Background Borjeson–Forssman–Lehmann syndrome (BFLS) is an X-linked recessive intellectual disability (ID) disorder caused by mutations in the PHF6 gene and characterised by variable cognitive impairment, a distinct facial gestalt, obesity, and hypogonadism. Female carriers are usually not affected or only mildly affected, and so far only two females with de novo mutations or deletions in PHF6 have been reported.
Methods and results We performed PHF6 mutational analysis and screening for intragenic deletions and duplications by quantitative real-time PCR and multiplex ligation dependent probe amplification (MLPA) in female patients with variable ID and a distinct appearance of sparse hair, remarkable facial features, hypoplastic nails, and teeth anomalies. We detected two truncating mutations and two duplications of exons 4 and 5. Furthermore, two female patients with PHF6 deletions and a similar phenotype were identified by routine molecular karyotyping. Recently, two patients with a clinical diagnosis of Coffin-Siris syndrome in early infancy had been found to harbour mutations in PHF6, and their phenotype in advanced ages is now described. Further studies revealed skewed X-inactivation in blood lymphocytes, while it was normal in fibroblasts, thus indicating functional mosaicism.
Conclusions Our findings indicate that de novo defects in PHF6 in females result in a recognisable phenotype which might have been under-recognised so far and which comprises variable ID, a characteristic facial gestalt, hypoplastic nails, brachydactyly, clinodactyly mainly of fingers IV and V, dental anomalies, and linear skin hyperpigmentation. It shows overlap with BFLS but also additional distinct features, thus adding a new facet to this disorder.
- Clinical genetics
Statistics from Altmetric.com
Borjeson–Forssman–Lehmann syndrome (BFLS; MIM #301900) is an X-linked disorder which was first described in 1962. It is characterised by moderate to severe intellectual disability (ID), epilepsy, hypogonadism, hypometabolism, pronounced obesity, swelling of subcutaneous tissue of the face, narrow palpebral fissures, and large but not deformed ears.1 Further clinical aspects can be moderately short stature, tapering fingers with hyperextensible joints, and short toes.2 Since the identification of mutations in the PHF6 gene (MIM *300414) as the underlying cause in 2002,3 17 families with 41 affected males and confirmed PHF6 mutation have been reported.2–,9
Female carriers from the known families are usually unaffected. Some have been reported to be mildly cognitively impaired or to show some of the phenotypic characteristics such as obesity, tapering fingers, prominent earlobes or facial aspects such as bitemporal narrowing or prominent supraorbital ridges.2 ,4,7–,9 In more than half of the families notably skewed X-inactivation in carrier females could be observed.2–,4 ,7–,9
So far, only two sporadic female patients with a de novo mutation or deletion in PHF6 have been reported. One girl was tested for PHF6 mutations because of a clinical suspicion of BFLS and was found to harbour the de novo p.Gly10Argfs*12 mutation.7 She had developmental delay with low normal IQ, behavioural anomalies, abnormal thyroid function, tall stature, macrocephaly, obesity, some facial characteristics with deep set eyes and fleshy earlobes, and broad feet with hammer toes. She showed a 93% skewed X-inactivation in DNA from blood.7 The second girl was tested unbiasedly on a single nucleotide polymorphism (SNP) array because of mild ID and found to harbour a de novo deletion of the terminal two exons of PHF6.10 She showed facial dysmorphism with bitemporal narrowing, a high nasal ridge, a short nose, and hypertelorism. Furthermore, she had slender and tapering fingers with hyperextensible joints, short distal phalanges, distal non-fixed flexion deformities, short toes with syndactyly IV/V, hypoplastic nails, wide sandal gaps, and linear hyperpigmentation of the skin. X-inactivation testing showed skewing of 96% in blood lymphocytes.10
We now report on seven females with de novo mutations, deletions or duplications in PHF6 and a distinct, recognisable phenotype, thus broadening the clinical and mutational spectrum of BFLS.
