Multi-layered mutation in hedgehog-related genes in Gorlin syndrome may affect the phenotype

Gorlin syndrome is a genetic disorder of autosomal dominant inheritance that predisposes the affected individual to a variety of disorders that are attributed largely to heterozygous germline patched1 (PTCH1) mutations. PTCH1 is a hedgehog (Hh) receptor as well as a repressor, mutation of which leads to constitutive activation of Hh pathway. Hh pathway encompasses a wide variety of cellular signaling cascades, which involve several molecules; however, no associated genotype–phenotype correlations have been reported. Recently, mutations in Suppressor of fused homolog (SUFU) or PTCH2 were reported in patients with Gorlin syndrome. These facts suggest that multi-layered mutations in Hh pathway may contribute to the development of Gorlin syndrome. We demonstrated multiple mutations of Hh-related genes in addition to PTCH1, which possibly act in an additive or multiplicative manner and lead to Gorlin syndrome. High-throughput sequencing was performed to analyze exome sequences in four unrelated Gorlin syndrome patient genomes. Mutations in PTCH1 gene were detected in all four patients. Specific nucleotide variations or frameshift variations of PTCH1 were identified along with the inferred amino acid changes in all patients. We further filtered 84 different genes which are closely related to Hh signaling. Fifty three of these had enough coverage of over ×30. The sequencing results were filtered and compared to reduce the number of sequence variants identified in each of the affected individuals. We discovered three genes, PTCH2, BOC, and WNT9b, with mutations with a predicted functional impact assessed by MutationTaster2 or PolyPhen-2 (Polymorphism Phenotyping v2) analysis. It is noticeable that PTCH2 and BOC are Hh receptor molecules. No significant mutations were observed in SUFU. Multi-layered mutations in Hh pathway may change the activation level of the Hh signals, which may explain the wide phenotypic variability of Gorlin syndrome.


Introduction
Gorlin syndrome, also referred to as the nevoid basal cell carcinoma syndrome (NBCCS; OMIM 109400), is characterized by several craniofacial abnormalities such as, macrocephaly, frontal bossing and coarse facial features. Recurrent pathological features, such as the development of basal cell carcinomas (BCCs) and odontogenic keratocysts, is a prominent characteristic of these patients [1,2]. Most individuals have skeletal anomalies (e.g., bifid ribs and wedgeshaped vertebrae). Ectopic calcification, particularly in the cerebral falx, is present in more than 90% of the affected individuals. Approximately 5% of all children with Gorlin syndrome develop medulloblastoma (primitive neuroectodermal tumor [PNET]), usually of the desmoplastic subtype. The condition has an autosomal dominant inheritance. Mutations affecting the transmembrane protein patched homolog 1 (PTCH1) are most frequently observed [3][4][5][6][7]. Additionally, mutation in a protein suppressor of fused (SUFU), encoded by SUFU gene [8,9], part of a corepressor complex, and PTCH2 are reported [10,11].
Heterozygous germline PTCH1 mutations have been detected in patients with Gorlin syndrome and the loss of function PTCH1 mutations are predicted to reduce suppression of Smoothened (SMO). In vertebrates, intracellular signaling activity is mediated through the zinc finger transcription factors GLIs. Suppression of SMO leading to GLI repressor formation was shown to block trans-localization of GLI into nucleus [12,13]. Mutation in PTCH1 gene causes functional loss of PTCH1 protein and gives rise to constitutive activation of hedgehog (Hh) signaling [14,15]. Hh signaling pathway governs complex developmental processes, including proliferation and patterning within diverse tissues including osteogenesis. These activities rely on a tightly regulated transduction system that converts graded Hh input signals into specific levels of pathway activity.
Recently, several other Hh receptors have been reported. Hedgehog-interacting protein 1 (HHIP1) is a Hh receptor whose function is similar to that of PTCH1 and PTCH2, i.e., to suppress Hh signaling [16]. Three other Hh co-receptors have been reported. CAM-related/downregulation by oncogenes (CDON), brother of CDON (BOC), and growth arrest-specific 1 (GAS1) are reported as positive transducers of Hh signaling during embryonic neuronal development [17,18]. However, the potential biological role of these co-receptors in skeletal development has not been reported.
