Whole-exome sequencing to identify novel mutations of nevoid basal cell carcinoma syndrome in a Chinese population

Abstract

BACKGROUND:

Nevoid basal cell carcinoma syndrome (NBCCS) is a rare autosomal dominant disease with a complex genetic etiology. Although three causative genes (PTCH1, PTCH2, SUFU) have been identified through linkage analysis and Sanger sequencing, the genetic background of NBCCS hasn’t been fully understood.

METHODS:

We performed a whole-exome sequencing (WES) in a Han Chinese NBCCS family and two unaffected volunteers to search for its causative gene. Bioinformatic analysis was used to select candidate genes and analyze the functional networks of each candidate gene.

RESULTS:

A total of 8 single-nucleotide variants (SNVs) were detected in PTCH1, PTCH2 and SUFU in all the 5 subjects, however none of them was considered the pathogenic genetic mutation in this NBCCS family. The following filtering process identified 17 novel candidate genes (GBP3, AMPD1, ASPM, UNC5C, RBM46, HSPA1L, PNPLA1, GPR126, AP5Z1, ZFHX4, KIF24, C10orf128, COX15, GPRC5A, UGGT2, RHBDF1, RPUSD1). Among them ZFHX4 had been already identified as a new basal cell carcinoma susceptibility loci through a genome-wide association study (GWAS) and was considered the most likely pathogenic gene for this NBCCS family. The functional network analysis revealed that ZFHX4 may be involved in notch signaling pathway.

CONCLUSIONS:

Our study reported the identification of 17 novel candidate genes in a Han Chinese family through WES. ZFHX4 may be a susceptibility gene for NBCCS in Chinese population.

Keywords

Gorlin syndrome nevoid basal cell carcinoma syndrome whole exome sequencing Single-nucleotide variants ZFHX4

1. Background

Nevoid basal cell carcinoma syndrome (NBCCS), also known as Gorlin syndrome, is a rare genetic disease characterized multiple basal cell carcinomas, palmar/plantar pits, jaw cysts, and bony deformities like kyphoscoliosis and frontal bossing [1]. This disease is transmitted in an autosomal dominant manner and has the characteristic feature of high penetrance and variable expressivity [2, 3]. There is a significant racial variation in the incidence of the disease worldwide and is most prevalent among Caucasians [4, 5], only a few reports has been reported in Asians so far.

Figure 1.

Clinical phenotype of the patient. (a) basal cell carcinomas (BCCs) around the eyelid. (b) a large BCC on the right temporal. (c) small asymmetric palmar-plantar pits on the right palm. (d) BCCs in the mouth. (e) dermal cysts on the left arm. (f) small asymmetric palmar-plantar pits on the left sole. (g) mandible cysts. (h) amellar calcification of the falx cerebri.

Clinical diagnosis of NBCCS relies on specific criteria, gene mutation analysis may confirm the diagnosis, genetic counseling is mandatory. NBCCS has a complex genetic etiology, linkage analysis and Sanger sequencing have identified three susceptible genes so far, including PTCH-1, PTCH-2 and SUFU [6, 7, 8], which are all involved in the Hedgehog (Hh) signaling pathway. However, mutations in theses genes may not account for all the NBCCS cases and there might be additional unidentified genetic defects responsible for this disease.

The next generation sequencing technologies, including whole genome and whole exome sequencing (WES) have been successfully applied in genetics research, especially in human diseases [9, 10]. This technique is a cost-effective one and especially suitable for the study of Mendel genetic disease: the candidate genes for monogenic disorders can be identified by exome sequencing of only a small number of individuals [11]. In this report, we applied WES in a Han Chinese NBCCS family and two unaffected volunteers to discover causative genes for this disease in this particular family.

2. Materials and methods

2.1 Case presentation

A 70-year-old man was referred to our department due to a large ulcerated lesion located on the right temporal. The lesion was insidious in onset and gradually increased in size. Our physical examination showed that there were multiple, discrete, pigmented, well defined, nontender mobile plaques and nodules of varying sizes present all over the body, especially around the eyelids and in the mouth; small asymmetric palmar-plantar pits could be found on his right palm and left sole; dermal cysts were found on his left thigh and left arm. In further examination, lamellar calcification of the falx cerebri was detected through brain CT scan; several mandible cysts were found through oral CT scan (Fig. 1). There was no bifid, fused or markedly splayed ribs found in chest X ray.

