RASopathies: Clinical Diagnosis in the First Year of Life

In the last 10 years, germline mutations in a number of genes coding transducers and modulatory proteins participating in the RAS-MAP kinase-signaling path- way have been causally linked to Noonan syndrome (NS) and a group of clinically related disorders, the so-called neuro-cardio-facio-cutaneous syndromes or RASopathies [Schubbert et al., 2007; Tidyman and Rauen, 2009; Tartaglia et al., 2010]. These disorders are characterized by facial dysmorphism, a wide spectrum of cardiac disease, postnatal reduced growth, ectodermal and skeletal defects, and variable cognitive deficits. Mutations in PTPN11, SOS1, KRAS, NRAS, RAF1 and BRAF genes have been documented to account for approximately 75% of individuals with NS [Tartaglia et al., 2001, 2007; Carta et al., 2006; Schubbert et al., 2006; Pandit et al., 2007; Razzaque et al., 2007; Roberts et al., 2007; Sarkozy et al., 2009a; Cirstea et al., 2010], the commonest and clinically most variable among these disorders [Allanson, 1987; Noonan, 1994; van der Burgt, 2007]. LEOPARD syndrome (LS) [Gorlin et al., 1969; Voron et al., 1976; Sarkozy et al., 2009b] and Noonan-like syndrome with loose anagen hair (NS/LAH) [Mazzanti et al., 2003], which phenotypically resemble NS, have been shown to be caused by mutations in the PTPN11, RAF1 and BRAF [Digilio et al., 2002; Legius et al., 2002; Pandit et al., 2007; Sarkozy et al., 2009a], and SHOC2 genes [Cordeddu et al., 2009], respectively, while defects in the CBL gene have been linked to a variable condition partially overlapping NS [Martinelli et al., 2010; Niemeyer et al., 2010; Perez et al., 2010]. Other disorders more severely affecting development and growth, but with clinical overlap with NS, include cardiofaciocutaneous syndrome (CFCS) [Reynolds et al., 1986; Roberts et al., 2006] and Costello syndrome (CS) [Costello, 1977; Hennekam, 2003; Gripp, 2005]. Both conditions are caused by defects in genes functionally related to those implicated in NS, the former being associated with mutations affecting the KRAS, BRAF, MEK1 and MEK2 genes [Niihori et al., 2006; Rodriguez-Viciana et al., 2006], and the latter being caused by a relatively narrow spectrum of HRAS mutations [Aoki et al., 2005].

In general, the clinical diagnosis of a condition with clinical traits fitting a specific RASopathy can be difficult in the first months of life, since most cardinal features utilized to categorize NS, LS, NS/LAH, CFCS and CS manifest later during childhood [Digilio et al., 2006a, 2007]. Nevertheless, a detailed definition of the neonatal phenotype and degree of variation characterizing these disorders would represent a clinically relevant tool to guide pediatricians and medical geneticists towards an earlier clinical diagnosis of general RASopathies and specific subtypes, prompting molecular characterization and more effective patient management and counseling.

Here, we review the clinical records referred to the first year of life of 57 patients clinically diagnosed as having a RASopathy and subsequently molecularly genotyped, in order to outline clinical features facilitating early diagnosis of these disorders, and add information on clinical history referred to the first months of life.

From 1990 to 2010, 134 patients with clinical features fitting the RASopathy spectrum were evaluated in their first year of life. Within this cohort, genomic DNA obtained from circulating leukocytes was available for molecular analysis in 96 subjects, for whom no obvious selection bias was apparent. The entire coding sequence of the PTPN11 gene was screened for mutations in all patients by single-strand conformation polymorphism analysis or denaturing high-performance liquid chromatography, as previously reported [Digilio et al., 2002; Tartaglia et al., 2002]. Fragments with an aberrant migration/elution pattern were sequenced. PTPN11 mutation-negative samples were successively screened for mutations in the coding region of the SOS1, RAF1, KRAS, NRAS, SHOC2, CBL, BRAF, MEK1, MEK2, and HRAS genes by denaturing high-performance liquid chromatography analysis and bi-directional direct sequencing. Disease causative changes were identified in 57 patients (table 1). PTPN11 mutations were detected in 31 subjects (54% of mutation-positive cases), allowing to obtain a molecularly confirmed diagnosis of NS and LS in 20 and 11 individuals, respectively. Heterozygous mutations in SOS1 (n = 8, 14%), RAF1(n = 4, 7%), BRAF(n = 1, 2%), NRAS(n = 1, 2%) and CBL(n = 1, 2%) were found in subjects with a phenotype fitting in with NS. A missense BRAF mutation was documented in 4 patients with a diagnosis of CFCS (7%), and one MEK2 mutation was found to occur in 1 additional CFCS case. Finally, mutations in HRAS and SHOC2 were observed in 4 patients with CS (7%), and 2 subjects with NS/LAH (4%), respectively. None of the analyzed patients carried a mutation in the KRAS gene.

