was carried out at the NICU. He was the third male child of a healthy non-consanguineous couple. At the time of his birth, his father and mother were 42 and 36 years old, respectively. Family history was unremarkable, especially for neurodevelopmental disorders, brain abnormalities, recurrent miscarriages, other birth defects and/or genetic illnesses. The child was born after an uneventful full-term gestation. Birth weight was 2.630 g (3rd centile), length 47 cm (4th centile), head circumference 32 cm (2nd centile), and APGAR scores 8/8 at 1′/5′, respectively.

Admission to the Neonatal Intensive Care Unit was requested for the combination of heart murmur, external genitalia anomalies and hypotonia, and respiratory distress. Physical examination on day 4 revealed triangular face, frontal bossing, clinically apparent hypertelorism, periorbital edema, depressed nasal bridge, bulbous nose, anteverted nares, low hanging columella, short philtrum, downturned corners of mouth, thin lips, hypoplastic chin, low set and posteriorly rotated dysmorphic ears, and low posterior hairline with apparent redundant neck skin (Fig. 1a). Ultrasound screening revealed atrial septal defect (ASD), hypoplastic corpus callosum and bilateral pyelectasis. Facial features worsened in the following weeks; in addition, the scrotum was severely enlarged, and penile hypospadias was noted (Fig. 1b). In addition, beside described clinical symptoms, the patient showed poor feeding, irritability, reduced activity and retching due to rapid-onset of hydrocephalus with increased Intracranial Pressure (ICP). For these reasons, he was transferred to another hospital for underwent a urgent surgical treatment consisting in the insertion of a drainage system (shunt), and therefore we lost the possibility of periodical follow up for the next 4 months. However, the parents continued to keep us informed. No significant abnormalities were detected when blood tests were performed, including complete blood cell count, blood gas analysis, conventional biochemical test, thyroid function test, amino acid analysis, and lysosomal enzyme activity analysis. The results of urinary organic acid and uronic acid analyses were also normal. In the clinical documentation sent by the parents, we found: RX examination with dorsal levoscoliosis, broad ribs and relatively long clavicles (Fig. 1c). Results of rapid standard karyotyping revealed a normal male (46,XY), while high resolution SNP-array analysis revealed a 0,6 Mb microduplication of uncertain clinical significance on the chromosome region 10q22.3, inherited from his healthy father (arr[hg19]10q22.3(81,362,685 × 2,81,385,244-81,986,097 × 3,81,990,582 × 2)pat). Even though the objective examinations were suggestive for Schinzel-Giedion Syndrome (Fig. 1d), we had to take into consideration all possible differential diagnoses and, for this reason, Whole Exome Sequencing Analysis was performed.

More than 15 million paired-end sequences were produced for this sample with a mean base quality score of 31 (i.e., probability of being an erroneous call less than 1/1000) and a 64X mean depth of coverage.

NGS analysis of the proband revealed that he was heterozygous for a de novo, previously described, mutation in SETBP1 (NM_015559) c.2608G > A, p.(Gly870Ser). The detected variant, already reported in dbSNP (rs267607040), was detected with a > 100X read depth (60 out of 159 reads supporting the alternate allele) and an elevated quality score. The identified candidate disease-causing variant was confirmed by direct Sanger Sequencing using specific primers (SETBP1, exon 4, Forward Primer: 5′- AGACTGATGGCCAACTCCC; Reverse Primer: 5′- GTTCTTTGTGCTGGTGTCGG) (Fig. 2).

In the literature, the frequency of the variant is reported to be very low in the ExAC global population (MAF: 0.00002) and absent in the European population. The variant affects a residue that is highly conserved in different species of vertebrates (Fig. 3) [30], and, according to the American College of Medical Genetics and Genomics guidelines [29], the variant was classified as pathogenetic. For the nucleotide variants nomenclature we used the format recommended by the Human Genome Variation Society (HGVS, http://​www.​hgvs.​org) recommendations and reported in the Leiden Open Variation Databases (LOVD) (https://​databases.​lovd.​nl/​shared/​individuals/​00289344).

Bioinformatic analysis excluded any other pathogenic or likely pathogenic heterozygous variants or biallelic variants in genes associated with compatible phenotypes. Given this result, the overall phenotype of the neonate was reviewed and, eventually, the molecular suspect of SGS was confirmed after a focused examination of the clinical findings.

Here, we report the case of a neonate suffering from SGS and carrier of the de novo recurrent variant c.2608 G > A, p (Gly870Ser) in SETBP1 (NM_015559). The SETBP1 gene encodes a protein which contains a SKI-homology domain, a SET-binding region which adds to three nuclear localization signals. Oakley and his collaborators [31] demonstrated that, by transducing murine bone marrow progenitors with high titer retrovirus expressing SETBP1, SETBP1 is able to bind to Hoxa9/10 promoters upregulating the two genes. In line with these findings, Piazza and his collaborators focused on the potential role of SETBP1 as a transcriptional modulator and showed that SETBP1 interacts with gDNA through its AT-hook domains by forming a multiprotein complex including HCF1, KMT2A, PHF8, and PHF6, which increases chromatin accessibility. Gene ontology analysis of deregulated SETBP1 target genes indicates that they are key controllers of visceral organ development and brain morphogenesis [32].