Figure 2.

Flow chart of the whole study.

The patient received a scalp lesion removal operation 50 years ago, he took another large hip lesion removal and skin graft surgery 6 years ago, histopathology in both times were suggestive of basal cell carcinoma (BCC). He had a history of cataract in both eyes.

On family history, both of the patient’s parents deceased for decades. His mother had a similar manifestation of “black skin nodules over the body” just like himself. His maternal grandmother died of abdominal cancer, he failed to provide more information on the cause of death of his father, maternal grandfather, paternal grandfather and paternal grandmother. This patient had a daughter, a son and a granddaughter, none of them carried the clinical phenotype of NBCCS (Fig. 2).

After his administration, the patient was diagnosed NBCCS. He received a lesion removal and skin graft operation and recovered well after the surgery.

2.2 Whole exome sequencing (WES) and read mapping

We harvested the peripheral blood from the NBCCS patient (sample 1), his unaffected daughter (sample 2), his unaffected son (sample 3), an unaffected female volunteer (sample 4) and an unaffected male volunteer (sample 5).

By using Qiagen DNA kit, genomic DNA was extracted from peripheral blood. IlluminaHiSeq 2500 in a paired end 2 $\times$ 100 nt multiplex procedure was used to capture all exons of the samples. The DNA library was constructed by fragmenting the genomic DNA; a Qubit ${}^{\@setsize{\scriptsize}{8pt}{\viipt}{\@viipt}\textregistered}$ 2.0 Fluorometer was used to determine the concentration of the library and the Agilent 2100 system was used to investigate the library’s quality; after these steps the Illumina PE Flow Cell v3 – HS system was used to sequence randomized DNA fragments. All sequencing processes were controlled by data collection software according to the HiSeq 2500 User Guide.

We aligned the paired-end reads to the reference human genome (hg19) using the third-party software BWA (Burrows – Wheeler Alignment, version 5.9). The average mapping ratio was as high as 99%. The Flagstat tool was utilized to assess the mapping information. Next, we analyzed the distribution of each sample’s reads in the target region and the enrichment of reads in the genome. SNVs (single nucleotide variations) were then processed using the GATK UnifiedGenotyper (GenomeAnalysisTK-3.1-1). Finally, we annotated the mutations using ANNOVAR software (GenomeAnalysisTK-3.1-1).

2.3 Identification of candidate genes

SNVs in PTCH1, PTCH2 and SUFU in all the samples were firstly analyzed. Since there was no positive finding, novel NBCCS pathogenic genes were identified through the following filtering process: low quality reads were firstly removed; SNVs located out of exon region and splicing site were then removed; synonymous SNVs and common SNVs (mutation frequence $\geqslant$ 0.95 in 1000 genomes project) were filtered out as well. We then used SIFT and PolyPhen-2 to predict the damage of amino acids structure in each SNV, SNVs scored “SIFT score $<$ 0.05 and PolyPhen-2 score $>$ 0.95” were kept for further filtering process. Finally, since NBCCS is an monogenic autosomal dominant genetic disease, SNVs met the criteria of “heterozygous in the patient and non-mutated homozygous in the control subjects” were kept. After these steps, we were able to obtain 17 candidate genes (Fig. 2).

Table 1
Summary of original exome sequencing data

Sample	Good	Mapped	Mapped	Unique mapped	Unique mapped	Reads on target	Coverage	Q20	Mean
	reads	reads	ratio (%)	reads	ratio (%)	region (%)	(%)	(%) ${}^{*}$	depth
1	91,499,824	91,427,801	99.92	81,915,441	89.53	84.66	99.97	98.04	157.9
2	70,532,508	70,464,557	99.9	62,743,581	88.96	84.62	99.85	98.11	120.14
3	83,304,366	83,258,379	99.94	75,248,867	90.33	84.8	99.97	98.30	145.2
4	79,221,816	79,181,323	99.95	70,657,777	89.19	84.78	99.83	98.19	135.77
5	86,683,994	86,638,744	99.95	76,217,354	87.93	83.83	99.97	98.24	144.56

${}^{*}$ Q20: Fred value $>$ 20. Fred value $=-$ 10*log10(p); P: Probability of measurement error.

