Targeted/exome sequencing identified mutations in ten Chinese patients diagnosed with Noonan syndrome and related disorders

Background Noonan syndrome (NS) and Noonan syndrome with multiple lentigines (NSML) are autosomal dominant developmental disorders. NS and NSML are caused by abnormalities in genes that encode proteins related to the RAS-MAPK pathway, including PTPN11, RAF1, BRAF, and MAP2K. In this study, we diagnosed ten NS or NSML patients via targeted sequencing or whole exome sequencing (TS/WES). Methods TS/WES was performed to identify mutations in ten Chinese patients who exhibited the following manifestations: potential facial dysmorphisms, short stature, congenital heart defects, and developmental delay. Sanger sequencing was used to confirm the suspected pathological variants in the patients and their family members. Results TS/WES revealed three mutations in the PTPN11 gene, three mutations in RAF1 gene, and four mutations in BRAF gene in the NS and NSML patients who were previously diagnosed based on the abovementioned clinical features. All the identified mutations were determined to be de novo mutations. However, two patients who carried the same mutation in the RAF1 gene presented different clinical features. One patient with multiple lentigines was diagnosed with NSML, while the other patient without lentigines was diagnosed with NS. In addition, a patient who carried a hotspot mutation in the BRAF gene was diagnosed with NS instead of cardiofaciocutaneous syndrome (CFCS). Conclusions TS/WES has emerged as a useful tool for definitive diagnosis and accurate genetic counseling of atypical cases. In this study, we analyzed ten Chinese patients diagnosed with NS and related disorders and identified their correspondingPTPN11, RAF1, and BRAF mutations. Among the target genes, BRAF showed the same degree of correlation with NS incidence as that of PTPN11 or RAF1.

NS is a rare disorder that was first defined by Dr. Jacqueline A. Noonan in 1963 [3]. NS is an autosomal dominant developmental disorder with an estimated prevalence of 1:1000 to 2500 births [4]. It is characterized by facial dysmorphisms, short stature, congenital heart defect, and variable degrees of developmental delay. Missense mutations in PTPN11, SOS1, RAF1, and KRAS account for approximately 50%, 10-13%, 3-17%, and <5% of all NS cases, respectively. De novo mutations account for 60% of all NS cases [5].
NSML was previously known as LEOPARD syndrome, which was derived from the primary symptoms that include multiple lentigines, electrocardiographic conduction defects, ocular hypertelorism, pulmonary stenosis, abnormal genitalia, growth retardation, and sensorineural deafness [6]. NSML is caused by carrying a heterozygous pathogenic variantin one of four specific genes, namely, PTPN11, RAF1, BRAF, and MAP2K1.
In the past, the standard genetic diagnostic process for NS was based on Sanger sequencing and single gene analysis for PTPN11. This can be followed by subsequent single-gene analyses for SOS1, RAF1, KRAS, NRAS, BRAF, and MAP2K1 when no mutation was identified for PTPN11. This inefficient procedure was timeconsuming and often led to additional economic burden for both the patients and clinicians. Recently, targeted/ whole exome sequencing (TS/WES) has increasingly been employed for clinical diagnosis and has changed the paradigm of molecular diagnostic testing because of advantages, such as cost-effectiveness, generation of high-quality outputs, simplicity, and automated operation [9][10][11]. TS/WES is employed to obtain more comprehensive and gene-level information and generate a more accurate diagnosis. In particular, TS/WES is useful for clinicians when the phenotypes of sporadic patients are variable and complicated.
In the present study, we identified mutations in the PTPN11, RAF1, and BRAF genes using TS/WES in patients who had above-mentioned clinical features.

Subjects
By retrospectively reviewing the results generated from targeted sequencing/whole exome sequencing between 2014 and 2016, ten patients with mutations in genes involved in Noonan syndrome and related disorders were identified and presented in this report (six males, four females). The mean age was 3.8 years (range: 5 months to 10 years). All patients received physical examination, neurological/neuropsychiatric assessment, biochemical testing, echocardiography, karyotype analysis, and tandem mass test. Family history was routinely been recorded. Whole-genome copy number variation (CNV) array and enzyme activity tests related to mucopolysaccharidosis/mucolipidosisi were performed in some of the patients.
All patients enrolled in this study have signed informed consent by their parents, including allowing pictures, medical data been published.

