by J. Koers
Recent front-end developments in mass spectrometry have focused on the introduction of samples at a rate that approaches the speed of the MS analysis itself. Two such recent developments are multiplexed liquid chromatography (LC) and laser diode thermal desorption (LDTD), both viable options for increasing throughput in drug discovery experiments. This article describes these technologies and compares their performance in two drug discovery applications, namely CYP450 inhibition assays and metabolic stability assays.
Mass spectrometry (MS) techniques have been used in drug discovery for decades. Historically, the instruments produced more data than the software could process. In recent years, the speed and analytical capacity of software has improved to such an extent that the data processing bottle-neck has all but disappeared. This has prompted instrument manufacturers to shift their efforts to increasing the efficiency and throughput of their systems. This has resulted in the development of increasingly fast front–end technologies, starting with liquid chromatography, then high-pressure liquid chromatography, followed by ultra high-pressure liquid chromatography.
Recent front-end developments have focused on increasing a mass spectrometer’s efficiency by introducing samples at a rate that approaches the mass spectrometer’s rate of analysis. Two recent developments are multiplexed liquid chromatography (LC) and laser diode thermal desorption (LDTD). Both create more capacity by increasing the time that the mass spectrometer is engaged in generating meaningful data, and minimising its idle time.
Selecting the optimal high-throughput technology for a laboratory requires a careful consideration of throughput requirements in the light of the features of each technology. When evaluating multiplexed LC versus LDTD, there are at least three main issues scientists should consider: speed, sensitivity and ease of system optimisation. This article describes the principles underlying both LC and LDTD, and compares their performance in two common drug discovery applications.
FRONT-END ALTERNATIVES MULTIPLEXED LC
A multiplexed LC system enables more than one HPLC system to feed sample injections into a single mass spectrometer by staggering the injections, thereby providing an increase in throughput.
Each channel operates independently; a user can simultaneously run the same method on all four channels, or different methods on each channel. Clearly a perfectly operating two-channel system will double the throughput of a single LC system, and a four-channel system will quadruple throughput, reducing typical average MS idle time from 75 percent to four percent. A multiplexed LC system thus provides maximum utilisation of the mass spectrometer with virtually no idle time.
There are several advantages to multiplexed LC. The most important is that the chromatographic separation is not sacrificed in order to improve sample throughput. Because the operation of each multiplexed LC system is staggered and in parallel, the MS is dedicated solely to a single sample stream during the critical elution step, thus maintaining data quality and sensitivity throughout the process. Multiplexing software controls the pumping operation, valve switching, cleaning and gradient procedures.
Additionally, each set of samples in the analytical run travels through a single LC system. The standards, quality control samples and study samples are all analysed on the same hardware. In essence, it is no different from a system with one LC connected to the mass spectrometer through a divert valve. Designed properly, intra-assay and inter-assay precision and accuracy on a multiplexed LC system will meet accepted criteria for a validated LC-MS/MS method.
LASER DIODE THERMAL DESORPTION
LDTD is fundamentally different from liquid chromatography in that a liquid phase is not required. The LDTD process begins by spotting 2 - 10 µL of sample into each well of a specially designed 96-well plate, each well of which has a metal bottom insert. The plate dries at room temperature. A laser diode heats the well to produce a rapid thermal desorption of the dried sample. Intact desorbed molecules are transported by a carrier gas to an atmospheric pressure chemical ionisation (APCI) region to undergo ionisation. APCI is achieved without the presence of a solvent or mobile phase, which changes the ionisation characteristics and allows the analysis of compounds that are not efficiently ionised under normal APCI.
Thermal desorption produces mainly intact desorbed molecules at temperatures below their reported melting point. The performance of LDTD is a result of the high heating rate, the carrier gas flow and the nanoscale deposit.
LDTD eliminates the risk of cross contamination or carryover normally associated with the use of chromatographic techniques such as HPLC, where injector, column and tubing are common to all injected samples. Other advantages include the absence of background noise normally induced by a mobile phase or an enhancement matrix (such as MALDI), and a simplified sample preparation -- only 2 - 10 µL of sample is required for each well, followed by a quick solvent evaporation. There is no need to add any additional matrix to the samples.
