CHRNG genotype–phenotype correlations in the multiple pterygium syndromes
- Julie Vogt1,2,
- Neil V Morgan1,
- Pauline Rehal2,
- Laurence Faivre3,
- Louise A Brueton2,
- Kristin Becker4,
- Jean-Pierre Fryns5,
- Sue Holder6,
- Lily Islam6,
- Emma Kivuva7,
- Sally Ann Lynch8,
- Renaud Touraine9,
- Louise C Wilson10,
- Fiona MacDonald2,
- Eamonn R Maher1,2
- 1Centre for Rare Diseases and Personalised Medicine and Department of Medical and Molecular Genetics, School of Clinical and Experimental Medicine, University of Birmingham, Birmingham, UK
- 2West Midlands Regional Genetics Service, Clinical Genetics Unit, Birmingham Women's Hospital, Birmingham, UK
- 3Centre Hospitalier Universitaire Dijon, Plateau Technique de Biologie, Dijon Cedex, France
- 4Labor Krone, Bad Salzuflen, Germany
- 5Center for Human Genetics, University of Leuven, Leuven, Belgium
- 6North West Thames Regional Genetics Service, The North West London Hospitals NHS Trust, Middlesex, UK
- 7Peninsular Regional Genetics Service, Clinical Genetics Department, Royal Devon and Exeter Hospital, Exeter, UK
- 8National Centre for Medical Genetics, Our Lady Hospital for Sick Children, Crumlin, Dublin 12, Ireland
- 9Service de Genetique, CHU-Hopital Nord, 42055 Saint Etienne Cedex 2, France
- 10North East Thames Regional Genetics Service, Clinical Genetics Department, Great Ormond Street Hospital, Great Ormond Street, London, UK
- Correspondence to Dr Julie Vogt, Clinical Genetics Unit, Birmingham Women's Hospital, Metchley Park Road, Edgbaston, Birmingham, B15 2TG, UK;
Contributors JV and ERM conceived, designed and coordinated the study. LF, LB, KB, J-PF, SH, LI, EK, SL, RT and LW contributed to patient enrolment in the study and provided patient information and samples. JV, ERM, NM, PR and FM analysed and interpreted the data. JV and ERM drafted the article. All authors have read and approved the final manuscript.
- Received 31 July 2011
- Revised 14 October 2011
- Accepted 28 October 2011
Background Germline mutations in the CHRNG gene that encodes the γ subunit of the embryonal acetylcholine receptor may cause the non-lethal Escobar variant (EVMPS) or the lethal form (LMPS) of multiple pterygium syndrome (MPS). In addition CHRNG mutations and mutations in other components of the embryonal acetylcholine receptor may present with fetal akinesia deformation sequence (FADS) without pterygia.
Methods In order to elucidate further the role of CHRNG mutations in MPS/FADS, this study evaluated the results of CHRNG mutation analysis in 100 families with a clinical diagnosis of MPS/FADS.
Results CHRNG mutations were identified in 11/41 (27%) of families with EVMPS and 5/59 (8%) with LMPS/FADS. Most patients with a detectable CHRNG mutation (21 of 24 (87.5%)) had pterygia but no CHRNG mutations were detected in the presence of central nervous system anomalies.
Discussion The mutation spectrum was similar in EVMPS and LMPS/FADS kindreds and EVMPS and LMPS phenotypes were observed in different families with the same CHRNG mutation. Despite this intrafamilial variability, it is estimated that there is a 95% chance that a subsequent sibling will have the same MPS phenotype (EVMPS or LMPS) as the proband (though concordance is less for more distant relatives). Based on these findings, a molecular genetic diagnostic pathway for the investigation of MPS/FADS is proposed.
- clinical genetics
- molecular genetics
- genetic screening/counselling
The multiple pterygium syndromes (MPS) are a spectrum of phenotypically heterogenous disorders recognisable by the presence of joint contractures (arthrogryposis), pterygia (skin webbing across the joints), typical facial appearance and a variety of other congenital anomalies. Both lethal (LMPS, MIM25390) and non-lethal (Escobar variant, EVMPS, MIM 265000) forms occur.1 In addition to pterygia, individuals with MPS may have features of the fetal akinesia deformation sequence (FADS) or developmental defects (joint contractures, micrognathia, cleft palate, and lethal pulmonary hypoplasia in LMPS). Cystic hygroma or increased nuchal translucency with the development of hydrops fetalis may be detected on ultrasonography in pregnancy.
