Women having germline mutations in their BRCA1 or BRCA2 genes have a 45-80 percent chance of developing breast cancer over their lifetime. Both BRCA1 and BRCA2 are thought to be tumor suppressor genes that require inactivation of both alleles for tumorigenesis to occur. Bi-allelic inactivation is considered to be an initiating event, however, recent research has shown that this may not be the case. Those studies showed that the wild-type (WT) allele of BRCA1 may undergo a loss of heterozygosity (LOH) as a late event, thus indicating that some other somatic event or events (i.e., only in tumor cells, not germline) initiate tumorigenesis prior to the inactivation of BRCA1/2. If the somatic genetic alterations in BRCA1/2 (i.e., mutations or LOH) are present in nearly all of the analyzed cells, then those mutations are termed clonal, meaning that those mutation(s) occurred at or near the onset of tumorigenesis.
Conversely, if a mutation is found in only a percentage of the cells, then that mutation occurred at some point after tumorigenesis, and is thus termed subclonal. In order to better understand the chronology of the somatic events which can lead to BRCA1/2 breast cancer, Felipe Geyer, MD, Memorial Sloan Kettering Cancer Center, New York, N.Y., and colleagues there and at Beth Israel Deaconess Medical School, Boston, undertook a study to evaluate the genomic landscape of tumors taken from BRCA1 and BRCA2 patients. In doing so, the researchers sought to ascertain whether the BRCA1/2 WT allele LOH and/or mutations of other tumor suppressor genes were clonal or subclonal in these samples.
Role of BRCA1/BRCA2
BRCA1 and BRCA2 are important tumor suppressor genes. These genes produce the proteins BRCA1 and BRCA2, which both function to help repair DNA. When asked by what mechanism they work, Geyer replied, “Both proteins assist in the repair of double-strand DNA breaks by participating in a nuclear complex that performs homologous recombination, the most accurate method for repairing that sort of DNA damage.”
For double strand break repairs, in homologous recombination, the repairing protein complex uses an intact homologous DNA sequence as a template for repairs. This sequence may be from a sister chromatid, the same chromosome or a homologous chromosome, depending upon the phase of the cell cycle when the repair occurs. “In addition to this important function, BRCA1 also serves other vital functions” he added. This protein also participates in ubiquitination, transcription and its regulation (Curr Opin Cell Biol 2003;15(3):345).
In this study, stored tumor and normal breast tissue samples were obtained from 29 BRCA1 and 10 BRCA2 breast cancer patients at the authors' institutes. From these, the DNA that was extracted was subjected to massively parallel sequencing via MSK-IMPAKT and/or an assay that targeted all the exons of 254 genes that were relevant to breast cancer or DNA repair. For the 111 genes covered in both sequencing assays, copy number alterations, insertions, deletions, and somatic single nucleotide variants were obtained. These data were subjected to analysis by the bioinformatics algorithms ABSOLUTE (Nat Biotechnol 2012;30(5):413-421) and FACETS (Bioinformatics 2012;28(20):2624).
“Using these programs, we were able to determine the fraction of cells harboring each mutation, or the cancer cell fraction; we were also able to obtain a copy number assessment and LOH status for these genes,” Geyer elaborated.
“In normal cells, we have two copies of each gene, however in cancer you can have fewer or more copies of these genes present in tumor cells,” he further explained. “An example particularly relevant to breast cancer where one can find more copies of the gene, or overexpression, is in the case of the HER2 gene.” The analyses done allowed the researchers to determine if the mutations present were clonal or subclonal, as well as assess if LOH had occurred for the BRCA1 and BRCA2 WT alleles.
All BRCA2 samples showed biallelic inactivation as the result of a LOH for the BRCA2 WT allele, whereas, for BRCA1 samples, 28 of 29 had LOH for the WT allele. While BRCA1 LOH was mostly clonal, in six BRCA1 samples, LOH of BRCA1 WT allele was subclonal and unlikely to be the initiating somatic genetic event. The one BRCA1 sample that did not have biallelic inactivation was from a breast cancer patient with ER+ lobular carcinoma that was diagnosed at 62 years old. This sample did not have the typical characteristics associated with defects in homologous recombination repair, and could be a sporadic breast cancer arising in a BRCA1 germline mutation carrier. In BRCA1 samples, somatic events included inactivating mutations or homozygous deletions to the following tumor suppressor genes: TP53, NF1, RB1, CDKN2A and PTEN. While the majority of BRCA1 samples showed additional mutations in cancer-related genes, 40 percent of the BRCA2 samples showed no mutations in those same genes.