Patients and methods
The clinical presentation in early infancy of this girl has been described elsewhere11; she had been diagnosed with Coffin–Siris syndrome (CSS) in early infancy because she had sparse hair, teeth anomalies (delayed dentition, small, wide-spaced, and irregularly shaped teeth), and hypoplastic nails and phalanges. Exome sequencing within a study to unravel the genetic causes of CSS did not detect a mutation in genes encoding for subunits of the SWI/SNF (SWItch/Sucrose Non-Fermentable) complex, but revealed the heterozygous de novo mutation c.914G>T, p.Cys305Phe in the PHF6 gene (P111).
Her age at last investigation was 11 years, when her height was 148.5 cm (0.7 SD), her weight was 39.4 kg (body mass index (BMI) 18), and her head circumference was 51 cm (−1.0 SD). She could speak in short sentences, write and read single letters, and do simple calculations with numbers up to 10. Her voice was deep and hoarse. On stairs her gait was unstable and she needed to hold on to support herself. Problems with aggressive behaviour, susceptibility to urinary tract infections, and frequent constipation were reported. Her hair at this time was no longer sparse. Facial dysmorphism included bitemporal narrowing, an accentuated arch to the eyebrows, synophrys, prominent supraorbital ridges, mild hypertelorism, a short nose with prominent columella, a wide mouth, and long and slightly posteriorly angulated ears (figure 1). Fingers were tapering and showed deformities with prominent and hyperextensible joints and extreme brachyclinodactyly V. Nails and distal phalanges of fingers and toes were hypoplastic (figure 2). Despite appropriate dental hygiene her teeth were prone to caries and some had to be extracted because of dental fistulas, probably resulting from infections of the apical periodontal area of the respective teeth. Also her permanent teeth were small, wide-spaced and irregularly shaped (figure 3), and her palate was narrow and high.
Conventional karyotyping on lymphocytes and fibroblasts as well as molecular karyotyping with an Affymetrix CytoScan HD-Array and sequencing of the ROR2 gene had been normal.
Again, the clinical presentation in early infancy of this girl has been described elsewhere,11 and she had been diagnosed with CSS in early infancy. Analyses within the study to unravel the genetic causes of CSS showed no that she did not carry a mutation in the genes encoding for the SWI/SNF complex, but revealed the de novo frameshift mutation c.677delG, p.Gly226Glufs*53 in PHF6 (P211).
Her latest assessment was at 9 years 11 months, when her height was 148 cm (1.7 SD), her weight was 30 kg (BMI 13.7), and her head circumference was 54 cm (1.4 SD). At this time she had mild ID with good speech abilities and could read, write, and calculate up to 100. A formal test at the age of 5 years 2 months revealed an IQ of 64 (SON-R 2½-7). She was reported to be very shy and did not tolerate being alone. Furthermore, sleeping difficulties were reported. She had mild clinodactyly and brachytelephalangy with hypoplastic nails, short toes with unilateral syndactyly of toes II and III (figure 2), hirsutism of her back, vertebral anomalies (wedge shaped vertebrae), and cone shaped epiphyses at digits II and V. During dental eruption temporary gingival protuberances had been observed, and her deciduous teeth had been reported to be small and pointed. Also her permanent teeth were small, wide-spaced, and irregularly shaped (figure 3). Facial dysmorphism included bitemporal narrowing, prominent supraorbital ridges, broad eyebrows, mild synophrys, mild hypertelorism, deep-set eyes, and full lips (figure 1). She had a hairy tragus and linear skin hyperpigmentation (figure 3).
Chromosome analysis was performed in blood lymphocytes and fibroblasts. Subtelomeric analysis and testing for fragile X syndrome were normal.
