Cardio-facio-cutaneous syndrome: Does genotype predict phenotype?

Abstract

INTRODUCTION

Cardio-facio-cutaneous (CFC) syndrome is a relatively rare sporadic multiple congenital anomalies/mental retardation condition with characteristic features that include congenital heart defects, a characteristic facial appearance, ectodermal abnormalities, gastrointestinal dysmotility that includes failure to thrive with severe feeding problems, moderate-to-severe intellectual disability, and short stature with relative macrocephaly. First described by Reynolds et al. [1986] in 8 patients, the syndrome has been the subject of many reports. The discovery of several causative genes (see below), has allowed a greater understanding of the breadth of clinical features. Little, however, has been published on genotype-phenotype correlations.

CFC shows considerable phenotypic overlap with Noonan and Costello syndromes, making clinical diagnosis challenging, especially in the young child. Over the past decade, it has been demonstrated that all 3 syndromes are caused by mutations in genes in the Ras-ERK signalling pathway; CFC by mutations in BRAF, MEK1, and MEK2; Noonan syndrome by mutations in PTPN11, SOS1, KRAS, RAF1, SHOC2, NRAS, and, occasionally, BRAF or MEK1; and Costello syndrome by HRAS mutations [Tartaglia et al., 2001; Aoki et al., 2005; Niihori et al., 2006; Rodriguez-Viciana et al., 2006; Nava et al., 2007; Pandit et al., 2007; Razzaque et al., 2007; Roberts et al., 2007; Tartaglia et al., 2007; Nystrom et al., 2008; Cordeddu et al., 2009; Cirstea et al., 2010]. Mutations in KRAS cause considerable phenotypic heterogeneity, and, in many individuals, the features are intermediate between those of Noonan and CFC syndromes [Zenker et al., 2007]. While individuals with mutations in SHOC2 have, in general, a distinct phenotype that represents a sub-type of Noonan syndrome and is easily recognized, several young children have presented with features quite characteristic of CFC [authors’ experience]. In up to a third of individuals with a clinical diagnosis of CFC, a mutation in one of the causative genes is not found [Rodriguez-Viciana et al., 2006; Nava et al., 2007; Narumi et al., 2007].

This clinical study of a cohort of 186 children and young adults with mutation-proven CFC is the largest to date and is focussed on the principal genes known to cause CFC, BRAF, MEK1 and MEK2. BRAF mutations are documented in 140 individuals (~75%), while 46 (~25%) have a mutation in MEK1 or MEK2. The age range is 6 months to 32 years, the oldest individual being a female from the original group reported in the seminal paper [Reynolds et al., 1986]. Fifty of the cohort are not previously published, but limited data on 136 are previously reported [Niihori er al., 2006; Narumi et al., 2007; Cave et al., 2007; Gripp et al., 2007; Armour and Allanson, 2008; Nystrom et al., 2008; Schultz et al., 2008]. While the methods of ascertainment vary from research group to research group, a core set of data have been gathered systematically, which provide new details of the breadth of phenotype and the frequency of particular features in each genotypic group.

METHODS

A core data set was established by a sub-group of authors (JA, BK and MZ). All international research consortia with an interest in CFC were approached and agreed to collaborate in assembling a large cohort of individuals with mutation-proven CFC. Each research consortium provided as complete a data set as possible for each patient. In many instances, the data provided exceeded information previously published. Ethics approval was obtained from the Research Ethics Boards of all collaborating institutions.

Individuals with a BRAF mutation were compared to individuals with a MEK mutation. Since the latter group was relatively small, aggregate data were chosen over separate MEK1 and MEK2 data. Results were expressed as a percentage: the number with a given feature compared to the total number for whom we had an informative answer (yes or no). Where no data were available, the individual was not included in the denominator.

Genotype-phenotype differences were evaluated using Fisher’s exact test with two-tailed significance. Bonferroni correction was used. Statistical significance was defined as a p-value less than 0.05

RESULTS

DISCUSSION

This is the largest study of CFC syndrome carried out to date, made possible by an international effort to share clinical and molecular data and collaborate on a number of research endeavors, including gene discovery and evaluation of genes in model organisms. Many of the individuals in this study have been previously reported [Niihori er al., 2006; Narumi et al., 2007; Cave et al., 2007; Gripp et al., 2007; Armour and Allanson, 2008; Nystrom et al., 2008; Schultz et al., 2008] but the systematic collection of clinical data for this study has, in many instances, increased what is known about those individuals. In addition, there are data on 50 unreported persons. The size of this cohort allows a robust genotype-phenotype comparison.

