THOR is a targetable epigenetic biomarker with clinical implications in breast cancer.

BACKGROUND: Breast cancer (BC) is the most frequently diagnosed cancer and a leading cause of death among women worldwide. Early BC is potentially curable, but the mortality rates still observed among BC patients demonstrate the urgent need of novel and more effective diagnostic and therapeutic options. Limitless self-renewal is a hallmark of cancer, governed by telomere maintenance. In around 95% of BC cases, this process is achieved by telomerase reactivation through upregulation of the human telomerase reverse transcriptase (hTERT). The hypermethylation of a specific region within the hTERT promoter, termed TERT hypermethylated oncological region (THOR) has been associated with increased hTERT expression in cancer. However, its biological role and clinical potential in BC have never been studied to the best of our knowledge. Therefore, we aimed to investigate the role of THOR as a biomarker and explore the functional impact of THOR methylation status in hTERT upregulation in BC. RESULTS: THOR methylation status in BC was assessed by pyrosequencing on discovery and validation cohorts. We found that THOR is significantly hypermethylated in malignant breast tissue when compared to benign tissue (40.23% vs. 12.81%, P < 0.0001), differentiating malignant tumor from normal tissue from the earliest stage of disease. Using a reporter assay, the addition of unmethylated THOR significantly reduced luciferase activity by an average 1.8-fold when compared to the hTERT core promoter alone (P < 0.01). To further investigate its biological impact on hTERT transcription, targeted THOR demethylation was performed using novel technology based on CRISPR-dCas9 system and significant THOR demethylation was achieved. Cells previously demethylated on THOR region did not develop a histologic cancer phenotype in in vivo assays. Additional studies are required to validate these observations and to unravel the causality between THOR hypermethylation and hTERT upregulation in BC. CONCLUSIONS: THOR hypermethylation is an important epigenetic mark in breast tumorigenesis, representing a promising biomarker and therapeutic target in BC. We revealed that THOR acts as a repressive regulatory element of hTERT and that its hypermethylation is a relevant mechanism for hTERT upregulation in BC.

Allelic expression imbalance of PIK3CA mutations is frequent in breast cancer and prognostically significant.

PIK3CA mutations are the most common in breast cancer, particularly in the estrogen receptor-positive cohort, but the benefit of PI3K inhibitors has had limited success compared with approaches targeting other less common mutations. We found a frequent allelic expression imbalance between the missense mutant and wild-type PIK3CA alleles in breast tumors from the METABRIC (70.2%) and the TCGA (60.1%) projects. When considering the mechanisms controlling allelic expression, 27.7% and 11.8% of tumors showed imbalance due to regulatory variants in cis, in the two studies respectively. Furthermore, preferential expression of the mutant allele due to cis-regulatory variation is associated with poor prognosis in the METABRIC tumors (P = 0.031). Interestingly, ER-, PR-, and HER2+ tumors showed significant preferential expression of the mutated allele in both datasets. Our work provides compelling evidence to support the clinical utility of PIK3CA allelic expression in breast cancer in identifying patients of poorer prognosis, and those with low expression of the mutated allele, who will unlikely benefit from PI3K inhibitors. Furthermore, our work proposes a model of differential regulation of a critical cancer-promoting gene in breast cancer.

Germline allelic expression of genes at 17q22 locus associates with risk of breast cancer.

INTRODUCTION: Translation of genome-wide association study (GWAS) findings into preventive approaches is challenged by the identification of the causal risk variants and the understanding of the biological mechanisms by which they act. We present using allelic expression (AE) ratios to perform quantitative case-control analysis as a novel approach to identify risk associations, causal regulatory variants, and target genes. METHODS: Using the breast cancer (BC) risk locus 17q22 to validate this approach, we measured AE ratios in normal breast tissue samples from controls and cases, as well as from unmatched blood samples. Then we used in-silico and in-vitro analysis to map and functionally characterised candidate causal variants. RESULTS: We found a significant shift in the AE patterns of STXBP4 (rs2628315) and COX11 (rs17817901) in the normal breast tissue of cases and healthy controls. Preferential expression of the G-rs2628315 and A-rs17817901 alleles, more often observed in cases, was associated with an increased risk for BC. Analysis of blood samples from cases and controls found a similar association. Furthermore, we identified two putative cis-regulatory variants - rs17817901 and rs8066588 - that affect a miRNA and a transcription factor binding site, respectively. CONCLUSION: We propose causal variants and target genes for the 17q22 BC risk locus and show that using AE ratios in case-control association studies is helpful in identifying risk and mapping causal variants.

quincunx: an R package to query, download and wrangle PGS Catalog data.

