Penny Coulson

Temporal stability and determinants of white blood cell DNA methylation in the Breakthrough Generations Study

Background: Epigenome-wide association studies (EWAS) using measurements of blood DNA methylation are performed to identify associations of methylation changes with environmental and lifestyle exposures and disease risk. However, little is known about the variation of methylation markers in the population and their stability over time, both important factors in the design and interpretation of EWAS. We aimed to identify stable variable methylated probes (VMP), i.e., markers that are variable in the population, yet stable over time. Methods: We estimated the intraclass correlation coefficient (ICC) for each probe on the Illumina 450K methylation array in paired samples collected approximately 6 years apart from 92 participants in the Breakthrough Generations Study. We also evaluated relationships with age, reproductive and hormonal history, weight, alcohol intake, and smoking. Results: Approximately 17% of probes had an ICC > 0.50 and were considered stable VMPs (stable-VMPs). Stable-VMPs were enriched for probes located in “shores” bordering CpG islands, and at approximately 1.3 kb downstream from the transcription start site in the transition between the unmethylated promoter and methylated gene body. Both cross-sectional and longitudinal data analyses provided strong evidence for associations between changes in methylation levels and aging. Smoking-related probes at 2q37.1 and AHRR were stable-VMPs and related to time since quitting. We also observed associations between methylation and weight changes. Conclusion: Our results provide support for the use of white blood cell DNA methylation as a biomarker of exposure in EWAS. Impact: Larger studies, preferably with repeated measures over time, will be required to establish associations between specific probes and exposures. Cancer Epidemiol Biomarkers Prev; 24(1); 221–9. ©2014 AACR.

Mitochondrial DNA Copy Number in Peripheral Blood Cells and Risk of Developing Breast Cancer

Increased mitochondrial DNA (mtDNA) copy number in peripheral blood cells (PBC) has been associated with the risk of developing several tumor types. Here we evaluate sources of variation of this biomarker and its association with breast cancer risk in a prospective cohort study. mtDNA copy number was measured using quantitative real-time PCR on PBC DNA samples from participants in the UK-based Breakthrough Generations Study. Temporal and assay variation was evaluated in a serial study of 91 women, with two blood samples collected approximately 6-years apart. Then, associations with breast cancer risk factors and risk were evaluated in 1,108 cases and 1,099 controls using a nested case-control design. In the serial study, mtDNA copy number showed low assay variation but large temporal variation [assay intraclass correlation coefficient (ICC), 79.3%-87.9%; temporal ICC, 38.3%). Higher mtDNA copy number was significantly associated with younger age at blood collection, being premenopausal, having an older age at menopause, and never taking HRT, both in cases and controls. Based on measurements in a single blood sample taken on average 6 years before diagnosis, higher mtDNA copy number was associated with increased breast cancer risk [OR (95% CI) for highest versus lowest quartile, 1.37 (1.02-1.83); P trend = 0.007]. In conclusion, mtDNA copy number is associated with breast cancer risk and represents a promising biomarker for risk assessment. The relatively large temporal variation should be taken into account in future analyses.

Crowdsourcing the General Public for Large Scale Molecular Pathology Studies in Cancer

