Genome-wide sequencing has identified de novo truncating mutations in ASXL3 in four patients with intellectual disability, feeding problems and distinctive facial features. Their presentation resembles that of Bohring-Opitz syndrome, which is associated with de novo nonsense mutations in ASXL1. This newly defined phenotype provides an important clinical resource for comparison with future cases in which mutations are found in ASXL3. The phenotypes for patients with mutations in each gene will undoubtedly be further delineated as more patients are reported.
Genome-wide sequencing strikes again
Dysmorphologists and clinical geneticists have worked over the past 40 years to define clinical syndromes by their similar facial features, physical attributes and clinical course. Cytogenomic and molecular testing have helped to define the underlying genetic etiology for many of these syndromes and clarified why related genes in the same pathway sometimes lead to similar clinical phenotypes. Now, with the advent of genome-wide sequencing, dysmorphologists have a powerful new tool at their disposal. Through sequencing, clinicians and researchers are able to study a cohort of patients with mutations in the same gene and to characterize further an identifiable syndrome. Once published, other similar cases can be tested, and gene variants discovered through whole-genome or exome sequencing can be compared with the previously published phenotype to define the relevant clinical features further.
This exciting story is repeated with the new report by Bainbridge et al. in Genome Medicine , which describes the discovery of de novo truncating mutations in ASXL3 in four patients with similar clinical presentations. The first two patients were identified through exome sequencing of the proband and both parents. The subsequent two cases were found through sequencing of the affected child only. A custom exon-capture strategy was used, followed by high-throughput sequencing with the Illumina HiSeq platform. All coding and near-intronic differences between the patients and the human reference genome were evaluated. The unaffected family members in the first two cases were used as controls to localize only de novo changes in the probands. De novo mutations were defined as those present in a patient but absent from all other members of their family, flagged as 'high quality', and supported by at least five reads with an allele percentage of 0.4 to 0.6%. The four de novo mutations that were found coded for premature stop codons and probably create a truncated ASXL3 protein. The mutations in each case were previously unreported and included insertions and deletions, which likely represent loss-of-function alleles, given that the mRNAs containing these stop codons are degraded by nonsense-mediated decay .
After identifying mutations in ASXL3 in these four patients, Bainbridge et al. used several large datasets to look for previously reported ASXL3 mutations. They found four other truncating mutations in ASXL3 in phenotypically normal patients. Three of these mutations were located in the 3′ end of ASXL3 , bringing into question whether these ASXL3 mutations are benign variants. The authors developed a model to determine the probability of observing multiple de novo mutations in ASXL3 in their four phenotypically similar patients. The results of their analysis show that this probability is approximately 4.0 × 10-17. This makes it highly unlikely that these changes occurred by chance, but functional studies are still required to confirm the pathological nature of these ASXL3 mutations .
The additional sex combs (Asx) gene has long been known to maintain homeotic gene activation and silencing in Drosophila, and three orthologs (Asxl1, Asxl2, Asxl3) are recognized in mice. ASXL1 is known to act on homeobox (Hox) genes as a repressor as well as an enhancer, and is important for development. Not much is known about the developmental function of the ASXL gene family in humans, but in Drosophila, Asx regulation is highly variable and tightly controlled directly after fertilization . In humans, ASXL3, like ASXL1 and ASXL2, codes for a putative polycomb protein that probably forms a complex with other proteins and acts as a histone methyltransferase .
De novo ASXL1 mutations have been identified in 9 out of 15 published cases of Bohring-Opitz syndrome (BOS) in which they were tested [4,5]. Somatic ASXL1 mutations have also been linked to myeloid malignancies and myelodisplastic syndromes . Following the example of other gene networks, such as the RAS signaling pathway genes and the RTK-P13K-AKT signaling pathway genes, it is plausible that somatic ASXL1 mutations could cause cancer, while germline mutations lead to developmental disorders [4,7,8]. To date, there have been two published cases of a neoplastic condition in BOS: a child with a medulloblastoma at 5 years old  and bilateral nephroblastomatosis found at autopsy in a 5-month-old infant . Both patients were negative for mutations in ASXL1 and ASXL3 [1,4]. It remains to be seen if patients with ASXL1 or ASXL3 mutations require tumor surveillance.
Patients with ASXL3 mutations versus ASXL1 mutations
Clinically, the four patients identified by Bainbridge et al. with de novo ASXL3 mutations presented with several features similar to BOS (Table 1). BOS is associated with high infant mortality, intrauterine growth retardation (89%), feeding difficulties often requiring G-tubes and leading to failure to thrive (100%), severe to profound developmental delays (100%), recurrent infections (62%), nonspecific brain abnormalities (78%), microcephaly (100%), wide prominent forehead (89%), deep palmar creases (57%) and a characteristic posturing known as BOS posture (100%) (percentages are based on the nine published ASXL1-mutation-positive cases). BOS posture is defined as flexion at the elbows with ulnar deviation, and flexion of the wrists and metacarpophalangeal joints. BOS patients also have distinctive facial features (Table 1) that include nevus flammeus typically over the glabella and eye lids (89%), facial hypotonia, prominent globes (100%) with high myopia (87%), hypertelorism (55%), upslanting palpebral fissures (67%), depressed nasal bridge (50%), anteverted nares (50%), broad alveolar ridges or high narrow palate (87%), micro/retrognathia (89%), low-set posteriorly rotated ears (67%), a low posterior hairline (67%) and hypertrichosis (89%) [4,5].