MPS are genetically heterogeneous. They are most commonly inherited in an autosomal recessive fashion, though autosomal dominant2 3 and X-linked recessive pedigrees have been described.4 Recently, genetic studies in autosomal recessive families have identified mutations in CHRNG, which encodes the γ subunit of the embryonic acetylcholine receptor in families with LMPS and/or EVMPS.5 6
Following the discovery of mutations in CHRNG in MPS, we undertook a genotype–phenotype study to investigate which clinical features might predict the likelihood of detecting a CHRNG mutation and whether the mutation spectrum differed between LMPS and EVMPS kindreds.
Patients and methods
One hundred families with at least one affected individual with clinical features of EVMPS, LMPS or FADS were recruited to this study. Entry criteria specified EVMPS/LMPS or FADS of unknown cause and so the cohort is representative of this group of families, but not of all cases of these disorders. Ethical approval and valid consent were obtained. DNA samples from the probands and, where possible, the parents were collected. Mutation analysis of all 12 coding exons of CHRNG was performed by direct sequencing. Homozygous CHRNG mutations c.320T→G (p.Val107Gly), c.136C→T (p.Arg46X), c.401_402delCT (p.Pro134Argfsx43), c.753_754delCT (p.Val253AlafsX44), c.459dupA (p.Val154SerfsX24), and possible mutation IVS4-9T→C (c.351-9T→C) in six families (15 affected individuals) were described previously.5
A standardised questionnaire was used to gather phenotypic information including details of family history, ethnicity, and perinatal and postnatal history. The results of relevant investigations, including brain scans and muscle biopsy results, were recorded where available. Genotype–phenotype correlations were evaluated by subdividing the cohort into non-lethal and lethal patients and CHRNG mutation positive/negative groups. In families with multiple affected siblings, an analysis was undertaken to estimate the likelihood of a further affected pregnancy with a similar phenotype.
Analysis was undertaken using Fisher's exact test. Statistical significance was taken at the 5% level.
Clinical classification and mutation detection rates
In 41 kindreds (containing 54 affected individuals) the proband presented with EVMPS and in 59 kindreds (84 affected individuals) the presentation was LMPS/FADS. Eleven (27%) of the EVMPS kindreds had a detectable CHRNG mutation. Five (8%) of the LMPS/FADS kindreds harboured a CHRNG mutation. In 31% of families with CHRNG mutations there was parental consanguinity (3/11 with EVMPS and 2/5 with LMPS/FADS) and in 69% the parents were unrelated (8/11 with EVMPS and 3/5 with LMPS/FADS).
Comparison of clinical features of 41 Escobar variant MPS kindreds and 59 LMPS/FADS kindreds
Joint contractures were universal and pterygia were present in 50/54 (93%) of the Escobar group and in 58/84 (69%) of the LMPS/FADS group. When the prenatal features were compared, cystic hygroma was present in 42/84 (50%), fetal hydrops in 39/84 (46%) and pulmonary hypoplasia in 30/84 (36%) of the LMPS/FADS cases. In contrast, cystic hygroma and fetal hydrops were rarely reported (1/54, 2%) in the non-lethal patients (p<0.0001 and p<0.0001) and pulmonary hypoplasia (2/54, 4%) was only occasionally recognised (p<0.0001). One child died at age 4 months following respiratory distress from birth and an associated diaphragmatic eventration, and another was a 3-day-old neonatal death. Reduced fetal movements were reported in 49/84 (58%) LMPS/FADS patients and 14/54 (26%) EVMPS patients (p=0.0002), while intrauterine growth retardation was present in 35/84 (42%) of the LMPS/FADS cases and 8/54 (15%) of the EVMPS cases (p=0.0012). Abnormalities were detected on ultrasound examination in 56/84 (67%) of the lethal cases compared with 3/54 (6%) of the non-lethal families (p<0.0001) (figure 1).
Cleft palate was identified in 32/84 (38%) of the lethal compared with 7/54 (13%) of the non-lethal group (p=0.0017). Central nervous system manifestations were more commonly noted in the lethal patients (14/84, 17%) compared with the non-lethal group (4/54, 7%) (p>0.1). Reduced muscle bulk was observed in 24/84 (29%) of the lethal families compared with 4/54 (7%) of the Escobar cases (p=0.0023). Poor postnatal growth was reported in 20/54 (37%) of the non-lethal MPS families.
Clinical features and CHRNG mutation detection
In the EVMPS cohort the clinical features of the CHRNG mutation positive and negative cases were generally similar (table 1).
Clinical features in the CHRNG mutation positive and CHRNG mutation negative LMPS/FADS patients
Central nervous system developmental anomalies were not observed in the CHRNG mutation group (0/7), but were identified in 14/77 (18%) (p>0.1) of those without a CHRNG mutation. A variety of central nervous system (CNS) malformations were present, and more than one structural abnormality was described in several patients. Of the various CNS anomalies, around 30% involved the cerebellum, with cerebellar hypoplasia in 5/77 (6%) of cases. Polymicrogyria was also documented in 4/77 (5%) and ventricular dilatation was reported in 3/77 (4%). Cleft palate was not observed in the CHRNG positive group but was reported in 32/77 (42%) of those without a CHRNG mutation (p=0.041) (table 2).