To the best of our knowledge, nine different point mutations have been reported in 26 SGS patients, including our patient, and all of them occurred within a hotspot of 12 base pairs in exon 4 of the SETBP1 gene (chromosome position: 42531904–42,531,918), corresponding to aminoacids 867 to 871 of the encoded protein. The nine variants to date identified are listed in Table 1 and depicted in Fig. 3. This recurrently mutated region, localized in the SKI homology region of the protein, is highly conserved in vertebrates (Fig. 4), suggesting that it might have an important biological role. It has been identified as a specific sequence that pilots the protein to the initial degradation step. It contains a consensus-binding region for β-TrCP1, the substrate recognition subunit of the E3 ubiquitin ligase, and might be critical for protein degradation through ubiquitin binding. When mutated, SETBP1 protein fails to bind to this E3 ligase subunit, and protein stability increases by interfering with ubiquitination, leading to the accumulation of SETBP1 protein in the cells [15, 32]. The disease causing the mutation identified in our patient had been previously reported and classified as likely pathogenetic in 6 of the 25 patients with molecularly confirmed SGS [2]. For this reason, from a molecular point of view, our report is useful to allow the definitive classification of this variant as pathogenetic. Interesting to note, although patients with a clinical diagnosis of SGS carrying a pathogenetic variant in SETBP1 are still few, the c.2608G > A, p.(Gly870Ser) variant is, together with the c.2612 T > C, p.I871T, the most frequent one with a prevalence of about 27% (7/26, Table 1).

We limited our literature review to all patients carrying the c.2608G > A, p.(Gly870Ser). The seven patients, including our patient, to date identified are listed in Table 2. The clinical comparison shows that the emerging phenotype remains comparable, despite a lack of clinical information in some reports (Table 2). In detail, the patients present large anterior fontanelle (6/7), prominent forehead/frontal bossing (7/7), midface hypoplasia/retraction (7/7), ocular hypertelorism (5/7), short and upturned nose (7/7) and low-set ear(s) (6/7). Other hallmarks of SGS syndrome, which are important clues for the diagnosis, are skeletal system malformations (7/7) in the form of synchondrosis, clinodactyly and camptodactyly, short first metacarpal bones, bowing of both tibiae, talipes and deformities of both feet, long bones and broad ribs, and urogenital and renal defects (6/7) in the form of hydronephrosis, micropenia and hypospadias. Neurological problems are also present, such as severe intellectual disability (7/7), seizures (7/7), hearing impairment (6/7), ventriculomegaly (6/7), and vision impairment (4/7). Other impairments observed since birth are characterized by feeding and respiratory problems, often associated with epileptic seizures (our patient and #4, Table 2), and neurodegenerative process caused by deficient white matter myelination and abnormal pattern of myelination (patient 2, Table 2), which is emerging as an important, but under-reported, central nervous system findings in patients with SGS [12, 33]. Accordingly, the overall clinical picture of the seven individuals carrying the c.2608 G > A, p (Gly870Ser) variant is homogeneous without significant differences from the remaining patients with SGS reported in medical literature. This clinical comparison allowed us to confirm that SGS is a severe disorder potentially presenting at birth (critically ill neonate). The molecular basis of SGS remains homogeneous with an apparently strict association between recognizable phenotype and mutation site(s).

More importantly, in our case, the molecular diagnosis of SGS was made before a definitive clinical diagnosis and it was possible by performing the Whole Exome Sequencing on samples obtained shortly after birth and rapid cytogenomic screening. The utility of fast exome/genome sequencing on critically ill neonates is startig to be included in the current clinical genomics applications [34]. Literature data are heterogeneous and, therefore, preliminary, but the diagnostic yield of such an approach is strongly influenced by the selection strategy of patients. For example, while a randomized study on unselected cases of patients referred to Neonatal Intensive Care Units had a positive rate of only 1 out of 29 [35], another study on 50 neonates selected with a phenotype-driven approach allowed a molecular diagnosis in 29 (54%) [36]. In our practice, we are still working on the appropriate, evidence-based approach of introducing genomic analysis in the routinary clinical activities of Neonatal Intensive Care Units for critically ill neonates. However, our experience shows that a phenotype-driven approach carried out by clinical geneticists working in a team with neonatologists and child neurologist should be preferred as it eases the post-analytical phase for a timely production of the written results. In our patient, although the facial gestalt at birth was evocative of a multiple malformation syndrome, the molecular results were crucial in establishing the diagnosis due to the presumably evolving nature of SGS. In our patient, the diagnosis of SGS allowed clinicians to (i) correctly focus their attention on borderline features, such as craniofacial anomalies which could have raised the suspicion of an underlying genetic neurological or metabolic disease, (ii) put the patient in a follow-up program for monitoring the syndrome-specific cancer risk, and (iii) provide the family with precise information for future family planning and prenatal diagnosis.

Our experience and emerging evidence from the literature suggest that Whole Exome/Genome diagnostic facilities should only be made accessible to the expert professionals in multidisciplinary teams.