Article Text
Abstract
Background: Haploinsufficiency of the gene encoding for transcription factor 4 (TCF4) was recently identified as the underlying cause of Pitt–Hopkins syndrome (PTHS), an underdiagnosed mental-retardation syndrome characterised by a distinct facial gestalt, breathing anomalies and severe mental retardation.
Methods: TCF4 mutational analysis was performed in 117 patients with PTHS-like features.
Results: In total, 16 novel mutations were identified. All of these proven patients were severely mentally retarded and showed a distinct facial gestalt. In addition, 56% had breathing anomalies, 56% had microcephaly, 38% had seizures and 44% had MRI anomalies.
Conclusion: This study provides further evidence of the mutational and clinical spectrum of PTHS and confirms its important role in the differential diagnosis of severe mental retardation.
Statistics from Altmetric.com
In 1978, Pitt and Hopkins described two unrelated patients with sporadic “mental retardation, wide mouth and intermittent overbreathing”.1 After this initial report, only four other sporadic cases and one sibling pair with a similar phenotype of severe mental retardation, wide mouth and breathing anomalies were reported as possible cases of Pitt–Hopkins syndrome (PTHS, OMIM 610954).2–5 Recently in two patients, two deletions, one of 1.2 Mb and one of 1.8 Mb, in 18q21.1, detected by molecular karyotyping using 100 K single-nucleotide polymorphism (SNP) arrays and bacterial artificial chromosome arrays, respectively, led to the identification of haploinsufficiency of TCF4, encoding a basic helix–loop–helix (bHLH) transcription factor, as the underlying cause of PTHS.6 7 A recurrent de novo missense mutation in the bHLH domain region of the TCF4 gene in three patients and one other missense mutation at the same position,6 7 as well as one splice-site mutation and three stop mutations7 in further patients were later identified. In the meantime, two other patients with the PTHS or overlapping phenotype and a deletion of TCF4 were reported.8 9
The TCF4 gene on chromosome 18 encodes a member of the bHLH transcription factor family (also called “E-proteins” as their basic domain binds to the Ephrussi-box (E-box) consensus binding site “CANNTG”).10 These transcription factors are able to bind DNA as homodimers or heterodimers with other classes of HLH proteins and play an important role in many developmental processes, including the differentiation of the vertebrate nervous system and the development of the cortex.11 12 TCF4 encodes at least two isoforms, differing in the presence of 4 amino acid residues (RSRS) 17 residues upstream of the HLH domain.13
Both null mutations and missense mutations located within the bHLH domain of TCF4 impaired its interaction in vitro with ASCL1, a tissue-specific HLH protein from the PHOX-RET pathway. As this pathway is involved in Hirschsprung disease (HSCR; OMIM 142623) and the Ondine hypoventilation syndrome (OMIM 209880) through its role in the development of noradrenergic derivatives, the finding of HSCR or severe constipation and breathing anomalies in patients with PTHS might be explained by impaired interaction with ASCL1.7 Investigations of tcf4 expression in Danio rerio embryos showed early expression in the pallium of the telencephalon, the diencephalon, the midbrain tegmentum, the hindbrain and the branchial arches, thus correlating with the phenotypical spectrum in humans with PTHS.8
Because of its phenotypical overlap with Angelman and Rett syndromes, we speculated that PTHS might become an important differential diagnosis with these conditions.
To further delineate the genotypic and phenotypic spectrum of PTHS and to establish its frequency in patients with severe mental retardation, we analysed 117 patients with overlapping clinical features.
METHODS
Ethics approval for this study was obtained from the ethics committee of the Medical Faculty, University of Erlangen-Nuremberg, and informed consent was obtained from the parents or guardians to study the patients.
Patients
Our study population contained 117 patients including two sibling pairs, who were referred to us with severe mental retardation and variable additional features reminiscent of the PTHS spectrum, such as microcephaly, dysmorphic facial gestalt or breathing anomalies. At least 70 of these patients had tested negative for Angelman and/or Rett syndrome.
