Abstract
To produce high-quality drug candidates and to support drug discovery decision making, compounds are subjected to predictive in vitro and in vivo absorption, distribution, metabolism, excretion, toxicology, drug metabolism, and pharmacokinetics assays, in which samples are typically analyzed using automated liquid chromatography–mass spectrometry (LC–MS). Depending on sample preparation and clean-up efforts, matrix components such as salts, nonprecipitated small molecules, and particulate matter can remain in the sample, thus having a detrimental effect on the performance of the autosampler, especially if the sample manipulation mechanism is syringe-based. This issue is amplified in high-throughput qualitative and quantitative LC–MS-based assays due to the large sample load and associated mechanical “wear-and-tear” of the syringe, resulting in poor data quality, increased costs, and lost time as a result of sample re-analysis. Described here are improvements made to the sample manipulation design and mechanism of a LEAP CTC HTS PAL autosampler. This setup completely eliminates the contact of samples and potentially abrasive and corrosive matrix components with both the syringe plunger and barrel through the use of a Teflon tubing loop, thus significantly extending the life of the injection device. In our high-throughput in vitro assay to assess the metabolic stability of compounds, we observed a 20-fold increase in the syringe lifetime using the improved sample manipulation design versus the standard setup with similar analytical performance such as reproducibility, accuracy, and precision. (JALA 2007;12:152–6)
Keywords
Introduction
Today's efforts to synthesize new chemical entities and to discover new drug candidates occur at an ever-increasing rate. As a result of the recent shift of the drug development paradigm to earlier assessment of a compound's drug properties and characteristics such as absorption, distribution, metabolism, excretion, toxicology, drug metabolism, pharmacokinetics, drug safety, and pharmacological selectivity, bioanalytical groups supporting this endeavor have to deal with a substantial sample load.1–4 These samples from in vitro and in vivo studies are analyzed using automated high-throughput analytical processes typically utilizing liquid chromatography–mass spectrometry (LC–MS) techniques to quantitatively measure the disappearance or in situ appearance of a specific analyte.5, 6 As a prerequisite, the analytical instrumentation (LC pump, autosampler, LC pre-column, LC analytical column, UV detector, and mass spectrometer) along with the analytical methods used needs to be robust, accurate, and reliable to reproducibly generate high-quality data. The large number of compounds being screened does not permit extensive sample preparation and clean-up; therefore, these samples may still contain a substantial amount of matrix components such as salts, nonprecipitated small molecules, and particulate matter. This can affect not only the LC-UV/MS analysis and data quality but also the analytical equipment, particularly an autosampler with syringe-based sample manipulation design and mechanism.
As part of the high-throughput screening model at our company, an in vitro assay is used to assess the metabolic stability of compounds. 7 After incubation of the compounds with microsomes, samples are prepared using protein precipitation 8 with a 2:1 (v/v) organic solvent (acetonitrile) to assay media ratio. The amount of the remaining non-metabolized compound in the supernatant is quantitatively determined by LC–MS analysis. The LC–MS system employs a LEAP CTC HTS PAL autosampler that uses a syringe to directly manipulate the samples. It was observed that after a syringe had handled about 500 samples and analyses, the reproducibility, accuracy, and precision of hourly “LC-UV/MS system control” samples drifted outside the range of acceptable data. These samples contain MS-active (verapamil) and UV-active (dimethoxy-naphthalene [DMN]) standards, which are analyzed in duplicate to monitor the performance of the analytical setup. Typical standard deviation values for the UV-active (DMN) and MS-active (verapamil) standards are in the range of 4 and 8%, respectively. Upon further investigation using visual microscopy, it appeared that the syringe experienced abrasion of the polyethylene or Teflon plunger tip and etching on the barrel most likely due to matrix components and/or mechanical “wear-and-tear.” This resulted in irreproducible sample volumes, which were detrimental to the bioanalytical assays. Using the hypothesis that the matrix was responsible for the deterioration of the syringe, a strategy came to mind to prevent the sample from entering the syringe barrel. This report describes improvements made to the sample manipulation design and mechanism by aspirating the sample into a Teflon tubing loop and thus eliminating any contact between the sample, its associated matrix components, and the syringe. This modified sample manipulation has resulted in a 20-fold increase in the lifetime of a syringe versus the standard setup with similar analytical performance such as reproducibility, accuracy, and precision. Approximately 10,000 samples can now be analyzed before the loop and syringe assembly needs to be replaced. This practical and inexpensive design without the need for additional hardware such as switching valves or dual-arm systems has been successfully implemented in both qualitative and quantitative high-throughput assays and is routinely being used. 7
Material and Methods
Instrumentation and Software
The autosampler used for the LC–MS system was a CTC HTS PAL (LEAP Technologies, Carrboro, NC). Custom sample manipulation software macros were programmed using PAL Cycle Composer software (V 1.5.2).