Table 2

Exonic SNVs in PTCH1, PTCH2 and SUFU

Gene	Chr	Pos	Ref	Alts	Alt	S1	S2	S3	S4	S5	Variant	Variant type	1000 G	rs number
											sites
PTCH1	Chr9	98209594	G	A	A	0/1	1/1	1/1	0/0	0/1	exon	nonsynonymous SNV	0.4	rs357564
PTCH1	Chr9	98211572	T	A	A	0/0	0/0	0/0	1/1	0/1	exon	nonsynonymous SNV	0.079	rs2236405
PTCH1	Chr9	98220322	A	C	C	0/0	0/0	0/0	1/1	0/1	exon	synonymous SNV	0.068	rs2066835
PTCH1	Chr9	98220550	A	G	G	0/0	0/0	0/0	0/1	0/0	exon	synonymous SNV	0.0046	rs2229062
PTCH1	Chr9	98238358	G	A	A	0/1	0/0	0/0	0/0	0/0	exon	synonymous SNV	0.092	rs2066836
PTCH2	Chr1	45292866	G	A	A	0/0	0/0	0/0	0/1	1/1	exon	synonymous SNV	0.15	rs2295996
PTCH2	Chr1	45293518	A	G	G	0/0	0/0	0/0	0/1	1/1	exon	synonymous SNV	0.45	rs7525308
SUFU	Chr10	104386934	T	C	C	0/1	0/1	0/1	0/0	0/1	exon	synonymous SNV	0.25	rs17114803

Chr: Chromosome ID. Pos: mutation position in each chromosome. Ref: reference base. Alt: altered base. Alts: altered bases, if there is more than one altered base. S1–S5: sample1-sample 5. 1000 G: allele frequencies in 1000 genomes project database.

2.4 Functional network analysis of the candidate genes

For the remaining 17 genes, a protein-protein interaction networks analysis (http://string-db.org/) was firstly carried out to identify the relationships between the product of these genes and the components of Hh signaling pathway. We then draw the genetic networks maps of each candidate gene to further analysis their genetic functions.

3. Results

3.1 Summary of whole exome sequencing

Overall, we obtained more than 87% uniquely mapped reads and more than 99% good matched reads. In each sample over 84% reads were on the target region. In total about 98.2% of the targeted bases were covered sufficiently to pass the quality assessment and were aligned to the human reference genome (hg19) with the BWA. The average sequencing depth in the exome region of the case was approximately more than 120 X (Table 1); the average sequencing depth in each autosome was approximately more than 120 X and the mean GC content was about 50% (Supplement Fig. 1a). The target SNV distribution and target SNV exonic function were presented in Supplement Fig. 1b and c.

In summary, we obtained a high quality WES data which could provide reliable information of the disease.

3.2 Comparison SNVs with PTCH1, PTCH2 and SUFU

In exon region, 5 SNVs were detected in PTCH1, 2 SNVs were detected in PTCH2 and 1 SNV was detected in SUFU respectively. However in the total 8 SNVs, 6 SNVs were synonymous variations, only 2 SNVs in PTCH1 gene were nonsynonymous variations. Between these 2 nonsynonymous SNVs, only 1 SNV was detected in the affected patient; however this SNV was a common variation, the mutation frequency in 1000 genomes project was 0.4, moreover both of his children carried this mutation (Table 2).

In summary, we detected no pathogenic SNVs in the three known suspected genes in this family.

