Biomarkers in Drug Discovery and Clinical Trials
Kirti Singh, Postdoctoral Scientist, Eli Lilly and Company
Novel robust biomarkers identification during drug development posess immense potential for treating incurable diseases. The increased popularity of biomarkers in pharmaceutical discovery is attributed to a rapidly evolving utilisation of biomarkers from early to late-phases drug discovery. This overview highlights biomarker categories based on clinical drug development, in addition to addressing the challenges in biomarker research.
Biomarkers are biological markers in the body that can be utilised to predict the presence or absence of abnormal biological processes. Biomarkers play a crucial role in disease prediction, detection, prognosis, and monitoring. Furthermore, biomarkers are also utilised to assess and monitor the response to a pharmacological intervention. Human body is composed of a plethora of biomolecules which can serve as biomarkers, including but not limited to DNA, mRNA, carbohydrates, lipids and proteins. Generally, protein and gene-based biomarkers are commonly utilised for diagnostic purposes. These biomarkers can be detected in body fluids, such as blood, urine, saliva, sweat, milk, cerebrospinal fluid or tissues. For example, blood glucose or haemoglobin A1c (HbA1c) is commonly used as a diagnostic biomarker for diabetes. Apart from complex biomolecules detected in body fluids and tissues, functional measurements to gauge physiological responses can also be categorised as biomarkers. For instance, heart rate and blood pressure are non-invasive functional biomarkers used to monitor cardiovascular diseases.
Any biomarker must possess few characteristics to be reliably used for monitoring and detection. Firstly, the biomarker must be safe and easy to measure. It should be reproducible, reliable, sensitive and specific to accurately reflect the outcome. In addition, for research purposes, it should be cost-effective and should be in the quantifiable detection range by current tools and technologies for diagnosis and monitoring of a particular disease. Biomarkers have been broadly classified by the US Food and Drug Administration (FDA) into seven different categories based on clinical application.
Risk/ Susceptibility Biomarker
A risk biomarker can be used to predict the increased or decreased likelihood of disease development in future for an individual who yet does not have a medical condition. These could be genetic mutations that makes certain individuals more susceptible to developing a disease and can be detected years or decades before disease manifestation, so that preventative measures could be employed. For instance, mutation in breast cancer genes 1 and 2 (BRCA1/2) is used to identify individuals predisposed to develop breast cancer. Elevated low-density lipoproteins (LDL) blood cholesterol could increase the susceptibility to cardiovascular diseases.
Diagnostic Biomarkers
As the name suggests, diagnostic biomarkers are used for accurate disease diagnosis by confirming the presence of biomarker. For example, glomerular filtration rate (GFR) is used as a diagnostic biomarker for chronic kidney conditions. Elevated chlorine content in sweat is used to confirm the presence of cystic fibrosis. A reliable and disease selective diagnostic marker is crucial for not only accurate diagnosis but also for clinical assessment. Further, an ideal diagnostic biomarker candidate should be highly specific and sensitive, which means 100 per cent positive result in diseased population and 100 per cent negative result in non- diseased subset to minimise both false-positive (non-diseased individual wrongly diagnosed) and false-negative (diseased individual not diagnosed) outcomes. Diagnostic biomarkers are often used to determine the inclusion and exclusion eligibility criteria for a target population in clinical trials.
Monitoring Biomarkers
Biomarkers which are repeatedly measured over a period of time to monitor disease progression, such as alterations in disease severity, occurrence of other associated co-morbidities, worsening of the previously determined diseased state are defined as monitoring biomarkers. These biomarkers could also be utilised to gauge the response to a treatment regimen on the disease state. These biomarkers usually project a rate of change in disease characteristics over time. In clinical trials, it is extremely important to establish biomarker monitoring during a patient's clinical course to assess the effect of intervention as the pre-treatment baseline characteristics may vary from patient to patient. Continuous monitoring during the course of treatment is necessary to detect when and how early therapeutic effects are observed. On the other hand, monitoring biomarkers could help identify non-responders with aggressive rate of disease progression or to detect toxicity.