Background Citizen science, scientific research conducted by non-specialists, has the potential to facilitate biomedical research using available large-scale data, however validating the results is challenging. The Cell Slider is a citizen science project that intends to share images from tumors with the general public, enabling them to score tumor markers independently through an internet-based interface. Methods From October 2012 to June 2014, 98,293 Citizen Scientists accessed the Cell Slider web page and scored 180,172 sub-images derived from images of 12,326 tissue microarray cores labeled for estrogen receptor (ER). We evaluated the accuracy of Citizen Scientists ER classification, and the association between ER status and prognosis by comparing their test performance against trained pathologists. Findings The area under ROC curve was 0.95 (95% CI 0.94 to 0.96) for cancer cell identification and 0.97 (95% CI 0.96 to 0.97) for ER status. ER positive tumors scored by Citizen Scientists were associated with survival in a similar way to that scored by trained pathologists. Survival probability at 15 years were 0.78 (95% CI 0.76 to 0.80) for ER-positive and 0.72 (95% CI 0.68 to 0.77) for ER-negative tumors based on Citizen Scientists classification. Based on pathologist classification, survival probability was 0.79 (95% CI 0.77 to 0.81) for ER-positive and 0.71 (95% CI 0.67 to 0.74) for ER-negative tumors. The hazard ratio for death was 0.26 (95% CI 0.18 to 0.37) at diagnosis and became greater than one after 6.5 years of follow-up for ER scored by Citizen Scientists, and 0.24 (95% CI 0.18 to 0.33) at diagnosis increasing thereafter to one after 6.7 (95% CI 4.1 to 10.9) years of follow-up for ER scored by pathologists. Interpretation Crowdsourcing of the general public to classify cancer pathology data for research is viable, engages the public and provides accurate ER data. Crowdsourced classification of research data may offer a valid solution to problems of throughput requiring human input.

Performance of automated scoring of ER, PR, HER2, CK5/6 and EGFR in breast cancer tissue microarrays in the Breast Cancer Association Consortium

Breast cancer risk factors and clinical outcomes vary by tumour marker expression. However, individual studies often lack the power required to assess these relationships, and large‐scale analyses are limited by the need for high throughput, standardized scoring methods. To address these limitations, we assessed whether automated image analysis of immunohistochemically stained tissue microarrays can permit rapid, standardized scoring of tumour markers from multiple studies. Tissue microarray sections prepared in nine studies containing 20 263 cores from 8267 breast cancers stained for two nuclear (oestrogen receptor, progesterone receptor), two membranous (human epidermal growth factor receptor 2 and epidermal growth factor receptor) and one cytoplasmic (cytokeratin 5/6) marker were scanned as digital images. Automated algorithms were used to score markers in tumour cells using the Ariol system. We compared automated scores against visual reads, and their associations with breast cancer survival. Approximately 65–70% of tissue microarray cores were satisfactory for scoring. Among satisfactory cores, agreement between dichotomous automated and visual scores was highest for oestrogen receptor (Kappa = 0.76), followed by human epidermal growth factor receptor 2 (Kappa = 0.69) and progesterone receptor (Kappa = 0.67). Automated quantitative scores for these markers were associated with hazard ratios for breast cancer mortality in a dose‐response manner. Considering visual scores of epidermal growth factor receptor or cytokeratin 5/6 as the reference, automated scoring achieved excellent negative predictive value (96–98%), but yielded many false positives (positive predictive value = 30–32%). For all markers, we observed substantial heterogeneity in automated scoring performance across tissue microarrays. Automated analysis is a potentially useful tool for large‐scale, quantitative scoring of immunohistochemically stained tissue microarrays available in consortia. However, continued optimization, rigorous marker‐specific quality control measures and standardization of tissue microarray designs, staining and scoring protocols is needed to enhance results.

High-throughput automated scoring of Ki67 in breast cancer tissue microarrays from the Breast Cancer Association Consortium

Automated methods are needed to facilitate high‐throughput and reproducible scoring of Ki67 and other markers in breast cancer tissue microarrays (TMAs) in large‐scale studies. To address this need, we developed an automated protocol for Ki67 scoring and evaluated its performance in studies from the Breast Cancer Association Consortium. We utilized 166 TMAs containing 16,953 tumour cores representing 9,059 breast cancer cases, from 13 studies, with information on other clinical and pathological characteristics. TMAs were stained for Ki67 using standard immunohistochemical procedures, and scanned and digitized using the Ariol system. An automated algorithm was developed for the scoring of Ki67, and scores were compared to computer assisted visual (CAV) scores in a subset of 15 TMAs in a training set. We also assessed the correlation between automated Ki67 scores and other clinical and pathological characteristics. Overall, we observed good discriminatory accuracy (AUC = 85%) and good agreement (kappa = 0.64) between the automated and CAV scoring methods in the training set. The performance of the automated method varied by TMA (kappa range= 0.37–0.87) and study (kappa range = 0.39–0.69). The automated method performed better in satisfactory cores (kappa = 0.68) than suboptimal (kappa = 0.51) cores (p‐value for comparison = 0.005); and among cores with higher total nuclei counted by the machine (4,000–4,500 cells: kappa = 0.78) than those with lower counts (50–500 cells: kappa = 0.41; p‐value = 0.010). Among the 9,059 cases in this study, the correlations between automated Ki67 and clinical and pathological characteristics were found to be in the expected directions. Our findings indicate that automated scoring of Ki67 can be an efficient method to obtain good quality data across large numbers of TMAs from multicentre studies. However, robust algorithm development and rigorous pre‐ and post‐analytical quality control procedures are necessary in order to ensure satisfactory performance.

