ISSN: 0970-938X (Print) | 0976-1683 (Electronic)
An International Journal of Medical Sciences
Research Article - Biomedical Research (2017) Volume 28, Issue 18
Xingwen Wang1, Yunyan Wu2, Qiang Li1, Dongxiao Lv1, Junqing Han1* and Ping Liu3*
1Department of Tumor Research and Therapy Center, Shandong Provincial Hospital Affiliated to Shandong University, Jinan, Shandong, PR China
2Department of Ophthalmology, Shandong Provincial Hospital Affiliated to Shandong University, Jinan, Shandong, PR China
3Department of Pharmacy, Shandong Provincial Hospital Affiliated to Shandong University, Jinan, Shandong, PR China
Accepted on August 30, 2017
The objective of the present study was to determine whether the alterations of cancer predisposition genes were different between radiosensitive and radioresistant patients with Nasopharyngeal Carcinoma (NPC). A total of 21 patients with nasopharyngeal carcinoma were included in this study. All patients were treated with standardized radiotherapy. Sixteen of the tumors were clinically radiosensitive and 5 were radioresistant. Genomic DNA, extracted from Formalin-Fixed Paraffin-Embedded (FFPE) tumor specimens obtained prior to treatment, was subjected to amplicon-based Next-Generation Sequencing (NGS) with primer sets targeting 50 critical human tumor suppressor genes and oncogenes. We identified 18 nonsynonymous mutations, 1 nonframeshift deletion and 1 frameshift mutation distributed across 15 genes, including 11 mutations have been reported in COSMIC (the Catalogue of Somatic Mutations in Cancer) or dbSNP database (database of single nucleotide polymorphisms), and 9 novel mutations. Most of these mutations have not been reported in NPC. More importantly, 5 radiosensitivespecific mutations targeting AKT1, PIK3CA, MET, TP53 and STK11 were observed, suggesting the genetic alterations of PI3K/AKT and p53 pathways were involved in the response to radiotherapy. Collectively, genetic mutations may differentiate tumor radiosensitivity and radioresistance, although validation of such mutations using a large sample size cohort is necessary before a solid conclusion can be reached.
Mutations, Amplicon, Next-generation sequencing, Nasopharyngeal carcinoma.
Nasopharyngeal Carcinoma (NPC) is extremely common in southern China, particularly in Guangdong, accounting for 18% of all cancers in China [1]. Epstein-Barr virus infection, heredity and environmental influences are major causes of NPC [2-4]. Most NPC patients are sensitive to radiotherapy, and combined radiochemotherapy may improve the survival of these patients [5,6]. However, radioresistance remains a serious problem affecting the treatment of NPC in many cases. If patients who tend to be resistant to radiotherapy can be identified before treatment and receive a higher dose of therapy without increased the rate of complications, the outcome of radiotherapy or combined radiochemotherapy may be significantly improved.
Microarrays and proteomics have been applied to identify radioresistance associated genes or proteins in multiple types of cancer, including NPC [7-15]. A number of studies have focused on the association between the radiosensitivity of various cancers and genetic changes of oncogenes and tumor suppressor genes, such as p53 [16-18], ATM (19), Epidermal Growth Factor Receptor (EGFR) [20,21]. Although mutations of RB2/p130 [22], p53 [23,24], PIK3CA [25-27] and EGFR [28] have been investigated in NPC, little is known about radioresistance-associated mutations in NPC.
With the advances in Next-Generation Sequencing (NGS) techniques, a large number of mutations have been identified in cancers [29]. Multiplex amplicon-based NGS, sequencing only the Regions of Interest (ROI), has several advantages: requires relatively small amounts of DNA, is relatively low cost and is compatible with Formalin-Fixed Paraffin-Embedded (FFPE)- DNA [30]. In the current study, genomic DNA was extracted from FFPE tissue samples from radiosensitive and radioresistant patients with NPC, and then the hotspot regions of 50 critical human tumor suppressor genes and oncogenes were amplified. Amplicon-based NGS was carried out to investigate any potential correlation of mutational hot spots with radiosensitivity.
