Cell-free DNA in plasma and ascites as a biomarker of bevacizumab response- a translational research sub-study of the REZOLVE (ANZGOG-1101) clinical trial

OBJECTIVE

To investigate cell-free DNA (cfDNA) in plasma and ascites and its association with clinical outcomes (paracentesis-free interval, overall survival) and CA125 level in participants with advanced ovarian cancer, treated with palliative intraperitoneal bevacizumab to delay re-accumulation of ascites.

METHODS

cfDNA was extracted from 0.3 to 1 mL samples from 20/24 participants of the REZOLVE trial. Standard and methylation-specific PCRs were performed to measure 3 biomarkers: total cfDNA (Alu), tumour-derived cfDNA (ctDNA, methylated IFFO1 promoter) and endothelium-derived cfDNA (ec-cfDNA, unmethylated CDH5 promoter). Values were correlated to clinical outcomes.

RESULTS

cfDNA was detected in all samples, with higher yield in ascites (mean 669 ng/mL) than plasma (mean 75 ng/mL, p < 0.0001). Ascites had a higher ctDNA proportion than plasma (74 % vs. 20 %, p < 0.0001) and plasma had a higher ec-cfDNA proportion than ascites (24 % vs. 16 %, p < 0.002). High ctDNA proportion (>75 %) in ascites was associated with a significantly shorter paracentesis-free interval (median interval 47.5 versus 84 days, hazard ratio (HR) 2.21, 95 % confidence interval (CI) 0.85 to 5.73, p = 0.039) and ctDNA presence in plasma was unfavourable for survival (median survival 56 versus 242 days, HR 3.21, 95 % CI 1.15 to 9.00, p = 0.008). A significant positive correlation was observed between ctDNA proportion in plasma and CA125 level (p = 0.012). No significant difference in total cfDNA, ctDNA nor ec-cfDNA was observed between participants who were responders versus non-responders.

CONCLUSION

Sufficient cfDNA was detected in both plasma and ascites to study three biomarkers. These samples can provide useful information and should be considered in the design of future ovarian cancer trials.

Circulating cell-free DNA is elevated in postmenopausal compared with pre- and perimenopausal women

OBJECTIVE

With the rising use of circulating cell-free DNA (cirDNA) liquid biopsies for disease screening, it is important to understand biological differences that may impact the accuracy of cirDNA-based clinical tests. Although a number of biological factors have been researched, the relationship between menopause and cirDNA has not been thoroughly investigated. We aimed to compare plasma cirDNA concentration and DNA fragment integrity in healthy women pre- and postmenopause.

METHODS

Blood was collected from healthy female volunteers 40 years and older. cirDNA was extracted from plasma (n = 52) and quantified by quantitative polymerase chain reaction (n = 47; 26 premenopause, mean age-46 y; 21 postmenopause, mean age-59 y). cirDNA concentration was quantitated using an ALU repetitive sequence with a 115-base-pair (bp) product (ALU-115), and long cirDNA fragments were quantitated using an ALU repetitive sequence with a 247-bp product (ALU-247). cirDNA integrity was expressed as a ratio of ALU-247 over ALU-115. Mann-Whitney U test was used to compare pre- and postmenopause qPCR results, and a two-tailed, unpaired t test was undertaken to compare the integrity ratio between the two groups.

RESULTS

Postmenopause plasma samples were found to have a significantly higher cirDNA concentration (P < 0.0001, premenopause: mean, 3.10 ± 1.84 ng/mL; median, 2.90 ng/mL; postmenopause: mean, 5.28 ± 2.76 ng/mL; median, 4.56 ng/mL) and significantly higher concentration of long-stranded cirDNA fragments (P = 0.0033, premenopause: mean, 1.06 ± 0.48 ng/mL; median, 0.96 ng/mL; postmenopause: mean, 1.69 ± 0.89 ng/mL; median, 1.48 ng/mL). There was no significant difference in the integrity ratio between the groups (P = 0.1788).

CONCLUSIONS

Plasma cirDNA concentrations are higher in postmenopausal women. This has important implications in cirDNA liquid biopsy development and screening, especially for diseases such as cancer where the majority of cases are diagnosed postmenopause.

Circulating cell-free endometrial DNA level is unaltered during menstruation and in endometriosis

STUDY QUESTION

Is circulating cell-free DNA (cirDNA) from the endometrium elevated during menstruation and in endometriosis?

SUMMARY ANSWER

Endometrial cirDNA does not increase during menstruation and is not elevated in endometriosis.

WHAT IS KNOWN ALREADY

Changes in cirDNA associated with common benign conditions are a potential source of false positives in cancer diagnostic applications, but also present an opportunity for biomarker development for diseases such as endometriosis. Elevated cirDNA has been reported in endometriosis patients compared to healthy community controls, but no difference in total or endometrial cirDNA has been found between patients with endometriosis and patients with other gynaecological conditions. Likewise, menstruation is a potential driver of changes in cirDNA levels and tissue profile, but total and endothelial cirDNA do not increase during menstruation.

STUDY DESIGN, SIZE, DURATION

For endometriosis comparisons, 59 participants with surgically confirmed endometriosis and 27 laparoscopic patients without endometriosis (hospital controls) were prospectively recruited, while 25 healthy community participants (healthy controls) were recruited in a university setting. Total and endometrial cirDNA and cirDNA fragmentation were measured across the three groups. For menstrual comparisons, 36 matched non-menstruating and menstruating samples were collected from healthy women recruited within a university setting, and the endometrial cirDNA was compared between the two groups.

