Molecular biomarkers, personalized medicine and other buzzwords are getting into the mainstream, so it is appropriate to take a look at the field and make a reality check: what exactly has been accomplished, where are the hot areas, what novel directions are likely to be ‘the next big thing’. The papers in this issue of Expert Review of Molecular Diagnostics will be of great interest to investigators working in this area. In a short overview, we have tried to provide a certain scaffold; it is by no means comprehensive and to a large degree presents a personal view of what we see now as the concept of personalized medicine and molecular biomarkers.
Definitions
Whenever a new field begins to emerge, it is useful to keep a close look at initial definitions and their modifications in order to ensure a coherent position on what is relevant and what is not. In 2001, the Biomarker Definitions Working Group defined biomarkers as ‘a characteristic that is objectively measured and evaluated as an indicator of normal biological processes, pathogenic processes, or pharmacologic responses to a therapeutic intervention’ Citation[1]. Objectivity of assessment is a significant limitation to the biomarker universe essentially excluding anything that involves a subjective finding by a human being. For example, immunohistochemical staining will produce a biomarker only when the subjective evaluation of intensity will transition from the old ‘plus-minus’ system to an objective conclusion. Similarly, even molecular imaging with specific tracers Citation[2] can relate to biomarkers only when assessment of the results is done objectively without subjective appraisals by the radiologist.
Types of biomarkers
Currently, biomarkers for companion diagnostics (CDx) are probably the hottest topic for pharmaceutical industry Citation[3], whereas molecular diagnostic enterprises are interested in diagnostic biomarkers (Dx) Citation[4]. Two other types – prognostic biomarkers and biomarkers for treatment monitoring – are somewhat lower on the list of priorities. It seems counterintuitive, but development of biomarkers for early detection of disease (screening biomarkers) is possibly less popular than discovery of risk-related biomarkers, although that may depend on the acknowledged problems with application of screening tests (e.g., the need to extensive and expensive clinical trials) Citation[5,6].
Areas of application
Potential areas of application are growing. While initially diagnostic biomarkers in infectious diseases Citation[7] and oncology Citation[8] have been the primary areas, molecular approaches to neurological diseases (e.g., Parkinson’s, Alzheimer’s and epilepsy) Citation[9–11], metabolic disorders (e.g., obesity) Citation[12] and psychiatry (e.g., schizophrenia) Citation[13] are evolving fast, and the need for both diagnostic and predictive (CDx) biomarkers is growing. Successful development of prognostic tests for breast cancer (OncotypeDx, Mammaprint, MapQuant Dx, Theros and PAM50) Citation[14] has added significant value supporting economics of reimbursement for future diagnostic tests in oncology. At the same time, it appears that psychological traits can also be described by objective and measurable changes in the human body Citation[15], and this relatively recent finding opens yet another area for biomarker-based test development. It is tempting to predict that biomarkers can be found for most, if not all, human conditions. The major problems are where to look for biomarkers and how to do that?
What molecular event can serve as a biomarker?
Here the space of potential objectively measurable changes is truly enormous and ranges from well-known biomarkers such as mutations in DNA and changes in expression patterns to much less explored alterations in protein phosphorylation Citation[16] and DNA methylation Citation[17] to recent findings of potential informative value of volatile organic compounds Citation[18], long noncoding RNA Citation[19], exosomal RNA Citation[20], metabolites Citation[21], autoantibodies Citation[22] and so on. The range of possibilities seems truly enormous until we consider limitations imposed by clinical utility and economic and practical realities.
Utility
Utility of a biomarker-based test is defined to a large degree by the invasiveness of sample collection procedure. Indeed, a cheek swab, exhaled air, stool or a measure of saliva or urine do not require any invasive procedures and are the most accessible samples for test development. A finger stick to collect a drop of blood is minimally invasive and together with noninvasive procedures can be done essentially everywhere. Venous blood, synovial fluid, cerebrospinal fluid and bronchoalveolar lavage require progressively more invasive procedures that have to be performed by a trained professional, so the applications of a biomarker to be tested in these samples are somewhat reduced. Tissue samples are usually collected by the most invasive procedures, so development of biomarkers for these samples are probably justified only in well-defined circumstances (e.g., to predict sensitivity of a tumor to a specific drug via testing for the presence of a CDx biomarker). The cost of sample collection and biomarker testing has to be considered as well. For obvious reasons, samples collected by an invasive procedure and tested by an expensive test will be at a disadvantage to a similar result achieved, for example, by analysis of saliva with a low-cost process. Practical considerations have to be taken into account as well. For example, collection of fecal samples is a noninvasive but unpleasant procedure, which can reduce compliance with testing, whereas blood-based tests may be noncompliant with religious beliefs of certain groups making testing more difficult.
Chemical & biological properties
The chemical and biological properties of a potential biomarker should be considered in the context of the sample where it will be tested. How stable is the objectively measured modification that will become a biomarker? This question is probably the most significant, because stability can have both positive and negative implications. Stability of the biomarker is necessary to minimize changes ex vivo during preparation, storage and transportation of the sample, so the biomarker can be readily measured in properly prepared and stored specimens even a long time after collection. At the same time, the biomarker should not be excessively stable in vivo because such stability will reduce its response to a degree when it will not reflect important physiological changes. To make things even more complicated, excessive reactivity of the biomarker in vivo is also detrimental because it may reduce its informative value for a disease or condition.
How many elements should be included in a biomarker? The answer to this question is determined by the required accuracy of the test; for example, prognostic biomarker OncotypeDx has 21 elements Citation[23], whereas there are 70 in Mammaprint Citation[24]. Inclusion of additional elements may increase the resulting accuracy, although at the cost of potential complications to test these elements and to incorporate them into the appropriate bioinformatics context.
Biomarker development starts with assessment of analytical performance of the biomarker detection technique, which should assure accurate detection of the chosen analyte. Clinical validity of the test should be then determined by its correlation with the outcome of interest, for example, response to therapy or detection of patients with the targeted disease. Finally, clinical utility determines the ability of the test to improve clinical management of the targeted disease.
We believe that DNA methylation is the best substrate for biomarker development because it is chemically stable, reflects gene expression, can be analyzed many years after sample collection and provides numerous potentially methylated sites that can act either as individual biomarkers Citation[25] or as multiplexed biomarker panels Citation[26]. Besides tissues, DNA methylation can be tested in blood using either DNA from nucleated cells Citation[15] or cell-free circulating DNA in plasma Citation[27,28], so different types of specimens can be analyzed using the same approach. Apparently, cell-free DNA is eliminated through kidneys, so cell-free DNA in urine can provide a noninvasive approach to test DNA methylation Citation[29]. We have shown that cell-free DNA in plasma has a half-life of <14 days and that its methylation reflects changes in the patient’s status Citation[30]. Our data indicate that premalignant, malignant, inflammatory and benign diseases can be detected via analysis of methylation in cell-free DNA, where disease-specific methylation patterns can be identified Citation[26,31,32]. Clinical testing of an appropriately selected combination of DNA fragments can be done in any clinical laboratory equipped with an instrument for real-time PCR or next-generation sequencing and there is little doubt that development point-of-care devices will soon make the procedure even easier to perform.
Our passion for epigenetic biomarkers does not mean that other types of tests will not be necessary. We are at the very early stages of a new, molecular era in pathology and pharmacology, so there is little doubt that very soon our current knowledge will require new measures to describe the molecular variety of human disease. Different combinations of different biomarkers appear to be the most promising way to advance the molecular classification of disease, molecular prediction of response to treatment and molecular monitoring of this response. The papers in this issue attest to this direction in molecular science.
Five-year view
Static biomarkers (e.g., DNA mutations) will soon become inadequate even for classification of disease, and dynamic biomarkers (e.g., epigenetic changes, protein modifications and so on) will be recognized as the most informative. This trend has started with expression profiling, which will become the norm for analysis of solid tissues and continues with a much more versatile epigenetic events. Protein interactions will emerge as the improvement of current proteome analysis and will further develop into multiprotein interactive mega markers. The trend of multicomponent markers will continue to grow, gradually pushing out monocomponent markers. The most sophisticated patients will demand precise molecular characterization of their disease and related information on its sensitivity to treatment. Biomarker-based tests will become obligatory for all new drugs and CDx will enhance efficacy of the new generation of treatments.