MOTIVATION: The Polygenic Score (PGS) Catalog is a recently established open database of published polygenic scores that, to date, has collected, curated and made available 721 polygenic scores from over 133 publications. The PGS Catalog REST API is the only method allowing programmatic access to this resource. RESULTS: Here, we describe quincunx, an R package that provides the first client interface to the PGS Catalog REST API. quincunx enables users to query and quickly retrieve, filter and integrate metadata associated with polygenic scores, as well as polygenic scoring files in tidy table format. AVAILABILITY AND IMPLEMENTATION: quincunx is freely available under an MIT License, and can be accessed from https://github.com/maialab/quincunx. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

The History of Lentil (Lens culinaris subsp. culinaris) Domestication and Spread as Revealed by Genotyping-by-Sequencing of Wild and Landrace Accessions.

Protein-rich legumes accompanied carbohydrate-rich cereals since the beginning of agriculture and yet their domestication history is not as well understood. Lentil (Lens culinaris Medik. subsp. culinaris) was first cultivated in Southwest Asia (SWA) 8000-10,000 years ago but archeological evidence is unclear as to how many times it may have been independently domesticated, in which SWA region(s) this may have happened, and whether wild species within the Lens genus have contributed to the cultivated gene pool. In this study, we combined genotyping-by-sequencing (GBS) of 190 accessions from wild (67) and domesticated (123) lentils from the Old World with archeological information to explore the evolutionary history, domestication, and diffusion of lentils to different environments. GBS led to the discovery of 87,647 single-nucleotide polymorphisms (SNPs), which allowed us to infer the phylogeny of genus Lens. We confirmed previous studies proposing four groups within it. The only gene flow detected was between cultivated varieties and their progenitor (L. culinaris subsp. orientalis) albeit at very low levels. Nevertheless, a few putative hybrids or naturalized cultivars were identified. Within cultivated lentil, we found three geographic groups. Phylogenetics, population structure, and archeological data coincide in a scenario of protracted domestication of lentils, with two domesticated gene pools emerging in SWA. Admixed varieties are found throughout their range, suggesting a relaxed selection process. A small number of alleles involved in domestication and adaptation to climatic variables were identified. Both novel mutation and selection on standing variation are presumed to have played a role in adaptation of lentils to different environments. The results presented have implications for understanding the process of plant domestication (past), the distribution of genetic diversity in germplasm collections (present), and targeting genes in breeding programs (future).

Allele-specific miRNA-binding analysis identifies candidate target genes for breast cancer risk.

Most breast cancer (BC) risk-associated single-nucleotide polymorphisms (raSNPs) identified in genome-wide association studies (GWAS) are believed to cis-regulate the expression of genes. We hypothesise that cis-regulatory variants contributing to disease risk may be affecting microRNA (miRNA) genes and/or miRNA binding. To test this, we adapted two miRNA-binding prediction algorithms-TargetScan and miRanda-to perform allele-specific queries, and integrated differential allelic expression (DAE) and expression quantitative trait loci (eQTL) data, to query 150 genome-wide significant ( P ≤ 5 × 10 - 8 ) raSNPs, plus proxies. We found that no raSNP mapped to a miRNA gene, suggesting that altered miRNA targeting is an unlikely mechanism involved in BC risk. Also, 11.5% (6 out of 52) raSNPs located in 3'-untranslated regions of putative miRNA target genes were predicted to alter miRNA::mRNA (messenger RNA) pair binding stability in five candidate target genes. Of these, we propose RNF115, at locus 1q21.1, as a strong novel target gene associated with BC risk, and reinforce the role of miRNA-mediated cis-regulation at locus 19p13.11. We believe that integrating allele-specific querying in miRNA-binding prediction, and data supporting cis-regulation of expression, improves the identification of candidate target genes in BC risk, as well as in other common cancers and complex diseases.

gwasrapidd: an R package to query, download and wrangle GWAS catalog data.

