Abstract
Many enzymes are therapeutic targets for drug discovery, whereas other enzymes are important for understanding drug metabolism and pharmacokinetics during compound testing in animals. Testing of drug efficacy and metabolism in an animal model requires the measurement of disease endpoints as well as assays of enzyme activity in specific tissues at selected time points during treatment. This requires the removal of tissue and biochemical assays. Techniques to noninvasively assess drug effects on enzyme activity using imaging technology would facilitate understanding of drug efficacy, pharmacokinetics, and drug metabolism. Using a commercially available cytochrome P−450 3A substrate whose oxidized product is a luciferase substrate, we show for the first time that cytochrome P−450 enzyme activity can be measured in vivo in real time by bioluminescent imaging. This imaging approach could be applicable to study drug effects on therapeutic target enzymes, as well as drug metabolism enzymes.
Introduction
In vitro cytochrome P−450 enzyme (CYP) assays have been well established during the past few decades (reviewed by Ref. [1]). More recently, a bioluminescent approach has been developed for determination of human CYP isoenzymes and caspase activity in vitro using luciferin derivatives as substrates [2]. Modifications of luciferin, the luciferase substrate, were made at the 6′-OH position of the molecule. In vitro, luciferin 6′ benzyl ether (LBE) can be specifically broken down by CYP3A4 or CYP3A7 into a benzyl metabolite and luciferin, and luciferin can then serve as a substrate for luciferase. Inhibition of the CYP3A4 enzyme by a specific inhibitor ketoconazole (KETO) was determined with IC50 = 0.06 μM using this method. At μM to mM levels, KETO inhibited the CYP3A4 activity in vitro about 88% [2]. The light production can report on the enzymatic activity of CYP3A4 and CYP3A7. In our proof of concept study, we attempted to measure the murine Cyp3a11 and Cyp3a13 (orthologs of human CYP3A4 and CYP3A7) enzyme activities using the modified luciferin LBE in vivo in real time. To determine the luciferin produced in vivo, mice must express the luciferase enzyme in liver where the Cyp3a11 and Cyp3a13 genes are primarily expressed [12]. Cyp1a2-luc transgenic mice [Crl:CD-1(ICR)BR-Tg(Cyp1a2-luc)Xen] constitutively express a high level of luciferase in the liver [13] and typical inducers of Cyp3a such as dexamethasone (DEX) and the CYP3A enzyme inhibitor ketoconazole (KETO) have no effect on the Cyp1a2-luc expression (data not shown). Therefore, the Cyp1a2-luc mouse can serve as a liver-specific luciferase expressing testing vehicle for assessing the potential use of the LBE substrate for assessing Cyp3a11 and Cyp3a13 enzyme activity in vivo.
In these experiments, mice were treated with either vehicle (DMSO) or DEX. Six hours after mice were treated with DMSO or the inducer DEX, the enzyme inhibitor KETO was given to both DMSO- and DEX-treated animals and the CYP3A4 enzyme substrate LBE was injected 4 min later. Animals were imaged immediately after injection of LBE using quantitative whole-body imaging technology [4]. DEX induced CYP3A enzyme activities approximately 2.4-fold (p < .1) above DMSO-treated animals at the 7-min time point following LBE injection (Figure 1B and C). KETO at 100 mg/kg (approximately 0.19 mM plasma concentration) strongly inhibited the DEX-induced CYP3A enzyme activities by 89% (p < .05) and inhibited baseline enzyme activity in DMSO-treated animals by 72% (p < .05) at the 10-min time point (Figure 1B and C) It has been reported that administration of KETO at a dose 200 mg/day (~0.003mM) resulted in a 49% decrease in clearance of docetaxel clearance by the CYP3A4 in human [7]. In vitro assay showed that KETO inhibited about 88% of CYP3A4 activity [2]. Our in vivo animal data correlate well with these data. It has also been demonstrated that DEX induced CYP3A proteins about 2.9-fold as determined by Western analysis [6], consistent with our bioluminescent data. We noted that a significant percentage of the animals (5 out of 8 at 80 mg/kg LBE; 2 out of 8 at 40 mg/kg LBE) died approximately 15 min after LBE injection. Animals that received the solvent for LBE (100 mM K2HPO4, pH 7.4) were unaffected, suggesting either LBE or its benzyl metabolite was lethal.
