The discovery of new biomarkers for a given pathway or drug class is an integral component of successful methods for measuring pharmacological and biological consequences. As new proteins continue to be surveyed as potential drug targets, the need for markers that are tightly linked to their functions or downstream consequences continues to increase. The need is particularly high for novel targets where the associated biochemical pathways are not well characterized, but many validated targets could also benefit from expanded portfolios of biomarker endpoints to track during drug discovery and development efforts. The most powerful approaches toward discovery of molecular biomarkers are derived from recent advances in the fields of proteomics and genomics. Although these fields are covered elsewhere in this volume (see 3.01 Genomics; 3.02 Proteomics; 3.05 Microarrays), a description of some examples of biomarker discovery approaches that utilize such technologies is of relevance to the topic of this chapter.
3.04.4.2.1 Gene expression profiling, genomic alterations, and tissue microarrays
Among the battery of emerging technologies often referred to as 'omics,' microarray-based gene expression profiling and other genomic/transcriptomic technologies have thus far been applied most extensively and, arguably, most successfully. Such large-scale profiling has shifted the emphasis of some research efforts from hypothesis testing to hypothesis generation; or, at the very least, biomarker generation. Indeed, individual gene transcripts as well as whole clusters of coregulated transcripts that are modulated by disruption of a given signaling pathway are candidate biomarkers (as well as potential drug targets in their own right if critically important to the same pathway). Microarray transcript profiling can be applied in preclinical models with the aims of better understanding of the pharmacodynamic effects of drug candidates (or other bioactive agents) and ultimately of identification of biomarkers that can be applied during clinical testing. Several examples of this approach have been described, including the use of tumor xenograft models to probe for gene expression biomarkers modulated by in vivo exposure to an antiangiogenic small-molecule kinase inhibitor31; exposure of bone marrow-derived macrophages to Gram-positive bacteria to identify gene expression change32; animal models for evaluating bladder tissue responses to inflammation-inducing agents33; and treatment of cancer cell lines with chemotherapy drugs and signal transduction inhibitors to probe for biomarkers of combination therapy.34 In one of these studies,31 a transcript whose level was modulated by the kinase inhibitor in the preclinical model was shown to encode a protein product which was modulated in posttreatment biopsies in some human patients that were treated with the same kinase inhibitor; this demonstrates the potential utility of biomarker discovery studies in translation to clinical applicability (a topic to be discussed later in this chapter).
These are only a few examples of how application of microarray-based or other large-scale expression profiling methods provide a powerful means to identify putative biomarkers, particularly in the context of well-controlled laboratory experiments (as care must be taken to minimize false-positive risk in experiments of this type, given the multidimensional complexity of the data output). However, it must be noted that many potential biomarkers have also been identified in the direct profiling of well-annotated human clinical specimens as opposed to laboratory experiments in model systems,35 and findings from either realm can be readily implemented in both to probe for biological or diagnostic relevance. An insightful commentary by Liu36 summarizes the interface between experimental and clinical gene expression biomarker discovery efforts, and contrasts purely correlative, empirical discovery approaches against bottom-up approaches that build on mechanistic insights into cellular or physiological processes. The latter approach is likely to be of the most relevance for target-based research efforts, at least in the near term. In any case, candidate biomarkers identified via transcriptome-wide survey methods are most efficiently probed in subsequent experiments via more narrowly focused assays, particularly if a relatively small number of markers are sufficient to serve as biological readouts (there are at present no hard and fast rules on what comprises an ideal number of analytes to measure in a biomarker panel, but lower is usually better in terms of cost and reproducibility so long as sufficient decision-enabling information is captured37). The most common assay format to routinely measure levels of specific transcripts would be variants of quantitative reverse transcription polymerase chain reaction (qRT-PCR) technology,38'39 which is a sensitive, specific, and versatile platform; however, if transcript level changes are well correlated with changes in the level of the corresponding protein (this may often not be the case40), then protein detection methods described earlier can also be employed.