MOTIVATION: The National Human Genome Research Institute Catalog of Published Genome-Wide Association Studies (GWAS) Catalog has collected, curated and made available data from over 7100 studies. The recently developed GWAS Catalog representational state transfer (REST) application programming interface (API) is the only method allowing programmatic access to this resource. RESULTS: Here, we describe gwasrapidd, an R package that provides the first client interface to the GWAS Catalog REST API, representing an important software counterpart to the server-side component. gwasrapidd enables users to quickly retrieve, filter and integrate data with comprehensive bioinformatics analysis tools, which is particularly critical for those looking into functional characterization of risk loci. AVAILABILITY AND IMPLEMENTATION: gwasrapidd is freely available under an MIT License, and can be accessed from https://github.com/ramiromagno/gwasrapidd. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Epigenetic therapy in urologic cancers: an update on clinical trials.

Epigenetic dysregulation is one of many factors that contribute to cancer development and progression. Numerous epigenetic alterations have been identified in urologic cancers including histone modifications, DNA methylation changes, and microRNA expression. Since these changes are reversible, efforts are being made to develop epigenetic drugs that restore the normal epigenetic patterns of cells, and many clinical trials are already underway to test their clinical potential. In this review we analyze multiple clinical trials (n=51) that test the efficacy of these drugs in patients with urologic cancers. The most frequently used epigenetic drugs were histone deacetylase inhibitors followed by antisense oligonucleotides, DNA methyltransferase inhibitors and histone demethylase inhibitors, the last of which are only being tested in prostate cancer. In more than 50% of the clinical trials considered, epigenetic drugs were used as part of combination therapy, which achieved the best results. The epigenetic regulation of some cancers is still matter of research but will undoubtedly open a window to new therapeutic approaches in the era of personalized medicine. The future of therapy for urological malignancies is likely to include multidrug regimens in which epigenetic modifying drugs will play an important role.

Candidate genetic modifiers for breast and ovarian cancer risk in BRCA1 and BRCA2 mutation carriers.

BACKGROUND: BRCA1 and BRCA2 mutation carriers are at substantially increased risk for developing breast and ovarian cancer. The incomplete penetrance coupled with the variable age at diagnosis in carriers of the same mutation suggests the existence of genetic and nongenetic modifying factors. In this study, we evaluated the putative role of variants in many candidate modifier genes. METHODS: Genotyping data from 15,252 BRCA1 and 8,211 BRCA2 mutation carriers, for known variants (n = 3,248) located within or around 445 candidate genes, were available through the iCOGS custom-designed array. Breast and ovarian cancer association analysis was performed within a retrospective cohort approach. RESULTS: The observed P values of association ranged between 0.005 and 1.000. None of the variants was significantly associated with breast or ovarian cancer risk in either BRCA1 or BRCA2 mutation carriers, after multiple testing adjustments. CONCLUSION: There is little evidence that any of the evaluated candidate variants act as modifiers of breast and/or ovarian cancer risk in BRCA1 or BRCA2 mutation carriers. IMPACT: Genome-wide association studies have been more successful at identifying genetic modifiers of BRCA1/2 penetrance than candidate gene studies.

TOX3 mutations in breast cancer.

TOX3 maps to 16q12, a region commonly lost in breast cancers and recently implicated in the risk of developing breast cancer. However, not much is known of the role of TOX3 itself in breast cancer biology. This is the first study to determine the importance of TOX3 mutations in breast cancers. We screened TOX3 for mutations in 133 breast tumours and identified four mutations (three missense, one in-frame deletion of 30 base pairs) in six primary tumours, corresponding to an overall mutation frequency of 4.5%. One potentially deleterious missense mutation in exon 3 (Leu129Phe) was identified in one tumour (genomic DNA and cDNA). Whilst copy number changes of 16q12 are common in breast cancer, our data show that mutations of TOX3 are present at low frequency in tumours. Our results support that TOX3 should be further investigated to elucidate its role in breast cancer biology.

Functional variants at the 11q13 risk locus for breast cancer regulate cyclin D1 expression through long-range enhancers.

Analysis of 4,405 variants in 89,050 European subjects from 41 case-control studies identified three independent association signals for estrogen-receptor-positive tumors at 11q13. The strongest signal maps to a transcriptional enhancer element in which the G allele of the best candidate causative variant rs554219 increases risk of breast cancer, reduces both binding of ELK4 transcription factor and luciferase activity in reporter assays, and may be associated with low cyclin D1 protein levels in tumors. Another candidate variant, rs78540526, lies in the same enhancer element. Risk association signal 2, rs75915166, creates a GATA3 binding site within a silencer element. Chromatin conformation studies demonstrate that these enhancer and silencer elements interact with each other and with their likely target gene, CCND1.