PARTICIPANTS/MATERIALS, SETTING, METHODS

cirDNA was extracted from venous blood plasma then quantitated by quantitative PCR of ALU repetitive element (115 bp) and TP53 gene sequence (105 bp) for total concentration. cirDNA derived from the endometrium was quantitated by methylation-specific droplet digital PCR of a FAM101A region (69 bp) after bisulfite conversion of the DNA. A cirDNA size fragmentation ratio was obtained by quantifying a long segment of ALU repetitive element (247 bp) and expressing the amount relative to the 115 bp ALU target.

MAIN RESULTS AND THE ROLE OF CHANCE

No differences in cirDNA level were found in any comparison populations in this study. Mean total cirDNA was unchanged between healthy controls (ALU-115-3.31 ng/ml; TP53-2.73 ng/ml), hospital controls (ALU-115-3.47 ng/ml; TP53-2.83 ng/ml) and endometriosis patients (ALU-115-3.35 ng/ml; TP53-2.66 ng/ml). Likewise, endometrial cirDNA was unchanged between healthy controls (18.3 copies/ml), hospital controls (20.6 copies/ml) and endometriosis patients (22 copies/ml). Endometrial cirDNA did not change during menstruation (non-menstruating: 38 copies/ml; menstruating: 33 copies/ml). Irrespective of endometriosis diagnosis, blood from patients undergoing laparoscopy (hospital controls: 0.77; endometriosis patients: 0.79), had a significantly higher cirDNA size ratio than community-recruited healthy controls (0.64), indicating increased abundance of long cirDNA fragments.

LIMITATIONS, REASONS FOR CAUTION

It was not possible to completely match the age, BMI and parity between the three cohorts investigated, however of these, only age has been shown to influence circulating DNA levels and not within the age range of our cohort. Blood from community-recruited healthy women and women undergoing laparoscopy was collected via antecubital vein venepuncture (processed within 3 h) and with either peripheral cannula or venepuncture (processed within 6 h), respectively, which could potentially impact the size distribution of circulating DNA fragments. For the collection of non-menstruating phase blood samples, we did not differentiate between follicular phase, ovulation and luteal phase. Thus, only the mensturating samples were collected at a consistent phase, and any fluctuations in cirDNA that occur at the other phases may have obscured small changes during menstruation.

WIDER IMPLICATIONS OF THE FINDINGS

There is no evidence that cirDNA has potential as a diagnostic biomarker for endometriosis. Endometriosis, representing a common benign gynaecological condition, and menstruation, representing a normal physiological occurrence in women, should not affect methylation-based diagnostics in other disease areas, including oncology.

STUDY FUNDING/COMPETING INTEREST(S)

N.L.Y.: Australian Government Research Training Program (RTP) Stipend through The University of New South Wales, Translational Cancer Research Network PhD Scholarship Top-Up Award via the Cancer Institute NSW, Beth Yarrow Memorial Award in Medical Science, UNSW Completion Scholarship; C.E.H.: Gynaecological Oncology Fund of the Royal Hospital for Women; K.W.: Ovarian Cancer Research Foundation and CAMILLA AND MARC. C.E.F.: UNSW Women's Wellbeing Academy and the Australian Human Rights Institute. We declare the following competing interest: K.W. holds stock in Guardant Health, Exact Sciences and Epigenomics AG. No other authors have competing interests.

TRIAL REGISTRATION NUMBER

N/A.

Circulating cell-free DNA undergoes significant decline in yield after prolonged storage time in both plasma and purified form

Objectives

Circulating DNA (cirDNA) is generally purified from plasma that has been biobanked for variable lengths of time. In long-term experiments or clinical trials, the plasma can be stored frozen for up to several years. Therefore, it is crucial to determine the stability of cirDNA to ensure confidence in sample quality upon analysis. Our main objective was to determine the effect of storage for up to 2 years on cirDNA yield and fragmentation.

Methods

We stored frozen EDTA plasma and purified cirDNA from 10 healthy female donors, then quantified cirDNA yield at baseline, and at regular intervals for up to 2 years, by qPCR and Qubit. We also compared cirDNA levels in non-haemolysed and haemolysed blood samples after 16 months of storage and tested the effect of varying DNA extraction protocol parameters.

Results

Storage up to two years caused an annual cirDNA yield decline of 25.5% when stored as plasma and 23% when stored as purified DNA, with short fragments lost more rapidly than long fragments. Additionally, cirDNA yield was impacted by plasma input and cirDNA elution volumes, but not by haemolysis.

Conclusions

The design of long-term cirDNA-based studies and clinical trials should factor in the deterioration of cirDNA during storage.

Comparison of total and endometrial circulating cell-free DNA in women with and without endometriosis

RESEARCH QUESTION

Do women with laparoscopically confirmed endometriosis have higher plasma concentrations of circulating cell-free DNA (cirDNA) than those without endometriosis?

