Temple-Baraitser Syndrome and Zimmermann-Laband Syndrome: one clinical entity?

Background KCNH1 encodes a voltage-gated potassium channel that is predominantly expressed in the central nervous system. Mutations in this gene were recently found to be responsible for Temple-Baraitser Syndrome (TMBTS) and Zimmermann-Laband syndrome (ZLS). Methods Here, we report a new case of TMBTS diagnosed in a Lebanese child. Whole genome sequencing was carried out on DNA samples of the proband and his parents to identify mutations associated with this disease. Sanger sequencing was performed to confirm the presence of detected variants. Results Whole genome sequencing revealed three missense mutations in TMBTS patient: c.1042G > A in KCNH1, c.2131 T > C in STK36, and c.726C > A in ZNF517. According to all predictors, mutation in KCNH1 is damaging de novo mutation that results in substitution of Glycine by Arginine, i.e., p.(Gly348Arg). This mutation was already reported in a patient with ZLS that could affect the connecting loop between helices S4-S5 of KCNH1 with a gain of function effect. Conclusions Our findings demonstrate that KCNH1 mutations cause TMBTS and expand the mutational spectrum of KCNH1 in TMBTS. In addition, all cases of TMBTS were reviewed and compared to ZLS. We suggest that the two syndromes are a continuum and that the variability in the phenotypes is the result of the involvement of genetic modifiers.

KCNH1 encodes a voltage-gated potassium channel that is predominantly expressed in the central nervous system, and mutations in this gene have been linked to both syndromes [6,7].
Here, we report on a Lebanese male patient with TMBTS having a mutation in KCNH1 that has previously been reported in a patient with ZLS. In addition, we have reviewed all published cases of TMBTS and highlight common features, as well as critical differences, between these two syndromes, and raise the issue of whether their classification into two entities is appropriate.

Clinical report
The male proband is the third child of healthy unrelated Lebanese parents. He was born at 36 weeks of gestation, after a complicated pregnancy characterized by the therapeutic administration, to the mother, of drugs against early contractions at 32 weeks of gestation. At birth, his weight was 2700 g (60 th percentile), his length 48 cm (75 th percentile) and his head circumference (OFC) 33 cm (60 th percentile). Family history was unremarkable. Marked hypotonia, constipation, and aplasia of thumb and great toe nails were noted in the first two to three days of life.
The propositus was referred for genetic examination at the age of 9 months. His weight was 9750 g (60 th percentile), length 71.5 cm (75 th percentile), OFC 42.7 cm (10 th percentile). He had a flat occiput, a frontal bossing, large ears, mild hypertelorism, epicanthal folds, a broad and depressed nasal bridge, a short columella, long philtrum, a broad mouth with downturned corners, a high arched palate, 2 upper and 2 lower incisors of normal shape, and full cheeks (Fig. 1). Widely spaced nipples and left chest depression were also noted. Both thumbs were held in an adducted posture and were terminally broad with aplasia of the nails bilaterally. Big toes were also broad, long, and with aplasia of nails. No hirsutism, no hypoplasia of the distal phalanges, no hypermobility, no camptodactyly, nor palmar creases were noted.
At 15 months old, his weight was 11 kg (75 th percentile), length 79 cm (50 th percentile), and OFC 45.7 cm (10 th percentile). Delays in developmental milestones were striking, as he could not stand up alone or walk with help, and could not follow or respond to simple commands. He had a myopathic face with poor visual contact, a wide open mouth and mild gingival enlargement (Fig. 1). Skeletal survey revealed nearly absent distal phalanges of the thumbs and great toes, very small femoral and humeral epiphyses, and an osteosclerosis of the anterior arc of the right 10 th rib (Fig. 2).
Magnetic resonance imaging, abdominal and heart ultrasound, brain stem auditory evoked responses, and EEG were normal. Complete blood count, hemoglobin electrophoresis, serum electrolytes, blood glucose levels, urinalysis, thyroid, liver and renal function tests were all unremarkable. Array CGH analysis and Chromosomal Microarray Analysis did not reveal any abnormalities (data not shown).