Heterogeneity of luminal breast cancer characterised by immunohistochemical expression of basal markers

Background:Luminal A breast cancer defined as hormone receptor positive and human epidermal growth factor receptor 2 (HER2) negative is known to be heterogeneous. Previous study showed that luminal A tumours with the expression of basal markers ((cytokeratin (CK) 5 or CK5/6) or epidermal growth factor receptor (EGFR)) were associated with poorer prognosis compared with those that stained negative for basal markers. Prompted by this study, we assessed whether tumour characteristics and risk factors differed by basal marker status within luminal A tumours.Methods:We pooled 5040 luminal A cases defined by immunohistochemistry (4490 basal-negative ((CK5 (or CK5/6))− and EGFR−) and 550 basal-positive ((CK5 (or CK5/6+)) or EGFR+)) from eight studies participating in the Breast Cancer Association Consortium. Case–case comparison was performed using unconditional logistic regression.Results:Tumour characteristics and risk factors did not vary significantly by the expression of basal markers, although results suggested that basal-positive luminal tumours tended to be smaller and node negative, and were more common in women with a positive family history and lower body mass index.Conclusions:Most established breast cancer risk factors were similar in basal-positive and basal-negative luminal A tumours. The non-significant but suggestive differences in tumour features and family history warrant further investigations.

Etiology of hormone receptor positive breast cancer differs by levels of histologic grade and proliferation

Limited epidemiological evidence suggests that the etiology of hormone receptor positive (HR+) breast cancer may differ by levels of histologic grade and proliferation. We pooled risk factor and pathology data on 5,905 HR+ breast cancer cases and 26,281 controls from 11 epidemiological studies. Proliferation was determined by centralized automated measures of KI67 in tissue microarrays. Odds ratios (OR), 95% confidence intervals (CI) and p‐values for case–case and case–control comparisons for risk factors in relation to levels of grade and quartiles (Q1–Q4) of KI67 were estimated using polytomous logistic regression models. Case–case comparisons showed associations between nulliparity and high KI67 [OR (95% CI) for Q4 vs. Q1 = 1.54 (1.22, 1.95)]; obesity and high grade [grade 3 vs. 1 = 1.68 (1.31, 2.16)] and current use of combined hormone therapy (HT) and low grade [grade 3 vs. 1 = 0.27 (0.16, 0.44)] tumors. In case–control comparisons, nulliparity was associated with elevated risk of tumors with high but not low levels of proliferation [1.43 (1.14, 1.81) for KI67 Q4 vs. 0.83 (0.60, 1.14) for KI67 Q1]; obesity among women ≥50 years with high but not low grade tumors [1.55 (1.17, 2.06) for grade 3 vs. 0.88 (0.66, 1.16) for grade 1] and HT with low but not high grade tumors [3.07 (2.22, 4.23) for grade 1 vs. 0.85 (0.55, 1.30) for grade 3]. Menarcheal age and family history were similarly associated with HR+ tumors of different grade or KI67 levels. These findings provide insights into the etiologic heterogeneity of HR+ tumors.