Prognostic Biomarkers
Prognostic biomarkers are utilised to indicate the increased or decreased probability of a clinical event, such as disease reoccurrence or progression in diseased population. For example, mutation in tumour suppressor gene, tumour protein 53 (p53) and chromosome 17p deletion are used as prognostic biomarkers to assess the likelihood of mortality in patients suffering from chronic lymphatic leukaemia. Another example, expression of prostate-specific antigen (PSA) and Gleason score are used to predict the likelihood of cancer progression in prostate cancer patients. The correlation of prognostic indicators with the likelihood of future event is highly context specific and may vary depending upon patient history, severity of disease, the strength of prognostic indicator along with other comorbid conditions etc. Hence, caution must be exercised before establishing any strong correlation.
Predictive Biomarkers
Predictive biomarkers help identify individuals who are more likely to respond, either favourably or unfavourably upon clinical intervention as compared to similar individuals without the biomarker. In clinical settings, predictive biomarkers are used either as eligibility criteria or to stratify the study population into biomarker positive and negative groups to identify the subset of diseased population for which the intervention is most effective. Furthermore, predictive biomarkers could also be patient characteristics, such as renal or hepatic function, cytochrome P450 mutation which makes a patient more likely to respond unfavourably due to toxicity.
Pharmacodynamic Biomarkers
Pharmacodynamic biomarkers serve as molecular indicators of biological response upon exposure to clinical intervention. The expected biological response could be beneficial or harmful. These biomarkers are critical for dose-selection, proof-of-concept studies, and early go/ no-go decisions in clinical trials. For instance, in patients with systemic lupus erythematosus, circulating B lymphocytes could serve as a pharmacodynamic biomarker to evaluate the effect of Belimumab, a monoclonal neutralising antibody against B-cell survival factor. Sweat chloride levels were utilised as clinical end point to determine the effect of Ivacaftor, a cystic fibrosis transmembrane conductance regulator (CFTR) channel modulator in cystic fibrosis patients.
Safety Biomarkers
Approximately, 30 per cent drugs fail in clinical trials due to toxicity-related adverse effects. Safety biomarkers are used to indicate the likelihood and extent of toxic effect due to drug administration and hence are crucial in clinical trials. Safety biomarker is a type of monitoring biomarker as repeated measurements are necessary to detect and manage potential toxicity by treatment interruption or dose modification. For example, in conjugation to serum creatinine levels, urinary biomarkers, such as total protein, albumin, Kidney injury molecule-1 (Kim-1), urinary clustarin, cystatin C and Trefoil factor 3 are used to monitor drug-induced nephrotoxicity. In addition, bilirubin and hepatic aminotransferase are commonly employed to assess hepatotoxicity.
Challenges in biomarker research and drug development
Over the past five decades, the applications of biomarkers in drug discovery and all phases of development have exponentially skyrocketed. The increased popularity of biomarkers in pharmaceutical discovery is attributed to a rapidly evolving utilisation of biomarkers from early to late-phases drug discovery. Over time the definition of biomarkers rapidly evolved and till date there is no standardised definition of biomarkers. Although, FDA has categorised biomarkers in 7 different categories, many other classification systems and sophisticated biomarker categories are emerging in literature, such as complex biomarkers, digital biomarkers etc. The explosive growth and integration of artificial intelligence (AI) and deep learning in various stages of drug development have accelerated drug-discovery efforts. The compound effect of AI-based algorithms sorting through huge human omics-datasets and the power of high-throughput screening (HTS) have expanded the horizons for biomarker discovery efforts. However, one of the major challenges is biomarker qualification and validation of analytical methods for biomarker quantification. Conventional immunoassay-based methodologies and histopathological assessments are used for biomarker discovery, however very low concentration biomarkers remain undetected with current tools and technologies. Further development and validation of analytical methods with lower limits of detects will not only accelerate drug-discovery efforts but constant innovation in the biomarker field will ultimately benefit patient and healthcare outcomes.
Author Bio
Kirti Singh is a postdoctoral scientist at Eli Lilly and Company, Indianapolis. She completed her Ph.D. in molecular and cellular pharmacology from Mercer University, Atlanta. Her doctoral research focused on studying the effects of reactive oxygen species on β2AR in airway obstructive disorder, particularly asthma. She has published her work on GPCRs in many high-impact factor, peer-reviewed journals.