COMPARISON
As mentioned, selection of a high-throughput front-end system largely rests on the importance of three variables: speed, sensitivity and ease of optimisation.
Because LDTD performs no physical chemical separation, it is faster than multiplexed LC but not as sensitive. LDTD can process a sample in seven to eight seconds. By comparison, a multiplexed LC system would require about 25 seconds. For most users, such a difference is not important — sample introduction taking under a minute would still be satisfactory.
While the LDTD approach provides more speed, the lack of physical separation has a drawback, namely lower sensitivity. The difference between the two techniques can be illustrated by the example of the analysis of two testosterone metabolites, namely, 4-hydroxy and 6-hydroxy testosterone, both of which have the same mass. When using LC, the metabolites are physically separated, such that there is a time separation between the two peaks, which are clearly distinguishable and quantifiable. Using LDTD, the 4-hydroxy and 6-hydroxy metabolites co-eluted (co-desorbed), making quantitation of each metabolite impossible [1].
The ease of system optimisation, i.e. the process of determining the ideal parameters, such as S-lens voltage, collision energy and molecular weight of precursor and product ions for an MS experiment, is another significant difference between multiplexed LC and LDTD.
With a multiplexed LC front-end, infusion optimisation is used to derive the optimum MS parameters. A compound is pumped into the MS at a constant rate and concentration. The infusion optimisation software alters the parameters in the MS and monitors the effects, thus automatically finding the optimum settings.
With an LDTD front-end, loop injection optimisation is simulated to derive optimum MS parameters. Because thermal desorption does not provide a uniform concentration of sample, it is more difficult for software to automate optimisation. Two variables are changing concurrently during loop injection optimisation, namely the sample concentration and MS parameters, making it less easy to determine the absolute optimum parameters.
Both multiplexed LC and LDTD work well. The question of which is most appropriate depends on a laboratory’s objective: are its goals best served through ultimate speed or ultimate sensitivity?
APPLICATIONS
Early assessment of absorption, distribution, metabolism and excretion (ADME) properties of drug candidates is an essential component of drug discovery. ADME characterisations can help identify positive and negative human pharmacokinetic properties. Two common assays, namely CYP450 inhibition and metabolic stability, can be used to compare the strengths of multiplexed LC and LDTD.
A key step in the drug discovery process is to identify possible drug-drug interactions. One of the most common causes of these interactions is competitive inhibition of compounds with cytochrome P450 isozymes. This is usually detected by performing cytochrome P450 (CYP450) inhibition assays. Such assays are well understood and standardised, and can be run easily once the mass spectrometer is optimised for the CYP450 probe compound associated with each of the isozymes [2]. One optimisation is required for a probe compound, and once known, the parameters do not change. Compound concentrations are at therapeutic levels. Since metabolite concentrations are very low, high sensitivity is required to monitor the turnover.
A metabolic stability assay investigates a compound’s metabolism over time to determine in vitro half-life. Concentrations are much higher in this assay than in the CYP450 assay, so sensitivity requirements are not as great. However, multiple compounds may be tested during a day, each requiring optimisation. Ease of optimisation is therefore of greater importance in this assay.
LDTD CYP450 APPLICATION
In a recent application of LDTD technology [1] designed to demonstrate throughput capability, researchers used a Phytronix Technologies LDTD source coupled to a Thermo Scientific TSQ Quantum Ultra triple stage quadrupole mass spectrometer.
The researchers looked at eight metabolites. Nine concentrations of inhibitor were tested; formed metabolites were determined at each inhibitor concentration. Enzyme activity was calculated and plotted. Of the eight metabolites, one was unidentifiable due to in-source fragmentation. In the assay performed with the LDTD source, all nine isoforms were pooled into one sample.
Sample-to-sample time was 15 seconds.
The CYP inhibition assay was also run on an LC-MS/MS system consisting of a standard, non-multiplexed HPLC coupled with an Applied Biosystems API4000. Samples were grouped into three sample pools due to differing lipophilicity. Sample-to-sample time was up to six minutes [1].
The researchers found that the data from the two methods were comparable and that the ability to pool samples using the LDTD system reduced time and cost. With a sample-to-sample speed of about 15 seconds, the team was able to complete a 96-well plate within 24 minutes. Because of possible in-source fragmentation, such as unconverted Phenacetin forming Paracetamol, the researchers concluded that LC-MS/MS should be used for greater selectivity [1].