DESIGN

Prospective study of women aged 18-45 years undergoing benign gynaecological laparoscopy at two tertiary hospitals. Venous blood was collected immediately before surgery, and women were allocated to the endometriosis or control groups based on surgical findings. Total plasma cirDNA and cirDNA integrity were measured by quantitative polymerase chain reaction (qPCR) targeting short (115 bases) and long (247 bases) ALU segments. Endometrial-derived cirDNA was measured by qPCR of bisulfite-treated cirDNA using primers selective for a FAM101A sequence uniquely unmethylated in endometrial tissue. Five cirDNA parameters were compared between the control and endometriosis cohorts: total cirDNA concentration, long-stranded cirDNA concentration, integrity ratio, endometrial cirDNA concentration and endometrial cirDNA proportion.

RESULTS

Twenty-eight endometriosis and 15 control samples were included. Women with and without endometriosis had cirDNA concentrations of 2.24 ± 0.89 ng/ml and 2.56 ± 0.92 ng/ml, respectively. Analysis by phenotype of endometriosis revealed a significantly higher endometrial cirDNA concentration in women with superficial disease (n = 10) compared with deep endometriosis (n = 18) (mean difference 0.14 ng/ml; 95% CI 0.15 to 0.26; P = 0.025), but not with controls.

CONCLUSIONS

No significant differences were found in any of the cirDNA parameters between women with and without endometriosis. The low statistical power and heterogenous pelvic pathology in the control group render it difficult to determine whether the negative results reflect a true lack of increase in cirDNA in endometriosis.

The influence of biological and lifestyle factors on circulating cell-free DNA in blood plasma

Research and clinical use of circulating cell-free DNA (cirDNA) is expanding rapidly; however, there remain large gaps in our understanding of the influence of lifestyle and biological factors on the amount of cirDNA present in blood. Here, we review 66 individual studies of cirDNA levels and lifestyle and biological factors, including exercise (acute and chronic), alcohol consumption, occupational hazard exposure, smoking, body mass index, menstruation, hypertension, circadian rhythm, stress, biological sex and age. Despite technical and methodological inconsistences across studies, we identify acute exercise as a significant influence on cirDNA levels. Given the large increase in cirDNA induced by acute exercise, we recommend that controlling for physical activity prior to blood collection is routinely incorporated into study design when total cirDNA levels are of interest. We also highlight appropriate selection and complete reporting of laboratory protocols as important for improving the reproducibility cirDNA studies and ability to critically evaluate the results.

Cell-free DNA is abundant in ascites and represents a liquid biopsy of ovarian cancer

OBJECTIVE

Malignant ascites is a common clinical feature of ovarian cancer and represents a readily accessible sample of tumour cells and tumour DNA. This study aimed to characterise the cell-free DNA (cfDNA) in ascites in terms of its size profile, stability and cell-free tumour DNA (cftDNA) content.

METHODS

Cell spheroids, loose cells and cell-free fluid was collected from ascites from 18 patients with ovarian cancer. cfDNA was isolated and assessed for size by electrophoresis, concentration by fluorometry,cftDNA content by methylation specific qPCR of HOXA9 and IFFO1 promoter regions and by targeted sequencing. Stability was assessed after ascites fluid was stored at 4 °C for 24 and 72 h before fractionating.

RESULTS

The concentration of cfDNA in ascites ranged from 6.6 to 300 ng/mL. cfDNA size distribution resembled blood plasma-derived cfDNA, with major peaks corresponding to mono- and di-nucleosome DNA fragments. High molecular weight cfDNA was observed in 7 of 18 patients and appeared to be associated with extracellular vesicles. IFFO1 and HOXA9 methylation was proportionately higher in cfDNA than spheroid- and loose-cell fractions and was not observed in healthy primary cells. Variant allele frequency was highest in cfDNA compared to single cells and spheroids from ascites. Though cancer cell numbers in ascites declined to near zero in recurrent ascites from one patient undertaking chemotherapy, cftDNA could still be sampled. cfDNA size, concentration and tumour content was stable over 72 h.

CONCLUSION

cfDNA in ovarian cancer ascites demonstrates inter-patient variability, yet is consistently a rich source of cftDNA, which is a stable substrate. This supports the wider clinical use of ascites in the molecular analysis of ovarian cancer.

Total and endothelial cell-derived cell-free DNA in blood plasma does not change during menstruation

Assays measuring cell-free DNA (cfDNA) in blood have widespread potential in modern medicine. However, a comprehensive understanding of cfDNA dynamics in healthy individuals is required to assist in the design of assays that maximise the signal driven by pathological changes, while excluding fluctuations that are part of healthy physiological processes. The menstrual cycle involves major remodelling of endometrial tissue and associated apoptosis, yet there has been little investigation of the impact of the menstrual cycle on cfDNA levels. Paired plasma samples were collected from 40 healthy women on menstruating (M) and non-menstruating (NM) days of their cycle. We measured total cfDNA by targeting ALU repetitive sequences and measured endothelial-derived cfDNA by methylation-specific qPCR targeting an endothelium-unique unmethylated CDH5 DNA region. CfDNA integrity and endothelial cfDNA concentration, but not total cfDNA, are consistent across time between NM and M. No significant changes in total (ALU-115 p = 0.273; ALU-247 p = 0.385) or endothelial cell specific (p = 0.301) cfDNA were observed, leading to the conclusion that menstrual status at the time of diagnostic blood collection should not have a significant impact on the quantitation of total cfDNA and methylation-based cancer assays.