3.3 Identification of candidate genes

Through the initial filtering process, 222 mutated genes were considered potentially hazardous in all the five subjects (Fig. 2). Considering the autosomal dominant inheritance pattern, we were finally able to narrow down to 17 genes as the candidate genes in this patient: GBP3, AMPD1, ASPM, UNC5C, RBM46, HSPA1L, PNPLA1, GPR126, AP5Z1, ZFHX4, KIF24, C10orf128, COX15, GPRC5A, UGGT2, RHBDF1, RPUSD1 (Table 3).

Table 3
List of the disease specific candidate genes

Gene	Chr	Position	Exon	Allele change	Amino acid variation	SIFT score	Polyphen2 HVAR score/prediction
GBP3	Chr1	89486370	exon2	C $>$ T	C12Y	0.01	1/Damanging
AMPD1	Chr1	115220086	exon9	C $>$ T	R454H	0	1/Damanging
ASPM	Chr1	197073196	exon18	G $>$ A	R1729W	0	0.999/Damanging
UNC5C	Chr4	96091414	exon15	C $>$ T	A841T	0.01	0.996/Damanging
RBM46	Chr4	155719189	exon3	T $>$ G	I126M	0.01	0.999/Damanging
HSPA1L	Chr6	31779728	exon2	C $>$ T/C $>$ G	A8T/A8P	0	1/Damanging
PNPLA1	Chr6	36262153	exon4	G $>$ A	A145T	0	1/Damanging
GPR126	Chr6	142688811	exon3	C $>$ T	T70M	0	1/Damanging
AP5Z1	Chr7	4830840	exon17	C $>$ T	R750W	0	1/Damanging
ZFHX4	Chr8	77763452	exon10	G $>$ A	R1432H	0.04	1/Damanging
KIF24	Chr9	34311195	exon2	G $>$ C	D50E	0	1/Damanging
C10orf128	Chr10	50374953	exon3	C $>$ G	D67H	0	1/Damanging
COX15	Chr10	101483799	exon5	G $>$ A	R222C	0	1/Damanging
GPRC5A	Chr12	13061553	exon2	C $>$ T	R124W	0	1/Damanging
UGGT2	Chr13	96506707	exon35	T $>$ A	E1344V	0	1/Damanging
RHBDF1	Chr16	112775	exon6	G $>$ A	R265W	0.01	1/Damanging
RPUSD1	Chr16	836149	exone6	A $>$ G	L247P	0	1/Damanging

Figure 3.

Protein-protein interaction networks between the candidate genes and Hedgehog (Hh) pathway.

3.4 Relationships between candidate genes and Hh pathway

The protein-protein interaction networks analysis revealed that 2 genes (AMPD1, GPR126) were involved indirectly in the Hh pathway: AMPD1 is linked to PTCH1 through CTNNB1, GPR126 is linked to PTCH2 and SUFU through HHIP. However none of the 17 candidate genes were directly involved in Hh pathway based on available research data (Fig. 3).

3.5 Functional networks of ZFHX4

Since ZFHX4 was discovered as a new basal cell carcinoma susceptibility loci in a genome-wide association study (GWAS), we specifically focused on the functional networks of this gene. The predicted functional partner of ZFHX4 include CHD4, C1QBP, NOTCH1, NOTCH2, NOTCH3, NOTCH4, ZEB1, TMEM14A, PIAS1, HIVEP3 (Fig. 4). These findings revealed a possible linkage between ZFHX4 and notch signaling pathway.

Figure 4.

Functional networks of ZFHX4 gene.

4. Discussion

In this study a WES was performed in a Han Chinese NBCCS family and two unaffected volunteers to validate the three known pathogenic genes and to identify other potential pathogenic genes for this disease. A total of 17 candidate genes were found and most of their gene functions haven’t been systematic studied yet.

Unlike other studies [6, 7, 8], WES was used to dissect the genetic background of this disease in our research. During our early literature review, we noticed that this technique hasn’t been used in NBCCS study yet. Traditional linkage and association genetic mapping studies have limited resolution in identifying and localizing the causal genes in the genome, because a mapping region often contains hundreds of candidate genes [12, 13]. The functional testing and validation for the most promising candidates among such large lists of genes are time consuming and challenging [11]. In recent years, high-throughput sequencing technique has become cost-effective, time saving and could provide more comprehensive genetic information in each subject. Focusing on the exome region can be an efficient strategy to dissect the Mendelian disorder such as NBCCS.