The development of high-throughput methods, i.e., next generation sequencing (NGS), has enabled genome-wide characterization of mutations. Several mutations in PTCH1 have been reported that mention it as the causative gene of Gorlin syndrome by conventional Sanger sequencing. However, Sanger sequencing of PTCH1 sometimes fails to determine mutations, probably because of technical problems [19][20][21]. Although the analysis of large genes such as PTCH1 is technically challenging, few reports have described the genome-wide characterization of mutation analysis of Gorlin syndrome. Accurate deleteriousness prediction for variants is crucial to screen out pathogenic mutations from background polymorphisms in NGS studies. Computational algorithms are key inputs for the characterization of variants owing to their ability to be employed on a scale that is consistent with the large number of variants being identified from the systematic screening of representative human populations [22].
In this context several sequence-and/or structure-based methods have been used to predict the potential impact of amino acid substitutions on protein structure and activity. PolyPhen-2 [23], MutationTaster2 [24,25], and others are able to predict 90% or more of the damaging mutations. For a given amino acid substitution in a protein, PolyPhen-2 extracts various sequence-and structure-based features of the substitution site and feeds them to a probabilistic classifier. [23] These prediction programs allow efficient filtering out of alterations that have a high disease-causing potential from the NGS data. MutationTaster2 integrates information from different biomedical databases and utilizes established analytic tools. Components of the analyses include evolutionary conservation and footprints. These prediction methods can help us narrow down candidate mutations and help identify the causative genes. Since these programs have both advantages and disadvantages, we tried PolyPhen-2 and MutationTaster2 to narrow down the candidate mutations. Here we report NGS performed to detect mutations which might cause functional alterations in Hh signaling. We found mutations in PTCH1 in all four patients, with additional mutations in several molecules involved in Hh signaling. Two of these pertained to Hh receptor PTCH2 and BOC. MutationTaster2 analysis indicates that these mutations may cause functional alteration.

Ethics statement
Written informed consent was obtained from all participants for this molecular genetics study. The study was approved by the Ethics Committee for clinical research at the Tokyo Dental College (Tokyo Japan) (no.527 and no.575) and complies with the tenets of the declaration of Helsinki.

Cell culture
Gorlin syndrome fibroblasts were obtained from oral tissues resected as part of treatment for diseases. Fibroblasts were collected from unaffected oral mucosa at the time of surgical operation. Fibroblasts were cultured in Dulbecco's modified Eagle's medium (DMEM) (Invitrogen, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (FBS) (Invitrogen, Carlsbad, CA, USA) and 1% Penicillin/Streptomycin (Invitrogen, Carlsbad, CA, USA) and maintained at 37˚C in 5% CO 2 .

Sequencing and data analyses
Genomic DNA was extracted using an Easy-DNA™ gDNA Purification Kit (Life technology). Genomic DNA was fragmented by Covaris. The size of the library fragments was mainly distributed between 200 and 250 bp. Enrichment of coding exons was performed using Sure Select XT Human All Exon v5 kit (Agilent Technologies, California, USA) for generation of exome libraries. The enriched exome was sequenced on Illumina Hiseq 2500 (Illumina Inc., San Diego, CA, USA) with 101-bp paired-end reads, which were aligned to the human reference genome (UCSC, hg19/ GRCh37) with Burrows-Wheeler Alignment software (version 0.7.10, http://bio-bwa.sourceforge. net/) for the subsequent variant analysis. Samtools (version 1.1, http://www.htslib.org/man/ samtools/) and Picard tools (version 1.115, http://broadinstitute.g.,ithub.io/picard/) were used to build indices and eliminate duplicates. Local realignment around indels (RealignerTargetCreator, IndelRealigner) and base score recalibration (BaseRecalibrator) were applied by GATK (The Genome Analysis ToolKit, Lite version 2.3.0) to ensure accuracy in the identification of indels and single nucleotide variants (SNVs). For each sample, a processed alignment file was saved in bam form. Previously identified common variants (frequency > 1%) and synonymous substitutions were filtered out using public databases including dbSNP 142, HapMap samples, and the 1000 Genome Project (http://www.1000genomes.org). The sequencing data were available in the DDBJ databank of Japan (Accession number: JGAS00000000099).