For the TP53 gene, somatic mutations were noted in 76 percent of the BRCA1 samples, and of this large percentage, 58 percent were clonal. Interestingly, only 10 percent of the BRCA2 samples had mutations in this gene. Geyer noted, “In the subset of BRCA1 tumors with subclonal LOH of the BRCA1WT allele, TP53 mutations were found to be clonal in 50 percent of them, and therefore, might have preceded BRCA1 LOH in tumor development.”
TP53 encodes tumor protein 53 (p53) a rather important protein that inhibits tumorigenesis via a variety of mechanisms. This protein performs the following antitumor functions:
- arrest the cell cycle at the G1/S phase when DNA damage is detected;
- activate DNA repair proteins;
- induce apoptosis if the DNA damage is irreparable; and
- play an essential role in the senescence response to short telomeres.
Note, that by halting the cell cycle, the activated repair proteins can then fix the DNA and eventually, the cell cycle can be continued; if the DNA can't be repaired, then apoptosis is induced. Mutations to TP53 could be especially troubling as the proteins produced by those mutated genes may become oncogenic (Cell Death and Differentiation 2015;22,1239-1249). The authors for that review note that several p53 mutants have gains of functions for several activities that are essential for neoplastic transformation. It should be noted that there is some debate regarding these findings.
The loss of function for homologous recombination repair in tumor cells means that those cells are vulnerable to double strand breaks in their DNA. Geyer explained, “This makes these tumors sensitive to certain therapies which may produce those double strand breaks in their DNA. One such therapy is the use of PARP inhibitors.”
PARP1 (poly ADP ribose polymerase type 1) is a protein that aids in the repair of single strand DNA breaks (or nicks). If these single strand breaks remain until DNA replication, then that act can form double strand breaks. Another strategy involves the use of DNA cross-linking agents. One class of compounds used for this type of cancer therapy is alkylating agents such as carmustine. Another family of DNA cross-linking agents is platinum complexes such as cisplatin or carboplatin. This last class of compounds is especially relevant to the BRCA1 gene.
In a retrospective study of non-small cell lung cancer patients undergoing second line platinum-based therapy, Souglakos et al. (J Thorac Oncol 2002;7(4):663-671) found that patients having decreased levels of mRNA expression for BRCA1 and/or ERCC1 (DNA excision repair protein ERCC1) in their primary tumors at the time of diagnosis had higher response rates and longer progression-free survival and overall survival. The patients who benefitted the most from cisplatin therapy were those having low expression rates for both BRCA1 and ERCC1. The tumors having low expression of those DNA repair-associated genes are thought to have enhanced sensitivity to the platinum-based DNA cross linking agents used. The authors suggested caution when considering the data, as their study was small; however, they do feel the results warrant a larger study to further explore their findings.
When the most accurate means of DNA repair is no longer available (homologous recombination), the tumors cells must then rely on other more error-prone methods. When asked about these methods, Geyer replied, “One such method for repairing double strand breaks is non-homologous end joining.”
This method is less accurate because it does not utilize a homologous sequence for repair, but instead relies upon short sequences called microhomologies. The use of these shorter sequences may lead to errors such as telomere fusion or translocation, two traits commonly associated with tumor cells. If that avenue of DNA repair is not available, an even more error-prone method for repair may be employed, microhomology-mediated end joining (MMEJ). In MMEJ, microhomologies (5-25 base pairs) located on either side of the break are utilized to guide the repair. This can lead to the deletion of the sequence between the homologies as well as translocation, inversion or other chromosomal abnormalities.
“Based on the errors in DNA repair that can occur when homologous recombination is not available, one can readily see how a tumor can consist of genetically-diverse cells,” Geyer noted. “In some instances, mutations can occur where the cells regain the ability to perform homologous recombination; those are called revertant mutations. Tumor cells that regain that ability would no longer be as sensitive to those therapies that cause double strand DNA breaks.
When asked about the future of this project, he stated, “One possibility we are looking to explore is the use of single-cell sequencing. As the costs associated with sequencing come down, at some point, we may be able to use this technique for 1,000-plus individual cells from a patient's tumor; this would give us a much clearer picture of the tumor cells' genetic makeup. The data obtained may direct the patient's treatment, or perhaps even suggest new targets for future therapies.”
Richard Simoneaux is a contributing writer.