This patient is the second child of healthy, non-consanguineous parents with a non-contributory family history. During pregnancy the mother had dental treatment, and polyhydramnios was noted. The patient was born at 41 weeks gestation with a weight of 2900 g (−1.6 SD) and a length of 49 cm (−1.5 SD). At the age of 3 days she had pneumonia. She showed delayed milestones with unstable gait and single words at age 3 years 7 months. Dental eruption started at 5 months and subsequently irregularly shaped and discoloured teeth were noted, leading to the suspicion of dentin hypoplasia. At least two deciduous teeth were lacking and second dentition was delayed. Furthermore, sparse hair and a pre-auricular pit were present. Transient conductive hearing loss occurred due to recurrent otitis media and was temporarily treated with tympanostomy tubes and hearing aids. X-rays at the age of 6 years showed brachymesophalangy of fingers II and V, a cone shaped epiphysis at the first ray of the right foot, and medial clefts in two vertebral bodies. During childhood and youth, behavioural problems with short attention span, hyperactivity, compulsive behaviour, sleeping difficulties, and lack of emotional detachment were reported. Menarche started at 16 years, and subsequent oligomenorrhoea has been noted. Ophthalmologic examination showed reduced vision, possibly due to retinal dystrophy. Seizures occurred at the age of 20 years. Conventional karyotyping and subtelomeric screening by fluorescence in situ hybridisation (FISH) at the age of 12 years were normal.
She was last reviewed aged 22, when her height was 163 cm (0.3 SD). She had mild obesity with a weight of 74.9 kg (BMI 28.2), and her head circumference was 56.5 cm (1.4 SD). Her expressive speech was limited to about 10 single words, but she had better perceptive abilities and used signs. Facial dysmorphism included bitemporal narrowing of the forehead, a prominent supraorbital region with highly arched eyebrows and mild synophrys, long shaped ears with prominent earlobes and pre-auricular pits, a short nose with broad and bulbous nasal tip, a thin upper lip, and a dental prosthesis because of hypodontia. The fingers were tapering with pronounced brachyclinodactyly of finger V, lacking distal intradigital flexion creases, and non-fixed flexion deformities, particularly of finger II. The toes were broad and short with hypoplastic nails. A single palmar crease and linear skin pigmentation at the upper legs and axillar regions were noted (figures 1⇑–3).
This patient is the second child of healthy non-consanguineous parents with no significant family history. She was born at 34 weeks gestation with a weight of 2450 g (1.2 SD), a length of 49 cm (0.5 SD), and a head circumference of 34 cm (1.4 SD). At the age of 3–4 months several single seizures occurred. Sitting age was over 1 year, walking age 3 years, and also her speech development was severely delayed. Her behaviour has been reported to be friendly and sociable. The patient has recurrent problems with constipation and has a neurogenic bladder which has been treated with an artificial urinary outlet. Early loss of teeth occurred, requiring dental prostheses.
Her last assessment was at age 32, when her height was 165.9 cm (0.2 SD), her weight was 65.9 kg (BMI 23.9), and her head circumference was 54 cm (−0.3 SD). She had a low anterior hairline, bitemporal narrowing, prominent supraorbital ridges, highly arched eyebrows with synophrys, deep set eyes, fleshy earlobes, and a bulbous nasal tip (figure 1). Her fingers were tapering with brachyclinodactyly of fingers IV and V and hypoplastic nails V, and her feet were broad and short with broad and short toes with hypoplastic nails (figure 2). Linear hyperpigmentation was prominent over her upper legs and in the axillar regions (figure 3).
Conventional karyotyping on DNA from blood lymphocytes and fibroblasts had been normal, as was molecular karyotyping with an Affymetrix 6.0 SNP-Array. Due to linear hyperpigmentation, mosaicism had been suspected. X-inactivation testing on blood lymphocytes showed skewing (100%), while it was normal in fibroblasts (70%).
This patient was the first child born at 39 weeks gestation to parents without a significant preceding family history. Her birth weight was 3200 g (−0.6 SD). At 26 months her height and weight were 92.2 cm (2.2 SD) and 12.4 kg (BMI 15), respectively, and her head circumference measurement at age 44 months was 53 cm (2.2 SD). Bone age was found to be advanced by 6 months. She was 6 years 3 months at the time of her last review. The patient had moderate ID with a deep voice and no active words at age 5 years, but was using more than 50 Makaton signs. She was reported to be a happy child with sociable behaviour. Sparse hair, particularly in the temporal regions, was noted. Further dysmorphism included bitemporal narrowing of the forehead, prominent supraorbital ridges, long shaped ears with prominent earlobes, and pronounced fifth finger clinodactyly. Linear skin hyperpigmentation was documented. Because of the ID, routine diagnostic molecular karyotyping with a 60 K oligo-array was performed and identified a 6 kb deletion at Xq26.2 that affects exons 4 and 5 of the PHF6 gene (Xq26.2(133,349,355-133,355,618)x1). X-inactivation studies of the androgen receptor (AR) gene locus on DNA from blood lymphocytes showed 100% skewing.
This patient was born after 36 weeks of gestation with a weight of 2610 g (−0.3 SD), a length of 47 cm (−0.5 SD), and a head circumference of 34 cm (0.5 SD). She had delayed development and began walking at age 2.5 years. During the first years frontotemporal alopecia as well as nystagmus, strabismus, small teeth, hypoplastic labia minora and clitoris, a persistent ductus arteriosus, and bilateral cutaneous syndactyly of toes II/III and IV/V were noted. X-rays at the age of 8 months indicated a possible partial premature lambdoid synostosis and absence of the middle phalanx of the second toe of the right foot. Age at last investigation was 6 years. Her height was 116 cm (0.04 SD), her weight was 20 kg (BMI 14.86), and her head circumference was 49 cm (−1.63 SD). She presented with moderate to severe ID and could speak about 50 words. Difficulties with swallowing coordination were noted. She was reported to be a sociable and happy child. Facial dysmorphism included bitemporal narrowing, highly arched eyebrows, a bulbous nasal tip, a thin upper lip, and irregularly shaped teeth (figures 1 and 3). Furthermore, long and tapering fingers with pronounced clinodactyly IV and V were noted (figure 2). She had kyphoscoliosis and a deep and hoarse voice.
Brain MRI and EEG examination showed no anomalies. Conventional karyotyping on blood lymphocytes and fibroblasts were normal. Also array comparative genomic hybridisation (CGH) and testing for chromosomal breakage, Smith–Lemli–Opitz syndrome (sterols), Charge syndrome (CHD7), STAR syndrome (FAM58A), and FGFR1, FGFR2, and TWIST were normal.
This girl was born after an uneventful pregnancy at 40 weeks gestation by caesarean section. Her birth weight was 3320 g (−0.4 SD), her length was 52 cm (0.1 SD), and her head circumference was 35.5 cm (0.5 SD). She showed delayed motor milestones, sitting at 18 months and walking at 3 years. She spoke her first words at age 3 years. At age 5.5 years testing revealed an IQ of 50 (SON-R 2½-7). Her last assessment was at 7.5 years, when her height was 130.5 cm (0.6 SD), her weight was 33 kg (BMI 19), and her head circumference was 51 cm (0.6 SD). She attended a special school and behaved sociably. Severe sleeping difficulties at night and compulsive eating behaviour were reported, however. She had muscular hypotonia, an ectopic left kidney, strabismus, hyperopia, and was treated for a blocked lacrimal duct. Furthermore she had bilateral sensorineural hearing loss. MRI of the brain showed enlarged ventricles; the EEG examination was normal. Facial dysmorphism included bitemporal narrowing, a broad nasal root, mild epicanthus and mild asymmetry of the palpebral fissures, a short nose with bulbous nasal tip, and a prominent philtrum. Fingers were slender with brachydactyly V. Feet and toes were short with small and dystrophic nails and unilateral syndactyly of toes III and IV.