METABOLIC STABILITY APPLICATION: HPLC AND MULTIPLEXED HPLC
Metabolic stability is defined as the percentage of parent compound lost over time in the presence of a metabolically active test system. For metabolic stability assays, the typical test systems use liver microsomes, liver S9, or hepatocytes (plated or suspended), depending on the goal of the assay [3].
To determine the metabolic stability of a new chemical entity, quantitation of the drug candidate from incubate supernatants is required. This is usually accomplished by high-performance liquid chromatography (HPLC) with mass spectrometry.
By understanding the metabolic stability of compounds early in discovery, compounds can be ranked for further studies, and the potential for a drug candidate to fail in development as a result of pharmacokinetic reasons may be reduced [3].
A recent study showed that throughput can be dramatically increased using automated optimisation and multiplexed LC when performing metabolic stability assays [4]. The combination of a Thermo Scientific Aria TLX-4 HPLC system, Thermo Scientific QuickQuan automated optimisation software and a TSQ Quantum Ultra triple stage quadrupole mass spectrometer was found to significantly decrease optimisation time and increase sample throughput without compromising quality.
Eight compounds were used to compare automated optimisation versus manual optimisation, and single-channel HPLC against multiplexed HPLC. Samples were analysed with a single HPLC/MS system and manual optimisation, and with a four-channel multiplexed HPLC/MS system with automated optimisation. Both achieved accuracies within 10% deviation from theoretical, and precisions below 10% relative standard deviation.
Manual optimisation took two hours of dedicated operator time – 15 minutes per compound. Automatic optimisation was achieved in a total of 16 minutes – two minutes per compound. A seven-fold improvement in optimisation time was achieved. Runtime on the single-channel HPLC system was 48 hours, whereas runtime on the four-channel multiplexed HPLC system was 12 hours., thus resulting in a four-fold improvement in runtime. The researches concluded that automated optimisation and multiplexed HPLC significantly increased throughput without compromising data quality.
CONCLUSION
LDTD and multiplexed LC are both viable options for increasing throughput in drug discovery experiments. Multiplexed LC offers higher sensitivity critical for applications such as CYP450 inhibition assays, as well as easier optimisation, which is useful when analysing multiple compounds per day. LDTD offers higher throughput, enabling the user to analyse more samples per day or obtain critical results quicker; however, automated optimisation is less accurate. Laboratories must assess both options in their quest for ultimate speed, ultimate sensitivity and ultimate ease of use.
REFERENCES
1. Wernevik J, Landin A, Tremblay P. LDTD-MS/MS and LC-MS/MS comparison for high throughput IC50 determination of CYP2A6, 2B6, 2C9, 2C19, 2D6 and 3A4 inhibition in Human Liver Microsomes (HLM). 28th Montreux Symposium on LC/MS, Montreux, Switzerland, Nov. 12-14, 2008.
2. Yan Z, Caldwell G. Evaluation of Cytochrome P450 Inhibition in Human Liver Microsomes. In Methods in Pharmacology and Toxicology, Optimization in Drug Discovery: In Vitro Methods; Yan, Z., Caldwell, G. Eds.; Humana Press Inc.: Totowa, NJ, 2004; pp 231.
3. Ackley D, Rockich K, Baker T. Metabolic Stability Assessed by Liver Microsomes and Hepatocytes. In Methods in Pharmacology and Toxicology, Optimization in Drug Discovery: In Vitro Methods; Yan, Z., Caldwell, G. Eds.; Humana Press Inc.: Totowa, NJ, 2004; pp 151.
4. Berube M. High Throughput LC-MS/MS Analysis of Metabolic Stability Incubations, Application Note 458; Thermo Fisher Scientific: Franklin, MA, April 2009.
TSQ Quantum Ultra, Aria and QuickQuan are trademarks of Thermo Fisher Scientific Inc. and its subsidiaries.
LDTD is a trademark of Phytronix Technologies.
THE AUTHOR
Jim Koers,
Marketing specialist,
Drug Discovery
Thermo Fisher Scientific,
San Jose, CA, USA
