biomarker classification based on mechanism of drug action and drug–disease interaction has been proposed (Danhof et al., 2005). Type 0 biomarkers (also called pharmacogenomic or predictive biomarkers) are measurable DNA/RNA characteristics that identify a PK or PD related genotype/phenotype of an individual, which determines drug response in that individual. For example, mutation of the epidermal growth factor receptor (EGFR) gene is reported to be associated with clinical responsiveness to the EGFR kinase inhibitor, gefitinib. Phosphorylated‐EGFR for gefitinib assessed by immunohistochemistry assays is a good example of type 0 biomarker. Other examples include phosphorylated‐CRKL for Gleevec and mRNA gene expression‐based biomarkers for anti‐cancer drugs in cell‐based assays. Type 1 biomarker is a measure of drug exposure (plasma or more PD‐relevant target tissue concentrations of a drug). Type 2 ‐ type 4 constitute target engagement or PD biomarkers at the different levels of PD modulation. Type 2 biomarkers reflect receptor occupancy and can be useful in cases where target occupancy correlates well with therapeutic response (Nordström et al., 1993). Type 3 biomarkers quantify target activation, which is determined by the intrinsic efficacy of the drug and receptor density. Intrinsic efficacy of the drug determines the extent of occupancy needed for target activation, while receptor density determines the system maximum, which, if different between sites, will determine the selectivity of drug action. An example of type 3 biomarkers is the quantitative electroencephalogram (EEG) parameters reflecting the OP3 opioid receptor activation for synthetic opioids (Van Der Graaf et al., 1997). Type 4 biomarkers refer to physiological measures in the integral biological system. Disease biomarkers serving as functional endpoints of disease progression (such as tumor size in cancer) at a physiological level constitute the type 5 biomarkers. Biomarkers need not be directly related to the clinical outcome. For example, tumor size reduction need not necessarily correlate to how well the patient feels. Those that are predictive of clinical outcome are considered as surrogate biomarkers. However, even generally accepted surrogate endpoints are unlikely to capture all the therapeutic benefits and the potential adverse effects that a drug will have in a diverse patient population (Lesko and AJ Atkinson, 2001). Accordingly, combinations of biomarkers will probably be needed to provide a more complete characterization of the spectrum of pharmacological response. A clinical endpoint (type 6 biomarker) is a characteristic or variable that reflects how a patient feels, functions, or survives and therefore, the ultimate measure of efficacy that quantifies the direct benefit to a patient. However, the long periods of time needed to achieve it make it an impractical measure during the short‐term clinical trials. Table 1.6 distinguishes Type 5 biomarkers from surrogate markers and clinical endpoints. Clinical endpoints include binary outcomes like cardiovascular events (presence or absence of stroke, myocardial infarction), death, or no death etc. Depending on the type of drug, some biomarkers are more relevant and readily available (Peck et al., 2003) than others. Biomarkers enable the demonstration of target engagement, modulation of pathophysiology or disease process and clinical efficacy needed for proof of mechanism, proof of principle, and proof of concept respectively, either in preclinical or clinical phases of drug development (see Figure 1.15). A quantitative relationship of target‐relevant drug concentration to PD biomarkers and to efficacy (surrogate marker/clinical endpoint) or long‐term safety, when supported by biology (pathway, disease) or by statistical evidence (if pathway is unknown) can guide dose‐finding and aid translation of nonclinical findings to humans (see Figure 1.15).
Data from genomics and proteomics are helping scientists to differentiate between healthy and disease states, leading to biomarker discovery (Colburn and Keefe, 2003). Most biomarkers are endogenous macromolecules which are measured in biological fluids such as whole blood plasma, serum, urine, saliva buccal mucosa samples, sweat or tissues, tumor etc. A biomarker must be measurable in preclinical models and patients (translatable). An analytical validation of a biomarker assay is needed to establish its performance characteristics (good precision, resolution, dynamic range, sensitivity) robustness and reproducibility (Colburn and Lee, 2003) (Figure 1.16). To improve the signal‐to‐noise ratio, biomarker levels measured under treatment should be normalized against baseline data either by subtraction or division. An ideal biomarker is one that permits dense, longitudinal and easy sampling (low volumes of samples collected at frequent timepoints) and non‐labile sample handing. It should ideally change rapidly in a short time so that it can be implemented in Phase I/II. It should be possible to sample, measure, and quantify the biomarker both at baseline and on‐treatment. These requirements are summarized in Figure 1.17a. PD biomarkers may be proximal or distal proteins (preferably in target tissue) in the targeted pathway that shows modulation upon treatment (Figure 1.17b). They allow ‘Go’/‘No Go’ development decisions based on evidence of achieving target engagement or desired biological activity. They can support dose selection for expansion cohorts as well as for pivotal trials and dose scheduling based on PD half‐life.
TABLE 1.6. Examples of different types of biomarkers.
| TYPE 5 biomarker | Surrogate marker | Clinical endpoint | |
|---|---|---|---|
| Definition | A characteristic that is objectively measured and evaluated as an indicator of normal biologic process or pharmacologic response | A biomarker intended to substitute for a clinical endpoint and expected to predict the effect of a therapy. Selection of surrogate is based on epidemiologic, therapeutic, pathophysiologic or other scientific evidence for predicting clinical endpoint | A characteristic or variable that reflects how a patient feels or functions or how long a patient survives |
| Value | Mainly in early efficacy and safety evaluation in in vitro studies, in vivo animal models, and early clinical trials to establish proof of concept. Need not be directly related to clinical outcome | Biomarkers that are readily observed and easily quantified. Predicts clinical outcome. Clinical relevance of the surrogate is generally well validated | Assess benefit (cure or reduced morbidity) of therapeutic intervention to the patient. Ultimate measure of efficacy, but difficult to quantify. |
| Examples | Blood cholesterol concentrations for assessing risk of heart disease. Receptor occupancy Extent of target modulation | Pupil dilation for narcotics, biochemical tumor markers for anticancer drugs, exercise tolerance tests in chronic stable angina, for myocardial infarction. QT interval as a surrogate for Torsades de Pointes. HbA1c for diabetes. Blood pressure, body weight for obesity; Viral load of HIV, hepatitis C or B virus for assessing level of infection. CD4 cell counts. | Chest pain for a medication aimed at prevention of heart attack Overall survival for cancer indications Recurrence of cancer, stroke Occurrence of infections in HIV |
| Biochemical and clinical biomarkers | |||
| Disease | Biochemical biomarker | Clinical biomarker | |
| Asthma and chronic obstructive pulmonary disease | Leukotrienes, chemokines, and cytokines | Pulmonary function tests, exacerbations | |