Effects of BRCA2 cis-regulation in normal breast and cancer risk amongst BRCA2 mutation carriers.

INTRODUCTION: Cis-acting regulatory single nucleotide polymorphisms (SNPs) at specific loci may modulate penetrance of germline mutations at the same loci by introducing different levels of expression of the wild-type allele. We have previously reported that BRCA2 shows differential allelic expression and we hypothesize that the known variable penetrance of BRCA2 mutations might be associated with this mechanism. METHODS: We combined haplotype analysis and differential allelic expression of BRCA2 in breast tissue to identify expression haplotypes and candidate cis-regulatory variants. These candidate variants underwent selection based on in silico predictions for regulatory potential and disruption of transcription factor binding, and were functionally analyzed in vitro and in vivo in normal and breast cancer cell lines. SNPs tagging the expression haplotypes were correlated with the total expression of several genes in breast tissue measured by Taqman and microarray technologies. The effect of the expression haplotypes on breast cancer risk in BRCA2 mutation carriers was investigated in 2,754 carriers. RESULTS: We identified common haplotypes associated with differences in the levels of BRCA2 expression in human breast cells. We characterized three cis-regulatory SNPs located at the promoter and two intronic regulatory elements which affect the binding of the transcription factors C/EBPα, HMGA1, D-binding protein (DBP) and ZF5. We showed that the expression haplotypes also correlated with changes in the expression of other genes in normal breast. Furthermore, there was suggestive evidence that the minor allele of SNP rs4942440, which is associated with higher BRCA2 expression, is also associated with a reduced risk of breast cancer (per-allele hazard ratio (HR) = 0.85, 95% confidence interval (CI) = 0.72 to 1.00, P-trend = 0.048). CONCLUSIONS: Our work provides further insights into the role of cis-regulatory variation in the penetrance of disease-causing mutations. We identified small-effect genetic variants associated with allelic expression differences in BRCA2 which could possibly affect the risk in mutation carriers through altering expression levels of the wild-type allele.

Allele-specific expression analysis methods for high-density SNP microarray data.

MOTIVATION: In the past decade, a number of technologies to quantify allele-specific expression (ASE) in a genome-wide manner have become available to researchers. We investigate the application of single-nucleotide polymorphism (SNP) microarrays to this task, exploring data obtained from both cell lines and primary tissue for which both RNA and DNA profiles are available. RESULTS: We analyze data from two experiments that make use of high-density Illumina Infinium II genotyping arrays to measure ASE. We first preprocess each data set, which involves removal of outlier samples, careful normalization and a two-step filtering procedure to remove SNPs that show no evidence of expression in the samples being analyzed and calls that are clear genotyping errors. We then compare three different tests for detecting ASE, one of which has been previously published and two novel approaches. These tests vary at the level at which they operate (per SNP per individual or per SNP) and in the input data they require. Using SNPs from imprinted genes as true positives for ASE, we observe varying sensitivity for the different testing procedures that improves with increasing sample size. Methods that rely on RNA signal alone were found to perform best across a range of metrics. The top ranked SNPs recovered by all methods appear to be reasonable candidates for ASE. AVAILABILITY AND IMPLEMENTATION: Analysis was carried out in R (http://www.R-project.org/) using existing functions.

A functional variant at a prostate cancer predisposition locus at 8q24 is associated with PVT1 expression.

Genetic mapping studies have identified multiple cancer susceptibility regions at chromosome 8q24, upstream of the MYC oncogene. MYC has been widely presumed as the regulated target gene, but definitive evidence functionally linking these cancer regions with MYC has been difficult to obtain. Here we examined candidate functional variants of a haplotype block at 8q24 encompassing the two independent risk alleles for prostate and breast cancer, rs620861 and rs13281615. We used the mapping of DNase I hypersensitive sites as a tool to prioritise regions for further functional analysis. This approach identified rs378854, which is in complete linkage disequilibrium (LD) with rs620861, as a novel functional prostate cancer-specific genetic variant. We demonstrate that the risk allele (G) of rs378854 reduces binding of the transcription factor YY1 in vitro. This factor is known to repress global transcription in prostate cancer and is a candidate tumour suppressor. Additional experiments showed that the YY1 binding site is occupied in vivo in prostate cancer, but not breast cancer cells, consistent with the observed cancer-specific effects of this single nucleotide polymorphism (SNP). Using chromatin conformation capture (3C) experiments, we found that the region surrounding rs378854 interacts with the MYC and PVT1 promoters. Moreover, expression of the PVT1 oncogene in normal prostate tissue increased with the presence of the risk allele of rs378854, while expression of MYC was not affected. In conclusion, we identified a new functional prostate cancer risk variant at the 8q24 locus, rs378854 allele G, that reduces binding of the YY1 protein and is associated with increased expression of PVT1 located 0.5 Mb downstream.