Table 1. Comparison of clinical features among patients with ASXL1 and ASXL3 mutations
Although all four subjects discussed by Bainbridge et al. shared clinical characteristics with each other and with BOS, these features are relatively nonspecific. The patients with an ASXL3 mutation had intrauterine growth retardation (3/4), severe feeding difficulties present from birth and requiring intervention (3/4), severe developmental delays (4/4), microcephaly (3/4), deep palmar creases (4/4), slight ulnar deviation of the hands (3/4) and a high-arched palate (3/4). These patients with ASXL3 mutations did not display many of the key features of BOS, including BOS posturing, myopia, nevus flammeus and a distinctive constellation of facial features. As with BOS, this novel syndrome is likely to be clinically heterogeneous. Not enough cases have been identified yet to define completely the clinical features or the facial appearance of this syndrome. Over time, as more cases are reported, this may become more evident.
Where do we go from here?
This intriguing new finding by Bainbridge et al., using genome-wide sequencing to identify a novel syndrome caused by de novo mutations in ASXL3, has added significantly to our knowledge of conditions associated with the ASXL gene family. As recently as 201l, the first association of mutations in the ASXL gene family was reported (ASXL1 causing BOS), and previous knowledge about the association of these genes with cancers in adults and with early development in mice and Drosophila may shed light on their role in developmental conditions. Because developmental disorders caused by mutations in the ASXL gene family in humans are still quite rare and are associated with high infant mortality, it is difficult to predict cancer risks. To delineate the syndromes caused by ASXL mutations, it would be reasonable to test other patients with phenotypic similarities to BOS for ASXL gene family mutations. Patients described by Hoischen et al. in 2011 who were negative for ASXL1 mutations were also found to be negative for ASXL3 mutations [1,4]. Undoubtedly, the publication of this article will lead to further reporting of patients with mutations in the ASXL gene family, and the clinical features of this ASXL3 mutation syndrome will become further defined. In the meantime, whole-genome sequencing will continue to provide insight into novel mutations causing previously undefined clinical conditions.
List of abbreviations used
BOS: Bohring-Opitz syndrome.
The authors declare that they have no competing interests.
Bainbridge MN, Hao H, Muzny DM, Musante L, Lupski JR, Graham BH, Chen W, Gripp KW, Wienker TF, Yang Y, Sutton VR, Gibbs RA, Ropers HH: De novo truncating mutations in ASXL3 are associated with a novel clinical phenotype with similarities to Bohring-Opitz syndrome.
Sinclair DA, Milne TA, Hodgson JW, Shellard J, Salinas CA, Kyba M, Randazzo F, Brock HW: The Additional sex combs gene of Drosophila encodes a chromatin protein that binds to shared and unique Polycomb group sites on polytene chromosomes.
Hoischen A, van Bon BWM, Rodriguez-Santiago B, Gilissen C, Vissers LE, de Vries P, Janssen I, van Lier B, Hastings R, Smithson SF, Newbury-Ecob R, Kjaergaard S, Goodship J, McGowan R, Bartholdi D, Rauch A, Peippo M, Cobben JM, Wieczorek D, Gillessen-Kaesbach G, Veltman JA, Brunner HG, de Vries BB: De novo nonsense mutations in ASXL1 cause Bohring-Opitz syndrome.
Boultwood J, Perry J, Pellagatti A, Fernandez-Mercado M, Fernandez-Santamaria C, Calasanz MJ, Larrayoz MJ, Garcia-Delgado M, Giagounidis A, Malcovati L, Della Porta MG, Jädersten M, Killick S, Hellström-Lindberg E, Cazzola M, Wainscoat JS: Frequent mutation of the polycomb-associated gene ASXL1 in the myelodysplastic syndromes and in acute myeloid leukemia.
Riviere J-B, Mirzaa GM, O'Roak BJ, Beddaoui M, Alcanrtara D, Conway RJ, St-Onge J, Worthylake T, Sullivan CT, Ward TR, Butler HE, Kramer NA, Albrecht B, Armour CM, Armstrong L, Caluferiu O, Cytrynbaum C, Drolet BA, Innes AM, Lauzon JL, Lin AE, Mancini GMS, Meschino WS, Reggin JD, Saggar AK, Lerman-Sagie T, Uyanik G, et al.: De novo germline and postzygotic mutations in AKT3, PIK3R2 and PIK3CA cause a spectrum of related megalencephaly syndromes.