Testing for fragile X syndrome and conventional karyotyping gave normal results. Because of ID, routine molecular karyotyping was performed with a HumanOmni1-Quad-Array and revealed a deletion of the whole PHF6 gene (arr(hg18)Xq26.2(133,325,966- 133,417,052) x1 dn; 91 kb). Additionally, a paternally inherited 105 kb deletion 5q12.2-q12.3(63.639.646-63.744.887)x1 was noted. X-inactivation testing on DNA from blood lymphocytes showed 100% skewing.
The present study is part of larger studies to unravel the genetic causes for ID and CSS, which were approved by the ethics committees of the Medical Faculty, University of Erlangen-Nuremberg, and of the Medical Faculty, University Duisburg-Essen (ethical votum 12-5089-BO for CRANIRARE and 08-3663 for MRNET), respectively. Informed consent was obtained from parents or guardians of all patients.
Patients and samples
DNA from blood lymphocytes was extracted with standard methods. DNA from fibroblasts in one of the early passages and of lymphoblastoid cell lines was extracted with the Qiagen DNAeasy system (Qiagen, Hilden, Germany). For saliva collection and DNA extraction the OraGene system (DNA Genotek, Ontario, Canada) was used.
Molecular studies of patient 1 and 2 are described elsewhere (P1 and P211). Patients 3, 4, and 7 were specifically selected for PHF6 testing due to their phenotypic resemblance to patient 1.
For PHF6 deletion and duplication screening a further 20 male patients and one female patient with a suspected clinical diagnosis of BFLS but no mutation in the PHF6 gene by previous sequencing were obtained.
In patient 5 routine diagnostic molecular karyotyping with a 60 K oligo-array was performed because of ID and identified a 6 kb deletion at Xq26.2. Further confirmation was undertaken with a 105 K X chromosome exon-oligo-array confirming the exon 4 and 5 deletion and providing the breakpoint coordinates, Xq26.2(133,349,355-133,355,618)x1, within the PHF6 gene. Parental analysis with a 60 K oligo-array showed the Xq26.2 deletion to be de novo. In patient 7 molecular karyotyping was performed with a HumanOmni1-Quad-Array and revealed a deletion of the entire PHF6 gene (arr(hg18)Xq26.2(133,325,966- 133,417,052)x1 dn; 91 kb). The deletion was confirmed in the patient by quantitative real-time PCR using three amplicons (as published previously12), and excluded in the parents.
DNA samples were derived from peripheral blood, and all coding exons with exon–intron boundaries of PHF6 were screened for mutations by bidirectional direct sequencing (ABI PRISM BigDye Terminator Sequencing Kit v.3; Applied Biosystems) with an automated capillary sequencer (ABI3130XL Genetic Analyser, Applied Biosystems). De novo mutations were confirmed by testing the parents. Primer sequences are available on request.
PHF6 deletion/duplication screening
Deletion and duplication screening in PH56 was performed with quantitative real-time PCR on DNA samples. The Universal ProbeLibrary Assay Design Tool (Roche Applied Science) was used to design four probes distributed over the PHF6 gene (GRCh37/hg19): PHF6_1 (chrX 133507639 133507702), PHF6_2 (chrX 133527515 133527593 PHF6_2), PHF6_3 (chrX 133546975 133547049), and PHF6_4 (chrX 133558330 133558422). Analysis was performed on a LightCycler 480 (Roche). Primer sequences are available on request.
Multiplex ligation dependent probe amplification (MLPA)
MLPA for exons 3, 4, 5, and 6 of PHF6 was performed with self-designed probes and with the SALSA MLPA reagents EK5 and the SALSA MLPA P200 Human DNA reference-1 probemix (MRC-Holland, Amsterdam, The Netherlands). Copy number values were calculated using values from three healthy female control individuals and the Seqpilot software (JSI Medical Systems, Kippenheim, Germany). Probe sequences and conditions are available on request.