Comparative validation of breast cancer risk prediction models and projections for future risk stratification

Background Well-validated risk models are critical for risk stratified breast cancer prevention. We used the Individualized Coherent Absolute Risk Estimation (iCARE) tool for comparative model validation of five-year risk of invasive breast cancer in a prospective cohort, and to make projections for population risk stratification. Methods Performance of two recently developed models, iCARE-BPC3 and iCARE-Lit, were compared with two established models (BCRAT, IBIS) based on classical risk factors in a UK-based cohort of 64,874 women (863 cases) aged 35-74 years. Risk projections in US White non-Hispanic women aged 50-70 years were made to assess potential improvements in risk stratification by adding mammographic breast density (MD) and polygenic risk score (PRS). Results The best calibrated models were iCARE-Lit (expected to observed number of cases (E/O)=0.98 (95% confidence interval [CI]=0.87 to 1.11)) for women younger than 50 years; and iCARE-BPC3 (E/O=1.00 (0.93 to 1.09)) for women 50 years or older. Risk projections using iCARE-BPC3 indicated classical risk factors can identify ~500,000 women at moderate to high risk (>3% five-year risk). Additional information on MD and a PRS based on 172 variants is expected to increase this to ~3.6 million, and among them, ~155,000 invasive breast cancer cases are expected within five years. Conclusions iCARE models based on classical risk factors perform similarly or better than BCRAT or IBIS. Addition of MD and PRS can lead to substantial improvements in risk stratification. Independent prospective validation of integrated models is needed prior to clinical evaluation risk stratified breast cancer screening and prevention.

Abstract P1-04-01: Epigenome-wide association study for breast cancer risk using whole genome and target captured bisulphite sequencing: A pooled case-control study nested in the breakthrough generations study

Background: The field of epigenetic epidemiology has rapidly advanced and recent work has discovered epigenetic markers of breast cancer risk in white blood cell (WBC) DNA. Using Epigenome-Wide Association Studies (EWAS) on the Illumina 450k methylation array, we and others have shown epigenome-wide hypomethylation (-0.2%, p -16 ) in incident breast cancer cases compared with controls in several prospective cohorts. We have proposed a mechanism that involves cancer risk exposures, lifetime and environmental events, that alter the epigenome and stably modifies an individual9s cancer risk. However, more work is needed to establish the clinical utility of this observation, the underlying causes of this variation and to determine whether the 1.7% of CpG sites targeted by the 450k array are representative of the remaining 98% of the epigenome that has not yet been interrogated. Our overall aim is to identify epigenetic traits within the epigenome that are associated with the risk of developing breast cancer. Methods: We conducted an EWAS using whole genome bisulphite sequencing (WGBS) of WBC DNA from incident breast cancer cases (n=548) compared to matched controls (n=548) from a prospective cohort (Breakthrough Generations Study) using a DNA pooling approach. Eight DNA pools were prepared in sequencing libraries and sequenced on the Hiseq2500 at PE100bp reads, resulting in ~10-fold coverage per CpG, per library, across ~20 million mappable CpG sites. Each pooled sample was also analysed in triplicate on the Illumina 450k methylation array for validation. We used Agilent target capture bisulphite sequencing (TCBS) for technical validation in a subset of breast cancer cases (n=48) and matched controls (n=48), individually barcoded and sequenced on the MiSeq at PE150bp, aiming for >1000 fold coverage of 425 kb targeted sequence. Results: Interrogation of specific genomic regions showed that gene-body methylation averages tended to be hypomethylated in cases, while CpG island averages identified both hypo- and hypermethylation. We have validated the same direction of change in 40/51 CpG islands that were covered by the Illumina 450K methylation array and have developed a target capture panel for validation of 960 gene body regions and 224 CpG island regions that were identified as significantly different between cases and controls (average -11%, FDR Conclusions: Results indicate that epigenome-wide hypomethylation and methylation in specific sites, particularly gene bodies, measured in pre-diagnostic blood samples may be predictive of breast cancer risk, and may thus be useful as a risk biomarker. Citation Format: Flanagan JM, Curry E, Stirling L, Flower K, Orr N, Tomczyk K, Coulson P, Jones M, Ashworth A, Swerdlow A, Brown R, Garcia-Closas M. Epigenome-wide association study for breast cancer risk using whole genome and target captured bisulphite sequencing: A pooled case-control study nested in the breakthrough generations study [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P1-04-01.