Few studies of genotype-phenotype correlation have been carried out to date. Nava et al. [2007], in a mixed cohort of children with CFC, Noonan and Costello syndromes, compared those with BRAF and MEK mutations, noting less frequent heart defects and milder motor delays in the latter group, two of whom had normal intelligence. The comparison with our study data is complicated, however, by the fact that 2 of the 3 children with a MEK mutation reported by Nava and colleagues carried a clinical diagnosis of Noonan syndrome. Schultz et al. [2008] reported BRAF and MEK mutations in 24 and 8 individuals with CFC, respectively, but failed to show phenotypic differences between the 2 mutation-specific groups. Dentici et al. [2009] reported 6 individuals with CFC and a MEK mutation and compared their features to individuals with MEK mutation in the literature. The 6 new cases did not differ with respect to phenotype.

Many of the clinical features described herein are in keeping with data from recently described series [Gripp et al., 2007; Narumi et al., 2007; Nava et al., 2007; Armour and Allanson, 2008] and the CFC index proposed by Kavamura et al. [2002]. Cardiac abnormalities were seen with similar frequency (see Table X). Arrhythmias were quite uncommon in this cohort with CFC: with 4 reports of supraventricular tachycardia, and one each of ventricular extrasystoles, AV block and Wolf-Parkinson-White syndrome, in contrast to the findings in Costello syndrome where they occur in almost half the affected individuals [Lin et al., 2011]. Intellectual disability was universal in 3 previous studies [Armour and Allanson, 2008; Nava et al., 2007; Narumi et al., 2007]. Our study had a small cohort which had undergone full psychometric testing, documenting normal intelligence in just one individual with a MEK mutation and borderline IQ in one with a BRAF mutation. Further details on specific aspects of cognition were not available. The structural central nervous system findings are similar in type and frequency to those in other studies. A recent review confirmed ventriculomegaly, hydrocephaly and cortical atrophy as the most frequent imaging findings [Papadopoulou et al., 2011]. It is difficult to compare skin findings between studies since the categories are presented differently. In large part, data seem comparable with prior studies, but a new study suggests that nevi and keratosis pilaris are more common than found in our cohort or previously reported (60% and 80% respectively) [Siegel et al., 2010]. The presence of normal or large birthweight with postnatal growth retardation and subsequent short stature is also fairly consistent across studies.

Previous studies report differences in the likelihood of polyhydramnios, hypotonia and failure to thrive (see Table X) [Armour and Allanson, 2008; Narumi et al., 2007; Nava et al., 2007; Gripp et al., 2007]. These are likely related, in part, to methods of ascertainment. Our study data support the high likelihood of significant gastrointestinal dysmotility and optic nerve hypoplasia documented by Armour and Allanson [2008], but not described in previous series.

Despite the fact that BRAF is a proto-oncogene and somatic mutations of BRAF have been identified in 7% of cancers [Makita et al., 2007], there are few published reports of neoplasia in CFC. The only malignancy in this series has been previously published by Al-Rahawan et al. [2007]. This 3-year-old boy with a MEK1 Y130C mutation had undergone a cardiac transplant at age 8 months for hypertrophic cardiomyopathy. He died shortly after an intra-cardiac mass was diagnosed as metastatic hepatoblastoma. It is unclear whether the post-transplant immune-suppressive therapy played a role in tumor development. There are 3 other individuals with molecularly confirmed CFC syndrome and malignancy. Acute lymphoblastic leukemia was diagnosed in 2 [van den Berghe and Hennekam, 1999; Niihori et al., 2006; Makita et al., 2007]. Both had BRAF mutations that have been reported in other individuals with CFC syndrome without accompanying malignancy. Non-Hodgkins lymphoma was reported in one [Ohtaki et al., 2010]. One boy with a BRAF mutation and a parasagittal meningioma is known to the support group CFC International. While multiple giant cell lesions are benign, they are tumor-like lesions probably driven by the proliferative effect of enhanced activity through the Ras-MAPK pathway, and are reported in association with a variety of pathway genes including BRAF [Neumann et al. 2009].

