The current state of serum biomarkers of hepatotoxicity

Abstract

The level of serum alanine aminotransferase (ALT) activity reflects damage to hepatocytes and is considered to be a highly sensitive and fairly specific preclinical and clinical biomarker of hepatotoxicity. However, an increase in serum ALT activity level has also been associated with other organ toxicities, thus, indicating that the enzyme has specificity beyond liver in the absence of correlative histomorphologic alteration in liver. Thus, unidentified non-hepatic sources of serum ALT activity may inadvertently influence the decision of whether to continue development of a novel pharmaceutical compound. To assess the risk of false positives due to extraneous sources of serum ALT activity, additional biomarkers are sought with improved specificity for liver function compared to serum ALT activity alone. Current published biomarker candidates are reviewed herein and compared with ALT performance in preclinical and on occasion, clinical studies. An examination of the current state of hepatotoxic biomarkers indicates that serum F protein, arginase I, and glutathione-S-transferase alpha (GSTα) levels, all measured by ELISA, may show utility, however, antibody availability and high cost per run may present limitations to widespread applicability in preclinical safety studies. In contrast, the enzymatic markers sorbitol dehydrogenase, glutamate dehydrogenase, paraxonase, malate dehydrogenase, and purine nucleoside phosphorylase are all readily measured by photometric methods and use reagents that work across preclinical species and humans and are commercially available. The published literature suggests that these markers, once examined collectively in a large qualification study, could provide additional information relative to serum ALT and aspartate aminotransferase (AST) values. Since these biomarkers are found in the serum/plasma of treated humans and rats, they have potential to be utilized as bridging markers to monitor acute drug-induced liver injury in early clinical trials.

Introduction

New drug products require rigorous preclinical and clinical testing and ultimately FDA and EMEA approval prior to being marketed. Many preclinical candidate compounds (PCCs) do not achieve ultimate regulatory approval because of induced organ toxicity. Of the PCCs that are discontinued for organ toxicity, up to half are due to hepatotoxic effects including necrosis, steatosis, cholestasis, proliferation, inflammation, and bile duct hyperplasia. Likewise, the most frequently cited reason for withdrawal of an approved drug is toxicity and often hepatotoxicity is the source of concern (Xu et al., 2004). Assessment of the potential for new therapeutics to cause liver injury in humans relies heavily on thorough evaluation of histomorphologic and clinical pathologic endpoints of hepatotoxicity in preclinical species. However, estimates indicate that current preclinical testing regimes successfully correlate to adverse events in human liver in only about 50% of the cases where clinical trials are performed (Olson et al., 2000). Biomarkers that are diagnostic of liver histomorphologic change in preclinical species that might bridge to humans are being evaluated, developed, and improved largely by industrial efforts, which is the subject of this review. Thus, novel biomarkers of liver injury are sought to lower the incidence of false negative results, thereby leading to more accurate prediction of drug-induced liver injury in preclinical and ultimately clinical studies. False positive preclinical indicators of hepatotoxicity are also of concern in pharmaceutical development since they could also impact potential development. Generally, false positive signals show reduced impact upon safety compared to false negative signals.

Ideal attributes of biomarkers of hepatic damage include organ specificity for liver, strong correlation with well-defined hepatic histomorphologic changes, out performance or added information to serum alanine aminotransferase (ALT) and/or aspartate aminotransferase (AST) values, adaptation of screening assays to high throughput modalities that are commercially available, sample accessibility by uninvasive procedures such as blood collection, and in the context of clinical translation, application across key preclinical species to humans, including mouse, rat, dog, and monkey. Once candidate markers are identified, rigorous technical validation or cross-validation and biological qualification is required for wide acceptance by scientific, medical and regulatory authorities. While no single biomarker has yet fully achieved this rigorous level of evaluation, these attributes are a common goal of the pharmaceutical community and regulatory agencies for all safety biomarkers in development. Recommendations for selection of established hepatotoxicity markers appropriate for preclinical screening have been previously discussed (Boone et al., 2005, Ramaiah, 2007). Preclinical assays currently include some combination of serum ALT and AST activities, alkaline phosphatase (ALP) activity, total bilirubin, gamma-glutamyl transferase (GGT) activity, bile acids and/or sorbitol dehydrogenase (SDH) activity (Table 1). Among these assays, some are more specific and/or sensitive than others for liver toxicity. For example, serum AST activity is associated with liver toxicity, but also can be elevated in association with heart and skeletal muscle injury (Burhop et al., 2004, Nathwani et al., 2005). Serum ALT activity is also associated with muscle necrosis. This review will cover the current state of hepatotoxicity biomarkers and their utility in the preclinical and clinical settings. Development of novel candidate hepatotoxicity markers will also be discussed as extensively as possible given current gaps in the published literature (Table 1).