Whole exome sequencing
Peripheral blood samples were collected from the patients and their parentsafter informed consent was obtained. Genomic DNA (gDNA) was extracted using Lab-Aid Nucleic Acid (DNA) Isolation Kit (Zeesan, China) according to the manufacturer's instructions.ClearSeq Inherited Disease or SureSelect Human All Exon V5 kit (Agilent, Santa Clara, CA, USA) were used for library preparation of targeted sequencing or whole exome sequencing, respectively. The resulting libraries were sequenced on a HiSeq 4000 platform (Illumina, San Diego, CA, USA) according to the manufacturer's instructions for paired-end 150-bp reads. The minimal data amount was2.5Gb per sample for TS and 8Gb per sample for WES.Fastq-format reads were aligned to the human reference genome (GRCh37/hg19) using BWA-0.7.10 [12]. BAM files were manipulated using Picard tools-1.124. Base calling was performed following GATK best practice version 3 [13]. Quality metrics were evaluated -the average depth was 80× per sample, with at least 97% of the target region covered by 10× reads or more. The vcf files were then annotated using SnpEff version 4.2 [14]. Variants with >1% frequency in the population variant databases -1000Genomes Project, Exome Variant Server (EVS) and Exome Aggregation Consortium (ExAC) or > 5% frequency in the local database with 150 exome datasets were filtered, and subsequentlyintergenic, intronic, and synonymous variants were filtered, except those located at canonical splice sites. Candidate variants were then evaluated in the context of clinical presentation and inheritance mode. Selected variants were validated by Sanger sequencing in the proband and parents. Paternity was confirmed for de novo variants.

Clinical presentations and comparison with literature
The detailed clinical features of the ten patients analyzed in our study are displayed in Table 1. Figure 1 shows the facial dysmorphisms of some of the patients(with consent obtained from parents for publication). All patients were sporadic cases.

Identification of disease-causing mutations
As shown in Table 2, TS/WES identified three genes harboring a total of ten mutations in the ten patients after filtering and manual review of the genes according High wide peaks of the vermilion - Thick curly hair or thin sparse hair --

Others
Developmental delay or cognitive deficit + + - ASD atrial septal defect, VSD ventricular septal defect, HCM hypertrophic cardiomyopathy, PDA patent ductus arteriosus, PVS pulmonary valve stenosis +present, −not present to clinical presentation. The genes that carried mutations were PTPN11 (3/10 = 30%), RAF1 (3/10 = 30%), and BRAF (4/10 = 40%). In this study, BRAF was found to be the most common pathological gene in the NS patients, followed by PTPN11 and RAF1. In our study, all detected mutations were de novo mutations and not present in their parents, with paternity confirmed. Patient 3, who presented multiple lentigines and carried a NSML-associated RAF1 mutation (c.770C > T, p.S257 L), was diagnosed with NSML [15][16][17][18], whereas patient 2, who carried the same mutation but lacked multiple lentigines, was diagnosed with NS ( Fig. 1). The diagnosis of patient 2 contradicted the previous claim that the S257 L mutation is always linked to hypertrophic cardiomyopathy.

Functional prediction of the novel mutant protein
We identified onenovel mutation in BRAF (c.1403 T > G, p.F468C) genes in patients 9. This variant has not been previously reported in the Human Gene Mutation Database, the 1000 Genomes Database, orGnomAD database at the time of writing of this manuscript. It was predicted to be "probably damaging" with a score of 0.996 for c.1403 T > G, p.F468C based on the PolyPhen-2 software, predicted to "affect protein function" with a score of 0.00 by the SIFT software, and classified as "disease-causing"   Fig. 2. Thewild-type residue was located in highly conserved domains. In BRAF, residue 468 is located in CR3, a highly conserved region that encodes a part of the kinase activity domain. The F468C mutation generates a smaller residue and potentially causes the loss of external interactions.