Abstract 731: Deciphering the origin of circulating cell-free DNA and its alterations during menstruation

Background and aim

Tumor derived cell-free DNA (cfDNA) in blood is a promising biomarker for cancer. Understanding factors that influence cfDNA concentrations and tissues of origins will assist with cancer biomarker development by excluding cfDNA changes driven by normal physiological processes. Endothelial cells make up 10% of the plasma cfDNA pool and undergo angiogenesis in the endometrium during the recurrent cycles of menstruation. Therefore, endothelial cfDNA changes are a potential source of false positive signals in cancer screening of pre-menopausal women. The aim of this study is to determine if cfDNA derived from the endothelium is increased during menstruation.

Methods

80 matched blood samples were collected from 40 healthy female donors at non-menstruating (NM) and menstruating (M) phases. CfDNA was extracted from 5 mL of plasma and quantified by qPCR by targeting a short (ALU115) and long region (ALU247) of ALU sequences. Methylation specific primers were designed to interrogate VE-cadherin DNA region at sequences downstream of the transcriptional start site (from +122 bp to +218 bp) to generate a 96 bp amplicon. The primer was tested against human primary aortic and blood-brain barrier endothelial cells as positive controls. To detect and measure endothelium derived sequences, cfDNA equivalent of 4 mL of plasma was bisulfite converted for quantification of endothelial specific VE-cadherin region by qPCR.

Results

VE-cadherin was identified from the literature and selected as a gene with methylation unique to endothelial cells. The VE-cadherin primer, designed specifically to amplify unmethylated version of the region, was tested against bisulfite-converted human aortic and blood-brain barrier genomic DNA with detection levels down to 111 pg and 125 pg input, respectively.

There was no statistically significant change in total cfDNA amount during menstruation and, based on the ratio of ALU115 and ALU247, no change in the integrity and size distribution of cfDNA was observed. We also found no correlation in total cfDNA amount between the NM and M phases. However, there was a positive correlation in the size ratio (r = 0.53), meaning that while individual cfDNA levels fluctuate over time, the size distribution stays relatively similar. Additionally, no significant changes were observed in the endothelium-derived cfDNA concentration between NM and M phases in the 40 matched plasma, including when the proportion is adjusted against the total cfDNA concentration. Positive correlation was also found between NM and M samples in the endothelium-derived cfDNA levels (r = 0.4).

Conclusion

This absence of change suggests a specific release pathway for cfDNA to enter the bloodstream rather than simply inefficiency in apoptotic debris clean-up and provides evidence that women can be screened for early detection of ovarian cancer using tests targeting cfDNA methylation regardless of their menstrual status.

Citation Format: Nicole Laurencia Yuwono, Kristina Warton, Claire Henry, Caroline Ford. Deciphering the origin of circulating cell-free DNA and its alterations during menstruation [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 731.

Abstract B28: Optimizing DNA processing and ovarian cancer methylation-specific PCR assays for the detection of early-stage ovarian cancer

Background: Ovarian cancer screening has the potential to greatly reduce mortality by shifting the time of diagnosis to an early stage when the tumor is confined to a single site and can be completely cured by surgical removal. Ovarian cancer DNA is present in blood and has a distinct methylation profile that distinguishes it from healthy blood cells, thus making it a target for ovarian cancer diagnosis via a blood test. However, while the presence of this DNA in the blood of cancer patients is well established, low abundance and DNA fragmentation present a barrier to developing clinical applications that target it.

Methods: We are systematically evaluating and optimizing each step of DNA extraction, processing, and analysis to increase the sensitivity of cancer DNA detection by PCR. At the same time, we are developing PCR assays to detect ovarian cancer specific methylation based on a previous tumor and blood screen that identified 73 methylated candidate biomarker regions.

Results: We have found that inhibitors of the PCR reaction introduced into the sample during DNA extraction and bisulphite treatment limit the total amount of DNA that can be analyzed, and thus the sensitivity of the test. Precipitation of the DNA samples prior to PCR analysis removes the inhibitors and increases useable sample volume and thus sensitivity. We have also developed methylation-specific PCR assays for DNA regions that are methylated in ovarian tumor tissue, and unmethylated in white blood cells.

Conclusions: We are progressing our candidate methylated DNA ovarian cancer biomarkers closer to testing in clinical samples. Our results on sample processing and DNA analysis methods to enable ovarian cancer DNA detection are relevant to all cancer types for which cell-free tumor DNA has clinical applications.

Citation Format: Kristina Warton, Nicole Yuwono, Claire Henry, Neville Hacker, Caroline Ford. Optimizing DNA processing and ovarian cancer methylation-specific PCR assays for the detection of early-stage ovarian cancer [abstract]. In: Proceedings of the AACR Special Conference on Advances in Ovarian Cancer Research; 2019 Sep 13-16, 2019; Atlanta, GA. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(13_Suppl):Abstract nr B28.

The untapped potential of ascites in ovarian cancer research and treatment

The build-up of fluid in the peritoneal cavity-ascites-is a hallmark of ovarian cancer, the most lethal of all gynaecological malignancies. This remarkable fluid, which contains a variety of cellular and acellular components, is known to contribute to patient morbidity and mortality by facilitating metastasis and contributing to chemoresistance, but remains largely under-researched. In this review, we will critically analyse the evidence associating ascites with metastasis and chemoresistance in ovarian cancer and provide an update on research in the field. We will argue the case for ascites as a unique and accessible substrate for tracking tumour progression and for translational research that will enhance our understanding of this cancer and lead to improvements in patient outcomes.

