Familial cleft tongue caused by a unique translation initiation codon variant in TP63

Variants in transcription factor p63 have been linked to several autosomal dominantly inherited malformation syndromes. These disorders show overlapping phenotypic characteristics with various combinations of the following features: ectodermal dysplasia, split-hand/foot malformation/syndactyly, lacrimal duct obstruction, hypoplastic breasts and/or nipples, ankyloblepharon filiforme adnatum, hypospadias and cleft lip/palate. We describe a family with six individuals presenting with a striking novel phenotype characterized by a furrowed or cleft tongue, a narrow face, reddish hair, freckles and various foot deformities. Whole-exome sequencing (WES) identified a novel heterozygous variant, c.3G>T, in TP63 affecting the translation initiation codon (p.1Met?). Sanger sequencing confirmed dominant inheritance of this unique variant in all six affected family members. In summary, our findings indicate that heterozygous variants in TP63 affecting the first translation initiation codon result in a novel phenotype dominated by a cleft tongue, expanding the complex genotypic and phenotypic spectrum of TP63-associated disorders.

TP63 is located on 3q28 and comprises 16 different exons [3,4]. It is a member of the large p53 family of transcription factors and a key regulator in epidermal development and differentiation [5]. Additionally, p63 has been linked to divers biological processes, e.g., differentiation, proliferation, adhesion, stem cell maintenance, apoptosis and aging [6,7]. The wide spectrum of cellular functions might be explained by the large number of p63 isoforms identified yet. Two different transcription start sites lead to two main classes of transcripts: TAp63 and ΔNp63. TAp63 contains three TA-specific exons, exon 1, 2 and 3, and encodes a transactivation domain (Fig. 1). The shorter isoform variant, ΔNp63, uses an independent promotor in exon 3' resulting in a truncated transactivating domain [6,8,9]. In addition, alternative splicing at the 3ʹ end of both the TAp63 and ΔNp63 transcripts generates at least three different C-terminal variants (α, β, and γ) [8,10]. The longest 3′ splice variant (including the exons [11][12][13][14] is the α isoform, which comprises a sterile alpha-motif (SAM) involved in protein-protein interactions and a transactivation inhibitory domain (TID) [10][11][12][13]. Two further variants, β (which lacks exon 13) and γ (that does not contain the exons 11-14 but has a γ-specific exon 10′), do not comprise the SAM and TID domains [14,15]. All isoforms contain the DNA-binding domain and oligomerization domain (OD) [5,8,10,16]. A wide spectrum of different heterozygous germline variants in TP63 have been identified as causative of the syndromes and malformations stated above. According to the literature (HGMD Database Professional 2020.4), most of the reported disease-causing variants (110 of 153) are missense variants. Variants associated with particular disorders usually cluster in specific functional domains, indicating a genotype-phenotype association [2,4,17]. For instance, variants causing the EEC3 syndrome are generally located in the DNAbinding domain, whereas variants leading to the AEC phenotype are regularly found either in the SAM and transactivation inhibitory domains or in the transactivating domain of the ΔNp63 isoforms ( Fig. 1) [2,5,18,19]. These findings suggest that . B Schematic diagram of the human TP63 gene structure. Alternative promoter use produces TA (transactivation) and N-terminally truncated (ΔN) isoforms, and alternative splicing produces C-terminal variants (α, β, γ). Colors within exons correspond to the different functional domains. Red arrow indicates the TP63 variant identified within this study. Orange arrow indicates the most likely alternatively used start codon at position 40, resulting in a shortened version of the TAp63 isoforms. Different TP63-associated disorders (ankyloblepharon-ectodermal defects-cleft lip/palate syndrome (AEC); acrodermo-ungual-lacrimal-tooth syndrome (ADULT); ectrodactyly, ectodermal dysplasia, cleft lip/palate syndrome 3 (EEC3); limb-mammary syndrome (LMS); split-hand/foot malformation type 4 (SHFM4)) and the typical location of their variant are indicated by black brackets. C Comparison of the six major isoforms encoded by TP63. Isoforms in the red box are potentially affected by the TP63 variant p.Met1?. Colors correspond to the different functional domains.  there might be specific molecular mechanisms for each TP63related condition.
Here, we describe a unique translation initiation codon variant in TP63 identified in a family with a striking novel phenotype. The variant, c.3G>T, affects the first transcription start site of p63 and is predicted to lead to impaired function of the TAp63 isoforms. All six patients carrying this variant show a cleft tongue.