Key issues
• Objective measurement of a biomarker is its defining characteristic. Without it, the test is not biomarker based.
• Diagnostic, predictive, prognostic biomarkers and biomarkers for treatment monitoring are the main types.
• Any molecule can fulfill the definition of a biomarker if it reflects the physiological state, changes when the state changes and is measured objectively.
• Utility of a biomarker is determined by a combination of clinical utility, ease and cost of sample collection and analysis. It should also account for possible cultural issues associated with collection and analysis of the specimen.
• Compliance with testing should be considered early in development.
• Stability of the sample and dynamic changes associated with physiological state and background variability are extremely important.
• Epigenetic changes and DNA methylation in particular appear to be the most versatile venue for biomarker development.
Financial & competing interests disclosure
V Levenson and A Melnikov are co-founders and employees (Chief Scientific Officer and Chief Technology Officer, respectively) of US Biomarkers, Inc., a biotechnology startup that aims to develop and commercialize DNA methylation-based biomarkers. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.
No writing assistance was utilized in the production of this manuscript.
References
- Biomarkers Definitions Working Group. Biomarkers and surrogate endpoints: preferred definitions and conceptual framework. Clin. Pharmacol. Ther. 69(3), 89–95 (2001).
- Herholz K. Use of FDG PET as an imaging biomarker in clinical trials of Alzheimer's disease. Biomark. Med. 6(4), 431–439 (2012).
- Moorcraft SY, Smyth EC, Cunningham D. The role of personalized medicine in metastatic colorectal cancer: an evolving landscape. Therap. Adv. Gastroenterol. 6(5), 381–395 (2013).
- Bhavsar T, McCue P, Birbe R. Molecular diagnosis of prostate cancer: are we up to age? Semin. Oncol. 40(3), 259–275 (2013).
- Qaseem A, Denberg TD, Hopkins RH Jr et al. Clinical guidelines committee of the American College of Physicians. screening for colorectal cancer: a guidance statement from the American College of Physicians. Ann. Intern. Med. 156(5), 378–386 (2012).
- Qaseem A, Barry MJ, Denberg TD, Owens DK, Shekelle P. Clinical guidelines committee of the American College of Physicians. screening for prostate cancer: a guidance statement from the clinical guidelines committee of the American College of Physicians. Ann. Intern. Med. 158(10), 761–769 (2013).
- Torres A, Ramirez P, Montull B, Menendez R. Biomarkers and community-acquired pneumonia: tailoring management with biological data. Semin. Respir. Crit. Care Med. 33(3), 266–271 (2012).
- Tockman MS, Gupta PK, Pressman NJ, Mulshine JL. Considerations in bringing a cancer biomarker to clinical application. Cancer Res. 52( Suppl. 52), S2711–S2718 (1992).
- van Dijk KD, Teunissen CE, Drukarch B et al. Diagnostic cerebrospinal fluid biomarkers for Parkinson's disease: a pathogenetically based approach. Neurobiol. Dis. 39(3), 229–241 (2010).
- Ho L, Sharma N, Blackman L, Festa E, Reddy G, Pasinetti GM. From proteomics to biomarker discovery in Alzheimer's disease. Brain Res. Brain Res. Rev. 48(2), 360–369 (2005).
- Pollard JR, Eidelman O, Mueller GP et al. The tarc/sicam5 ratio in patient plasma is a candidate biomarker for drug resistant epilepsy. Front. Neurol. 3, 181 (2012).
- Pyrzak B, Ruminska M, Popko K, Demkow U. Adiponectin as a biomarker of the metabolic syndrome in children and adolescents. Eur. J. Med. Res. 15( Suppl. 15), 147–151 (2010).
- Mill J, Tang T, Kaminsky Z et al. Epigenomic profiling reveals DNA-methylation changes associated with major psychosis. Am. J. Hum. Genet. 82(3), 696–711 (2008).