DNA extraction and Whole Genome Sequencing (WGS)
Whole genome sequencing was carried-out on the patient and his parents using the HiSeq 2500 sequencer (Illumina, San Diego, CA, USA). Libraries were generated from 1 μg of genomic DNA [8] using the Illumina TruSeq DNA PCR-Free Sample Preparation Kit. Genomic DNA was sheared using the Covaris system (Woburn, MA, USA). Isolated DNA fragment ends were blunted, A-tailed and ligated with sequencing adaptors with index sequences. Excess adapters and enzymes were removed using AMPure beads (Beckman Coulter Genomics, Danvers, MA, USA). Indexed libraries were size selected to 350 bp range using bead-based capture and the concentration of amplifiable fragment was determined by qPCR relative to sequencing libraries with known concentration. Normalized libraries were clustered on a c-BOT machine and 125 bp paired-end sequencing was performed on the HiSeq2500 system.

WGS data analyses
Raw data was mapped to the human genome reference build 19 (http://www.broadinstitute.org/ftp/pub/seq/references/Homo_sapiens_assembly19.fasta) using BWA aligner [9] version 0.7.7-r441 and variant call was performed using GATK [10] version 3.3.2. The rare variant analysis was performed using the xbrowse tool (https:// xbrowse.broadinstitute.org/). For the parents and the child, a 'De novo Dominant' inheritance model was selected, with severity of the variant effect set to 'moderate to high impact' (Nonsense, essential splice sites, missense frameshift and in frame), call quality as high (genotype quality > 20 and allele balance ratio > 25 %) and allele frequency < 1 % in 1000 genomes and The Exome Aggregation Consortium (ExAC) v0.3 datasets. Functional consequences of amino acid substitutions have been predicted using various tools [11][12][13][14].

Results
Whole Genome Sequencing identified 3 missense mutations in TMBTS patient (Table 1). We validated and confirmed the de novo origin of these variants by Sanger sequencing.
The mutation in KCNH1 (c.1042G > A) has a damaging effect according to all different effect predictors tested. STK36 has a missense mutation (c.2131 T > C), which also has damaging effects according to half of the effect predictors tested. ZNF517 has a missense mutation (c.726C > A) predicted as disease causing by one of the effect predictors.
The KCNH1 mutation results in a substitution of Glycine by Arginine. Same mutation is found in both isoforms of this protein: p.