Familial cosegregation of the trait and mutation was documented in 8 families (14% of genotyped cases), including 4/20 NS cases and 1/11 LS cases with PTPN11 mutations, 2/8 cases with SOS1 mutations, and in the single case with CBL mutation. In 6 of the 8 families, the mutated allele was transmitted by the affected mother.

Clinical data were reviewed by analyzing available clinical records. Information was collected on pregnancy, delivery, growth parameters at birth, clinical features, including major and minor anomalies or defects depicted by cerebral ultrasound examination or MRI, 2-dimensional color-Doppler echocardiography, renal ultrasonography, audiological evaluation by BSERA, neurological assessment for hypotonia and developmental delay. Facial dysmorphism was considered to be marked, in the presence of 6 or more features, including hypertelorism, downslanting palpebral fissures, epicanthal folds, short broad nose, deeply grooved philtrum, high wide peaks of the lip vermilion, micrognathia, low-set and/or posteriorly angulated ears with thick helices, and low posterior hairline, while they were considered mild in the presence of 5 or less features [Allanson et al., 2010]. Clinical diagnosis before and after genetic testing was recorded.

The clinical features recorded in the mutation-positive subjects with RASopathy are summarized in table 2. Mean age at the time of clinical diagnosis was 4.7 months. The comparison of clinical diagnosis before genetic testing and after molecular diagnosis is reported in table 3. Mildly or markedly expressed facial anomalies were found in all patients (fig. 1). Newborns/infants with CS had coarse facial appearance. Additional clinical features invariably documented in each RASopathy included: (1) polyhydramnios and short neck/pterygium in CS; (2) weight below the <3rd centile in NS/LAH, CS, and RAF1 mutation-positive NS; (3) thoracic anomalies in BRAF mutation-positive CFCS; (4) congenital heart defects (CHD) in NS and LS; (5) developmental anomalies in NS/LAH, CFCS, and CS; (6) feeding difficulties in patients with NS/LAH, and CS; (7) laryngomalacia in RAF1 mutation-positive NS; (8) hair anomalies in patients with NS/LAH.

In table 4, the prevalence of clinical features in the present series is compared to percentages found in published reports with wider age distribution.

Occasional findings included: (1) cerebral anomalies (corpus callosum hypoplasia and cerebellar hypoplasia, Arnold-Chiari malformation); (2) ocular anomalies (coloboma, microphthalmia, iris heterochromia, cataract, and nystagmus) (table 2).

The retrospective analysis of clinical features in patients with molecularly confirmed RASopathies, documented during the first year of life, suggests that some characteristics can be used as hints for suspecting a RASopathy, in general, during the first months. In addition, some clinical signs may lead to the suspect of individual disorders or, within a disorder, to individually affected disease genes. Nevertheless, this is a selected cohort, so that the frequency of clinical features may not be applied for general infants with a RASopathy, but specifically for those within the first year of life. Some of the cases included in this report initially received a clinical diagnosis which was changed after the results of genetic testing (table 3). This occurred more often in patients with LS, originally classified as having NS, and in patients with BRAF mutation, which were diagnosed as having an unspecified Noonan-like syndrome.

CHD was the clinical finding leading to medical attention in patients with NS and LS, being invariably present, independently of the affected disease gene and type of mutation. Patients with CFCS and CS displayed CHD in 75% of the cases. In these infants, major reasons for the patients’ referral were either feeding difficulties (80–100%) or developmental motor delay (100%).

Anatomic types of CHD were variable in the different disorders and molecular subgroups. Pulmonary valve stenosis was more common in newborns and infants with a diagnosis of NS associated with mutations in PTPN11 and SOS1, and in CFCS subjects with BRAF mutation, while hypertrophic cardiomyopathy was significantly associated with LS-causing PTPN11 mutations and NS-causing RAF1 mutations. Such associations are in line with previously observed genotype-phenotype correlations [Tartaglia et al., 2002; Sarkozy et al., 2003; Digilio et al., 2006b, 2009; Limongelli et al., 2007; Pandit et al., 2007]. Of note, a wide spectrum of CHDs was documented in the neonatal series, including also atrioventricular canal defect, mitral anomalies, and aortic coarctation. The prevalence of CHDs in this report was slightly higher compared to published series comprising patients with a wider age distribution (table 4). This could be due to the fact that CHDs are often the first clinical symptoms attracting medical attention and suspicion of a RASopathy in the neonatal period and during infancy. On the other hand, the finding that all patients with NS displayed CHDs and facial anomalies actually means that the presence of a heart defect and some facial features suggestive of NS were necessary for making an early diagnosis of NS, while an early diagnosis can be missed in NS patients without a heart defect.

