Addressing ADME Bottlenecks

New technologies speed ADME screening

By Vaughn P. Miller

BIOCIUS Life Sciences

The drug discovery community recognizes that evaluation and optimization of ADME (absorption, distribution, metabolism and excretion) profiles early in the drug discovery process can significantly contribute to reducing the attrition rate of compounds later in development due to poor pharmacokinetic (PK) properties1,2. In the past decade, the number of compounds available for ADME screening has greatly increased due to advances in combinatorial chemistry and high throughput screening. The large number of compounds generated today within medicinal chemistry project teams makes in vivo animal ADME or PK profiling nearly impossible to accomplish in a timely and cost effective manner. Therefore, in vitro ADME assays have been implemented early in the drug development process to act as surrogates for in vivo testing. These assays allow for optimization of candidate structure-activity relationships in parallel with the optimization of structure-ADME properties. Candidate compounds are rank ordered according to potency and ADME properties; those with inferior rankings are further optimized or dropped to minimize the probability of failure in later stages of drug development.

The most commonly performed in vitro ADME assays in early drug discovery are: metabolic stability, permeability (using cell lines or a parallel artificial membrane permeability assay, or PAMPA), CYP inhibition, plasma protein binding and assessment of physical chemical properties (i.e. solubility). While high throughput potency assays generally utilize reporter-based plate reader technologies that have been optimized for speed and efficiency, ADME assays require the individual characterization and quantity measurement of each drug candidate in each assay. Therefore, mass spectrometry — particularly liquid chromatography/tandem mass spectrometry (LC-MS/MS) — is the analysis method of choice for ADME assays because of its high sensitivity and selectivity in measuring a broad range of drug candidates.

However, LC-MS/MS-based ADME laboratories are often challenged to meet the expectation of timely ADME data delivery to project teams that require this information for the fast, iterative process of hit-to-lead and lead optimization. Usually in vitro assays are performed in parallel using highly automated 96- or 384-well plate robotic systems capable of unattended operation and the generation of an enormous number of samples. For example, a typical metabolic stability assay for a set containing 48 compounds would generate 768 samples, all of which require analysis. Using a standard LC-MS/MS cycle time of 2.5 minutes per sample, the total analysis time for this compound set would exceed 32 hours3. In addition, several hours of instrument time (two to four hours) needs to be spent to optimize MS/MS transitions and instrument parameters (MRM) for each test compound before sample analysis can start. Consequently, a single LC-MS/MS system has the capacity to analyze relatively few compounds. If analysis efficiencies are not improved, more mass spectrometry instrumentation will be required to keep pace with the growing number of in vitro ADME samples. It is not uncommon for a centralized bioanalytical group of a pharmaceutical company to analyze hundreds of thousands of in vitro ADME samples on a yearly basis4.

As illustrated in the metabolic stability example above, the two most significant bottlenecks hindering efficient ADME analysis by LC-MS/MS are MS/MS method development and sample analysis cycle times. Recently, new mass spectrometry-based technologies have been introduced to address these bottlenecks. These technologies can be organized into two distinct groups:

1)improvements to conventional LC-MS/MS and

2)novel technologies that incorporate new ionization sources, sample injection methods or mass analyzers.

Improvements to conventional LC-MS/MS have primarily focused in two areas: automation of MS/MS method development and shortening the LC time per sample by using ultra-high pressure liquid chromatography (UHPLC) or multiplexing of LC columns. A triple-quadrupole mass spectrometer used in multiple reaction monitoring (MRM) mode requires that MS/MS method parameters such as polarity, source conditions, parent to fragment ion transition and collision energy are optimized for each compound in order to reach acceptable sensitivity for assay analysis. Until recently, this optimization was performed manually using infusions of test compounds in a process that takes five to ten minutes per compound. This is a significant bottleneck for groups that need to analyze scores — if not hundreds — of discrete compounds in a short period of time. Now most ADME analysis groups utilize automated solutions from the major mass spectrometer vendors that optimize the MS/MS methods through the use of autosamplers and software, and then incorporate them into a centralized database without the need for human intervention. For example, “infusion quality” MS/MS methods are available from a number of MS instrument vendors, including ThermoFisher, AB Sciex and Agilent. While these technologies have greatly streamlined the workflow of MS/MS method development, several minutes per sample is still required to perform these automated processes.

Technologies that shorten the LC time required per sample prior to introduction into the mass spectrometer have made significant improvements in ADME analysis cycle times. UHPLC systems that incorporate sub-2µm LC columns and high flow rates have been shown to achieve faster analysis times while retaining chromatographic resolution. In 2004, Waters launched the first UHPLC system and since then most of the major LC vendors have developed similar systems. Several papers have shown ADME profiling assays using these systems with sample cycle times ranging from 1.0 to 1.5 minutes per injection4.