CLINICAL REPORT
We describe six patients with cleft tongue (HP:0000221) who were referred to our Institute of Human Genetics for clinical and diagnostic evaluation. The cleft tongue (HP:0000221) was already present at birth and did not change over time. Their sense of taste was not affected and no signs of submucous cleft palate were observed. The patients reported no difficulty in swallowing. The clinical characteristics of the patients described in the present report are summarized in Table 1. The family pedigree is illustrated in Fig. 2A and is consistent with autosomal dominant inheritance. The corresponding images are shown in Fig. 2B. The main clinical features are summarized and compared to the typical clinical findings of other TP63-associated disorders in Table 2.

Patient 1
The boy (III.1 in Fig. 2A) was born spontaneously after 39 weeks of gestation. Birth weight was 3350 g (−0.3 SD), length 53 cm (−0.5 SD) and occipitofrontal head circumference (OFC) 35 cm (−0.2 SD). Apgar scores were 10 at 5 and 10 min, respectively. Due to bilateral talipes equinovarus (HP:0001762) he required a series of castings followed by night splinting. His psychomotor development was slightly delayed. He walked independently at the age of 15 months and started to speak first words at the age of 2 years. At clinical evaluation at 12 years and 4 months, height was 167 cm (1.6 SD), weight 65.9 kg (1.8 SD), and OFC was 55.2 cm (0.4 SD). He showed reddish hair (HP:0002297), freckles (HP:0001480), and a narrow face (HP:0000275). His tongue showed a deep single median furrow looking like a cleft tongue (HP:0000221). Serum creatine kinase levels were unremarkable. Fig. 2A was born at term. His birth weight was 3630 g (median), length 56 cm (1.5 SD) and OFC 36 cm (0.3 SD). Apgar scores were 10 at 5 and 10 min. Pigeon toes (HP:0001760) were noted shortly after delivery. His development was slightly delayed. He walked independently at the age of 16 months and started to speak first words at the age of 20 months. He received speechand physiotherapy. At the age of 2 years and 4 months, he was diagnosed with Langerhans' cell histiocytosis, which could be treated successfully with radiation. At 10 years and 6 months, his height was 145 cm (0.2 SD), weight 37.9 kg (0.3 SD), and OFC was 55 cm (0.9 SD). He showed the following facial features: narrow face (HP:0000275), reddish hair (HP:0002297), freckles (HP:0001480), and a median furrowed tongue/cleft tongue (HP:0000221). Serum creatine kinase levels were unremarkable.

Patient 4
Patient I.2 showed also a median furrowed tongue/cleft tongue (HP:0000221). She had a pointed foot as a child and reported muscular hypotonia (HP:0001252). She reported that her already deceased mother had similar symptoms, too.