In addition to gene expression changes, another type of biomarker of relevance to target selection, particularly in heterogeneous, mutationally driven diseases such as cancer, is change in protein expression due to changes in gene copy number or other mechanisms. Genes that are selectively lost (genomic deletion) or selectively increased in copy number (genomic amplification) in disease tissue are candidates for causative agents in cancer, and genome-level survey approaches are a means to identify such candidates via comparative genomic hybridization or related genomic methods.41,42 Confirmation that expression of a given protein is frequently altered in a large number of cases of a particular type of cell or tissue can be determined in a high-throughput fashion through profiling of micro- or macroarrays containing many representative specimens. Tissue arrays usually consist of microscope slides mounted with dozens of small tissue specimens representing a survey of similar disease type (e.g., breast cancer biopsies) or of various normal organs and tissues (for instance, to probe the distribution of expression of a novel target, with regard to potential toxicity).43,44 Tissue arrays are typically probed for proteins via immunohistochemistry methods, and for gene transcripts or copy number via in situ hybridization. Applications for which tissue arrays are invaluable include confirmation of whether a potential target is overexpressed (or, conversely, never expressed) in a certain disease indication, or likewise that a gene is frequently deleted or amplified, as well as the aforementioned query of the normal tissue distribution of targets in multiple species. An application conceptually related to tissue arrays is that of reverse phase protein microarrays; these consist of arrays of protein lysates rather than intact tissue specimens, and allow for immunoblot probing of protein expression (or activation state, in the case of phosphorylated proteins, for instance) across hundreds of lysates, including serial dilutions to aid in quantitation.45,46 Reverse phase microarrays are particularly useful for profiling cell lines or tissues that are too limited in quantity for histological handling.
Along with genomic approaches, proteomics has also enabled much progress to be made toward biomarker discovery. The application of multiplex protein profiling has been previously mentioned, and most other proteomic discovery efforts rely on the quantitation of relative differences in abundance of specific proteins. The detection methods are numerous and will not be described in detail in this chapter; indeed, many of the analytic approaches are still in the initial testing phase and will require additional testing. However, the potential utility of proteomic profiling in identification of potential drug targets as well as putative biomarkers is illustrated in innovative applications of proteomic mapping. For example, in an effort to identify proteins induced at the tissue-blood interface in vivo (and thus potentially accessible to targeted antibodies or other biological therapeutics), multidimensional protein identification technology (MudPIT; a 'shotgun' proteomics approach that utilizes chromatographic separation followed by tandem mass spectrometry for identification of peptide sequences to confirm protein identity) was applied to luminal endothelial cell plasma membranes isolated from rat lungs and from cultured microvascular endothelial cells.47 A large set of proteins was identified, and subsequent validation experiments established specific proteins as putative targets (as well as endothelium-specific biomarkers) in lung and tumor tissues.48 Such open-ended profiling using proteomic technologies holds much promise for both target and biomarker discovery, but careful application, cautious interpretation, and confirmatory follow-up experiments (using both the initial proteomic screening assay as well as measurements of specific proteins via immunoassays) are essential to optimal success.
As proteomics technologies continue to emerge as powerful biomarker discovery tools, so also does the field of metabolomics (also known as metabonomics). Metabonomic research is focused on the measurement of low-weight components, primarily in biological fluids such as urine and serum, and thus is not of direct relevance in searches for protein or nucleic acid biomarkers that may also be drug targets. However, metabonomics profiles may indeed be of utility in monitoring the biological consequences of modulating the activity of a target. The basic approach relies on nuclear magnetic resonance (NMR) spectroscopy to generate spectrometric readouts of the relative abundance of low molecular weight metabolites.49 The method can be cost-effective and relatively high throughput with the availability of suitable NMR capability, and is most reliable in comparing metabolite patterns before and after a specific stimulus in samples collected from the same subject. To date, the most common application of metabonomics profiling has been in the toxicology field with the objective of identifying metabolite signatures of toxic stresses50 (more on this subject below); however, examples of applications in the idenfication of efficacy and mechanistic markers can be found, such as a study of a diabetes drug (the peroxisome proliferator-activated receptor gamma agonist rosiglitazone) in a mouse model.51 In this study, a subset of specific metabolites - the lipid metabolome - was profiled via thin-layer chromatography coupled with capillary gas chromatography, and a number of alterations were detected. For broader metabolome-wide profiling, it is more typical for a pattern of peaks, or 'fingerprint' (similar to those obtained in mass spectrometry-based proteomic profiling), to be the primary output of metabolome data analysis. However, identification of specific peaks of interest can be done via a variety of means, such as structural similarity matching against databases of known molecules or by further isolation of the peak fraction and subsequent direct analysis.49 As with genomics and proteomics, metabonomic methods hold much promise but also warrant much rigor in their application (a cautionary perspective on the importance of appropriate methodological and statistical rigor in marker discovery studies is provided by Ransohoff,52 with a key message being the necessity of performing independent confirmatory experiments to firmly establish linkage between multianalyte changes and the phenomenon in question). The convergences of all of these methodologies along with bioinformatics comprise the nascent field known as systems biology. As this field matures, its impact on biomarker discovery applications will likely continue to increase.
Was this article helpful?
Learning About 10 Ways Fight Off Cancer Can Have Amazing Benefits For Your Life The Best Tips On How To Keep This Killer At Bay Discovering that you or a loved one has cancer can be utterly terrifying. All the same, once you comprehend the causes of cancer and learn how to reverse those causes, you or your loved one may have more than a fighting chance of beating out cancer.