The PolyPhen-2 grades the mutations as probably damaging, possibly damaging, or benign, along with a numerical score ranging from 0.0 (benign) to 1.0 (damaging). Though the possibly damaging score is generally interpreted as an indication of a mild effect or low penetrance, it is intended as a measure of prediction confidence rather than the effect size. Identified variants were also compared with known alterations reported in mutation databases such as the Human Gene Mutation Database (HGMD).

Mutations in PTCH1
Clinical features of the four Gorlin syndrome patients enrolled in this study are summarized in Table 1. We used the criteria described by Kimonis et al., which comprises of six major and six minor criteria [29], for the diagnosis of Gorlin syndrome. All four patients met at least two major criteria. All individuals had odontogenic keratocysts (Fig 2) and had a familial history of the syndrome. Case 1 had BCCs, skin pit, spinal deformity, and calcification of the falx. Case 2 had BCCs and skin pit. Case 3 showed skin pit, rib anomaly, and calcification of the falx. Case 4 only had KCOT and family history.
To identify the alterations in PTCH1, mutation analysis was performed using MutationTa-ster2. We observed an average of 95.33% of bases at >30X coverage from exome sequence. Specific SNVs causing inferred amino acid changes were detected in two patients and frameshift mutations were observed in two patients.
Case 1: we identified p.Leu446 frameshift mutation located in exon 9. A previous report showed similar alteration showing only one amino-acid change [30] located in the 2 nd transmembrane domain (p.438-457). Case 4: The p.Leu505Arg changes were predicted to be functionally pathogenic by Muta-tionTaster2, and have not been reported in either HGVD. p.Leu505Arg mutation was located near the 4 th transmembrane domain (p.502-522).
We found mutations in the PTCH1 gene at the transmembrane domain or a close-by site in all patients and these mutations have never been reported.
Additional mutations were discovered in Hh-related genes Next we focused on the 84 Hh-related genes listed in Table 2. To discriminate mutations from polymorphisms, we narrowed down our search as illustrated in the flowchart (Fig 2). Specific SNVs causing inferred amino acid changes were detected in 22 genes (Table 3). Each individual had different types of mutations in eight genes, BMP2, CDON, CSNK1A1L, DISP2, FAT4, LRP2, NPC1, and PTCHD3, but all patients had at least one of these additional mutations.
First, we investigated whether mutations were located in either the PTCH2 or SUFU genes, since both were reported as rare causative genes of Gorlin syndrome. We identified p. Arg175Leu and p.Arg74His in case 1 and p.Arg673Leu in case 3 in the PTCH2 gene (Table 4). P.Arg175Leu and p.Arg74His were categorized as "disease causing" by MutationTaster2 and as "probably damaging" and "possibly damaging" by Polyphen-2. We detected c.1299T>C in all individuals in SUFU. Although c.1299T>C in SUFU was categorized as "polymorphism" by MutationTaster2, we qualified this alteration as non-causative.
Second, to identify the impact of an amino acid substitution on the structure or function of proteins expressed in patients with inherited diseases, we analyzed 22 genes with 61 variants using MutationTaster2 and PolyPhen-2.
Case 2: We identified mutations in 15 genes ( Table 3). The mutation located in WNT9B was predicted to be functionally pathogenic. The p Arg258His in WNT9B was located on chromosome 17q21. WNT9b is expressed in the ectoderm of the early facial prominence in mice [31,32]. The p Arg258His mutation was predicted to be "possibly damaging" by polyphen-2 (Table 4). Case 3: We identified mutations in 16 genes (Table 3). Among these, mutations in RUNX2 and BOC were identified as disease causing genes. The p.Phe29Leu in intron of RUNX2 gene was predicted to be "disease causing" by MutationTaster2. p.Gly309Arg was detected in BOC, one of the Hh receptors, which regulates Shh-dependent cell fate specification and axon guidance [17]. The p.Gly309Arg mutation was predicted to be "probably damaging (0.998)" by PolyPhen-2.
