Abstract
Background/Aim: Historically, breast cancer has been treated according to an evaluation of biomarkers, such as the estrogen receptor and HER2 status. Recently, molecular profiling has been used to detect driver mutations and select anti-cancer treatment strategies. In addition to detecting pathogenic mutations, the total mutation count (tumor mutation burden) has been considered as another biomarker. Materials and Methods: We performed molecular profiling of 143 breast cancer tissues obtained from resected tissues via surgical operation. Results: Suspected germline mutations were detected in 10% of the patients with a higher somatic mutation ratio. Conclusion: As hypermutated breast cancers are more likely to benefit from certain anti-cancer treatment strategies, molecular profiling can be used as a biomarker.
All cancers arise as a result of mutations (germ-line/somatic) in the genome (1). Historically, breast cancer has been treated according to a biomarker evaluation, including the estrogen receptor and HER2 status (2, 3). Currently, multigene assays are widely used to predict the risk of relapse after surgery (4). While some of these variabilities are explained by traditional clinicopathological factors (including patient age, tumor stage, histological grade, and estrogen receptor status), molecular profiling studies have defined breast cancer subtypes with distinct clinical outcomes (5).
To better understand the genomic backgrounds of patients, we performed a molecular profiling study of breast cancer patients in a single institution.
Materials and Methods
This study was performed in accordance with the Declaration of Helsinki. Additionally, the requisite permissions were obtained from the ethical committee of Tokyo Women's Medical University, and written informed consent was obtained from all participants. The inclusion criteria of this study were that patients were clinically diagnosed with breast cancer and underwent surgical operation between June 2013 and September 2015. The clinical features of the patients were obtained from medical records.
From frozen tissues obtained through surgical operation, the genomic DNA was extracted using the QIAamp DNA extraction kit (QIAGEN, Hilden, Germany), according to the manufacturer's instructions. The purity and concentration of the extracted DNA were measured using a NanoDrop 1000 spectrophotometer (Thermo Scientific, Wilmington, DE, USA) and a Qubit fluorometer (Life Technologies, Carlsbad, CA, USA).
A HaloPlex HS target enrichment system (Agilent Technologies, Santa Clara, CA, USA) was used for this study. The library was designed using SureDesign (6). According to previous studies (1, 7), we selected 59 targeted genes, and all coding exons and intron–exon boundaries of them were included in the panel (Table I). Approximately 57.6 ng of DNA in a volume of 32 μl was used. A260/A280 ratios of 1.8-2.0 were considered acceptable. DNA fragmentation was assessed using agarose gel electrophoresis (2%). The DNA library was constructed using molecular indices and barcodes for simultaneous processing of multiple samples and removal of duplicated reads, respectively. Finally, the enriched library concentration was estimated using the 2200 TapeStation (Agilent Technologies). High-throughput sequencing was performed using 251-bp paired-end reads on a Mi-Seq platform (Illumina, Inc., San Diego, CA, USA), and 12 samples were sequenced at once. The raw data were aligned to the human genome build19. After annotation, FASTQ files were generated automatically. Variant calls were performed using SureCall (Agilent Technologies), and variants were extracted as excel data. Finally, filtering and curation were performed manually. The sample data were omitted from statistical analysis when the percentage of analyzable target regions covered by at least 10 reads was below 90. Prediction scores, including CADD_phred, were obtained using wANNOVAR (8).
The list of the genes included in the NGS panel.
Suspected germline variants identified in patients.
Germline variants were assumed to be observed in the NCCN guideline germline list (BRCA1, BRCA2, CDH1, NF1, PTEN, STK11, and TP53) (9). The filtering policy used was that each variant had to fulfill the following three conditions: a) a variant frequency below 1% in ExAC_EAS (The Genome Aggregation Database [former name Exome Aggregation Consortium (ExAC)]) (10), 1000G_EAS and ALL (1000 Genomes) (11), HGVD (Human Genetic Variation Database) (12), and dbSNP (13); b) a higher than 30% variant allele frequency in the total depth; c) should meet at least one of the following categories: i) variants registered in ClinVar (14) with “pathogenic” or “likely pathogenic”, ii) variants related to loss of function (nonsense or frameshift), or iii) BRCA1/2 variants not registered in ClinVar and HGMD, but a CADD_phred was higher than 15.