Amplification of the CAG repeat in exon 1 of the AR gene was performed by PCR with fluorescence tagged primers. Subsequently, digestion with the methylation sensitive enzyme HpaII was performed, and fragments were analysed with an automated capillary sequencer (ABI 3100; Applied Biosystems). Genescan and Genotyper Softwares (Applied Biosystems) were used to determine fragment sizes and intensities, and the degree of X-inactivation was calculated as described elsewhere.13
X-inactivation testing in patients 5 and 7 was initiated at the respective centres within routine diagnostics.
All the females in this report were intellectually disabled with varying severity from mild to severe. They showed strikingly similar facial features with bitemporal narrowing of the forehead, prominent supraorbital ridges with an accentuated arch to the eyebrows, narrow palpebral fissures, mild hypertelorism in the younger girls, a bulbous nasal tip and prominent columella, long and slightly posteriorly angulated ears, and a relatively large mouth (figure 1). Toes and fingers were short with hypoplastic distal phalanges or nails and a pronounced brachyclinodactyly of finger V (figure 2). Additional consistent findings were an abnormal dental status with small, irregularly shaped teeth that were prone to caries and premature loss, and also linear skin hyperpigmentation in six out of seven patients (figure 3). Clinical details of all patients are summarised in table 1.
As reported in more detail elsewhere, exome sequencing in patient 1 and her parents revealed the de novo missense mutation c.914G>T, p.Cys305Phe in exon 9 of the PHF6 gene (P111).
Direct sequencing of the PHF6 gene in patients 2 and 6 with a similar phenotype revealed the de novo frameshifting mutation c.677delG, p.Gly226Glufs*53 in exon 7 (P211), and the c.955C>T, p.Arg319* mutation in exon 9, respectively. Sequencing of PHF6 in patients 3 and 4, also selected due to their clinical resemblance to patient 1, revealed normal results. Subsequent screening for small, intragenic deletions and duplications by quantitative real-time PCR detected a de novo dosage gain of the probe PHF6_2, covering the boundary between intron 3 and exon 4 in both patients. These duplications and their de novo occurrence were confirmed by MLPA in the patients and by excluding them in the parents. With MLPA probes for exons 3 to 6, the duplication was determined to affect exons 4 and 5 in both patients. This duplication is, if in tandem, predicted to result in a frameshift. An overview on the identified defects in PHF6 is displayed in figure 4.
Deletion and duplication screening by quantitative real-time PCR on DNA samples of a larger group of 20 male patients and one female patient with suspected BFLS but previously normal PHF6 sequencing revealed no further aberrations.
X-inactivation testing in all seven patients revealed 100% skewing in DNA from blood lymphocytes. In patients 1, 2, 4, and 6, from whom DNA from fibroblasts was available, X-inactivation in these samples was normal (table 2).
We report on seven female patients with de novo mutations, deletions or duplications in PHF6, a gene that has previously been implicated in X-linked recessive BFLS. All the patients in this cohort display a distinct, recognisable phenotype that overlaps but is not identical with the phenotype observed in male patients with BFLS.
Accordingly, our patients were not initially considered to have a clinical diagnosis of BFLS. The deletions in patients 5 and 7 were detected by unbiased screening with molecular karyotyping because of ID. In patients 2, 3, 4, and 6 the clinical diagnosis was made based on clinical resemblance to patient 1 after detecting the PHF6 mutation in her. Targeted testing of PHF6 in these four patients revealed a de novo defect in all, thus confirming the distinctiveness and recognisability of the phenotype.
Our female patients display a distinct, recognisable facial gestalt (figure 1). Male patients with BFLS have some similar facial characteristics, including deep-set eyes, prominent supraorbital regions, and large ears with prominent earlobes.5 In both males and females the facial phenotype is also characterised by broad and prominent zygomata as well as by bitemporal narrowing. However, the facial appearance in the female patients described here with de novo PHF6 defects seems to be more distinct, particularly those with frontotemporal sparse hair, a more pronounced supraorbital region with bitemporal narrowing, prominent, highly arched eyebrows with synophrys and with less prominent ears. The high nasal root and bulbous nasal tip, observed in most of our patients, is not obvious in any of the published reports of males with PHF6 mutation.
The variability of ID ranging from mild to severe (table 1) in male patients with BFLS and this cohort of female patients shows some overlap. Additional features such as behavioural anomalies in some patients, short toes, and hypogonadism, which results in cryptorchidism and small genitals in males and hypoplastic clitoris (P1, P6) or oligomenorrhoea (P3 and P4) in females, are again overlapping (table 1).
In contrast, mild short stature or obesity, which are common in male individuals with BFLS, are not striking features in our patients. Their heights are in normal ranges, and only one of the two adult women (P3) and one of the younger patients (P7) have mildly increased BMIs (table 1). Most of our patients are too young, however, for us to make a definite statement on obesity, as this frequently only becomes apparent during adolescence.
In male patients with BFLS the fingers are described as tapering and hyperextensible. Clinodactyly of finger V or small nails have been reported in only two patients, respectively.4 ,5 ,8 In contrast, these are consistent and pronounced findings in our female cases. They show notable brachyclinodactyly, particularly of fingers IV and V, as well as non-fixed flexion deformities in the older patients, small or hypoplastic nails, and sometimes even severe brachytelephalangy (figure 2). In general, the foot anomalies seem to be more severe than the hand anomalies, this also being the case for male patients. Of note, while there is often generalised brachydactyly, the index digit appears to be particularly affected in many of the feet and some of the hands as well. Dental anomalies or sparse hair are only rarely reported in male patients with BFLS, yet are common findings and pronounced in our patients (table 1, figure 4 and 3). Dental anomalies seem to affect both deciduous and permanent teeth and are characterised by small and irregularly shaped, wide-spaced teeth. Furthermore, discolouring, susceptibility to caries, and occurrence of dental fistulas, probably resulting from periodontal infections, are observed. Hypodontia can result from both incomplete number of teeth as well as from premature loss.
Interestingly, the characteristic facial phenotype in the female patients seems to evolve with age, with clinically strikingly overlaps with CSS (MIM #135900) in younger children. Thus, it might not be easily distinguishable from CSS patients caused by mutations in the SWI/SNF complex components in early infancy. Two of the girls (P1 and P2) were suspected of having CSS in early infancy because of hypoplastic distal phalanges, sparse hair, and dental anomalies. CSS as well as overlapping Nicolaides–Baraitser syndrome (MIM #601358) and unspecific ID have recently been shown to be caused by mutations in genes encoding components of the SWI/SNF chromatin remodelling complex.14–,18 This complex regulates gene expression by using energy from ATP hydrolysis to alter chromatin structure around regulated genes in order to facilitate access of other transcription factors.19 PHF6 is widely expressed, and the protein contains two plant homeodomain (PHD)-like zinc finger domains.3 Similar domains are common in chromatin remodelling proteins that interact with post-translational modified histones.20–,22 In combination with its localisation in the nucleus, this points to a role for PHF6 in transcription.3 To date, however, little is known about the precise function of PHF6. Recently, its role in modifying chromatin structure and therefore possibly in gene regulation was confirmed by identification of an interaction with the NuRD complex, a multifunctional epigenetic regulator.23 Similar chromatin remodelling capacities between PHF6 and the SWI/SNF complex might therefore provide a possible explanation for the overlapping phenotype of patients with CSS and young girls with BFLS.