Fine scale mapping of the breast cancer 16q12 locus.

Recent genome-wide association studies have identified a breast cancer susceptibility locus on 16q12 with an unknown biological basis. We used a set of single nucleotide polymorphism (SNP) markers to generate a fine-scale map and narrowed the region of association to a 133 kb DNA segment containing the largely uncharacterized hypothetical gene LOC643714, a short intergenic region and the 5' end of TOX3. Re-sequencing this segment in European subjects identified 293 common polymorphisms, including a set of 26 highly correlated candidate causal variants. By evaluation of these SNPs in five breast cancer case-control studies involving more than 23 000 subjects from populations of European and Southeast Asian ancestry, all but 14 variants could be excluded at odds of <1:100. Most of the remaining variants lie in the intergenic region, which exhibits evolutionary conservation and open chromatin conformation, consistent with a regulatory function. African-American case-control studies exhibit a different pattern of association suggestive of an additional causative variant.

Extent of differential allelic expression of candidate breast cancer genes is similar in blood and breast.

INTRODUCTION: Normal gene expression variation is thought to play a central role in inter-individual variation and susceptibility to disease. Regulatory polymorphisms in cis-acting elements result in the unequal expression of alleles. Differential allelic expression (DAE) in heterozygote individuals could be used to develop a new approach to discover regulatory breast cancer susceptibility loci. As access to large numbers of fresh breast tissue to perform such studies is difficult, a suitable surrogate test tissue must be identified for future studies. METHODS: We measured differential allelic expression of 12 candidate genes possibly related to breast cancer susceptibility (BRCA1, BRCA2, C1qA, CCND3, EMSY, GPX1, GPX4, MLH3, MTHFR, NBS1, TP53 and TRXR2) in breast tissue (n = 40) and fresh blood (n = 170) of healthy individuals and EBV-transformed lymphoblastoid cells (n = 19). Differential allelic expression ratios were determined by Taqman assay. Ratio distributions were compared using t-test and Wilcoxon rank sum test, for mean ratios and variances respectively. RESULTS: We show that differential allelic expression is common among these 12 candidate genes and is comparable between breast and blood (fresh and transformed lymphoblasts) in a significant proportion of them. We found that eight out of nine genes with DAE in breast and fresh blood were comparable, as were 10 out of 11 genes between breast and transformed lymphoblasts. CONCLUSIONS: Our findings support the use of differential allelic expression in blood as a surrogate for breast tissue in future studies on predisposition to breast cancer.

Association of ESR1 gene tagging SNPs with breast cancer risk.

We have conducted a three-stage, comprehensive single nucleotide polymorphism (SNP)-tagging association study of ESR1 gene variants (SNPs) in more than 55,000 breast cancer cases and controls from studies within the Breast Cancer Association Consortium (BCAC). No large risks or highly significant associations were revealed. SNP rs3020314, tagging a region of ESR1 intron 4, is associated with an increase in breast cancer susceptibility with a dominant mode of action in European populations. Carriers of the c-allele have an odds ratio (OR) of 1.05 [95% Confidence Intervals (CI) 1.02-1.09] relative to t-allele homozygotes, P = 0.004. There is significant heterogeneity between studies, P = 0.002. The increased risk appears largely confined to oestrogen receptor-positive tumour risk. The region tagged by SNP rs3020314 contains sequence that is more highly conserved across mammalian species than the rest of intron 4, and it may subtly alter the ratio of two mRNA splice forms.

Somatically acquired hypomethylation of IGF2 in breast and colorectal cancer.