This study has limitations. Data are provided by 9 different research consortia, each of which has its own ascertainment method. While a standard set of clinical data on each subject has been sought, the quantity of data on each varies as it was not possible to re-evaluate everyone to ensure all details could be provided. In addition, a small subset of data published by Armour and Allanson [2008] was derived from parental questionnaire, introducing the possibility of recall bias. Those parents were members of CFC International. Medical records were not systematically collected for this study and some children had seen pediatric subspecialists while others had not. Parents who seek membership of such support groups may have children with greater needs or may be more inclined to seek out subspecialty resources. Lastly, the small size of the cohort with a MEK mutation does not allow meaningful comparison between MEK1 and MEK2 phenotypes. We are unable to assess differences between these 2 genotypes, between BRAF and MEK1 or BRAF and MEK2. Our study of CFC continues and we hope to be able to address this deficiency in the future.

This study reports the most frequent medical issues in 186 individuals with mutation-proven CFC syndrome. Knowledge of the causative mutation allows confidence in the diagnosis and, more importantly, comparison of the 2 genotypic groups. While not reaching statistical significance, it appears a mutation in MEK1 or MEK2 is associated with a higher likelihood of prematurity, absolute macrocephaly, ventricular septal defect, keratosis pilaris, pectus deformity, cryptorchidism and a renal anomaly. Conversely, there is a lower likelihood of atrial septal defect and hypertrophic cardiomyopathy, curly or sparse hair, severe intellectual disability, serious and long-lasting gastrointestinal dysmotility leading to failure to thrive and the need for assisted feeding, optic nerve hypoplasia/dysplasia, and kyphosis. It is important to note that only the difference in frequency of pulmonary stenosis reached statistical significance. With time and the increasing availability of reasonably-priced molecular testing, children and adults with milder features will come to attention and these genotype-phenotype data will evolve.

ACKNOWLEDGMENTS

Biography

Footnotes

Contributor Information

Judith E Allanson, Department of Genetics, Children’s Hospital of Eastern Ontario, Ottawa, ON, Canada. Ph: 613-737-2233. Fax: 613-738-4220. ac.no.oehc@nosnalla.

Göran Annerén, Department of Immunology, Genetics and Pathology, Uppsala University, Sweden. es.uu.pgi@nerenna.narog.

Yoki Aoki, Department of Medical Genetics, Tohoku University School of Medicine 1-1, Seiryomachi, Aobaku, Sendai 980-8574, JAPAN. Phone:81-22-717-8140, FAX:81-22-717-8142. pj.ca.ukohot.dem@yikoa.

Christine M Armour, Medical Genetics Unit, Kingston General Hospital, Kingston, ON Canada, ten.irak.hgk@cruomra.

Marie-Louise Bondeson, Department of Immunology, Genetics and Pathology, Uppsala University, Sweden. es.uu.pgi@nosednoB.esiuoleiraM.

Helene Cave, Hôpital Robert Debré, Département de Génétique; Université Paris 7-Denis Diderot, Paris, France. rf.phpa.bdr@evac.eneleh.

Karen W Gripp, Division of Medical Genetics, Alfred I. duPont Hospital for Children, Wilmington, DE 19803, USA. gro.sruomen@ppirgk.

Bronwyn Kerr, Genetic Medicine, St Mary’s Hospital, Manchester, UK. ku.shn.tfmc@rreK.nywnorB.

Anna-Maja Nystrom, Department of Immunology, Genetics and Pathology, Uppsala University, Sweden.

Katia Sol-Church, Nemours Biomedical Research and Nemours Center for Applied Clinical Genetics, Alfred I. DuPont Hospital for Children, Wilmington, DE 19803, USA. gro.sruomen@ruhclosk.

Alain Verloes, Hôpital Robert Debré, Département de Génétique; Université Paris 7-Denis Diderot, Paris, France. rf.phpa.bdr@seolrev.niala.

Martin Zenker, Institute of Human Genetics, University Hospital Magdeburg, Leipziger Str. 44, 39120 Magdeburg, Germany, phone: ++49 391 6715064, fax: ++49 391 6715066, ed.ugvo.dem@reknez.nitram.

REFERENCES