- Paik S. Is gene array testing to be considered routine now? Breast 20( Suppl. 20), S87–S91 (2011).
- Mathews HL, Konley T, Kosik KL et al. Epigenetic patterns associated with the immune dysregulation that accompanies psychosocial distress. Brain Behav. Immun. 25(5), 830–839 (2011).
- Hricik T, Federici G, Zeuner A et al. Transcriptomic and phospho-proteomic analyzes of erythroblasts expanded in vitro from normal donors and from patients with polycythemia vera. Am. J. Hematol. 88(9), 723–729 (2013).
- Kisiel JB, Garrity-Park MM, Taylor WR, Smyrk TC, Ahlquist DA. Methylated Eyes Absent 4 (EYA4) gene promotor in non-neoplastic mucosa of ulcerative colitis patients with colorectal cancer: evidence for a field effect. Inflamm. Bowel Dis. 19(10), 2079–2083 (2013).
- de Meij TG, Larbi IB, van der Schee MP et al. Electronic nose can discriminate colorectal carcinoma and advanced adenomas by fecal volatile biomarker analysis: proof of principle study. Int. J. Cancer doi:10.1002/ijc.28446 (2013) ( Epub ahead of print).
- Calore F, Lovat F, Garofalo M. Non-coding RNAs and cancer. Int. J. Mol. Sci. 14(8), 17085–17110 (2013).
- Tanaka Y, Kamohara H, Kinoshita K et al. Clinical impact of serum exosomal microRNA-21 as a clinical biomarker in human esophageal squamous cell carcinoma. Cancer 119(6), 1159–1167 (2013).
- Yang M, Soga T, Pollard PJ. Oncometabolites: linking altered metabolism with cancer. J. Clin. Invest. 123(9), 3652–3658 (2013).
- Chatterjee M, Tainsky MA. Autoantibodies as biomarkers for ovarian cancer. Cancer Biomark. 8(4–5), 187–201 (2010).
- Lyman GH, Cosler LE, Kuderer NM, Hornberger J. Impact of a 21-gene RT-PCR assay on treatment decisions in early-stage breast cancer: an economic analysis based on prognostic and predictive validation studies. Cancer 109(6), 1011–1018 (2007).
- Mook S, Schmidt MK, Weigelt B et al. The 70-gene prognosis signature predicts early metastasis in breast cancer patients between 55 and 70 years of age. Ann. Oncol. 21(4), 717–722 (2010).
- Levenson VV, Melnikov AA. The MethDet: a technology for biomarker development. Expert Rev. Mol. Diagn. 11(8), 807–812 (2011).
- Liggett TE, Melnikov A, Yi Q et al. Distinctive DNA methylation patterns of cell-free plasma DNA in women with malignant ovarian tumors. Gynecol. Oncol. 120(1), 113–120 (2011).
- Ponomaryova AA, Rykova EY, Cherdyntseva NV et al. Potentialities of aberrantly methylated circulating DNA for diagnostics and post-treatment follow-up of lung cancer patients. Lung Cancer 81(3), 397–403 (2013).
- Ma Y, Wang X, Jin H. Methylated DNA and microRNA in body fluids as biomarkers for cancer detection. Int. J. Mol. Sci. 14(5), 10307–10331 (2013).
- Bryzgunova OE, Skvortsova TE, Kolesnikova EV et al. Isolation and comparative study of cell-free nucleic acids from human urine. Ann. N. Y. Acad. Sci. 1075, 334–340 (2006).
- Liggett TE, Melnikov AA, Marks JR, Levenson VV. Methylation patterns in cell-free plasma DNA reflect removal of the primary tumor and drug treatment of breast cancer patients. Int. J. Cancer 128(2), 492–499 (2011).
- Liggett T, Melnikov A, Yi QL. Differential methylation of cell-free circulating DNA among patients with pancreatic cancer versus chronic pancreatitis. Cancer 116(7), 1674–1680 (2010).
- Melnikov A, Scholtens D, Godwin A, Levenson V. Differential methylation profile of ovarian cancer in tissues and plasma. J. Mol. Diagn. 11(1), 60–65 (2009).