Facial features were present in all patients, being either mildly or markedly expressed. We were unable to identify a distinct pattern suggestive for any molecular subtype or disorder, with the only exception of the occurrence of ‘coarse’ appearance in infants with CS. This finding is in agreement with the conclusions of a recently published study [Allanson et al., 2010]. Protruding tongue was found in all patients with NS/LAH and CS, and in a large proportion of NS newborns and infants with a SOS1 mutation.

Feeding problems caused by failure to thrive were highly common in patients with CFCS, CS, and NS/LAH. They consisted of poor suck and slow feeding with recurrent vomiting, requiring tube feeding or gastrostomy during the first year of life. The degree of growth deficiency appears related to severity of the feeding problems and CHD, in particular in the presence of hypertrophic cardiomyopathy.

Delayed developmental milestones and hypotonia were found in approximately half of mutation-positive patients, and, in general, developmental delay was milder in babies with PTPN11 mutations, and more severe in those with CFCS and CS. Length of hospitalization and the degree of poor growth were likely related to delayed motor development, but a possible relationship with the molecular defect cannot be excluded. This is consistent with previous surveys in older patients pointing to a more severe impairment in the presence of mutations affecting the downstream components of the RAS-MAPK pathway [Cesarini et al., 2009; Pierpont et al., 2009]. Discordance in prevalence of delayed developmental milestones in different ages is found in patients with LS associated with PTPN11 mutations (table 4), with a lower occurrence in older patients.

Epilepsy was found in 40% of infants with CFCS associated with BRAF mutations, in 25% of CS individuals, and only in one case of NS due to PTPN11 mutation.

Among ocular defects, coloboma and microphthalmia were seen in patients with PTPN11 mutations, while cataract, nystagmus and visus defects occurred in CFCS patients with BRAF and MEK2 mutations and in patients with Costello syndrome.

Laryngomalacia was present in patients with RAF1 and HRAS gene mutations.

Ectodermal and skin features are hallmarks for differentiating RASopathies during childhood and adolescence (table 4). Consistent with the distinctive hair anomalies (i.e., easily pluckable, sparse, thin, slow-growing hair characterized by an anagen stage of hair follicle development and bulbs lacking internal and external root sheats) associated with the p.Ser2Gly SHOC2 substitution [Mazzanti et al., 2003; Cordeddu et al., 2009], abnormal hair features compatible with a loose anagen hair condition was invariantly observed during the first year of life in patients with the invariant SHOC2 mutation. Of note, thin, sparse and/or wispy hair occurred preferentially in association with NRAS, BRAF, MEK1, and HRAS mutations, while normal hair was generally observed in association with PTPN11, SOS1 and RAF1 mutations. Hair was often lacking in the first year of life in patients with CFCS caused by BRAF or MEK2 mutations, in which an eczematous skull skin was also present (table 4). These findings indicate that hair characteristics might significantly change with time, as in the case of CS, which is characterized by the distinctive curly and kinky hair structure observed during childhood and adulthood.

Lentigines and café-au-lait spots were found in infants with LS-causing PTPN11 mutations, while keratosis pilaris over the face and the extensor surfaces and eczema were documented to be more common in those with SOS1, SHOC2 and BRAF mutations. Generalized neonatal ichthyosis was found in single patients with SHOC2 and MEK2 mutations. Of note, all these skin features were present in the first months of life, with the only exception of lentigines, which developed in 9 of the 11 cases with molecularly confirmed LS after the second year of life (table 4) [Digilio et al., 2006a].

In conclusion, analysis of clinical features in the present series of patients with RASopathies assessed during the first year of life suggests that the following characteristics can be useful for early prediction of the molecular subtype of the disorder: (1) anatomic type of CHD, (2) severity of feeding difficulties and developmental milestones, (3) hair and skin anomalies. In contrast, facial features are useful in making the general diagnosis of RASopathy, but are not helpful in determining the specific molecular subtype. While larger cohorts are required to attain a more accurate clinical definition of each condition during the neonatal period and infancy, the documented correlations denote that a subset of features can direct towards a prompt diagnosis, and a more effective patient management and genetic counseling.

We are indebted to the patients and families who participated in the study. This research was funded by grants from Telethon-Italy (GGP10020), ‘Associazione Italiana Sindromi di Costello e cardiofaciocutanea’ and ERA-Net for research programs on rare diseases 2009 (European network on Noonan syndrome and related disorders) to M.T.