Molecular testing
DNA samples derived from peripheral blood were screened for TCF4 mutations by bidirectional direct sequencing of the coding exons 2–19 and the non-coding exon 20 and intronic flanking regions (ABI BigDye Terminator Sequencing Kit V.2.1; Applied Biosystems, Foster City, California, USA) using an automated capillary sequencer (ABI 3730; Applied Biosystems). Primer pairs and PCR conditions are available on request.
Paternity was verified in samples taken from patients 7 and 12 and their parents. Probes were verified by genotyping with 14 polymorphic microsatellite markers (PowerPlex 16 System, Promega, Madison, Wisconsin, USA) to exclude any possibility of mistakes. Reverse transcriptase (RT)-PCR was performed for patients 12 and 14, using primers located in exons 13 and 17 and exons 5 and 11, respectively, on cDNA obtained from mRNA (Superscript II Reverse Trancriptase; Invitrogen, Carlsbad, California, USA). For patient 12, mRNA was extracted from lymphoblastoid cell lines (RNeasy Mini Kit; Qiagen, Valencia, California, USA) and peripheral blood (PAXgene system; Preanalytix, Franklin Lakes, New Jersey, USA), using commercial kits in accordance with the manufacturers’ instructions. Aberrant transcripts were extracted from agarose gel (QiaQuick gel extraction ki; Qiagen) and sequenced after reamplification.
Bioinformatic analyses
Disordered protein segments and linear protein interactions motifs were identified using the software programs DisEMBL14 and ELM,15 respectively. The effect of the G358V mutation on the aggregation tendency of TCF4 was assessed using AGGRESCAN,16 a web-based software program for the prediction of aggregation-prone segments in protein sequences and the analysis of the effect of mutations on aggregation propensities of proteins.
Functional testing
Functional consequences of the G358V mutation were tested with a transcriptional reporter assay as described previously.7 Immunofluorescence was performed with a primary antibody against TCF4 (ab2233-100; Abcam, Cambridge, Massachusetts, USA) and a CY3-labelled secondary anti-goat antibody (C2821-1ML; Sigma-Aldrich, St Louis, Missouri, USA) on JEG3 cells previously transfected with either wild-type TCF4 or mutant TCF4.
RESULTS
Clinical findings
All 16 patients with proven TCF4 mutations in this study had severe mental retardation with very little speech (2 patients had <5 words) or no speech (14 patients) and with limited walking abilities. They resembled each other with a specific facial phenotype characterised by deep-set eyes, broad and often beaked nasal bridge with down-turned, pointed nasal tip and flaring nostrils, wide mouth with widely spaced teeth, cupid’s-bow upper lip, a protruding lower face, and mildly cup-shaped and fleshy ears (fig 1). Microcephaly was observed in 56% of the patients and breathing anomalies in 56%. Seizures occurred in 38% of the patients between the age of 0 and 5 years. MRI anomalies such as bulging of caudate nuclei, ventricular asymmetry, agenesis or hypoplasia of the corpus callosum and atrophy of the frontal and parietal cortex were observed in 44% of the patients. Hypotonia and constipation were common findings in 13 and 11 of the patients, respectively. Personality was described as happy in 15 of the patients. Single palmar creases were reported in 11 patients and scoliosis in 4 patients. Other less common anomalies were myopia and fetal pads in four and six patients, respectively (table 1).
Molecular testing
Sequencing of TCF4 in 117 patients with severe mental retardation and clinical findings overlapping with PTHS revealed novel mutations in 16 patients. Large deletions of TCF4 could be excluded by the identification of at least one heterozygous SNP in all but five of the remaining patients. Owing to their facial phenotype, there is a strong suspicion that two of the patients with normal TCF4 sequencing and exclusion of large deletions of the TCF4 gene do have PHTS. Single exon deletions or mutations in regulatory elements of TCF4 and locus heterogeneity may explain these patients and are also possibilities in the other, less characteristic patients.