Comparison of the clinical features of combined EVMPS and LMPS/FADS patients according to the presence or absence of a detectable CHRNG mutation
Twenty-one of 24 (86%) CHRNG mutation positive MPS patients had pterygia compared with 86/114 (75%) in the mutation negative group (p=0.28). Cleft palate was observed in 1/24 (4%) of the CHRNG positive patients and 38/114 (33%) of the CHRNG negative MPS patients (p=0.0025). CNS malformations were not associated with a CHRNG mutation (0/24 vs 20/114) (p=0.024).
The 20 CHRNG mutations detected in our cohort are listed in table 3. In 12 patients the mutations were homozygous while four patients were compound heterozygous for the CHRNG mutations. Seven of the 20 mutations had not been reported previously. Mutations were distributed throughout the gene, with a hotspot identified in exon 5 (figure 2). The recurrent mutation c.459dupA (p.Val154SerfsX24) was identified in six kindreds from different ethnic backgrounds affected with both EVMPS and LMPS. It was found in the homozygous state and with a second CHRNG mutation in two compound heterozygous individuals. The other less frequently observed recurrent CHRNG mutation c.753_754delCT (p.Val253AlafsX44) in exon 7 also occurred as a homozygous mutation and in compound heterozygous individuals. It was present in three families with both the non-lethal and lethal phenotypes.
Haplotype analysis with microsatellite markers (D2S1363, D2S2193 and D2S2344) suggested that the three families with the CHRNG c.459dupA mutation did not share a common founder mutation (table 4).
Intrafamilial variation in CHRNG mutation positive families
In our cohort there were 41 kindreds in which the proband presented with an EVMPS phenotype. All 10 affected siblings born subsequently also had an EVMPS phenotype. In one CHRNG positive family, although the first and second affected siblings died in the neonatal period and early infancy, the third affected sibling did not have lethal MPS.5 However, as this was a single family, the significance of this is unclear. It is uncertain how the lethality was influenced by the severity of the congenital malformations and the availability of medical care. This family also had four additional cousins with an Escobar type MPS phenotype and one deceased cousin. In another family, the cousin of two siblings with non-lethal MPS was severely affected and died in infancy. Within the CHRNG mutation negative group in one kindred an affected child who died at 3 years old (about which little detail is available) had an affected second sibling who survived. There was also one family with three non-lethal MPS siblings. This family had two cousins who died in early childhood about which there are no further details.
Our cohort included 59 kindreds in which the proband presented with LMPS/FADS. All of the 25 affected siblings had an LMPS/FADS phenotype (25/25). This included six families with three lethally affected siblings. In one CHRNG mutation positive kindred a sibling died in the neonatal period while the other was diagnosed with fetal hydrops in early pregnancy. Although the presentations were variable, they were both in the lethal spectrum. Overall, the chance of a similar MPS phenotype arising in further affected siblings is estimated to be 10/12+25/25=35/37=95% (95% CI 81% to 99%).
The identification of mutations in the γ subunit of the acetylcholine receptor (CHRNG) gene in a subset of patients with MPS has facilitated diagnosis and genetic counselling of affected families. In addition the finding also established that the pterygia and other developmental anomalies in these cases are a consequence of fetal akinesia in early prenatal life. Depending on the timing and the severity of the disturbance to the neuromuscular junction (NMJ) formation, mutations in acetylcholine receptor subunit genes may cause a spectrum of conditions ranging from lethal MPS/FADS7 to postnatal congenital myasthenic syndrome.8 CHRNG mutations cause a distinctive phenotype because although fetal akinesia in early to mid gestation is severe, the γ subunit is not a component of the adult acetylcholine receptor (the switch from embryonal to adult occurs in the third trimester); therefore, postnatal muscle weakness and fatigability seen in other acetylcholine receptor deficiency conditions, such as the congenital myasthenic syndromes, is not a feature of CHRNG mutation positive EVMPS.
We found CHRNG mutations in only a minority of EVMPS and LMPS/FADS kindreds (23% and 8%, respectively). Two limitations to our estimate of the prevalence of CHRNG mutations are that as LMPS/FADS families with alternative diagnoses (detected in about 50%9) were excluded, the overall CHRNG mutation detection rate in all cases would be lower. However, conversely, heterozygous exonic deletions and mutations outside of the coding exons and flanking sequences would not have been detected by our mutation detection strategy.