None of the SNVs in the three known causative genes (PTCH1, PTCH2, SUFU) was considered pathogenic in our study. This is probably a consequence of genetic heterogeneity. As a complex genetic disease, the diagnosis criteria of NBCCS is complicated [14] and the clinical phenotype are diverse among different patients; for example, almost 10% of the patients did not manifest basal cell carcinoma [15]. Moreover, there is a significant racial variation in the incidence of the disease worldwide, the incidence in Asians are significantly lower than in Caucasus. These implied that there might be several causative genes of NBCCS among different patients.

Hh signal pathway abnormalities is the most important reason of NBCCS [16], our further functional networks analysis revealed that only two genes – AMPD1 and GPR126 have indirect relations with the components of Hh pathway. However the relationships were weak owing to the lack of sufficient information about their gene function.

GWAS have yielded a substantial number of variants that predispose to BCC, including MC1R, ASIP, TYR, SLC45A2, OCA2, IRF4 PADI6/RCC2, RHOU, TERT, KRT5, CDKN2A/B, KLF14, TP53, TGM3, RGS22, MYCN, CASP8-ALS2CR12, ZFHX4 and GATA3 [17]. Among them, ZFHX4 was the only one found in the 17 candidate genes in our study.

ZFHX4 gene was mapped to human chromosome 8q21, it was about 180 kb long, containing 12 exons that encodes a 3599-amino acid protein carrying 4 homeodomains and 22 zinc fingers. Although little is known about the function of this gene yet, our functional networks analysis proved that this gene may be involved in notch signaling pathway. Notch pathway is linked to the development of various cancers including skin cancers [18] and the notch pathway could be inactivated by zinc via a PI3K-Akt-dependent way [19]. Medulloblastoma – another clinical manifestation in NBCCS patients, was related to NuRD complex recruitment [20]. ZFHX4 was also found to interacts with the NuRD complex and regulates the glioblastoma tumor-initiating cell state [21]. This reveals a possible link between ZFHX4 and the development of medulloblastoma. In other studies, ZFHX4 was found to be susceptibility loci for several diseases, including Congenital heart defects [22], intellectual disability and/or multiple congenital anomalies [23], Hereditary congenital ptosis [24] and Peters anomaly [25]. Further studies are needed to investigate the function of ZFHX4.

We believe that this study may increase the understanding of NBCCS genetic background, especially in Asians. ZFHX4 may have the potential to become the bio-marker of NBCCS, which may be helpful in antenatal diagnosis and genetic consulting.

There were still some limitations which need to be addressed. Firstly, in this Han Chinese NBCCS family, only one living family member was affected; although we narrow the number of candidate genes down into 17 through the filtering process, a WES on two or more affected members may lead to a even less candidate gene number. Secondly, more functional analyses were warranted to elucidate the biological plausibility of this NBCCS susceptibility gene in further research.

5. Conclusions

In conclusion, we reported the identification of 17 novel candidate genes in a Han Chinese family through WES. Among them ZFHX4 was the most likely pathogenic gene for the disease in this particular family. However, the exact biological mechanism of these genes involved in NBCCS pathogenesis still needs to be clarified.

Footnotes

Supplementary data

Supplementary Fig. 1.

Distribution of the coverage, variant location and exonic function. (a) Distribution of the coverage for each sample. The x-axis represents the position of different chromosomes. The continuous red line represents the exon coverage range in each chromosome. The discontinous red line represents the mean Guanine-Cytosine (GC) content in each chromosome. The gray line represents the GC content in each chromosome. (b) Variant location in each sample. (c) Exonic function in each sample.

References

Devi

et al., Gorlin Syndrome, Indian J Dermatol 58(3) (2013), 241.

Bare

J.W.

Lebo

R.V.

and Jr

E.E.

, Loss of heterozygosity at chromosome 1q22 in basal cell carcinomas and exclusion of the basal cell nevus syndrome gene from this site, Cancer Res 52(6) (1992), 1494–1498.

