Background
Patients presenting with developmental delay and multiple dysmorphic features are a common diagnostic challenge in the genetics clinic. Over the past decade, many new genetic syndromes have been identified within this area. A significant number of these have been linked to genes involved in histone modification and chromatin remodeling. These include: Kabuki syndrome types 1 and 2 [MIM:147920 and 300867] [
1,
2], Kleefstra syndrome [MIM: 610253] [
3],
KAT6B-related disorders [MIM: 606170 and 603736] [
4], Weaver syndrome [MIM: 277590] [
5],
HDAC8-related disorders [MIM:30882 and 309585] [
6‐
8], and Wiedemann-Steiner syndrome [MIM: 605130] [
9]. These along with Rubenstein-Taybi [MIM: 180849] [
10] and Sotos Syndrome [MIM 117550] [
11] make up a broad range of conditions cause by defects in chromatin remodeling genes. Similar to the loss of epigenetic control seen in Rett Syndrome [MIM: 312750], these disorders are thought to result from global changes in gene expression throughout development leading to abnormalities in multiple body systems. The majority of individuals with these disorders have impaired brain development leading to developmental delay and/or intellectual disability.
As these chromatin remodeling defect disorders are rare, with some having only a small number of cases reported, the complete phenotypic spectrum of many of them has not been well described. Thus while careful phenotyping remains critical for clinical diagnosis, it will often be insufficient to distinguish between related disorders. Genome-wide clinical tests such as SNP-based chromosomal microarray testing (SNP-CMA), clinical exome sequencing (CES), and clinical genome sequencing are incredibly powerful tools at identifying disease-causing variants in these genes: SNP CMA detection rate for ID ranges between 10-24% [
12], while the diagnostic yield of exome sequencing, in patients with normal CMA results, ranges 10-40% [
13].
Also of note is the strong pattern of
de novo variants observed in many chromatin remodeling disorders [
1,
2,
4,
9,
14]. Complete parent-proband trio sequencing is warranted in cases with developmental delay and dysmorphic features, as it has the power to directly identify
de novo variants. In addition to expediting the process of identifying
de novo variants in the known chromatin remodeling genes, there are many histone modification genes which have not been associated with human disease [
15]. With complete trio clinical exome sequencing, it is possible to identify candidate novel disease gene associations using clinical information and predictive molecular tools.
The two patients presented in this report were seen at different medical institutions and by separate medical teams. Based on the reported clinical findings, there was no
a priori expectation from within the clinical laboratory that these two individuals were connected in any way. Clinical exome sequencing was performed on full trios in both cases using clinically validated protocols (Additional file
1), detecting unique
de novo likely pathogenic variants in the
KMT2A (
MLL) gene in each patient.
Fusions between the
KMT2A gene with a variety of other genes are commonly observed in leukemic cells [
16,
17], giving the gene its original name: “myeloid/lymphoid or mixed lineage leukemia gene” or
MLL. KMT2A is widely expressed, detectable in most human tissues [
18]. It contains 36 exons and has three known mRNA isoforms (NM_001197104.1, NM_005933.3, and NM_024891.2). It is a homologue of the
d. melanogaster gene trithorax. Mice heterozygous for a knockout mutation of the homologous
Ktm2a gene exhibit retarded growth, skeletal and hematopoietic abnormalities [
19,
20]. The
KMT2A gene product KMT2A contains several functional domains. One domain is a SET domain which acts as a histone H3 lysine 4-specific methyltransferase, thus regulating a variety of developmental genes including those in the HOX family [
21].
Wiedemann-Steiner Syndrome has been described as a clinical entity defined by the presence of hypertrichosis cubiti (hairy elbows) and variable presentation of additional features such as facial dysmorphism, short stature, intellectual disability, and developmental delay [
22‐
25]. In an exome sequencing study of WSS,
de novo DNA variants in the
KMT2A gene were identified in five out of six patients, strongly implicating this gene as the major disease gene for WSS [
9].
Conclusions
A subset of WSS is caused by heterozygous
de novo variants in the
KMT2A (
MLL) gene [
9]. This subset is characterized by mild to moderate developmental delay, dysmorphic facial features (including: long eyelashes, thick or arched eyebrows, downslanting palperbral fissures, broad nasal bridge, and Cupid’s bow abnormality of the upper lip), and hypertrichosis cubiti (excessive hair on the elbows). A “slim and muscular build” was noted in 3/5 initial
KMT2A-related WSS cases
. Other features observed in some WSS patients include high narrow palate, tapering fingers, 5
th finger clinodactyly and hypotonia. The clinical spectrum of features associated with WSS is wide and may continue to expand as additional patients such as these are identified.