Circulating cell-free DNA from plasma undergoes less fragmentation during bisulfite treatment than genomic DNA due to low molecular weight

BACKGROUND

Methylation patterns in circulating cell-free DNA are potential biomarkers for cancer and other pathologies. Currently, bisulfite treatment underpins most DNA methylation analysis methods, however, it is known to fragment DNA. Circulating DNA is already short, and further fragmentation during bisulfite treatment is of concern, as it would potentially reduce the sensitivity of downstream assays.

METHODS

We used high molecular weight genomic DNA to compare fragmentation and recovery following bisulfite treatment with 2 commercially available kits (Qiagen). The bisulfite treated DNA was visualised on an agarose gel and quantified by qPCR. We also bisulfite treated, visualised and quantitated circulating DNA from plasma.

RESULTS

There was no difference in DNA fragmentation between the two kits tested, however, the Epitect Fast kit gave better recovery than the standard Epitect kit, with the same conversion efficiency. We also found that bisulfite treated circulating DNA migrates as distinct bands on agarose gels, suggesting that, in contrast to genomic DNA, it remains largely intact following treatment. Bisulfite treatment of 129 and 234 base PCR products confirmed that this was due to the short length of the circulating DNA fragments. Compared to double stranded DNA, bisulfite treated single stranded DNA gives a very weak signal on gel electrophoresis.

CONCLUSIONS

DNA fragmentation during bisulfite treatment does not contribute to loss of sensitivity in methylation analysis of circulating DNA. The absence of DNA fragments below approximately 170 bases from agarose gel images of purified circulating DNA raises the possibility that these fragments are single stranded following the DNA extraction step.

Methylated circulating tumor DNA in blood: power in cancer prognosis and response

Circulating tumor DNA (ctDNA) in the plasma or serum of cancer patients provides an opportunity for non-invasive sampling of tumor DNA. This 'liquid biopsy' allows for interrogations of DNA such as quantity, chromosomal alterations, sequence mutations and epigenetic changes, and can be used to guide and improve treatment throughout the course of the disease. This tremendous potential for real-time 'tracking' in a cancer patient has led to substantial research efforts in the ctDNA field. ctDNA can be distinguished from non-tumor DNA by the presence of tumor-specific mutations and copy number variations, and also by aberrant DNA methylation, with both DNA sequence and methylation changes corresponding to those found in the tumor. Aberrant methylation of specific promoter regions can be a very consistent feature of cancer, in contrast to mutations, which typically occur at a wide range of sites. This consistency makes ctDNA methylation amenable to the design of widely applicable clinical assays. In this review, we examine ctDNA methylation in the context of monitoring disease status, treatment response and determining the prognosis of cancer patients.

Methylation of cell-free circulating DNA in the diagnosis of cancer

A range of molecular alterations found in tumor cells, such as DNA mutations and DNA methylation, is reflected in cell-free circulating DNA (circDNA) released from the tumor into the blood, thereby making circDNA an ideal candidate for the basis of a blood-based cancer diagnosis test. In many cancer types, mutations driving tumor development and progression are present in a wide range of oncogenes and tumor suppressor genes. However, even when a gene is consistently mutated in a particular cancer, the mutations can be spread over very large regions of its sequence, making evaluation difficult. This diversity of sequence changes in tumor DNA presents a challenge for the development of blood tests based on DNA mutations for cancer diagnosis. Unlike mutations, DNA methylation that can be consistently measured, as it tends to occur in specific regions of the DNA called CpG islands. Since DNA methylation is reflected within circDNA, detection of tumor-specific DNA methylation in patient plasma is a feasible approach for the development of a blood-based test. Aberrant circDNA methylation has been described in most cancer types and is actively being investigated for clinical applications. A commercial blood test for colorectal cancer based on the methylation of the SEPT9 promoter region in circDNA is under review for approval by the Federal Drug Administration (FDA) for clinical use. In this paper, we review the state of research in circDNA methylation as an application for blood-based diagnostic tests in colorectal, breast, lung, pancreatic and ovarian cancers, and we consider some of the future directions and challenges in this field. There are a number of potential circDNA biomarkers currently under investigation, and experience with SEPT9 shows that the time to clinical translation can be relatively rapid, supporting the promise of circDNA as a biomarker.

Methylation-capture and Next-Generation Sequencing of free circulating DNA from human plasma

BACKGROUND

Free circulating DNA (fcDNA) has many potential clinical applications, due to the non-invasive way in which it is collected. However, because of the low concentration of fcDNA in blood, genome-wide analysis carries many technical challenges that must be overcome before fcDNA studies can reach their full potential. There are currently no definitive standards for fcDNA collection, processing and whole-genome sequencing. We report novel detailed methodology for the capture of high-quality methylated fcDNA, library preparation and downstream genome-wide Next-Generation Sequencing. We also describe the effects of sample storage, processing and scaling on fcDNA recovery and quality.

RESULTS

Use of serum versus plasma, and storage of blood prior to separation resulted in genomic DNA contamination, likely due to leukocyte lysis. Methylated fcDNA fragments were isolated from 5 donors using a methyl-binding protein-based protocol and appear as a discrete band of ~180 bases. This discrete band allows minimal sample loss at the size restriction step in library preparation for Next-Generation Sequencing, allowing for high-quality sequencing from minimal amounts of fcDNA. Following sequencing, we obtained 37 × 10(6)-86 × 10(6) unique mappable reads, representing more than 50% of total mappable reads. The methylation status of 9 genomic regions as determined by DNA capture and sequencing was independently validated by clonal bisulphite sequencing.