Collection of NPC specimens
A total of 21 Formalin-Fixed Paraffin-Embedded (FFPE) samples were taken form Shandong Provincial Hospital Affiliated to Shandong University, China after approval by the local ethics committee. All patients were given written informed content. All patients were treated with two cycles of radiotherapy (50-70 Gy total dose) after cisplatin-docetaxel induction chemotherapy. Four response categories are proposed: complete response (CR, which means the disappearance of all detectable tumor), partial response (PR, which is defined as more than 50% decrease in tumor size), stable disease (SD, which means tumor remains the same size), and Progressive Disease (PD). Radiosensitive patients are those reached CR 2 to 4 w after completion of radiotherapy, and radioresistant patients are clarified as those of PR or SD or even with PD 2 to 4 w after completion of radiotherapy. FFPE specimens were obtained prior to treatment, and 16 patients were found radiosensitive (Samples 1-16) and 5 patients (Samples 17-21) were found radioresistant during follow-up.
DNA isolation and amplicon library preparation
Genomic DNA was extracted separately from FFPE samples using QIAamp DNA FFPE Tissue Kit (QIAGEN, Hilden, Germany) according to the manufacturer’s protocol. The concentrations of isolated FFPE-DNA samples were determined on a Qubit 2.0 Fluorometer (Life Technologies).
The amplicon library was prepared using DNA Seq Library Preparation Kit for Amplicon Sequencing-Illumina Compatible (Gnomegen, San Diego, CA), which contains 207 primer pairs and targets 50 proto-oncogenes and tumor suppressor genes (Table 1). Target sequences were amplified with 10 ng FEPEDNA. Subsequently, end repair, barcode ligation and a simple PCR amplification were performed following the manufacturer’s instructions with several purification steps (Gnomegen). The barcode sequences were unique for each sample. Size selection was then performed to obtain clean library products. Following Quality Control, 21 purified and quantified libraries were pooled to form a library product.
ABL1 | AKT1 | ALK | APC | ATM | BRAF | CDH1 | CDKN2A | CSF1R | EZH2 |
CTNNB1 | EGFR | ERBB2 | ERBB4 | FBXW7 | FGFR1 | FGFR2 | FGFR3 | FLT3 | GNA11 |
GNAS | GNAQ | HNF1A | HRAS | IDH1 | IDH2 | JAK2 | JAK3 | KDR | KIT |
KRAS | MET | MLH1 | MPL | NOTCH1 | NPM1 | NRAS | PDGFRA | PIK3CA | PTEN |
PTPN11 | RB1 | RET | SMAD4 | SMARCB1 | SMO | SRC | STK11 | TP53 | VHL |
Table 1. The panel of 50 proto-oncogenes and tumor suppressor genes.
Illumina sequencing and sequence assembly
Sequencing was performed on an Illumina Hiseq2000 sequencing platform (Illumina, San Diego, CA) with 150 bp pair reads. After sequencing, the raw data were separately trimmed to remove 5’- and 3’- adapters and low quality regions (<Q20). After discarding reads shorter than 50 bp, reads were mapped against a reference genome using Burrows-Wheeler Alignment tool (BWA). Variances were identified by VarScan software (http://varscan.sourceforge.net) with the following parameters: minimum base quality of 20, minimum coverage of 30X, minimum variation frequency of 3%, no strand bias, and a P value of less than 0.05.
Sequencing data
We applied NGS to investigate mutations of 50 protooncogenes and tumor suppressor genes on 5 radioresistant (Samples 17-21) and 16 radiosensitive (Samples 1-16) patients. We generated 1.3-2.4 million sequencing reads per sample (Q20>96% and Q30>93%). As shown in Table 2, the align rate was >98%, on target rate was >93%, the average read depth (the number of times a nucleotide is sequenced) per sample was 5000X, and coverage uniformity was >90%. Thus, the sequencing data allowed the detection of variances.