The imprinted insulin-like growth factor 2 (IGF2) gene is expressed predominantly from the paternal allele. Loss of imprinting (LOI) associated with hypomethylation at the promoter proximal sequence (DMR0) of the IGF2 gene was proposed as a predisposing constitutive risk biomarker for colorectal cancer. We used pyrosequencing to assess whether IGF2 DMR0 methylation is either present constitutively prior to cancer or whether it is acquired tissue-specifically after the onset of cancer. DNA samples from tumour tissues and matched non-tumour tissues from 22 breast and 42 colorectal cancer patients as well as peripheral blood samples obtained from colorectal cancer patients [SEARCH (n=case 192, controls 96)], breast cancer patients [ABC (n=case 364, controls 96)] and the European Prospective Investigation of Cancer [EPIC-Norfolk (n=breast 228, colorectal 225, controls 895)] were analysed. The EPIC samples were collected 2-5 years prior to diagnosis of breast or colorectal cancer. IGF2 DMR0 methylation levels in tumours were lower than matched non-tumour tissue. Hypomethylation of DMR0 was detected in breast (33%) and colorectal (80%) tumour tissues with a higher frequency than LOI indicating that methylation levels are a better indicator of cancer than LOI. In the EPIC population, the prevalence of IGF2 DMR0 hypomethylation was 9.5% and this correlated with increased age not cancer risk. Thus, IGF2 DMR0 hypomethylation occurs as an acquired tissue-specific somatic event rather than a constitutive innate epimutation. These results indicate that IGF2 DMR0 hypomethylation has diagnostic potential for colon cancer rather than value as a surrogate biomarker for constitutive LOI.

Allele-specific up-regulation of FGFR2 increases susceptibility to breast cancer.

The recent whole-genome scan for breast cancer has revealed the FGFR2 (fibroblast growth factor receptor 2) gene as a locus associated with a small, but highly significant, increase in the risk of developing breast cancer. Using fine-scale genetic mapping of the region, it has been possible to narrow the causative locus to a haplotype of eight strongly linked single nucleotide polymorphisms (SNPs) spanning a region of 7.5 kilobases (kb) in the second intron of the FGFR2 gene. Here we describe a functional analysis to define the causative SNP, and we propose a model for a disease mechanism. Using gene expression microarray data, we observed a trend of increased FGFR2 expression in the rare homozygotes. This trend was confirmed using real-time (RT) PCR, with the difference between the rare and the common homozygotes yielding a Wilcox p-value of 0.028. To elucidate which SNPs might be responsible for this difference, we examined protein-DNA interactions for the eight most strongly disease-associated SNPs in different breast cell lines. We identify two cis-regulatory SNPs that alter binding affinity for transcription factors Oct-1/Runx2 and C/EBPbeta, and we demonstrate that both sites are occupied in vivo. In transient transfection experiments, the two SNPs can synergize giving rise to increased FGFR2 expression. We propose a model in which the Oct-1/Runx2 and C/EBPbeta binding sites in the disease-associated allele are able to lead to an increase in FGFR2 gene expression, thereby increasing the propensity for tumour formation.

Common germline genetic variation in antioxidant defense genes and survival after diagnosis of breast cancer.

PURPOSE: The prognosis of breast cancer varies considerably among individuals, and inherited genetic factors may help explain this variability. Of particular interest are genes involved in defense against reactive oxygen species (ROS) because ROS are thought to cause DNA damage and contribute to the pathogenesis of cancer. PATIENTS AND METHODS: We examined associations between 54 polymorphisms that tag the known common variants (minor allele frequency > 0.05) in 10 genes involved in oxidative damage repair (CAT, SOD1, SOD2, GPX1, GPX4, GSR, TXN, TXN2, TXNRD1, and TXNRD2) and survival in 4,470 women with breast cancer. RESULTS: Two single nucleotide polymorphisms (SNPs) in GPX4 (rs713041 and rs757229) were associated with all-cause mortality even after adjusting for multiple hypothesis testing (adjusted P = .0041 and P = .0035). These SNPs are correlated with each other (r2 = 0.61). GPX4 rs713041 is located near the selenocysteine insertion sequence element in the GPX4 3' untranslated region, and the rare allele of this SNP is associated with an increased risk of death, with a hazard ratio of 1.27 per rare allele carried (95% CI, 1.13 to 11.43). This effect was not attenuated after adjusting for tumor stage, grade, or estrogen receptor status. We found that the common allele is preferentially expressed in normal lymphocytes, normal breast, and breast tumors compared with the rare allele, but there were no differences in total levels of GPX4 mRNA across genotypes. CONCLUSION: These data provide strong support for the hypothesis that common variation in GPX4 is associated with prognosis after a diagnosis of breast cancer.