We identified 12 novel stop mutations, including two splice-site mutations, distributed over the gene, one novel frameshift mutation located towards the C-terminus and resulting in an elongation of the putative protein, two novel missense mutations in exon 18 coding for the HLH domain, and one novel missense mutation in exon 14 (table 2, fig 2). De novo occurrence was proven in all but three stop mutations, for whom parental samples were not available. Assumed probe relationships were confirmed in the patients and both parents in the exceptional missense mutation in exon 14 and the unusual splice-site mutation IVS14+3A→G. The exceptional missense mutation in exon 14 was also excluded in 192 healthy control chromosomes. RT-PCR performed on mRNA from a lymphoblastoid cell line of patient 12 with the IVS14+3A→G mutation revealed an aberrant transcript with skipping of exon 14, which leads to a frameshift that results in a premature stop after three amino acids (fig 3B). RT-PCR in patient 14, who had the splice-site mutation IVS9+2insGT, also revealed an aberrant transcript, skipping exon 9, which leads to a frameshift resulting in a premature stop after 14 amino acids (fig 3C).
Bioinformatic analyses
Computational analyses indicated that the N-terminal 550 amino acids of TCF4 are predominantly disordered and do not adopt a globular (domain-like) tertiary structure. There was also no evidence that position 358 is part of a specific linear protein-interaction motif that might be destroyed by the mutation G358V (data not shown).
An analysis of the TCF4 aggregation properties revealed that the G358V mutation leads to an increase in aggregation tendency as indicated by the larger hotspot area at the site of the mutation (fig 3A). Compared with the size of adjacent hotspots, the hotspot emerging at the site of the G358V mutation is more than twice as large.
Functional testing
Functional testing of the missense mutation G358V with a transcriptional reporter assay showed only inconsistently a mild reduction in transcriptional activity compared with the wildtype (data not shown). Immunofluorescence with antibodies against TCF4 revealed no visible aggregates (data not shown).
DISCUSSION
Clinical spectrum
Including our 16 novel patients, 27 patients with molecularly confirmed PTHS are currently known.6–8 All of these patients are severely mentally retarded with no or only very limited speech and limited mobility. The earliest reported walking age was 2 years in patient 8, whereas some of the older patients can not walk without support (eg, patient 2) or not at all (eg, patient 1). Gait is often unstable and ataxic. Furthermore, muscular hypotonia seems to be a common feature particularly at younger ages.
Breathing anomalies occurred in 18 of 27 (67%) patients, with age of onset varying from a few months to teenage years. These episodes were characterised by daytime periods of hyperventilation followed by apnoea. Milder anomalies may also occur such as “playing with breath” without apnoea in patient 9 or a singular occurrence of hyperventilation after narcosis in patient 16. The oldest patients without breathing anomalies were 18 years old.
Microcephaly, both congenital and acquired, occurred in about 63% of all known patients. While birth weight was in the lower normal range in the patients in our first study,7 in this cohort birth weight was normal or high normal in all cases for whom this information was available.
Further common symptoms were seizures with onset from birth up to 9 years of age (44%), constipation (67%) and minor anomalies such as single palmar creases and supernumerary phalangeal flexion creases (63%). Hirschsprung disease, which had occurred in one of the first patients,7 was not noticed in any of the present cohort.
In 14 of 22 (64%) patients who had MRI studies performed, anomalies such as bulging of caudate nuclei, ventricular asymmetry, agenesis or hypoplasia of the corpus callosum, and atrophy of the frontal and parietal cortex or arachnoidal cysts were reported. Patients with early-onset seizures had mild broadening of the ventricular system (P4), mild atrophy of frontal and parietal cortex (P7) and a retrocerebellar arachnoidal cyst or mega cisterna magna (P16). Additional less common clinical findings were scoliosis (26%), myopia (19%), strabismus (30%), hypogenitalism (19%) and accessory nipples (19%). Hands and feet were often described as slender and small, with single palmar creases in many and fetal pads in some patients. The lack of gross malformations is in accordance with the embryonic tcf4 expression in D rerio, which is restricted to the brain and branchial arches.8
Most of the patients showed a happy and placid personality; violent or autoaggressive behaviour was only reported in three patients. Stereotypic movements were observed in 30% of patients and in patient 13, the loss of hand use was reported. These features, in addition to microcephaly, breathing anomalies, severe mental retardation and seizures, resemble the features seen in both Rett and Angelman syndromes. Accordingly, most of the patients referred to us with suspected PTHS had already been tested for these diseases previously. As we found mutations in about 14% of the patients in this study, the important role of PTHS as a differential diagnosis of Rett and Angelman syndromes is further confirmed. The most consistent aspect distinguishing PTHS from the other two syndromes is the characteristic facial gestalt including a coarse face, high cheekbones, a beaked nasal tip, a protruding lower face, a wide mouth with cupid’s-bow shaped upper lip and wide-spaced teeth. However, the facial phenotype can be subtle as that seen in patient 5, who does not have the characteristic beaked nose and resembles the other patients mostly through the shape of her face, with high cheekbones.