Bialer

M.G.

et al., Prenatal diagnosis of Gorlin syndrome, Lancet 344(8920) (1994), 477.

Fujii

and Miyashita

, Gorlin syndrome (nevoid basal cell carcinoma syndrome): update and literature review, Pediatr Int 56(5) (2014), 667–674.

M.L.

, Nevoid basal cell carcinoma syndrome (Gorlin syndrome), Orphanet J Rare Dis 3 (2008), 32.

Johnson

R.L.

et al., Human homolog of patched, a candidate gene for the basal cell nevus syndrome, Science 272(5268) (1996), 1668–1671.

Smyth

et al., Isolation and characterization of human Patched 2 (PTCH2), a putative tumour suppressor gene in basal cell carcinoma and medulloblastoma on chromosome 1p32, Hum Mol Genet 8(2) (1999), 291–297.

Pastorino

et al., Identification of a SUFU germline mutation in a family with Gorlin syndrome, Am J Med Genet 149A(7) (2009), 1539–1543.

Majewski

et al., What can exome sequencing do for you? J Med Genet 48(9) (2011), 580–589.

10.

Yang

et al., Exome sequencing identified NRG3 as a novel susceptible gene of Hirschsprung’s disease in a Chinese population, Mol Neurobiol 47(3) (2013), 957–966.

11.

S.B.

et al., Targeted capture and massively parallel sequencing of 12 human exomes, Nature 461(7261) (2009), 272–276.

12.

Altshuler

Daly

M.J.

and Lander

E.S.

, Genetic mapping in human disease, Science 322(5903) (2008), 881–888.

13.

Redon

et al., Global variation in copy number in the human genome, Nature 444(7118) (2006), 444–454.

14.

Kimonis

V.E.

et al., Clinical manifestations in 105 persons with nevoid basal cell carcinoma syndrome, Am J Med Genet 69(3) (1997), 299–308.

15.

M.L.

et al., Nevoid basal cell carcinoma syndrome. Clinical findings in 37 Italian affected individuals, Clin Genet 55(1) (1999), 34–40.

16.

Bresler

S.C.

Padwa

B.L.

and Granter

S.R.

, Nevoid Basal Cell Carcinoma Syndrome (Gorlin Syndrome), Head Neck Pathol 10(2) (2016), 119–124.

17.

Stacey

S.N.

et al., New basal cell carcinoma susceptibility loci, Nat Commun 6 (2015), 6825.

18.

Alketbi

and Attoub

, Notch Signaling in Cancer: Rationale and Strategies for Targeting, Curr Cancer Drug Targets 15(5) (2015), 364–374.

19.

Baek

S.H.

et al., Zinc-induced downregulation of Notch signaling is associated with cytoplasmic retention of Notch1-IC and RBP-Jk via PI3k-Akt signaling pathway, Cancer Lett 255(1) (2007), 117–126.

20.

Spina

et al., Critical role of zinc finger protein 521 in the control of growth, clonogenicity and tumorigenic potential of medulloblastoma cells, Oncotarget 4(8) (2013), 1280–1292.

21.

Chudnovsky

et al., ZFHX4 interacts with the NuRD core member CHD4 and regulates the glioblastoma tumor-initiating cell stat, Cell Rep 6(2) (2014), 313–324.

22.

Osoegawa

et al., Identification of novel candidate gene loci and increased sex chromosome aneuploidy among infants with conotruncal heart defects, Am J Med Genet A 164A(2) (2014), 397–406.

23.

Vultovan Silfhout

A.T.

et al., Clinical significance of de novo and inherited copy-number variation, Hum Mutat 34(12) (2013), 1679–1687.

24.

Nakashima

et al., Genome-wide linkage analysis and mutation analysis of hereditary congenital blepharoptosis in a Japanese family, J Hum Genet 53(1) (2008), 34–41.

25.

Happ

et al., 8q21.11 microdeletion in two patients with syndromic peters anomaly, Am J Med Genet A 170(9) (2016), 2471–2475.