Prioritizing Variants in Complete Hereditary Breast and Ovarian Cancer Genes in Patients Lacking Known BRCA Mutations - Caminsky - 2016 - Human Mutation
Introduction
Currently,
the lifetime risk for a woman to develop breast cancer (BC) is 12.3%
and 1.3% in the case of ovarian cancer (OC [Howlander et al.,
2014]).
Approximately 5–10% of all BC cases are hereditary in nature, versus
25% for OC, where relative risk (RR) of BC or OC with one affected first
degree family member is estimated at 2.1 and 3.1, respectively
[Stratton et al.,
1998; Walsh et al.,
2011]. Two highly penetrant genes,
BRCA1 and
BRCA2,
are associated with a large proportion of hereditary breast and ovarian
cancer (HBOC) cases. However, the estimated rate of linkage to these
genes is
significantly higher than the proportion of pathogenic
mutations identified in HBOC families [Ford et al.,
1998], suggesting unrecognized or unidentified variants in
BRCA1/2.
Clinical
BRCA1/2
testing is restricted primarily to coding regions. Limitations on how
variants can be interpreted, lack of functional validation, and
mutations in other genes contribute to uninformative results.
The
heritability that is not associated with BRCA genes is likely
due to other genetic factors rather than environmental causes,
specifically moderate- and low-risk susceptibility genes [Antoniou and
Easton,
2006]. Hollestelle et al. [
2010]
point out the challenges in estimating increased risks associated with
mutations in these genes, as the disease patterns are often incompletely
penetrant, and require large pedigree studies to confidently assess
pathogenicity.
Next-generation sequencing
(NGS) of gene panels for large cohorts of affected and unaffected
individuals has become an increasingly popular approach to confront
these challenges. Numerous HBOC gene variants have been catalogued,
including cases in which RR
(risk ratio) has been determined; however,
the literature
is also flooded with variants lacking a clinical interpretation [Cassa
et al.,
2012].
It is not feasible to functionally evaluate the effects all of the
variant(s) of uncertain significance (VUS) identified by NGS. Further,
in silico assessment of protein coding variants has not been entirely reliable [Rogan and Zou,
2013; Vihinen,
2013]. Several approaches have been developed to better assess variants from exome and genome-wide NGS data [Duzkale et al.,
2013; Kircher et al.,
2014].
Nevertheless, there is an unmet need for other methods that quickly and
accurately bridge variant identification and classification.
To
begin to address this problem, we sought to provide potentially novel
interpretations of noncoding sequence changes, based on disruption or
acquisition of interactions with proteins that recognize nucleic acid
binding sites. Information theory (IT) based analysis predicts changes
in sequence binding affinity, and it has been applied and validated for
use in the analysis of splice sites (SSs),......
In the present study, we have sequenced 13 additional genes that have been deemed HBOC susceptibility loci (BARD1, EPCAM, MLH1, MRE11A, MSH2, MSH6, MUTYH, NBN, PMS2, PTEN, RAD51B, STK11, and XRCC2 [Minion et al., 2015]). These genes encode proteins with roles in DNA repair, surveillance, and cell cycle regulation (Fig. 1; for further evidence supporting this gene set see Supp. Table S1 [Apostolou and Fostira, 2013; Al Bakir and Gabra, 2014]),
and are associated with specific disease syndromes that confer an
increased risk of BC and OC, as well as many other types of cancer
(Supp. Table S2). High-risk genes confer >4 times increased risk of
BC compared to the general population. BRCA1 and BRCA2 are estimated to increase risk 20-fold [Antoniou et al., 2003]....
Figure 1. (below) Common genomic pathways among 20
HBOC genes, including risk and relevant literature. The left, top, and
right circles indicate sequenced genes that play important roles in the
mismatch repair (MMR), Fanconi anemia, and DNA double-strand break
repair pathways, respectively. The bottom circle contains genes involved
in cell cycle control.
Genes considered to present a high risk of
breast and/or ovarian cancer when mutated are bolded, moderate-risk
genes are underlined, and low-risk genes are in normal font. The
estimated number of articles listing a gene's association with breast or
ovarian cancer (based on a systematic search in PubMed [performed June
2015]) is indicated in superscript. **
MUTYH is only high risk in the case of bi-allelic mutations.
*EPCAM
is not involved in any pathways, but is associated with hereditary
nonpolyposis colorectal cancer (HNPCC) by virtue of the fact that 3′
deletions of EPCAM can cause epigenetic silencing of MSH2, causing Lynch syndrome protein. See Supp. Table S1 for citations and further evidence supporting this gene set.