Visualization of the murine CYP3A activities in vivo. (A) Concept of directly imaging the drug effect on its target enzyme. Inactive modified luciferin specific for a target enzyme is injected into a universally expressed luciferase transgenic animal. The enzyme acts on the luciferin derivative and releases the active luciferin. The luciferase then acts on the released luciferin and produces light in the tissues in which the target enzymes are active. The light production is related to the target enzymatic activities. (B) Male Cyp1a2-luc mice [13] at ages of 8–12 weeks old were treated intraperitoneally with either DMSO or DEX at 100 mg/kg dissolved in DMSO. Six hours after injection, each group of mice were separated into two groups and injected intraperitoneally with either DMSO or KETO at 100 mg/kg dissolved in DMSO. Five minutes following this injection, the mice were injected with LBE at 40–80 mg/kg dissolved in 100 mM K2HPO4, pH 7.4 and imaged for 3 min at each time point following LBE injection using the IVIS Imaging System 100 (Xenogen, Alameda, CA). Images shown are from groups that received LBE at the dose of 40 mg/kg. (C) Photons per second emitted from the liver region were quantified using LivingImage software (Xenogen) and plotted as mean ±standard error (n =3–4). * and ** indicate significant differences from DMSO control value for each time point at p <.1 and p <.05, respectively. +and ++indicate significant differences between DEX and a combination of DEX and KETO treatments at p < .1 and p < .05, respectively.
Our results show for the first time that drug metabolism enzyme activity can be measured in vivo in real time using bioluminescent imaging as the conversion of LBE to an active luciferase substrate by the CYP3A enzyme. Modulation of CYP3A enzyme activities by a gene inducer and an enzyme inhibitor was as predicted. Although the LBE substrate proved lethal to the test animals at the doses used, this toxicity, although problematic, does not alter the conclusion that bioluminescent detection can be used to assay enzymatic activity in vivo noninvasively by bioluminescent imaging. Alternative chemical modifications of the LBE substrate could reduce toxicity of this molecule or the benzyl metabolite. It is likely that the benzyl group is toxic to animals. Modification of the luciferin at 6′-OH with a nontoxic molecule such as 5- or 6-carbon sugars may reduce the toxicity while the specificity to CYP3A remains. This concept of using cleavable luciferins could be applicable to therapeutic target enzymes and make it possible to assay drug efficacy in vivo in real time.
Enzymes that are involved in the disease process are often therapeutic targets. Examples include protein kinases and thymidine phosphorylase for cancers, dipeptidyl peptidase IV for Type 2 diabetes, and betasecretase for Alzheimer's disease, and so forth (reviewed by Refs. [3,5,10,11]). Lead drug candidates identified by in vitro screening and optimization are further evaluated in animal models for efficacy, specificity, pharmacology, and toxicology. Determination of in vivo efficacy and specificity in animal disease models is usually a difficult and complicated process. Direct detection of the inhibitory effects of a compound on its target enzyme in vivo could facilitate evaluation of its efficacy, specificity, and pharmacodynamics.
The imaging of reporter mice is a new approach to look at drug actions at the level of gene expression and hormone receptor activity [8]. Determination of drug actions on their specific target enzymes in vivo is important because most drugs function at the level of protein activity. The results described here outline a new general approach to assess drug metabolism in animal models (Figure 1A). Modified luciferins specific for CYP isoenzymes and caspases are now commercially available. In addition, luciferin derivatives have been developed for carboxylic esterase, arylsulfatase, alkaline phosphatase, and carboxypeptidases A, B, and N [9]. As enzymatic cleavage of these compounds leading to the release of active luciferin has been demonstrated, these could also serve as in vivo substrates. In principle, this approach could be used for any enzyme targets if inactive luciferin derivatives can be catalyzed into active luciferin by these enzymes. Solubility and toxicity of these luciferin derivatives should be considered, optimized, and tested. A luciferase transgenic animal that expresses high levels of luciferase universally will be ideal for this application so that enzymatic activity could be detected in any tissues. Alternatively, tissue-specific luciferase transgenic animals can be used to determine drug effects in specific tissues.