Case 4: We identified mutations in 17 genes (Table 3). p.Pro913Ser was detected in BOC. This mutation was predicted as "disease causing" by MutationTaster2. (Table 4)

Discussion
We present here two major findings. First, the four individual patients who qualified the diagnostic criteria for Gorlin syndrome had novel PTCH1 gene mutations. These mutations were located near transmembrane lesions. Second, exome analysis of all four patients revealed additional mutations in Hh receptor PTCH2 or BOC.
Whole-exome sequencing (WES) has become the standard method for identification of causative genes of rare monogenic diseases [33,34]. An efficient and popular strategy is "intersection filtering," in which the whole-exomes of several unrelated affected individuals are sequenced, and their sequence variants are filtered and compared [35]. This approach is useful for both rare inherited and sporadic diseases. We performed an exome analysis to detect mutations in the predicted gene PTCH1, in patients with Gorlin syndrome. Although numerous Hedgehog-related multi-layered gene mutation potentially affects phenotype   Hedgehog-related multi-layered gene mutation potentially affects phenotype germline mutations of PTCH1 have been reported in these patients, neither clustering of mutations nor genotype-phenotype correlations have been identified. However, the onset of related malignancies in patients with Golrin syndrome is based on the classical 2 hit suppressor gene model: baseline heterozygosity preceding the germline PTCH1 mutation is the first hit. The second hit is UV or ionizing radiation. Congenital malformations may be attributable to alterations in the concentration of the PTCH1 gene product in the dosage-sensitive Hh signaling pathway. Therefore, it is highly likely that genetic variations contributed to PTCH1 phenotype variations. Other important facts of genetic analysis of Gorlin syndrome was that many cases that researchers cannot find mutations in the PTCH1 gene probably due to technical difficulties or existence of mutations outside of the regions analyzed or in genes other than PTCH1 [21]. These facts suggest the need to consider possible mutations in genes other than PTCH1.
In fact, SUFU [8] and PTCH2 [10,11] mutations have been reported as another possible causative genes of Gorlin syndrome. However, few studies have investigated PTCH1 and other possible mutations simultaneously.
Because increasingly sophisticated methods particularly emphasize the fact that primary mutation discovery in humans will remain crucial to scientific progress, analysis of the WES and whole-genome sequencing (WGS) data allows for determination of the phenotypic consequences of genetic variation. Therefore, it is reasonably assumed that the mutation analysis of genes other than PTCH1 using WES may define potential diagnostic, preventive, and therapeutic opportunities for Gorlin syndrome and illuminate disease mechanisms.
In this study, we first used NGS to screen PTCH1 genes in four individuals with Gorlin syndrome. We detected four novel mutations, of which p.Leu446 frameshift mutation was identified in case 1, c.652G>A in case 2, p.Asp460 frameshift mutation in case 3, and p.Leu505Arg missense mutation in case 4. Although no hotspot was identified, several mutations were found to be localized near the transmembrane lesions of PTCH1 in all four cases.
It is pertinent to note the diverse phenotypes shown in Table 1 and Fig 2. We think that these phenotypic divergences might indicate the existence of other gene mutations. First, we analyzed genes closely related to the Hh pathway, as these have wide range of biological functions and are likely to have impact on pathophysiology of Gorlin syndrome. We performed WES to simultaneously screen all exons of the 84 genes listed in Table 2. These 84 genes consisted of Hh ligands, receptors, transcription factors, and WNT, Hippo, and TGF/BMP signaling, all of which have a crosstalk with Hh signaling [36][37][38][39]. As expected, we observed several sequence variants. However, the consequences of these variants or mutations are usually difficult to assess in terms of their pathogenicity and require more detailed analyses. Thus we used open access programs to assess disease-causing potential of the DNA sequence variants.
First, we set a relatively high standard for variant candidates. The NGS coverage level often determines whether variant discovery can be made with a certain degree of confidence at particular base positions. It is generally recommended that for detecting human genome mutations, coverage should be at least from 10× to 30×. Therefore we preset the minimum sequence coverage to be 30.
In addition to coverage threshold, we applied MutationTaster2 and PolyPhen-2 analyses, which predict functional annotation of gene mutations, extract protein sequence annotations and structural attributes, and build conservation profiles. Through the use of two of these software programs, we were able to detect gene mutations which probably give rise to abnormal or malfunctioning proteins in three additional genes, PTCH2, BOC, AFT4, and WNT9b.
