ADMETox profiling has assumed centerstage in the drug discovery process since the late 1990s. ADMETox studies map the absorption, distribution, metabolism, excretion, and toxicity readouts of drugs or drug compounds in the context of the human body.
In spite of the increased spending in R&D, many promising drugs failed, and new drug approvals steadily were declining. “Most of these drugs failed in late-stage clinical trials and had unfavorable ADME properties. So ADME profiling became critical in an effort to identify issues early on and increase the drug productivity rate,“ says Philip S. Burton, Ph.D., CEO, CSO of ADMETRx (www.admetrx.com). The firm was one of the presenting companies at IBCs (www.ibcusa.com) recent preclinical development forum at Boston.
“An analysis of clinical failure rates in 2000 showed that failure due to ADME/PK (pharmacokinetics) fell to 10% of the overall rate due to improved ADME profiling. However, the problem was that even with improved ADME, compounds were continuing to fail due to Tox and efficacy issues. So Tox has now assumed greater importance, in particular idiosyncratic toxicities that crop up during Phase III trials.“
Many drug discovery companies are using the fail fast/fail cheap model in high-throughput screening (HTS) mode to weed out unpromising compounds early on in the discovery process. HTS, however, is typically not deployed in ADMETox areas, and only a small percentage of prescreened compounds are subject to ADME profiling. “The fail fast/fail cheap approach can eliminate promising candidates as well due to the selection criteria applied in this process. Companies are missing out on these opportunities as early failures are rarely if ever revisited,“ says Dr. Burton.
In silico Modeling
The in silico approach has been gaining traction in ADMETox research for the past five years. In silico tools enable researchers to do extensive modeling and virtual screening to generate a virtual ADMETox profile for thousands of compounds, eliminating poor candidates upfront. Actual assays are then performed from the short-listed pool of compounds, increasing the success rate of finding a good candidate with desirable ADMETox characteristics.
Bio-Rad (www.biorad.com) recently launched The KnowItAll Informatics System, Metabolomics Edition to aid in silico metabolomics research, which specifically focuses on the M and T of ADMETox.
The initial focus of the KnowItAll platform was analytical chemistry integrated with cheminformatics. At the end of 2003, Bio-Rad added in silico ADMETox prediction tools to this platform. “The new KnowItAll Metabolomics Informatics edition is empowered by the integration of Infometrix Pirouette chemometrics technology for multivariate analysis of NMR, spectrometry, and chromatography data,“ explains Gregory M. Banik, Ph.D., general manager of the informatics division.
The in silico modeling approach, using KnowItAll, needs to be initially validated against compounds with known experimental property values. Then, the models can be used for compounds for which experimental property values have not been measured to make predictions of the ADMETox characteristics or even build a virtual library for screening.
“Our platform offers several global models that can be used alone or together in a consensus fashion, which can decrease the error and increase the accuracy of prediction. It uses data mining and multivariate analysis tools that evaluate multiple endpoints simultaneously to rank order compounds, based on a composite score,“ states Dr. Banik. Bio-Rad plans to further enhance the ability to build and integrate local models in the KnowItAll system by integrating CSIROs (www.csiro.com) MolSAR molecular modeling technology.
“ADMETox in vitro testing has changed from an HTS approach to a more focused approach,“ comments Adrian Sheldon, Ph.D., associate director of the in vitro ADMET group at Charles River Laboratories (www.criver.com), which was another company discussing the topic at the IBC meeting. “However, the goals of in vitro ADME testing are still the same, that is rank order compounds for success and detect any compound issues.“
“We have several in vitro assays for determining drug-drug interactions,“ states Howard Haspel, Ph.D., study director of the In vitro ADMET group. “Under FDA guidelines, compounds require earlier testing for human specific metabolites. We have developed a broken cell method for rapid Phase I-Phase II metabolite profiling and species-specificity profiling that provides snapshot data in less than two weeks.
“Determining compound properties can be tricky and can result in unreliable data. We use integrated controls to assign a confidence score to potentially problematic compounds and support the results with other assay methods. ADMETox data generated by Charles River Laboratories have been used in over a dozen FDA IND filings by our clients,“ says Dr. Haspel.
ADMETRx offers in vitro ADME and physicochemical properties profiling along with consulting and integrative modeling approaches to pick the best bet compounds. It thus assists companies to make better compound fate decisions, based on an expert domain knowledge approach and a multicriteria decision-based strategy.
“The ADMETRx approach identifies and fills data gaps and integrates all this information to select optimal candidates for advancement,“ says Jay T. Goodwin, Ph.D., president & COO. ADMETRx addresses many of the challenges faced in drug discovery, such as achieving sufficient oral bioavailability, assessing CNS penetration, and evaluating potential for drug-drug interactions, while balancing these determinants with intrinsic activity, he maintains.
