A Computational Solution to Bolster Epigenetic Clock Reliability for Clinical Trials and Longitudinal Tracking

Epigenetic clocks are widely used aging biomarkers, but they are calculated from methylation data for individual CpGs that can be surprisingly unreliable. We report that technical noise causes six major epigenetic clocks to deviate by 3 to 9 years between replicates. We present a novel computational solution: we perform principal component analysis followed by biological age prediction using principal components, extracting shared age-related changes across CpGs while ignoring noise from individual CpGs. Our novel principal-component versions of six clocks show agreement between most technical replicates within 1 year, and increased stability in short- and long-term longitudinal studies. This requires only one additional step compared to traditional clocks, does not require prior knowledge of CpG reliabilities, and can improve the reliability of any existing or future epigenetic biomarker. The extremely high reliability of principal component epigenetic clocks makes them particularly useful for personalized medicine and clinical trials evaluating novel aging interventions.

Development of a Fast and Reproducible Assay for the Clinical Implementation of Epigenetic Biomarkers to Predict Decitabine Response in Patients with Chronic Myelomonocytic Leukemia

Chronic myelomonocytic leukemia (CMML) is an aggressive myelodysplastic/myeloproliferative neoplasm represented by abnormal, clonal monocytosis, resistance to conventional chemotherapy, and an elevated risk of transforming into acute leukemia. While allogeneic hematopoietic stem cell transplantation is the only curative treatment option, this is not available to everyone due to either lack of compatible donors or the existence of comorbidities that make the patient ineligible for such procedure. Hypomethylating agents (HMAs) such as Decitabine and Azacitidine have shown efficacy in ∼50% of the patients. However, response to these agents can take 6 months or longer and patients are required to stay on the drug for this period of time, even if only about 50% of them will benefit from it. Therefore, in order to provide the best treatment option at an appropriate timing, robust biomarkers for predicting the response to HMAs are urgently needed. Using genome-wide methylation analysis (ERRBS: enhanced reduced representation bisulfite sequencing), we previously identified 21 epigenetic biomarkers which accurately predicted the response to decitabine in patients with CMML (Meldi K, et al. J Clin Invest. 2015). However, due to the complexity of this assay as well as the genome-wide nature of it, ERRBS is not suitable for the clinical setting. Thus, we sought to develop a novel, targeted assay that can be used in the clinical setting for the implementation of this biomarker as well as to validate the performance of the epigenetic biomarker in an independent cohort of patients. For this purpose, primers targeting the 21 regions included in the epigenetic biomarker were designed with overhang adapter sequences compatible with illumina multi-index system. Genomic DNA was extracted from CD34+ human blood cells and bisulfite conversion was performed using EZ DNA Methylation Kit (ZYMO Research). Multiplex PCR parameters were established by calculating the Gibbs Free Energy (dG) for each primer combination using PrimerSuite program. Four primer pools were determined based on the dG and optimum annealing temperatures at single PCR. Multiplex PCR was performed using FastStart High Fidelity PCR System (Roche) and the combined PCR amplicons were purified by AMPure XP (Beckman Coulter). Purified amplicons were indexed by Nextera XT index kit (illumina) and KAPA HiFi HotStart ReadyMix (KAPA Biosystems) and sequenced on an Illumina MiSeq sequencer (2x150 bp). Sequencing reads were mapped to human reference genome (hg19) using Bowtie2, and cytosine methylation states were determined by Bismark. Twenty out of 21 biomarkers were successfully amplified and sequenced. Using a standard curve generated with in vitro methylated DNA, we demonstrated that the assay accurately captured the different methylation percentages in the standard curve and was highly reproducible, with correlation coefficient r=0.98 between two independent operators (Fig.1). Moreover, analysis of 26 CMML patient samples previously analyzed by ERRBS showed strong correlation between ERRBS and the targeted panel assay (Pearson's r: 0.83-0.98). To test the performance of predicting decitabine response, we analyzed both the ERRBS and the targeted panel assay data of the 26 patients (12 responders, 14 non-responders) using supervised machine learning algorithm, SVM classifier. The two assays showed 100% concordance of predictions with an accuracy of 96%, indicating high reproducibility and predictive performance of the novel targeted panel assay. In summary, we have developed a fast, robust, and highly reproducible MiSeq-based, targeted panel assay for decitabine response prediction in patients with CMML. Figure 1 Figure 1. Disclosures Santini: Astex: Membership on an entity's Board of Directors or advisory committees; Takeda: Membership on an entity's Board of Directors or advisory committees; Novartis: Honoraria, Membership on an entity's Board of Directors or advisory committees; Menarini: Membership on an entity's Board of Directors or advisory committees; Gilead: Membership on an entity's Board of Directors or advisory committees; Geron: Membership on an entity's Board of Directors or advisory committees; BMS/Celgene: Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding.

An Examination of Risk Factors for Tobacco and Cannabis Smoke Exposure in Adolescents Using an Epigenetic Biomarker

Objective: Evolving patterns of nicotine and cannabis use by adolescents require new tools to understand the changing epidemiology of these substances. Here we describe the use of a novel epigenetic biomarker sensitive to both tobacco and cannabis smoke in a longitudinal sample of high-risk adolescents. We examine risk factors for positivity for this epigenetic biomarker in comparison to positivity for conventional serum biomarkers of nicotine and cannabis use.Method: Eastern Iowa 10th graders who had a friend or family member who smoked were eligible to participate in a longitudinal study over 10–12th grades. Subjects provided self-report data on nicotine, tobacco, and cannabis use patterns as well as blood samples that were used for serum cotinine and THC assays. DNA was prepared for analysis of methylation at the CpG cg05575921, a sensitive indicator of smoke exposure. Relationships between positivity for each these biomarkers and a variety of risk factors, including demographics, family and peer relationships, psychopathology, willingness to smoke, and perceptions of typical cigarette and cannabis users, were examined at the 10th (n = 442), 11th (n = 376), and 12th (n = 366) grade timepoints.Results: A increasing proportion of subjects were positive for cotinine (5–16%), THC (3–10%), and cg05575921 methylation (5–7%) across timepoints, with some overlap. Self-reported combusted tobacco and cannabis use was strongly correlated with all biomarkers, whereas cg05575921 methylation was not correlated with reported e-cigarette use. Dual users, defined as those positive for nicotine and THC in the 12th grade showed the greatest cumulative smoke exposure, indicated by cg05575921 methylation. Subjects reported more positive attitudes toward cannabis users than cigarette smokers, and willingness to smoke and positive perceptions of tobacco and cannabis smokers were significant risk factors for biomarker positivity across timepoints.Conclusion: We conclude that measurement of cg05575921 methylation in adolescents is a useful tool in detecting tobacco smoking in adolescents, and may be a novel tool for the detection of cannabis smoking and cannabis and tobacco co-use, though non-combusted forms of nicotine use do not appear to be detectable by this method.