MutationTaster2 is a web-based application for rapid evaluation of the disease-causing potential of DNA sequence alterations. MutationTaster2 integrates information from different biomedical databases and employs established analytic tools [24,27,28].
We observed that some of the mutations in PTCH2 were already reported to be disease causing, such as p.Arg175Leu and p.Arg74His in case 1 [1,2]. We also detected p.Arg673Leu in PTCH2 in case 3. These three mutations were determined as "disease causing" by MutationTa-ster2. Polyphen-2 showed p.Arg175Leu as "probably damaging" and p.Arg74His in case 1 and p.Arg673Leu in case 3 were "possibly damaging." Thus, these are considered as possible disease-causing mutations.
PTCH2 was reported to be a rare causative gene for Gorlin syndrome. Individuals who have mutations only in PTCH2 showed relatively milder symptoms. It is intriguing that two out of four Gorlin patients had mutations in both PTCH1 and PTCH2. These two genes have a redundant role because PTCH2 gene has similar functions to those of PTCH1; however, Ptch2 -/mice are viable and fertile, whereas Ptch1 -/mice die at E9.5 with ectopic Hh signaling throughout the embryo [40,41]. Alfaro et al. showed that Ptch1 -/-; Ptch2 -/neuralized embryoid bodies (NEBs) have a higher level of Shh pathway activation than Ptch1 -/-NEBs [42]. It is conceivable that double mutations in PTCH1 and PTCH2 may lead to more severe symptoms.
p.Gly309Arg and p.Pro913Ser were detected in biregional Codon-binding protein (BOC), which positively regulates Shh signaling and promotes Shh-dependent cell fate specification and axon guidance [18,43]. The p.Gly309Arg mutation was predicted to be "probably damaging (0.998)" by polyphen-2. Cdo and Boc bind Shh through a high-affinity interaction with a specific fibronectin repeat that is essential for Hh activity [18]. Cdo −/− ; Boc −/− double mutants were reported to resemble Gli2 −/− mutants and lack the expression of Shh in the presumptive floor plate [44]. Although the biological significance of BOC is not well-characterized, to the best of our knowledge, BOC at least has a function as a Hh receptor.
PTCH2 and BOC are receptors of Hh and they regulate Hh pathway; therefore, they must have some effects on Hh signaling. It is noteworthy that all four patients had disease causing mutations in two of these Hh receptors.
We found a p.Ile1900Leu mutation in FAT4 gene in case 1. We also found a p.Arg258His mutation in the WNT9b gene in case 2 and a p.Phe29Leu in the intron of RUNX2 in case 3. Fat4 is well known as a regulator of cell proliferation and planar cell polarity, and Fat4-null mice exhibit a short body with complicated morphological abnormalities, including abnormal orientation of inner ear hair cells, cystic dilation of kidney tubules, and defects in organ shapes [45]. WNT9b is the target gene of Hh and controls cell proliferation and cell polarity. WNT9b null mice display CLP and skeletal abnormalities in the upper jaw, which confirms that WNT9b is required for the proper development of the lip and palate [46]. An accumulating body of evidence suggests a crosstalk between Wnt and Hh [47]. Thus, mutations in FAT4 and WNT9b may also play a supporting role in the development of Gorlin syndrome. The p.Phe29-Leu in the intron of RUNX2 gene was predicted to be "disease causing" by MutationTaster2. Several reports have indicated that mutations in introns could have some impact such as splicing aberrance or promotor enhancer dysfunction.
In this study, we focused on the genes that have been previously reported to have relationship with Hh signaling. Mutations were also detected in Hh unrelated genes, although its functional importance is unclear. We think that it is certainly possible that mutation in other genes may affect Hh signaling. Further investigation is required to elucidate the passivity. Only four individuals showed these mutations; however, if the number of samples increased, more significant information will be shown, and additional candidate gene related to Gorlin syndrome may be discovered.
Collectively, our findings indicate that NGS is essential to understand the genetic background of Gorlin syndrome. It may be important to identify mutations in Hh receptors other than PTCH1, such as PTCH2 and BOC. Hereditary disorders are typically considered to be caused by a mutation on a single gene. However, to elucidate the genetic basis of a syndrome, WGS or WES is indispensable.