Sample | Aligned reads (M) | Align rate | On target reads (M) | On target rate | Mean depth (1000x) | Specificity | Coverage uniformity |
---|---|---|---|---|---|---|---|
1 | 1.929 | 99.34% | 1.889 | 97.91% | 8.847 | 97.13% | 91.41% |
2 | 1.678 | 99.31% | 1.637 | 97.54% | 7.661 | 96.79% | 90.55% |
3 | 1.322 | 99.40% | 1.301 | 98.47% | 6.128 | 97.84% | 89.10% |
4 | 1.568 | 99.28% | 1.528 | 97.44% | 6.994 | 96.40% | 90.85% |
5 | 1.556 | 99.41% | 1.504 | 96.67% | 7.149 | 96.12% | 78.11% |
6 | 2.423 | 99.30% | 2.362 | 97.48% | 11.123 | 96.67% | 90.52% |
7 | 1.568 | 99.16% | 1.514 | 96.59% | 7.105 | 95.75% | 81.89% |
8 | 1.966 | 99.11% | 1.894 | 96.38% | 8.898 | 95.50% | 82.63% |
9 | 1.71 | 99.27% | 1.678 | 98.13% | 7.893 | 97.41% | 93.80% |
10 | 1.67 | 98.78% | 1.636 | 97.98% | 7.605 | 96.69% | 87.29% |
11 | 1.68 | 99.27% | 1.645 | 97.92% | 7.762 | 97.15% | 86.70% |
12 | 2.012 | 99.24% | 1.967 | 97.74% | 9.192 | 96.97% | 85.00% |
13 | 1.674 | 98.69% | 1.646 | 98.34% | 7.574 | 96.71% | 89.75% |
14 | 2.273 | 99.95% | 2.249 | 98.91% | 3.076 | 89.14% | 74.06% |
15 | 1.962 | 99.92% | 1.936 | 98.67% | 3.239 | 86.72% | 76.50% |
16 | 1.844 | 99.97% | 1.808 | 98.08% | 3.229 | 88.53% | 77.91% |
17 | 2.017 | 99.98% | 1.978 | 98.04% | 3.333 | 88.10% | 78.40% |
18 | 1.602 | 99.97% | 1.573 | 98.19% | 3.051 | 88.14% | 79.51% |
19 | 1.925 | 99.98% | 1.902 | 98.82% | 3.187 | 85.83% | 75.21% |
20 | 1.971 | 99.11% | 1.918 | 97.30% | 9.112 | 96.43% | 93.22% |
21 | 1.854 | 99.36% | 1.73 | 93.29% | 8.248 | 92.60% | 87.56% |
Table 2. Quality control of sequencing data.
Frequency and distribution of variances
After data processing, 67 variances were detected in the 21 samples (Table S1). Then, we excluded 27 variances within Untranslated Region (UTR) and intron, as well as 20 synonymous mutations (no amino acid change occurred), and obtained 20 distinct variants (18 nonsynonymous, 1 nonframeshift deletion and 1 frameshift) of 15 genes (PTEN, HRAS, AKT1, IDH2, CDH1, TP53, SMAD4, STK11, PIK3CA, FGFR3, KDR, APC, MET, FGFR1 and NOTCH1) (Table 3). Of the 20 distinct variants, 10 were found in COSMIC database (the Catalogue of Somatic Mutations in Cancer) [31] and 8 variants were found in dbSNP database (database of single nucleotide polymorphisms) [32]. Nine novel variants, which were not found in COSMIC database or dbSNP, were then subjected to Polyphen2 (http://genetics.bwh.harvard.edu/ pph2/). The results showed that 8 novel variants were predicted to be “probably damaging” or “possibly damaging”.
Screening for radioresistance-associated variances
By comparing the 20 variances occurred in radiosensitive and radioresistant samples, we found that 5 variants occurred only in radiosensitive samples, but not in radioresistant samples. These 5 variants may be associated with radioresistance (Table 3). No obvious difference was observed in other 15 variants between the two groups of samples. Table 1.