Exome sequencing results for these trios are suggestive of a molecular diagnosis of Wiedemann-Steiner Syndrome (WSS) in both patients. Our patients shares several of the features of
KMT2A-associated WSS, including postnatal growth retardation, developmental delay, wide nasal bridge, broad/bulbous nasal tip, and downslanted palpebral fissures [
9,
30]. They do not however have clear hypertrichosis cubiti, the clinical feature most readily associated - but not pathognomonic - with WSS. In them the excess of hair is manifested by thick eyebrows and hair, long thick eyelashes, and the sacral hypertrichosis observed in patient 2 (Table
1). In one of our patients and one previously reported individual with WSS [
9], there is history of recurrent infections, though it remains unclear whether their immune dysfunction is related to
KMT2A mutation. Patient 2 has several clinical features not previously observed in individuals with WSS, including: unilateral microphthalmia, micrognathia, 3–4 finger syndactyly, and premature eruption of adult teeth.
The
de novo variant identified in the
KMT2A gene in patient 1 is a missense c.4342T>C variant. To date, all
KMT2A variants reported in WSS patients are premature truncation variants, suggesting haploinsufficiency as the disease mechanism. As the c.4342T>C variant does not result in protein termination, the effect of this variant on KMT2A protein abundance and/or activity cannot be confidently predicted. However, this missense variant is located within a PDH homeodomain zinc finger domain, a domain thought to coordinate protein-protein interactions involved in transcriptional activation [
31]. The web-based tool Human Splicing Finder v2.4.1 [
32] was unable to provide a meaningful prediction as to whether this variant impacts splicing.
Given that typical human exomes carry between zero and five high confidence
de novo coding variants [
9,
13,
14,
33‐
39] and the inclusive approach to generating the primary gene list (over 1,000 genes included in each case), the identification of a previously unreported
de novo missense variant in the
KMT2A gene in a single case is not by itself a significant finding. However, combined with the phenotypic overlap between individuals with
de novo variants in
KMT2A with WSS and these two unrelated patients, these findings strongly implicate a causal relationship between the observed variants and the clinical presentation of these individuals. Functional analysis or identification of other patients with the same variants and similar phenotypes would provide additional support.
This report highlights the value of full trio clinical exome sequencing for individuals with multiple congenital anomalies and developmental delay whose features are not consistent with one particular syndrome, supporting the model of medical genetics practice recently suggested by Shashi and colleagues [
39]. Without parental sequences, the variants in
KMT2A would not have been singled out from among many similar heterozygous candidate variants identified within the primary gene list. Thus full trio exome sequencing greatly improved the interpretability of the test in these patients.
Financial considerations are also an important factor in molecular testing. Full trio clinical exome sequencing is comparable in cost to gene panel testing [
40] and, if pursued as a second-line test after clinical microarray analysis (SNP-CMA), is likely a far more efficient use of resources than iterative single gene testing in cases with developmental delay and dysmorphic features.
Consent
For both patients, a parent or legal guardian consented to the following statement: “We [the UCLA Clinical Genomics Center] will use your results to improve Clinical Exome Sequencing by comparing your data to others”. Additional written consent was acquired for both patients for the use of their photographs for research publication.
Ethics statement
As the genetic testing data were obtaining using a clinical test and appropriate written consent for testing was obtained, this report is exempt from ethics approval for medical research of human subjects. All authors have received training and are compliant with the Health Information Portability and Accountability Act of 1996 (HIPAA).
Acknowledgements
We thank the patients and their families for their essential contributions to this work. Technical assistance was provided by Nora Warschaw, Traci Toy, Robert Chin, Thien Huynh and Jean Reiss at the UCLA Molecular Diagnostics Laboratories and all members of the UCLA Clinical Genomics Center. Variant Annotator X (VAX) software was used with the permission and guidance of its author, Michael Yourshaw and computational assistance was provided by Bret Harry. The Genomics Data Board at UCLA is a multi-disciplinary body comprised of medical geneticists, genetic counselors, molecular geneticists and cytogeneticists, bioinformatics specialists, and other physicians and scientists which is responsible for interpreting clinical exome sequencing results. We would like to thank all participating members of the Genomics Data Board for their vital contributions to this work, specifically Drs. Kingshuk Das, Cristina Palmer, Ascia Eskin, Sibel Kantarci, and Julian Martinez-Agosto.
This work was partially presented at the American Society of Human genetics annual meeting, in Boston, MA (October 22–26, 2013. Poster #3082 F).
Competing interests
SPS, JLD, KD, HL, FQ-R, and WWG work for a fee for service laboratory providing diagnostic testing. The remaining authors declare that they have no competing interests.
Authors’ contributions
SPS performed analysis and prepared the manuscript. RL and NM provided phenotype information and photographs for patient #2. HL performed analysis and interpretation of molecular testing. ND served as liaison between sites and contributed to the description of phenotypes for both cases. JM provided phenotype information and photographs for patient #1. PFO provided genetic counseling and phenotype information for patient #1. JLD, EV, SFN, and WWG participated in the study design and provided clinical laboratory testing for both cases. FQ-R conceived of the study, and participated in its design and coordination, and the Genomic data board. All authors read and approved the final manuscript.