Multiplexing LC-MS/MS using a “staggered parallel” design is now used in many ADME laboratories. The efficiency of LC-MS/MS analysis is improved by connecting multiple LC systems to a single mass spectrometer in such a way so that only the chromatographic area that contains the analyte(s) of interest enters the mass spectrometer for analysis. Home built or commercial systems have been shown to reduce the effective sample analysis time for ADME assays on the mass spectrometer to approximately one minute4,5.

Improvements to conventional LC-MS/MS have significantly reduced the two major bottlenecks associated with ADME assay analysis. However, it is clear that innovative new technologies are necessary to produce a severalfold further reduction in analysis preparation and run times or elimination of the bottlenecks altogether. In the past few years, new technologies have emerged which eliminate chromatography in ADME assay analysis, thereby greatly increasing the speed of sample analysis.

Direct analysis in real time (DART) is an open-air ionization technique that has been developed for the direct analysis of samples without any sample preparation or chromatography. Commercial DART systems are available. While many intriguing applications have been developed for the analysis of surfaces using DART, the use of this technology for the analysis of ADME assays is just starting to be explored. Yu et al. reported that DART analysis of in vitro metabolic stability assays could be accomplished in approximately 50 sec/sample6.

Another novel ionization-based technology that does not utilize chromatography is laser diode thermal desorption (LDTD) interfaced with atmospheric pressure chemical ionization (APCI) and a triple-quadrupole mass spectrometer. Dried samples on a special metal alloy plate are analyzed by the system very quickly. Sample analysis speeds of approximately 19 seconds/sample for CYP inhibition assays have been reported7. While LDTD significantly decreases analysis times, it requires considerable off-line sample preparation which decreases its impact on laboratory efficiency.

Using well known solid phase extraction (SPE) technology to replace LC prior to mass spectrometry analysis has been shown to be an effective and efficient methodology for ADME assay analysis by many laboratories4. Recently an ultra-fast online SPE-MS/MS system has been described which produces sample analysis cycle times of approximately seven seconds and generates results comparable to LC-MS/MS analysis for many in vitro ADME assays4,8.

In the past few years, the ADME community has begun to explore the capabilities of the newly commercialized high resolution mass spectrometers (i.e. time-of-flight or TOF), which have the quantitative performance of most triple quadrupole instruments. These instruments are capable of scanning a full spectrum of masses (for example: 100 to 800 Da) within the timeframe of a typical analysis run, rather than just the few masses monitored by a triple quad, thereby allowing the analysis of a broad spectrum of analytes using generic MS conditions. A TOF instrument has the ability to eliminate the bottleneck of MS/MS method development through the use of universal or generic MS conditions. The potential to use this type of mass analyzer for simultaneous parent quantitation and metabolite identification is very exciting and has been demonstrated9. Very recently, an SPE-TOF system was shown to provide LC-MS/MS comparable data for metabolic stability and PAMPA assays with analysis times of seven sec/sample using generic MS conditions10.

The challenge for those attempting to efficiently analyze large numbers of unique analytes in discovery ADME assays has spurred the development of many new technologies. There are readily available technologies that can very quickly analyze ADME assays (>10x faster than LC-MS/MS) by greatly reducing — in some cases eliminating — the bottlenecks of long analysis times and MS/MS method development. These technologies will likely continue to evolve, be carefully evaluated, and find utility within the general bioanalytical community for applications far outside the realm of the typical drug discovery ADME group.

References

1.Hop CE, Cole MJ, Davidison RE, et. al. Curr. Drug Metab. 2008; 9 (9):847-53.

2.Kola I and Landis J, Nat. Rev. Drug Discov. 2004; 3, 711.

3.Xu, R, Manuel, M, Cramlett J, and Kassel DB, J. Chromatogr. A, 2010; 1217:1616-1625.

4.Shou, WZ and Zhang J, Expert Opin. Drug Metab. Toxicol. 2010; 6(3):321-336.

5.Briem S, Pettersson B, Skoglund E, Anal. Chem. 2005, 77 (6):1905-10.

6.Yu S, Crawford E, Tice J, Musselman B, Wu JT. Anal. Chem. 2009; 81:193-202. (DART)

7.Wu J, Hughes CS, Picard P, Letarte S, Gaudreault M, Levesque J-F, Nicoll-Griffith DA, Bateman KP, 2007; Anal. Chem. 79 (12):4657-4665.

8.Brown A, Bickford S, Hatsis P, Amin J, Bell L, Harriman S, Rapid Commun Mass Spectrom. 2010 24(8):1207-10.

9.O’Connor D and Mortishire-Smith R, Anal. Bioanal. Chem. 2006; 385:114-121.

10.Hatsis P, Romm M, Miller V, Amin J, LaMarr W, Ozbal C, and Harriman H, Proc. 58th ASMS Conf. Mass Spectrometry and Allied Topics, Salt Lake City, UT, 2010.

Vaughn P. Miller, Ph.D. is director ADME at BIOCIUS Life Sciences, a provider of automated sample processing technology. He can be reached at vmiller@biocius.com