CONCLUSIONS

Our optimized methods provide high-quality methylated fcDNA suitable for whole-genome sequencing, and allow good library complexity and accurate sequencing, despite using less than half of the recommended minimum input DNA.

Abstract A4: Whole genome sequencing of free circulating DNA from plasma of HGSOC patients

Genomic analysis of free circulating DNA (fcDNA) has a number of potential clinical applications, including cancer studies. Tumor-derived fcDNA harbors the same molecular aberrations, including mutations and methylation, as the derivative tumor. As it can be collected by non-invasive means, fcDNA is particularly promising as a cancer detection tool. In addition, molecular characterization of fcDNA in cancer patients is holds promise for determining tumor sub-types and monitoring response to treatment. However, because fcDNA concentrations are relatively low in the circulation, whole-genome analysis for molecular characterization is quite challenging. Standardization of fcDNA collection and processing for downstream analysis is imperative to help address these challenges.

Our studies focus on the development of an early detection test for high-grade serous ovarian cancer (HGSOC), using whole-genome methylation analysis in fcDNA from HGSOC patients. DNA methylation plays a key role in the development of many cancer types, and thus carries great potential as a cancer diagnostic biomarker. We propose that DNA methylation changes in HGSOC tumors can be detected in the patient's fcDNA and can be used as a blood-based test to detect HGSOC. We have carried out Next Generation Sequencing (NGS) of methylation-enriched fcDNA from 3 HGSOC patients and 5 healthy controls with the aim of identifying a list of differentially methylated loci in fcDNA that can distinguish between HGSOC and control plasma.

We also sought to generate standard protocols for collection and processing of fcDNA for whole-genome studies. First we analyzed effects of blood storage time on fcDNA recovery and quality and determined that storage for ≥8 hrs prior to plasma separation leads to increased DNA concentrations, as well as an appearance of a band that co-migrates with high-molecular weight genomic DNA. This increased DNA concentration likely represents genomic DNA contamination as a result of leukocytes lysis during storage.

We next analyzed the effects of processing and scaling on enriched fcDNA recovery and quality. We used a methyl-binding protein-based (MBD2) protocol to enrich for methylated sequences in fcDNA isolated from plasma samples. Following application of a modified protocol due to low concentrations of fcDNA in plasma, we obtained a 10.2-14.9% enrichment of methylated fragments. In preparing the sample libraries for NGS, the predominant fcDNA size of ~180 base pairs allows minimal sample loss during the size restriction step. NGS analysis provided 37-86x10^6 unique mappable reads per fcDNA sample, representing &gt;50% of the total mappable reads. These read counts indicate a satisfactory level of library complexity was achieved from minimal fcDNA input.

Our standardization protocols provide the ability to comprehensively sequence patient fcDNA using half of the recommended sample input. This optimized approach allows in-depth whole genome characterization and comparison of fcDNA isolated from HGSOC and healthy controls to develop diagnostic tools and investigate biological mechanisms driving tumor progression.

Citation Format: Kristina Warton, Lin Vita, Nicola J. Armstrong, Warren Kaplan, Kevin Ying, Helena Mangs, Neville F. Hacker, Robert L. Sutherland, Susan J. Clark, Goli Samimi. Whole genome sequencing of free circulating DNA from plasma of HGSOC patients. [abstract]. In: Proceedings of the AACR Special Conference on Advances in Ovarian Cancer Research: From Concept to Clinic; Sep 18-21, 2013; Miami, FL. Philadelphia (PA): AACR; Clin Cancer Res 2013;19(19 Suppl):Abstract nr A4.

Prognostic and diagnostic significance of DNA methylation patterns in high grade serous ovarian cancer

OBJECTIVE

Altered DNA methylation patterns hold promise as cancer biomarkers. In this study we selected a panel of genes which are commonly methylated in a variety of cancers to evaluate their potential application as biomarkers for prognosis and diagnosis in high grade serous ovarian carcinoma (HGSOC); the most common and lethal subtype of ovarian cancer.

METHODS

The methylation patterns of 10 genes (BRCA1, EN1, DLEC1, HOXA9, RASSF1A, GATA4, GATA5, HSULF1, CDH1, SFN) were examined and compared in a cohort of 80 primary HGSOC and 12 benign ovarian surface epithelium (OSE) samples using methylation-specific headloop suppression PCR.

RESULTS

The genes were variably methylated in primary HGSOC, with HOXA9 methylation observed in 95% of cases. Most genes were rarely methylated in benign OSE, with the exception of SFN which was methylated in all HGSOC and benign OSE samples examined. Methylation of DLEC1 was associated with disease recurrence, independent of tumor stage and suboptimal surgical debulking (HR 3.5 (95% CI:1.10-11.07), p=0.033). A combination of the methylation status of HOXA9 and EN1 could discriminate HGSOC from benign OSE with a sensitivity of 98.8% and a specificity of 91.7%, which increased to 100% sensitivity with no loss of specificity when pre-operative CA125 levels were also incorporated.

CONCLUSIONS

This study provides further evidence to support the feasibility of detecting altered DNA methylation patterns as a potential diagnostic and prognostic approach for HGSOC.