We did not identify a clear difference in mutation spectrum between EVMPS and LMPS/FADS cases. Neither the mutation type nor its position in the gene appeared to correlate with the severity of the phenotype. Indeed, the same mutation was identified in both the lethal and Escobar variant cases in different families. For example, the c.459dupA mutation was found in individuals with non-lethal MPS as well as in lethal cases. A possible explanation for this is the effect of genetic and environmental modifiers and this interpretation would be consistent with the more pronounced interfamilial (vs intrafamilial) phenotypic variability. The CHRNG mutations c.459dupA and c.753_754delCT were recurrent in our study group. However, there was no evidence to support a common haplotype.
Cystic hygroma and fetal oedema are frequent in LMPS/FADS.10 11 Pterygia have been observed before the development of fetal oedema, and although pterygia are regarded as a consequence of hypokinesia,12 fetal oedema may further inhibit movement and exacerbate akinesia and muscle hypoplasia, predisposing to the associated congenital anomalies and the typical craniofacial features.11 The absence of cystic hygroma and fetal hydops from the EVMPS cohort suggests that these are markers for more severe fetal akinesia. Pulmonary hypoplasia was not observed in the index EVMPS cases but was detected in 2/12 siblings—though this may reflect more detailed search after the birth of an affected child.
Cleft palate was identified in over a third of the LMPS/FADS patients but was less common in the EVMPS cases. Cleft palate was observed in 1/17 (6%) of the CHRNG positive patients and 6/37 (16%) of the CHRNG negative EVMPS patients.
The presence of CNS anomalies in LMPS/FAD suggests that a CHRNG mutation is unlikely. Thus, whereas CNS malformations were not reported in those with a CHRNG mutation, about a fifth of the CHRNG negative group had one or more CNS anomalies (most commonly cerebellar hypoplasia, polymicrogyria or ventricular dilatation). As reported previously, growth retardation is frequently observed in LMPS/FADS,13–15 but poor postnatal growth was reported in over a third of the non-lethal MPS families (table 1). Almost all those MPS patients with CHRNG mutations had pterygia (21/24) and thus pterygia predicted the presence of a CHRNG mutation.
Phenotypic variation of affected children within families as well as between families is well documented in MPS.16 We have attempted to address the clinically important question of the chance of a lethally affected sibling being born to a family with a child with EVMPS, or the chance of a non-lethally affected baby arising in a family who have previously had a lethally affected baby. Our analysis suggests that there is generally very good concordance between siblings, although the prediction may be less accurate for more distantly related individuals.
After identifying mutations in CHRNG as a cause of LMPS/FADS, we subsequently demonstrated that mutations in RAPSN and DOK7 could cause a similar phenotype.5 7 17 Although the number of cases identified with RAPSN and DOK7 mutations are small (n=3 and n=3, respectively), we note that unlike CHRNG mutation positive patients, in whom pterygia were almost universally present, those with mutations in RAPSN and DOK7 did not have pterygia (this may reflect more severe akinesia in early fetal life in the former).
Intrauterine growth retardation was observed in all except the RAPSN group. This may indicate that mutations in RAPSN have less severe effects on early fetal development and fetal growth. Siblings described in a further family reported with RAPSN mutations were not lethally affected, although they had severe respiratory distress. CHRNA1 and CHRND have been found in cases of LMPS/FADS. Prenatal fetal akinesia was detected in all of these patients. Pterygia and the presence of congenital anomalies such as cleft palate and cardiac defect were reported. CNS malformations were not a feature.18
Based on current knowledge of the underlying molecular pathology, we suggest that CHRNG mutation analysis should be offered to all families with EVMPS consistent with autosomal recessive inheritance. For cases of LMPS/FADS with CNS anomalies the detection rate for CHRNG, CHRNA1, CHRND, RAPSN, and DOK7 mutations is likely to be small and so we suggest that other diagnoses should be investigated. In cases of LMPS/FADS with pterygia then CHRNG mutation analysis should be performed first (followed by CHRNA1 and CHRND). If pterygia are not present but an embryonal acetylcholine receptor defect is suspected, then RAPSN and DOK7 mutation analysis may be pursued before CHRNG, CHRNA1 and CHRND (though the mutation detection rate for each individual gene is not high). The predicted gene specific mutation detection rates in different subgroups of LMPS/FADS patients are summarised in figure 3 (note the wide confidence intervals due to the small number of data points in some cases). Nevertheless, as second generation sequencing techniques are advanced and utilised for diagnostic testing, and further EVMPS/LMPS/FADS genes are identified, it will become less important to prioritise individual genes for analysis, and a parallel sequencing of a large number of genes is likely to be adopted because of the extensive locus heterogeneity in LMPS/FADS.
We thank the families and referring clinicians for their help with this research.
Competing interests None.
Ethics approval Ethics approval was provided by South Birmingham Ethics Committee.
Provenance and peer review Not commissioned; externally peer reviewed.