Patient 5
The girl (III.4 in Fig. 2A) was born after 38 weeks of gestation by Caesarean section. Her birth weight was 3500 g (0.8 SD), length was 52 cm (0.75 SD) and OFC 34 cm (−0.2 SD). Apgar scores were 10 at 5 and 10 min, respectively. She received physiotherapy due to muscular hypotonia (HP:0001252) and uses orthopedic insoles due to flat valgus foot. Otherwise, her psychomotor development was normal. She walked independently at the age of 15 months and started to speak first words at the age of 1 year. At clinical evaluation at 12 years and 4 months, her height was 167.2 cm (1.6 SD), weight 65.9 kg (1.8 SD), and OFC was 55.2 cm (0.4 SD). She showed the following characteristic facial features: narrow face (HP:0000275), reddish hair (HP:0002297), freckles (HP:0001480), and a median furrowed tongue/cleft tongue (HP:0000221).  Table 2. Summary of the clinical findings in our patients with the main clinical feature of cleft tongue (CT) (HP:0000221) compared to the typical clinical findings of different TP63-associated disorders: ankyloblepharon-ectodermal defects-cleft lip/palate syndrome (AEC); acro-dermo-unguallacrimal-tooth syndrome (ADULT); ectrodactyly, ectodermal dysplasia, cleft lip/palate syndrome 3 (EEC3); limb-mammary syndrome (LMS); split-hand/ foot malformation type 4 (SHFM4) and isolated orofacial cleft 8 (OFC8) [1]. Human Phenotype Ontology (HPO) [40].  [20]) and predicted to have a severe impact on protein function. This heterozygous missense variant, c.3G>T, was located in exon 1 of the TP63 gene ( Fig. 1)  gov/orffinder/) the most likely alternatively used start codon is the next downstream AUG codon at position 40, resulting in a shortened version of the isoform originally annotated as TAp63 (Fig. 1B). Sanger sequencing was used to validate the WES data and to prove that the other affected family members, II.1, II.3 and I.2, do carry the heterozygous variant as well ( Fig. 2A). No additional patient material was available for detailed isoform characterization.