Discussion
In this study, we verified the prevalence of PTPN11, RAF1, and BRAF mutations in Chinese patients diagnosed with NS and related disorders via TS/WES. We identified a total of ten mutations in the ten patients. All patients who carried PTPN11 and BRAF mutations were diagnosed with NS. Two patients who carried the same RAF1 mutation presented different features and were separately diagnosed with NS and NSML. PTPN11 is thought to be the most common pathogenic gene that causes NS, followed by RAF1. BRAF mutations are very rarely found in NS cases [1,15]. PTPN11 encodes a key protein, a member of the protein tyrosine phosphatase (PTP) family, whichresponds to growth factors, hormones, and cell adhesion molecules [19]. RAF1 is a downstream factor of RAS signaling in the MAPK pathway that encodes a protein with 648 amino acids and comprises three domains, namely, CR1, CR2, and CR3. NS and NSML are both associated with mutations in PTPN11 and RAF1. However, some of the mutations potentially drive the NS phenotype, while other mutations are predicted to produce the NSML phenotype [20].
PTPN 11K70R has not been published in the literature, but in Clinvar, itis classified as "Likely Pathogenic".This variant has been identified in 5 affected individuals and segregates with symptoms of Noonan syndrome in one family. As lack of clinical data from other study, we cannot compare the phenotype among the patients who had the same K70R mutation.
The RAF1 mutations 770C > T (p. S257 L) and 781C > A (p. P261T) detected in this study were both clustered in the CR2 domain, which is important for regulatory phosphorylation and binding with the 14-3-3 protein. In a previous study, RAF1 was thought to be associated with HCM because all patients that carried the S257 L mutation were diagnosed with HCM, and two of them died from severe HCM [18]. This genotypephenotype correlation appeared to be domain-specific, since the region encoding the 14-3-3 consensus site was affected in the HCM patients. In our study, both patients 2 and 3 carried the S257 L mutation, which was associated with both NS and NSML [18]. Patient 3 displayed typical HCM echocardiography and multiple lentigines in the face, so an NSML diagnosis should be considered. However, patient 2 presented normal interventricular septum(IVS) and mildly thickened left ventricular posterior wall (LVPW), so an HCM diagnosis cannot be confirmed at this point. The patient did not present lentigines, so he was diagnosed with NS.
Sarkozy reported a female whose early clinical presentationwas typical of NS but eventually developed hearing loss and lentigines, which are typical phenotypesof NSML, as the disease progressed [7]. Lentigines usually Fig. 2 3D structural models of BRAF containing the mutant sites appear at an early age(eg, 4-5 years old), and increase until puberty. Similarly, the penetrance of left ventricular hypertrophy (LVH) is also age-dependent. The LVH of HCM often becomes apparent during adolescence or young adulthood. Patient 2 was only 15 months old upon admission and can thus develop LVH later in life. Therefore, patient 2 requires further follow-up to determine whether a novel phenotype will emerge.
The BRAF gene is thought to be the primary cause of CFCS. BRAF mutations account for around 50%-75% of all CFCS cases, but is implicated in only a small fraction of NS and NSML cases (<2%) [7,[22][23][24]. Sarkozy and Koudova identified some individuals who were clinically diagnosed with NS or NSML that carried BRAF mutations [21,25]. However, NS-or NSML-related BRAF mutations aren't as same as those that occur in CFCS, suggesting a genotype-phenotype correlation. Unfortunately, the mechanisms underlying this phenomenon remain to be elucidated. The c.770A > G(p.Q257R) mutation is the most widespread CFCS pathogenic variant [8] and was also detected in patient 8. Assuming a genotype-phenotype correlation, patient 8 should present features of CFCS. However, he had characteristic facies, cardiac defects, short stature, abnormal brain MRI, failure to thrive, and relative developmental delay, but lacked typical cutaneous abnormalities and musculoskeletal and ocular abnormalities; hence, he was diagnosed with NS instead of CFCS. This specific case expanded the mutational spectrum of the BRAF gene in NS and highlighted the genetic heterogeneity of BRAF.
We detected two mutations at residue 468 in the BRAF gene. Patient 6 carried a c.1403 T > C (p.F468S) mutation, which has been reported in a previous study [26]. Patient 9 carried a c.1403 T > G (p.F468C) mutation affecting the same protein. However, F468Cwas never been reported in NS or related disorderspreviously.Interestingly, itwas detected in paraffin-embedded tumoursepecimens of a hairy cell leukemia (HCL) patient [27] and a colorectal cancer patient [28].There is evidence from in vitro and in vivotransfection experiments [29] that F468C mutation leads to increased activity of BRAF and may thus be disease-defining mutation of HCL or colorectal cancer. By sequencing BRAF genefrom normal gastric biopsies of the HCL patient, germline mutation is excluded [27].Our report is the first time to detect F468C germline mutation in a non-cancer patient.Patients 6&9presented similar clinical characteristics, which supported the idea that the phenotype resulting from BRAF mutations is allele-specific and suggested that residue 468 may be a "hotspot" mutation site in Chinese patients.
The ten patients in this study shared features, such as congenital heart defect, short stature, and special facies, that led to difficulties in defining CFCS, NSML, or NS using clinical criteria. Next-generation sequencing (NGS) is a rapid and economical technique that provides molecular-based diagnosis for clinically overlapping conditions. NGS facilitates early disease diagnosis, especially for patients with mild/moderate, atypical features, and can potentially direct clinicians towards more reliable genetic counseling and clinical treatment of the patients.

Conclusions
Overall, we verified the prevalence of PTPN11, RAF1, and BRAF mutations in NS and related disorders in the Chinese population. BRAF showed the same degree of correlation with NS incidence as that of PTPN11 or RAF1. The same mutation can result in different phenotypes, suggesting that the phenotypes arising from RAF1 or BRAF defects are likely to be allele-specific.