Somatic mutations were defined as follows: a) variant frequencies below 1% in ExAC_EAS, 1000G_EAS, 1000G_ALL, HGVD, and dbSNP; b) variants including any nonsynonymous, frameshift, inframe, and splicing acceptor/donor site mutations; c) variant allele frequency in the total depth of 25%-50%; d) a variant read depth of higher than 10; and e) CADD_phred higher than 15.
We analyzed the relationship between the clinical features of patients and the results of molecular analyses using the Wilcoxon signed-rank test and the Jonckheere–Terpstra trend test. Statistical analyses were performed using R (15). A p-value lower than 0.05 was considered statistically significant.
Results
Patients. The 142 tissues from 142 patients (their average ages at this study were 63.1±13.4; range=30-93 years) were included in this study.
Molecular analyses. In two samples, the percentage of analyzable target regions covered by at least 10 reads was below 90; thus, these two samples were removed from the statistical analysis. The minimum and maximum numbers of the total variant call in all samples were 112 and 157, respectively. After filtering, 303 variant calls (193 variant types in 43 genes) remained (Data are available upon request). One of the variants in TP53 was drug treatable/actionable. Among the identified variants, 14 types of variants were considered as germline variants, owing to the higher allele frequencies in the read depth (Table II). One patient had two suspected germline variants.
Statistical analysis. Fourteen patients with suspected germline variants had 4.5 variants on average. In contrast, 126 patients showed 1.9 variants on average. There was a statistically significant difference between the two groups (Table III). Regarding the clinical features, there was no statistical significance in correlation with the number of total variants (Table IV).
Comparison based on suspected germline mutations.
Discussion
In this study, 14 types of suspected germline variants were identified in 14 samples from 140 patients, indicating that 10% of the patients showed possible hereditary/familial cancers. Four patients showed suspected BRCA1/2 germline variants. It is known that mutations in BRCA1/2 are responsible for the hereditary breast and ovarian cancer (HBOC) syndrome (16). TP53 variants were identified in three patients. This gene is related to Li-Fraumeni syndrome. Breast cancer can occur in patients with Li-Fraumeni syndrome. Similarly, PTEN, NF1, and CDH1 can also be related to familial breast cancer (17). Patients with suspected germline variants showed higher statistically significant variant calls. This is reasonable because loss of function of tumor suppressor genes causes multiple gene mutations.
Recently, it has been considered that the more somatic mutations a tumor has, the more neoantigens it is likely to form (18). This condition is known as tumor mutational burden (TMB) and can represent a useful estimation of the tumor neoantigenic load. TMB is an emerging biomarker in cancer and has been associated with microsatellite instability, defective DNA replication/repair, and response to programmed cell death -1 (PD-1) and programmed death-ligand 1 (PD-L1) blockade immunotherapy (19). As hypermutated breast cancers are more likely to benefit from PD-1 inhibitors, mutation analysis can be used as a biomarker (20).
There are some limitations in this study. Because this study was not designed to analyze germline DNA extracted from normal tissues, including blood, it was impossible to confirm true germline mutations. The study design was planned in 2016. At that time, driver genes for breast cancer were not fully understood and some important genes were not included in the gene panel. Therefore, most of the conclusions were drawn from the information available by our study design. However, the final conclusion does not contradict previous studies.
Results of statistical analyses of total variant call numbers.
Footnotes
Authors' Contributions
TK, HK and TY designed the study. TA, YO, and BW carried out all experiments. TY supervised all experiments performed. TU, KY-S and MW participated in the data interpretation. TK drafted the manuscript. TY prepared the final manuscript. As the principal investigator, TO supervised the study. All Authors read and approved the final manuscript.
Conflicts of Interest
The Authors declare they have no conflicts of interest regarding this study.
- Received June 5, 2020.
- Revision received June 26, 2020.
- Accepted June 29, 2020.
- Copyright© 2020, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved
