Screening for novel oncology biomarker panels using both DNA and protein microarrays. John Anson, PhD VP Biomarker Discovery

Screening for novel oncology biomarker panels using both DNA and protein microarrays John Anson, PhD VP Biomarker Discovery

Outline of presentation Introduction to OGT and our approach to biomarker studies OGT s genomic biomarker programme Sense proteomic biomarker discovery platform Prostate cancer biomarker discovery programme

Oxford Gene Technology Founded by Ed Southern in 1995 60 people >20% revenue growth p.a. Focused on microarrays/dna chips

Innovative clinical genetics and diagnostic solutions to advance molecular medicine Clinical and Genomic Solutions Array services and cytogenetics products Biomarker Discovery Genomic- and protein-based diagnostics Technologies For Molecular Medicine IP Licensing

OGT products, services and research

Biomarker discovery strategic focus Disease Area Focus: Prostate, Ovarian and Colorectal Cancer

Conventional biomarkers studied DNA Single nucleotide polymorphisms (SNPs) RNA Messenger RNA (mrna) markers Protein Metabolites

New Wave of genomic biomarkers DNA Single nucleotide polymorphisms (SNPs) Epigenetic (methylation) Copy number variation (CNV) RNA Messenger RNA (mrna) markers MicroRNA (mirna) markers Proteins Metabolites

Biomarker discovery and validation at OGT OGT s current genomic biomarker programmes include: Prostate Cancer mirna markers Colorectal Cancer methylation markers Ovarian Cancer methylation markers Genomic biomarkers for: Early disease identification and prognosis Differentiation of aggressive vs non-aggressive tumours Prediction of patient response to medications The best drug at the correct dose for the right patient personalised medicine

Screening for autoantibodies

Clinical significance of autoantibodies Autoantibodies are formed in many diseases including cancer and heart disease The appearance of autoantibodies may precede disease symptoms by many years Easy to detect from blood with high sensitivity Identification of disease specific autoantibodies can lead to a new generation of diagnostic and prognostic indicators Non-invasive serum sampling is the future of cancer diagnostics. By detecting autoantibodies in serum using a novel functional protein microarray, the Sense approach can improve both the specificity and sensitivity of these tests. Prof. Norman J Maitland University of York

Autoantibody based tests in the news The London Times, June 2010

Detecting autoantibodies in human serum 1. Apply patient serum Y Y 2. Incubate and wash Y Y Y Y 3. Detect with α-human antibody

Sense protein array configuration Content Expressed in baculovirus, C-terminal BCCP fusion Approximately 1346 proteins corresponding to 1296 unique genes (lot-specific) Arrays 1346 unique protein features printed in quadruplicate 48 pairs of Cy3 labelled fiduciary markers Positive controls Dilution series of human IgG paired spot per dilution acts as control for secondary antibody incubation IgHG1 human immunoglobulin on array in quadruplicate acts as control for expression and secondary incubation

Technical performance comparison of 2 different serum samples 1 1 2 2 Standard data extraction methods used

Reproducibility a key to robust autoantibody detection fiduciary markers protein feature replicates in quadruplicate Inter array CV s of <10%

OGT s internal biomarker discovery objectives Primary: to identify panels of biomarkers which can distinguish specific disease conditions against a background of other disease conditions e.g. prostate cancer cf. prostatic hypertrophy and prostatitis colorectal cancer cf. benign colorectal conditions Secondary: to identify panels of markers which can distinguish between different disease stages e.g. distinguish between early and late stage cancer Potential future studies: to identify panels of biomarkers which are indicative of disease outcome e.g. aggressive versus indolent prostate cancer

Exploring the biomarker discovery space Biomarkers from different cohorts occupy a different but overlapping space

Avoiding bias in study design Samples randomised prior to assay Operational bias minimised by the use of SOPs for all procedures including sample collection, transport, storage, assay and analysis Samples and data blinded at all stages of the study Use of standards and QC used to normalise and validate assays Samples used for biomarker discovery are not used to validate the results