A further frequent finding in six of our seven female patients was linear skin hyperpigmentation, which suggested possible mosaicism, and we were able to confirm discrepancies of X-inactivation patterns in different tissues of identical patients (table 2). Activity of the mutated PHF6 allele in certain tissues might therefore be responsible for the manifestation of a rather severe phenotype in our female patients. However, some of the healthy carrier females in BFLS families do have skewed X-inactivation, while others do not.7 The available data on X-inactivation testing in female carriers is not extensive enough to deduce a possible correlation between skewed X-inactivation and severity of clinical signs.
Regarding the PHF6 mutations themselves there is no clear genotype–phenotype correlation. Only one particular splice site mutation resulting in a protein with retained function was discussed as a possible explanation for the mild phenotype in one family.9 The majority of mutations reported in families with BFLS are missense mutations which are distributed over the gene/protein without clustering to specific regions such as the functional domains (figure 4). A number of stop and frameshifting mutations have been reported, thus pointing to a loss-of-function mechanism. The PHF6 aberrations found in our female patients are novel and include a whole gene deletion, an intragenic deletion, two duplications predicted to result in frameshifting, a stop mutation, a frameshifting mutation, and a missense mutation located in exon 9 which encodes one of the PDH1-like zinc fingers. All of these mutations are predicted to have a severe effect on the protein, likely resulting in loss of function. On the one hand, this might provide a possible explanation for the quite severe phenotype in the female patients. On the other hand, one might also speculate that the mosaicism in our patients might have a specific effect on the phenotypic expression and be responsible for some of the symptoms that do not commonly occur in the male patients. This idea is supported by the observation that in one published case a girl with a de novo mutation in PHF6 had no linear skin hyperpigmentation and a typical BFLS appearance,7 while in a second published case a girl with a de novo deletion in PHF6 had linear skin pigmentation and a phenotype with a facial gestalt, dental anomalies, and pronounced brachyclinodactyly,10 manifestations more similar to those in our patients.
We were intrigued by detecting the same duplication of exons 4 and 5 in two patients. To investigate if this duplication might be due to a recurrent breakpoint and therefore be more frequent, we performed duplication/deletion screening in 21 patients with suspected BFLS but previously normal PHF6 sequencing. In this group we did not detect further intragenic dosage alterations.
We conclude that our findings show that de novo defects in PHF6 in females result in a recognisable phenotype which might have been under-recognised so far, and for which functional mosaicism of X-inactivation might be a contributing factor. It overlaps with the male BFLS phenotype but has specific, distinct features, thus adding a new facet to the phenotypic spectrum. Our findings might indicate that de novo mutations in supposedly X-linked recessive genes are possibly more common than previously expected.
We thank the participating patients and their families. We furthermore thank Daniela Falkenstein, Sabine Kaya, Regina Kubica, and Christine Zeck-Papp for excellent technical assistance.
Contributors The study was designed and results were interpreted by CZ, H-JL, and DW. LB, TC, FD, HE, GG-K, LG-N, DH, JH, WJ, and ARa collected and contributed clinical data, DNA samples and array results. CZ, CK, ARe, BW, MZ, and H-JL performed and contributed to the molecular studies. The manuscript was written by CZ, H-JL, and DW, and all authors contributed to the final version.
Funding IZKF Erlangen, Deutsche Forschungsgemeinschaft (DFG), and German Federal Ministry of Education and Research (BMBF). CZ was supported by the IZKF (Interdisziplinäres Zentrum für Klinische Forschung) Erlangen and by a grant from the Deutsche Forschungsgemeinschaft (ZW 184/1-1). This work was part of the CRANIRARE-2 Network funded by the German Federal Ministry of Education and Research (BMBF) by grant number 01GM1211B to DW, and part of the MRNET consortium also funded by a grant from the German Ministry of Research and Education to DW (01GS08167).
Competing interests None.
Patient consent Obtained.
Ethics approval Ethics committees of the Medical Faculty, University of Erlangen-Nuremberg, and of the Medical Faculty, University Duisburg-Essen.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.