Mutational spectrum
In this study, we could identify 16 novel mutations in TCF4, distributed over the gene. (table 2, fig 2); 13 of these are frameshift, nonsense or splice-site mutations, therefore further confirming TCF4 haploinsufficiency as the disease causing mechanism. One of these splice-site mutations is an A→G exchange at position IVS14+3, where both A and G are possible.17 Nevertheless, because this mutation occurred de novo, we assumed abnormal splicing. We found an aberrant transcript by RT-PCR using cDNA from the patient, even though this was weaker than the wild-type allele, and confirmed skipping of exon 14 by sequencing the aberrant transcript (fig 3B). One proven de novo frameshift mutation is located in the C-terminal exon 19 and results in an elongation of the putative protein by 37 amino acids. Owing to its location near to the C-terminus, nonsense-mediated mRNA decay, as assumed for stop mutations, might not occur, but impairment of protein function due to changes in protein structure is likely, particularly as the functional HLH domain is located in the C-terminal part of the protein.
Amiel et al6 had identified three missense mutations at the same amino acid position (576 or 580, depending on the isoform) within the bHLH domain in four patients, thus indicating a mutational hot spot at this position. Surprisingly, in 22 patients with defects in TCF4 (7 and this study) we found this particular missense mutation only once (table 2). However, we found two further missense mutations in the bHLH domain located only two amino acid positions upstream of the reported recurrent mutation site, which lead to a change from arginine to proline or histidine at position 574 or 578, respectively. Both mutations at position 576/580 and 574/578 affect evolutionarily highly conserved glutamic and arginine residues constituting the E-box recognition motif.6 We showed previously that such an impairment of the functional bHLH domain reduces interaction with ASCL1 in transactivating an E-box-containing reporter construct, to a similar degree as haploinsufficient stop mutations.7
Interestingly, we also found the first missense mutation outside the bHLH domain. This G358V mutation in exon 14 affects an evolutionarily highly conserved position. The mutation was excluded in both parents and 192 healthy control chromosomes, and sampling errors were excluded by short tandem repeat marker analysis. Nevertheless, as this mutation is not located within a known functional domain, the pathogenic mechanism remains unclear. Reporter-assay testing of interaction with ASCL1 showed a mild but not significant decrease in transcriptional activity (data not shown), which is not surprising as the mutation is not located within the DNA or protein-binding domain. Computational analyses indicated that the N-terminal 550 amino acids of TCF4 are predominantly disordered and do not adopt a globular (domain-like) tertiary structure. Such regions were reported previously to mediate either specific interactions via short linear sequence motifs15 or nonspecific interactions that might lead to aggregation.16 However, computerised investigation of whether position 358 is part of a specific linear protein interaction motif gave no evidence for such a role of this sequence region (data not shown). In contrast, an analysis of the TCF4 aggregation properties revealed that the G358V mutation leads to an increase of the aggregation tendency, thus offering an explanation for the slightly reduced transcriptional activity observed. The small overall effect, however, suggests that only small-sized or transient/reversible aggregates are formed. This is also consistent with the immunofluorescence data that did not reveal visible TCF4 aggregates in cells transfected with the mutant protein compared with the wild type.
In conclusion, we have further delineated the mutational and clinical spectrum of PTHS and confirmed its important role in the differential diagnosis of severe mental retardation.
Acknowledgments
We are grateful to the patients and families who participated in this study. We thank D Schweitzer for skilful technical assistance. This work was supported by grant RA 833/7-1 to A Rauch funded by the Deutsche Forschungsgemeinschaft (DFG). M Peippo at The Department of Medical Genetics, Väestöliitto is funded by Finland’s Slot Machine Association (RAY).
REFERENCES
Footnotes
Competing interests: None.
Patient consent: Parental consent obtained.