Gene | Accession | Chromosome | Mutations | Cosmic | dbSNP | Polyphen-2 | |
---|---|---|---|---|---|---|---|
Radiosensitive | AKT1 | NM_001014431 | chr14 | E17K | COSM33765 | rs121434592 | |
TP53 | NM_001126118 | chr17 | S51fs | COSM131026, COSM131025, COSM131024 | |||
STK11 | NM_000455 | chr19 | F354L | COSM21360 | rs59912467 | ||
PIK3CA | NM_006218 | chr3 | R108H | COSM27497, COSM342716 | |||
MET | NM_000245 | chr7 | N375S | COSM710 | rs33917957 | ||
Others | PTEN | NM_000314 | chr10 | Q110E | no entry | 0.001 (benign) | |
HRAS | NM_001130442 | chr11 | G13D | COSM490 | rs104894226 | ||
IDH2 | NM_001290114 | chr15 | P32H | no entry | 0.998 (probably) | ||
CDH1 | NM_004360 | chr16 | P372H | no entry | 0.999 (probably) | ||
TP53 | NM_001126118 | chr17 | P33R | COSM250061 | rs1042522 | ||
SMAD4 | NM_005359 | chr18 | D424N | no entry | 0.575 (possibly) | ||
PIK3CA | NM_006218 | chr3 | Y68C | no entry | 1 (probably) | ||
PIK3CA | NM_006218 | chr3 | E545A | COSM297145, COSM12458 | rs121913274 | ||
PIK3CA | NM_006218 | chr3 | L719V | no entry | 0.849 (possibly) | ||
FGFR3 | NM_000142 | chr4 | V266M | COSM29434 | |||
KDR | NM_002253 | chr4 | Q472H | COSM149673 | rs1870377 | ||
APC | NM_000038 | chr5 | A1582T | no entry | 1 (probably) | ||
APC | NM_000038 | chr5 | A1582P | no entry | 1 (probably) | ||
FGFR1 | NM_001174066 | chr8 | 43_44del | no entry | rs138489552 | ||
NOTCH1 | NM_017617 | chr9 | E1567D | no entry | 0.549 (possibly) |
Table 3. Identified distinct variants.
Radioresistance remains a major obstacle to the treatment of NPC. Genetic changes have been reported to be associated with the radiosensitivity of various cancers. Identification of NPC radioresistance-associated mutations will help finding biomarkers to predict the response to radiotherapy and explaining the molecular mechanisms of NPC radioresistance. As far as we known, this study represents the first sequencing study of multiple oncogenes and tumor suppresser genes in radiosensitive and radiotolerant NPC patients.
In this study, NPC samples were tested with a 50-gene hotspot panel, of which, mutations in PIK3CA, NRAS, KIT, PDGFRA, ABL, HRAS, BRAF and EGFR [24,26-28,33,34] have been reported in NPC. Here, we identified 18 nonsynonymous mutations, 1 nonframeshift deletion and 1 frameshift mutation distributed across 15 oncogenes and tumor suppresser genes, of which 8 mutations were previously observed in dbSNP and 10 mutations, were already known Cancer-Associated Mutations (COSMIC). By analyzing with Polyphen-2 Software, 8 of the 9 novel mutations were predicted to be damaging. These results suggested that we have identified novel mutations which may be associated with tumorigenesis. Alterations in most of the 15 genes have not been reported in NPC, except PI3CA and TP53. Our study also provided novel information regarding gene alterations during NPC tumorigenesis.
The PI3K/AKT pathway is known to regulating radioresistance in cancer cells [35-37] and experimental animals [38]. p53 is a well-known regulator of radiosensitivity in tumor cells [39,40]. By comparing radioresistant and radiosensitive samples, 5 radiosensitive-specific mutations, AKT1 (COSM33765), PIK3CA (COSM27497, COSM342716), MET (COSM710), TP53 (COSM131026, COSM131025 and COSM131024) and STK11 (COSM21360), were identified, suggesting the genetic alterations of PI3K/AKT and p53 pathways were involved in the response to radiotherapy. However, these mutations should be validated by large sample size cohort. In vitro and in vivo experiments with these mutations may further elucidate the underlying mechanisms.
In summary, the present study has identified several candidate variants in NPC patients that may be used to predict increased radiosensitivity although the sample size was relatively small. Our findings reported may be used to distinguish radiosensitive and radioresistant patients before treatment, thus driving personalized therapeutic strategies and improving the therapy outcome.
This work was supported by the grant from the Science and Technology Development Plan Project of Shandong Province (No. 2012G0021822).
No conflict of interest to declare.