Discussion
We report on a male Lebanese patient in which a de novo missense heterozygous mutation c.1042G > A in the KCNH1 gene led to TMBTS.
KCNH1 is a member of voltage-gated potassium channel proteins. It is recognized as an important regulator of cell proliferation in bone-marrow derived mesenchymal stem cells, and is involved in fundamental cellular and developmental processes [15,16].
Mutations in KCNH1 have been recently associated with TMBTS [6]. Moreover, de novo gain-of-function mutations in KCNH1 have also been reported in individuals with ZLS [7].
Generally, TMBTS and ZLS can be distinguished by their characteristic phenotypic features, which include absence or dyplasia of all nails and hypertrichosis in ZLS vs hypoplasia or aplasia of only the great toe and thumb's nails in TMBTS (Table 2). With this in mind, we considered that our patient had TMBTS. These syndromes are currently considered to be two separate entities, but their common characteristics suggest that these two syndromes may be different presentations of the same disorder. In fact, many common characteristics of patients with TMBTS and ZLS have been noted, such as, seizures, hypertrichosis, hypotonia, aplasia of nails, etc., which sometimes occur in some but not all patients ( Table 2). It is noteworthy to mention that many clinical databases do not even mention TMBTS as a differential diagnosis for ZLS because of the absence of hypertrichosis, even though not all reported patients with ZLS present this characteristic.
Interestingly, the same mutation (c.1042G > A) identified in our patient has never been reported with TMBTS, but was previously detected in patients with ZLS (patient 7 in Abo-Dalo et al. or subject 3 in Kortüm et al.) [17,18]. This substitution leads to a gain of function effect and mutants carrying this mutation exhibit an accelerated channel activation and a slower deactivation [18]. Along with the previously identified p.(Ile494Val) misense variant in KCNH1, which was shared among individuals with TMBTS and ZLS, the genetic defect identified in our patient, i.e., p.(Gly348Arg) was found in patients bearing different phenotypes and thus supposedly different syndromes. This provides stronger evidence that both syndromes clearly overlap and could be a phenotypic continuum. In fact, the mutation c.1042G > A was found in a patient with ZLS who does not present with hypertrichosis, similar to the patient reported herein. However, the patient had in addition, aplasia of all nails of hands and feet, thoracic scoliosis, and infrequent seizures, which were not present in our patient who had a delay in epiphyseal maturation, (Table 3) a feature never reported before in both entities, and gingival enlargement. The latter is a characteristic not reported previously in TMBTS affected individuals, however it is a frequent feature in patients with mutations of KCNH1 (Bramswig et al.). Genetic modifiers, possibly involving the Na + and Ca 2+ channels, might block the KCNH1 channels and result in the gingival enlargement as it is observed in individuals treated with Na + blocker phenytoin or Ca 2+ channel blocker nifedipine [18].
On the other hand, the patient reported by Kortum et al., developed seizures in adolescence, therefore one could speculate a late occurence of epilepsy in the patient described here with the same mutation. Yet, Bramswig et al. described 3 individuals presenting with an identical KCNH1 variant but with different clinical features with regard to epilepsy [19]. Consequently, the presence of a pathogenic KCNH1 variant alone could not allow for a prediction of occurence of epileptic seizure.
Other genetic modifiers could be responsible for the observed differences in clinical phenotype. We looked deeper at the results of the WGS and noticed mutations in two other genes STK36 and ZNF517, which were classified in some databases as possibly damaging. However, their significance remains to be elucidated. Recently, de novo mutations in STK36 have been identified in patients with epileptic encephalopathies [20]. Although our patient who has missense mutation in STK36 does not present with epilepsy at present, he might develop it in adolescence as in patient 3 in Kortum et al. Thus, concordant to previous reports, our data supports the evidence that the mutated KCNH1 is a major cause of TMBTS and ZLS, while other     genes can act as disease modifying roles. Understanding the molecular mechanisms by which these genes exert disease modifying roles might help in the better understanding of the pathogenesis of these syndromes. Finally, both ZLS and TMBTS patients with KCNH1 mutations show similar phenotypes. Nevertheless, two other ZLS patients were also described with mutations in the ATP6V1B2 gene that encodes a component of the vacuolar ATPase (V-ATPase). These mutations present a more pronounced phenotype characterized mostly by hypertrichosis and a coarser facial phenotype ( Table 2). But due to the limited number of individuals described, a conclusion about whether probands with mutations involving ATP6V1B2 lead to a more severe syndrome might not be accurate. On the other hand, Kortüm et al. screened a cohort of 24 ZLS patients, of which only 8 had mutations in KCNH1 and ATP6V1B2 suggesting further the genetic heterogeneity in the ZLS disorder [18].

Conclusions
In summary, this study shows that the same KCNH1 mutation can lead to both ZLS and TMBTS. The phenotypic variability could be the result of a modifier gene or genes, and identification of such genes would be of great importance. A careful analysis of genetic polymorphisms in various loci should be taken into consideration for clinical diagnosis. Further investigations are needed to confirm if ATP6V1B2 mutations lead to a more severe phenotype.
Abbreviations ID, intellectual disability; TMBTS, Temple-Baraitser syndrome; WGS, whole genome sequencing; ZLS, Zimmermann-Laband syndrome Table 3 Clinical comparison between the patient here described with TMBTS and the patient described by Kortüm