Peakdales (www.peakdale.co.uk) core expertise is innovative heterocyclic and medicinal chemistry. They use this approach coupled with computer-aided drug design to construct compounds with a superior starting point for drug discovery testing. Peakdale also offers chemogenetic arrays and collection of compounds designed to target receptors, families, enzymes, GPCRs, kinase library, and protease assays for enabling researchers in the ADMETox arena.
“Peakdale adopts a traffic light system—red, amber, or green representing outright issues, possible problems, or no problems respectively—to make a go-or-no-go decision for hit or lead compounds,“ says Terry Hart, Ph.D., director of medicinal chemistry services at Peakdale Molecular.
“Outright issues could be irreversible protein binding, mutagenic toxicity, herg, or apoptosis, while addressable problems could typically be due to bioavailability, Cyp inhibition, selectivity, or minor solubility issues. Hence, it is crucial to perform ADMETox profiling early in drug discovery so one is cognizant of what is ahead if a compound gets the go-ahead with possible problems.“
Caliper Life Sciences (www.caliperls.com) offers assay services and informatics tools through Novascreen Biosciences (www.novascreen.com). Novascreen recently announced the availability of General Side Effect II panel (GEN SEP II), a specialized side effect screening service that can be used with its Side Effect Database (SED).
The GEN SEP II assay panel spans a broad therapeutic range and encompasses a wide variety of molecular targets, including GPCRs, transporters, enzymes, nuclear receptors, ligand, and voltage-gated ion channels, to provide valuable information on a compounds efficacy and safety. The NovaScreen SED houses assay profiles through the same assay panel as the GEN SEP II on over 2,000 known biologically active compounds, like marketed/ failed/withdrawn drugs, reference agents, and natural products.
“The GEN SEP II and SED enable companies to map their compound in the context of the database to assist in identifying and eliminating side effect and safety concerns,“ says Seth Cohen, Ph.D., director of application sciences. “A key challenge is how to integrate and interpret data in a broader setting. This is where predictive ADMETox and in silico approaches are starting to live up to their potential.“
Automation in ADMETox
Automation is beginning to play a bigger role in ADMETox assays. The emphasis is on how to use automation beyond screening assays to lead optimization assays. Beckman Coulter (www.beckmancoulter.com) is trying to enable this shift with their Biomek FX ADMETox Workstation that was introduced about a year ago to meet the demand for PAMPA (parallel artificial membrane permeability analysis) automation.
The workstation is validated for automated in vitro drug permeability measurements using pIONs (www.pion-inc.com) PAMPA method to assess the passive absorption of a drug candidate through an artificially created lipid membrane.
The Biomek software on the workstation allows integration and scheduling of methods across different components. The pION PAMPA Evolution 96 System predicts the absorption of gastrointestinal-tract candidate drug molecules at different pH conditions. Data is assessed as a spectrometric readout and analyzed using pION software. PAMPA can also be used to model absorption across the blood-brain barrier using a different lipid formulation, also available from pION.
“A typical run time from set up to final result is less than two hours for a 96-well plate in a walk-away mode. We are currently developing a multiplate assay format,“ says David Daniels, Ph.D., applications marketing manager at Beckman.
“In addition, we now offer a drug solubility assay along with PAMPA on the same deck and workstation. The combined assay offering covers the need for the A in drug screening. The company is developing assays for the M and T of ADMETox, partnering with Promega to provide automated methods for their CYP450-GLO and the CytoTox assays, respectively.“
Caliper also provides equipment for conducting ADMETox assays in an automated format. “We provide labware tools validated with widely used vendor consumables and work with kit vendors to provide precreated methods templates for several in vitro ADMETox assays in an automated high-throughput format,“ says Dr. Cohen.
Aurora Discovery (www.auroradiscovery.com) offers ADMETox researchers assay miniaturization capability in a high-throughput format with their low-volume, 1,536-well and 3,456-well ChemLib cyclo-olefin microplates. These plates allow companies to do basic research and screen a larger percentage of their compounds in ADME assays. The ChemLib plate is supported by Auroras microfluidic workstations that enable low-volume dispensing.
“The tools offered by Aurora put ADMETox assays on a fast track,“ says Peter Coassin, CTO, who talked about his companys efforts in this field at IBCs Screentech Conference. “Data for hundreds of compounds in several assays can be generated and reported back to the chemistry groups within a couple of days. The assay miniaturization technology and plate tool were developed along with Vertex Pharmaceuticals (www.vpharm.com).
“Vertex uses high-density plates extensively for their CYP450 inhibition assays. These assays raise flags early on potential drug-drug interaction liabilities. HT-CYP450 inhibition data can be used as predictors of clinically significant outcomes,“ explains Coassin.
“Vertex is one of a few companies that perform ADMETox profiling for a majority of compounds synthesized in discovery,“ says Gregor Zlokarnik, Ph.D., research fellow. “We typically run around 200 compounds a week through multiple CYP450 assays in high-density plates. We profile compounds over several logs in concentrations and generate IC50 curves for multiple enzyme isoforms. Additional benefits of using these plates are usage of less material and the ability to generate more data points in the same run.“