Validation of RNA-based molecular clonotype analysis for virus-specific CD8+ T-cells in formaldehyde-fixed specimens isolated from peripheral blood

Recent advances in the field of molecular clonotype analysis have enabled detailed repertoire characterization of viably isolated antigen-specific T cell populations directly ex vivo. However, in the absence of a biologically contained FACS facility, peripheral blood mononuclear cell (PBMC) preparations derived from patients infected with agents such as HIV must be formaldehyde fixed to inactivate the pathogen; this procedure adversely affects nucleic acid template quality. Here, we developed and validated a method to amplify and sequence mRNA species derived from formaldehyde fixed PBMC specimens. Antigen-specific CD8+ cytotoxic T-lymphocyte populations were identified with standard fluorochrome-conjugated peptide-major histocompatibility complex class I tetramers refolded around synthetic peptides representing immunodominant epitopes from HIV p24 Gag (KRWII[M/L]GLNK/HLA B*2705) and CMV pp65 (NLVPMVATV/HLA A*0201 and TPRVTGGGAM/HLA B*0702), and acquired in separate laboratories with or without fixation. In the presence of proteinase K pre-treatment, the observed antigen-specific CD8+ T-cell repertoire determined by molecular clonotype analysis was statistically no different whether derived from fixed or unfixed PBMC. However, oligo-dT recovery methods were not suitable for use with fixed tissue as significant skewing of clonotypic representation was observed. Thus, we have developed a reliable RNA-based method for molecular clonotype analysis that is compatible with formaldehyde fixation and therefore suitable for use with primary human samples isolated by FACS outside the context of a biological safety level 3 containment facility.

An in vitro study of the effects of exposure to a GSM signal in two human cell lines: Monocytic U937 and neuroblastoma SK-N-SH

The use of mobile phones is increasing, which also increases the population's exposure to global system of mobile communications (GSM) signals. Questions of safety and possible biological effects are of concern and to date, remain largely unanswered. In order to examine possible biological effects of a GSM-like signal at a cellular level, we exposed two human cell lines (one of neuronal (SK-N-SH) and the other of monocytoid (U937) origin) to a 900 MHz RF signal, pulsed at 217 Hz, producing a specific absorption rate (SAR) of 0.2 W/kg. Putative effects were assessed by comparing radiofrequency-exposed cells to sham-exposed cells using a variety of assay techniques. For the cell line SK-N-SH, effects were specifically assessed by gene microarray, followed by real-time PCR of the genes of interest, Western blot analysis was used to measure heat shock protein levels, and flow cytometry to measure cell cycle distributions and apoptosis. Effects of radiofrequency on the cell line U937 were assessed by cell viability and cell cycle analysis. From our study of these two cell lines, we found no significant difference between sham-exposed versus radiofrequency-exposed cells in any of the assays or conditions examined.

A novel gene family induced by acute inflammation in endothelial cells

The aim of this study was to characterise a novel family of inflammatory genes induced by pro-inflammatory cytokines in primary human endothelial cells. Using a genome-wide array screen two previously uncharacterised genes, NLF1 and NLF2 were identified that were upregulated over 30 fold by treatment with interleukin 1beta for 2 h. They were also found to respond to tumour necrosis factor alpha, suggesting a general role in inflammation. Expression of both genes peaked 2 h after addition of interleukin 1beta, with similar kinetics to the fastest nuclear factor kappaB (NF-kappaB) induced genes. The activation of both genes by interleukin 1beta was abrogated by the proteasomal inhibitor, lactacystin which blocks activation of NF-kappaB by preventing IkappaB degradation. Furthermore, two sequences with homology to NF-kappaB binding sites in the promoter of NLF1 were found to be essential for rapid elevation in expression in response to interleukin 1beta. NLF1 and NLF2 transcripts were found predominantly in endothelial cells, and the encoded proteins were localised to the nuclear compartment suggesting a role in the regulation of transcription. Transfection of recombinant NLF into endothelial cells resulted in upregulation of the Rho kinases, Rnd1 and Gem GTPase. We propose that NLF1 and NLF2 belong to a novel gene family encoding nuclear factors with a role in regulating genes which control cellular architecture. This might increase vascular permeability in acute inflammation.

Recombinant CLIC1 (NCC27) assembles in lipid bilayers via a pH-dependent two-state process to form chloride ion channels with identical characteristics to those observed in Chinese hamster ovary cells expressing CLIC1

CLIC1 (NCC27) is an unusual, largely intracellular, ion channel that exists in both soluble and membrane-associated forms. The soluble recombinant protein can be expressed in Escherichia coli, a property that has made possible both detailed electrophysiological studies in lipid bilayers and an examination of the mechanism of membrane integration. Soluble E. coli-derived CLIC1 moves from solution into artificial bilayers and forms chloride-selective ion channels with essentially identical conductance, pharmacology, and opening and closing kinetics to those observed in CLIC1-transfected Chinese hamster ovary cells. The process of membrane integration of CLIC1 is pH-dependent. Following addition of protein to the trans solution, small conductance channels with slow kinetics (SCSK) appear in the bilayer. These SCSK modules then appear to undergo a transition to form a high conductance channel with fast kinetics. This has four times the conductance of the SCSK and fast kinetics that characterize the native channel. This suggests that the CLIC1 ion channel is likely to consist of a tetrameric assembly of subunits and indicates that despite its size and unusual properties, it is able to form a completely functional ion channel in the absence of any other ancillary proteins.