DISCUSSION
In this study, we present a three-generation family with six individuals showing a striking novel phenotype, characterized by a furrowed or cleft tongue, a narrow face, reddish hair, freckles and various foot deformities. Despite intensive research, we were not able to identify a similar case in the literature and to make a specific clinical diagnosis, so we decided to perform variant analysis by whole-exome sequencing. We selected the three most distantly related affected family members to reduce the amount of shared benign variants and to maximize the segregation filtering power. Focusing on heterozygous variants consistent with a dominant transmission in line with the family history, we identified a heterozygous variant, c.3G>T, in TP63 affecting the translation initiation codon (p.1Met?). Heterozygous variants in TP63 have been linked to several syndromic and isolated diseases [5,18] with partly overlapping phenotypes and various combinations of the following features: ectodermal dysplasia (e.g., nail dysplasia, sparse hair, hypopigmentation, freckling, tooth abnormalities), split-hand/foot malformation/syndactyly, lacrimal duct obstruction, hypoplastic breasts and/or nipples, ankyloblepharon filiforme adnatum, hypospadias and cleft lip/palate [1,2]. The large number of p63 isoforms might explain the wide spectrum of TP63-associated diseases (Fig. 1). Two different transcription start sites lead to two main classes of transcripts: TAp63 and ΔNp63 [6,8,9]. Furthermore, alternative splicing at the 3ʹend of both the TAp63 and ΔNp63 transcripts generates at least three different C-terminal variants (α, β, and γ) [8,10]. Cleft tongue has not been linked as a characteristic feature to TP63-related disorders until now [21][22][23]. However, other TP63 variants have been described to cause forms of orofacial clefts including cleft lip, cleft palate and cleft uvula [24][25][26]. Interestingly, especially loss-of-function variants in TP63 seem to result in a phenotype dominated by oral clefting [5]. So far, no TP63 variant has been described that affects the translation initiation codon of the first transcription start site, neither as disease-causing (HGMD Database Professional 2020.4) nor in healthy controls (https://gnomad.broadinstitute.org; last access date 10/02/2021, [20]). The most likely used next start codon is at position 40, leading to a shortened version of the TAp63 isoforms. For further verification, we validated the WES data and performed segregation analysis. The TP63 variant, c.3G>T, was found in all six affected family members by Sanger sequencing, supporting our theory that this unique variant causes a novel TP63-associated phenotype.
The specific phenotype of the reported patients might indicate that the affected TAp63 isoforms are especially expressed during a specific period of the development during embryogenesis. Tongue development is a complex process starting around week 4 of embryonic development [27]. The muscles of the tongue predominantly derive from the myoblasts, which originate in the occipital somites, and tongue connective tissue and vasculature are derived from cranial neural crest cells [28,29]. Reciprocal interactions between cranial neural crest cells and myogenic cells are essential in the coordination of tongue development, and deep single median furrows on the dorsal surface of the tongue or a cleft tongue can occur if this complex process is disturbed [30]. Expression analysis of different TP63 isoforms indicates that TAp63 is the dominant isoform expressed in skeletal muscle and plays a critical role specifically in late myogenic differentiation stages [31], providing an explanation for the observed phenotypic differences between patients described in this study and previously described TP63-associated phenotypes. Specifically, TAp63γ contributes to muscle growth, remodeling and functional differentiation by controlling specific sub-sets of target genes, and it is involved in the formation of atrophic myotubes and reduced myoblasts fusion index [32]. Dysfunction of TAp63, therefore, might directly affect myoblast fusion during tongue development, resulting in a deepfurrowed/cleft tongue, and, furthermore, provide an explanation for the muscular hypotonia that is present in varying degrees in all our patients. Further investigations are required to verify this hypothesis.
In general, the occurrence of a cleft tongue with or without other abnormalities is rare. Most of the other syndromes related with cleft/furrowed or bifid tongue, such as orofaciodigital syndrome 1 [MIM 311200], Robinow syndrome [MIM 180700] or distal arthrogryposis type 5D [MIM 615065], are as well associated with other forms of orofacial clefts, e.g., cleft lip and/or cleft palate. Among the genetic causes associated with these phenotypes, pathogenic variants in DVL1/DVL3 and WNT5A (all (i.a.) for Robinow syndrome) and TCTN3 (orofaciodigital syndrome 4) have been identified. Encoded proteins are involved in different cellular pathways which, like the sonic hedgehog (SHH) and WNT pathways, are key signaling pathways during development and maintenance of various tissues and, similar to p63, regulate expression of a huge variety of target genes [33][34][35]. Interestingly, it has been shown that p63 can modulate both pathways and act as an activator of SHH signaling as well as a repressor of canonical WNT signaling in different tissues and developmental stages [36,37]. However, these regulations can be exerted by different isoforms of TP63, including ΔNp63 variants, which are not affected by the variant identified in this study [38]. Therefore, it remains highly speculative whether the observed phenotype is caused by direct interaction of p63 with these pathways.
Additionally, we found by extensive research of the literature a further hint in line with our assumption. Pries et al. reported in 1974 three patients with EEC syndrome showing a furrow in the midline of the tongue [39], too. Consistent with the clinical diagnosis of EEC syndrome, these patients showed further malformations including ectodermal dysplasia as well as cleft lip and palate. One of the patients with furrowed tongue reported by Pries et al. had additional split-hand/foot malformation. Overall, they were more severely affected than the patients described here. The genetic cause of these cases is unfortunately unknown. We can only speculate what kind of variant might have caused their specific phenotype, but still, the phenotypic overlap to the patients presented in this study is quite striking. Interestingly, muscular hypotonia was not mentioned as a clinical feature in these three cases. Actually, it remains unclear whether the muscular hypotonia that is present in varying degrees in all of our patients can be explained by the TP63 variant. Despite intensive efforts, we were not able to identify any additional genetic cause of this. Presenting these cases will contribute to identify further families with hereditary cleft tongue due to variants in TP63, which will help to gain further insights into p63associated pathomechanisms and to answer these open questions.
In summary, our findings indicate that heterozygous variants in TP63 affecting the first translation initiation codon result in a novel phenotype dominated by a cleft tongue, expanding the complex genotypic and phenotypic spectrum of TP63-associated disorders.

DATA AVAILABILITY
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.