Study design the ideal sample cohort Target disease serum samples Normal healthy controls Diseases of the same organ or tissue Diseases of the same type e.g. cancer Diseases with related symptoms e.g. inflammation Unrelated disease controls

Prostate cancer screening > current state of the art Prostate Specific Antigen (PSA) test Good sensitivity (~90%), but poor specificity (<50%) 30 million tests annually in US alone (market size approx $3 billion) BUT... PSA testing can t detect prostate cancer and, more important, it can t distinguish between the two types of prostate cancer the one that will kill you and the one that won t....the test is hardly more effective than a coin toss Quotes from a recent article in NY Times by Richard Ablin, the scientist who discovered PSA in the 1970 s

Prostate cancer pilot study > novel biomarker panels discovered Data generated on 3 array types comprising 925 unique proteins printed in quadruplicate Sample cohort: case (n=73) and control (healthy (n=37) & BPH (n=23)) Panels identified using genetic programming Sets of panels identified with 2 to 15 members and ranked by sensitivity/specificity Maximum sensitivity/specificity score is 1.89 Sensitivity and specificity both >90% Patent application filed

Prostate cancer follow on study > generalised study design Sample cohort for discovery Target disease samples Normal non-diseased control samples Interfering disease samples Control samples up to 20 different conditions Total of ~1,700 samples for discovery and validation Sample cohort for validation Biomarkers validated against naïve sample sets Target disease samples Normal non-diseased control samples Other disease samples

Prostate cancer follow on study > target sample cohorts Prostate cancer Matched normals >50 years old Matched normals <30 years old Benign prostatic hypertrophy Prostatitis Other cancers Other diseases In total 400 case and ~1300 controls

Prostate cancer follow on study > study design workflow Randomised discovery cohort (>1500) Randomised validation cohort Discovery set Naïve in-house validation set Naïve sets of samples which have not been used for biomarker discovery are used for validation Assay Data analysis Training and test set Biomarker panel with classification Assay Data analysis In-house validated panel Power analysis to determine number of validation samples Assay Data analysis Diagnosed naïve samples Validated panel

Sense Superior Performance Designed In Biomarker discovery and analytical validation demand a highly reproducible platform, with low technical variation Thus enabling the subtle biological variation between case and control to be detected with greater ease Robust statistical analysis and unparalleled array consistency deliver novel, superior and validated protein biomarker panels; the Sense advantage is designed in: Robust statistical analysis CV s of less than 10% between arrays real biological variation is easier to spot Unparalleled array consistency - reproducible arraying of relevant protein content in quadruplicate for better statistical analysis this means less false positives/negatives 30

OGT can help! OGT has made major investments in infrastructure to support high throughput genomics and biomarker discovery and validation projects. Run your samples using our HT facility We have an unparalleled track record in delivering high quality data for CNV analysis and other microarray applications Partner with us to access powerful genomic and proteomic biomarker technologies

Acknowledgements collaborators and customers

Thank you

CytoSure: For research use only This product is provided under an agreement between Agilent Technologies, Inc., and OGT. The manufacture, use, sale or import of this product may be subject to one or more of U.S. patents, pending applications, and corresponding international equivalents, owned by Agilent Technologies, Inc. The purchaser has the non-transferable right to use and consume the product for RESEARCH USE ONLY AND NOT for DIAGNOSTICS PROCEDURES. It is not intended for use, and should not be used, for the diagnosis, prevention, monitoring, treatment or alleviation of any disease or condition, or for the investigation of any physiological process, in any identifiable human, or for any other medical purpose. This document and its contents are Oxford Gene Technology IP Limited. All rights reserved. OGT, Genefficiency, Oligome, and Oxford Gene Technology are trademarks of Oxford Gene Technology IP Limited. Sense is a registered trademark of Sense Proteomic Limited.