Crystal structure of a soluble form of the intracellular chloride ion channel CLIC1 (NCC27) at 1.4-A resolution

CLIC1 (NCC27) is a member of the highly conserved class of chloride ion channels that exists in both soluble and integral membrane forms. Purified CLIC1 can integrate into synthetic lipid bilayers forming a chloride channel with similar properties to those observed in vivo. The structure of the soluble form of CLIC1 has been determined at 1.4-A resolution. The protein is monomeric and structurally homologous to the glutathione S-transferase superfamily, and it has a redox-active site resembling glutaredoxin. The structure of the complex of CLIC1 with glutathione shows that glutathione occupies the redox-active site, which is adjacent to an open, elongated slot lined by basic residues. Integration of CLIC1 into the membrane is likely to require a major structural rearrangement, probably of the N-domain (residues 1-90), with the putative transmembrane helix arising from residues in the vicinity of the redox-active site. The structure indicates that CLIC1 is likely to be controlled by redox-dependent processes.

The nuclear chloride ion channel NCC27 is involved in regulation of the cell cycle

NCC27 is a nuclear chloride ion channel, identified in the PMA-activated U937 human monocyte cell line. NCC27 mRNA is expressed in virtually all cells and tissues and the gene encoding NCC27 is also highly conserved. Because of these factors, we have examined the hypothesis that NCC27 is involved in cell cycle regulation. Electrophysiological studies in Chinese hamster ovary (CHO-K1) cells indicated that NCC27 chloride conductance varied according to the stage of the cell cycle, being expressed only on the plasma membrane of cells in G2/M phase. We also demonstrate that Cl- ion channel blockers known to block NCC27 led to arrest of CHO-K1 cells in the G2/M stage of the cell cycle, the same stage at which this ion channel is selectively expressed on the plasma membrane. These data strongly support the hypothesis that NCC27 is involved, in some as yet undetermined manner, in regulation of the cell cycle.

Functional characterization of the NCC27 nuclear protein in stable transfected CHO-K1 cells

NCC27 belongs to a family of small, highly conserved, organellar ion channel proteins. It is constitutively expressed by native CHO-K1 and dominantly localized to the nucleus and nuclear membrane. When CHO-K1 cells are transfected with NCC27-expressing constructs, synthesized proteins spill over into the cytoplasm and ion channel activity can then be detected on the plasma as well as nuclear membrane. This provided a unique opportunity to directly compare electrophysiological characteristics of the one cloned channel, both on the nuclear and cytoplasmic membranes. At the same time, as NCC27 is unusually small for an ion channel protein, we wished to directly determine whether it is a membrane-resident channel in its own right. In CHO-K1 cells transfected with epitope-tagged NCC27 constructs, we have demonstrated that the NCC27 conductance is chloride dependent and that the electrophysiological characteristics of the channels are essentially identical whether expressed on plasma or nuclear membranes. In addition, we show that a monoclonal antibody directed at an epitope tag added to NCC27 rapidly inhibits the ability of the expressed protein to conduct chloride, but only when the antibody has access to the tag epitope. By selectively tagging either the amino or carboxyl terminus of NCC27 and varying the side of the membrane from which we record channel activity, we have demonstrated conclusively that NCC27 is a transmembrane protein that directly forms part of the ion channel and, further, that the amino terminus projects outward and the carboxyl terminus inward. We conclude that despite its relatively small size, NCC27 must form an integral part of an ion channel complex.

Inhibition of antigen and calcium ionophore induced secretion from RBL-2H3 cells by phosphatase inhibitors

The role of serine/threonine protein phosphatases PP1 and PP2A in mast cell secretion was investigated using the phosphatase inhibitors okadaic acid and calyculin A. Calyculin A (5-25 nm) inhibited antigen-induced secretion from a rat mucosal mast cell line (RBL-2H3) when added in conjunction with the activator. Okadaic acid (250-1000 nm) inhibited secretion only when added before activation and did so in a time- and concentration-dependent manner. Both inhibitors caused the cells to become rounder, but only calyculin A induced membrane blebbing and a loss of adherence. Okadaic acid also inhibited secretion induced by the calcium ionophore A23187, in the presence or absence of PMA, indicating that the phosphatase inhibitors act on a component of the secretory pathway downstream of calcium mobilization. Okadaic acid increased the phosphorylation of a number of proteins, as did an analogue methyl okadaate, which also inhibited secretion, but less effectively. Okadaic acid induced the phosphorylation of triton-insoluble proteins of 55, 18 and 16 kDa. The 55 kDa protein was identified as vimentin and okadaic acid induced its partial translocation to the triton-soluble fraction. Our data indicate that full secretory function in mucosal mast cells requires phosphatase activity.

Molecular cloning and expression of a chloride ion channel of cell nuclei

Ion channels are known to be present on the plasma membrane of virtually all cells and have been found on the membranes of various intracellular organelles. However, until recently they were believed not to occur at the nuclear membrane. In this study we describe the molecular cloning and characterization of a nuclear ion channel protein, designated nuclear chloride channel-27 (NCC27), from the human myelomonocytic cell line, U937. NCC27 is a novel chloride ion channel protein that was found to localize principally to the cell nucleus. Its only known homologue is a bovine chloride ion channel protein (p64) believed to localize to internal organelles. NCC27 therefore represents the first human member of a new class of organellar chloride ion channel proteins.