Sample Manipulation Assembly
The threaded needle holder, plunger, and magnetic syringe holders were built in-house using anodized aluminum based on the dimensions of the commercially available model. The syringe used was an LC syringe (22 gauge, blunt needle point, polyethylene plunger tip, #L100.K100) from Microliter Analytical Supplies (Suwanee, GA). The Teflon tubing (inside diameter 0.025 in., outside diameter 0.06 in., #4E02501810) was from Zeus (Orangeburg, SC). The needle (22 gauge, blunt needle point, #7770-02) was from Hamilton (Reno, NV). The needle was connected to the Teflon tubing using a VacuTight fitting (#P846) and ferrule (#P840) from Upchurch Scientific (Oak Harbor, WA).
The syringe needle was manually bent by 180° using a template made in-house (Fig. 1). The Teflon tubing was manually pushed over the needle end and the syringe tip.
Syringe needle bending plate to bend the syringe needle 180° around the post at the bottom of the plate.
HPLC-grade water and HPLC-grade methanol were obtained from EM Science (Darmstadt, Germany). The autosampler wash solvent consisted of water:methanol (80:20, v/v).
Discussion
The commercially available sample manipulation mechanism of the autosampler consists of a robotic x, y, z-stage that moves a removable magnetic holder with the mounted syringe (Fig. 2) between the stationary sample vials, LC injection port, and wash stations. By vertical movement (z-direction) of the plunger via a wormgear, solvent is aspirated into, held in, or dispensed from the syringe barrel. This solvent, whether wash solution or sample containing matrix components, can affect and potentially damage the barrel and plunger tip, resulting in non-reproducible sample volumes.
The original LEAP CTC HTS PAL autosampler syringe assembly where the sample enters the syringe.
Our strategy was to implement a design that completely eliminated the contact of the sample with the syringe. Furthermore, we wanted to maintain the concept of a removable magnetic syringe holder by simply replacing it with our design, and thus being able to switch between standard and modified versions within seconds.
The improved sample manipulation mechanism consists of a standard syringe that is connected to the needle via a Teflon tubing loop. The length and consequently the volume of the tubing can be manually adjusted so that the sample does not enter the syringe barrel. This assembly is mounted on a modified removable magnetic holder (Fig. 3a–d) and will be referred to as the “unit.” To accommodate the tubing, the syringe is offset from the center, which is addressed by the modified plunger holder. An example of the functionality is illustrated by the following procedure using a custom software macro. This particular protocol is routinely used in both qualitative and quantitative high-throughput in vitro LC–MS assays such as the assessment of the metabolic stability of compounds. 7 In this case, the modified design resulted in a 20-fold increase in the lifetime of a syringe compared to the commercially available model with similar analytical performance such as reproducibility, accuracy, and precision as determined via statistical data obtained by hourly “LC-UV/MS system control” standards. Notably, the volume of sample and wash solvents manipulated by means of Teflon tubing and software macro as well as any protocol action is customizable, which enables the presented design to be applied to many different assays (Table 1).
Example of a custom software macro used with the modified sample manipulation design
(a) Modified sample manipulation design where the sample does not enter the syringe. (b) Details of the modified sample manipulation design. (c) Computer-aided design (CAD) drawing of the modified sample manipulation design (front view). (d) CAD drawing of the modified sample manipulation design (back view).
Conclusions
An improved sample manipulation design and mechanism for a LEAP CTC HTS PAL autosampler has been developed. The setup eliminates any contact of the sample, including potentially destructive matrix components, with the syringe barrel and plunger tip, thus significantly increasing the lifetime of the syringe with analytical performance such as reproducibility, accuracy, and precision similar to that of the original syringe assembly. The modified unit is routinely used in high-throughput qualitative and quantitative LC–MS assays with excellent data reproducibility, precision, and accuracy.
Footnotes
Acknowledgment
The authors would like to thank Ms. Bethanne Warrack for critical proof reading of the manuscript.