click to enlarge Table 1: Dialysis Assay Evaluation Conditions (All tables: Taylor Technology, Inc.)

A rapid equilibrium dialysis (RED) drug-protein binding assay using LC-MS/MS was developed using a novel technique that resulted in significantly improved assay precision. While RED-based LC-MS/MS assays offer a speed advantage, the analysis is dependent on accurate sampling and volumetric errors can significantly skew the results. The addition of carefully selected reference compounds against which the compound of interest can be normalized substantially reduces sampling errors as the aliquot volume no longer impacts the precision measurement.

Human plasma containing one or several test compounds, a reference compound, and quality control (QC) compounds is filled into one of the two chambers; aqueous buffer is filled into the other chamber. Peak areas for test compounds, QC compounds, and the reference compound are determined by an LC-MS/MS method. These are then expressed as peak area ratios relative to the peak area of the reference compound.

Compared to an equilibrium dialysis assay, this assay exhibits three unique advantages: reduced dependence on dialysis time; aliquot errors in sample preparation having little effect to the final result; and simultaneous generation of four QC data points per protein-binding result of each test compound. The first two merits automatically correct major experimental errors and the third merit provides additional confidence in data when all four QCs meet literature or historic values.

click to enlarge Table 2: Detected PB data compared to literature values.

Introduction
Drug-protein binding (PB) is an important factor affecting disposition profiles and pharmacological efficacy of a drug. It is well known that only unbound drug molecules are able to transport through membranes, undergo metabolism, and interact with pharmacological receptors. Depending on the drug, PB may be a critical factor in allometric scaling and pharmacokinetic/pharmacodynamic (PK/PD) modeling in special populations, making it highly desirable to evaluate the degree of PB efficiently and reliably for drug candidates in both drug discovery and development processes.

Albumin liquid chromatography, ultra-filtration, and equilibrium dialysis are the three major technologies applied in the industry. The albumin liquid chromatography method is considered high throughput; however, the deviation between predicted and true PB values is difficult to estimate because the albumin column can’t operate at physiological pH (7.4) and does not represent all of the proteins in human plasma. This kind of deviation is compound dependent, especially for test compounds with pKa values close to 7.4 and/or compounds capable of binding to various types of proteins. The ultra-filtration method suffers from non-specific binding, which impacts the accuracy of the assay. The equilibrium dialysis method is considered an accurate assay because it mimics the in vivo environment, but it requires very accurate sample aliquot collection.

click to enlarge Table 3: A special example that verapamil and a test compound have the same free fraction 10%.

Faced with these challenges, pharma companies are increasingly employing the services of contract research organizations (CROs) whose core business is developing, validating, and utilizing assays in the development of new drug products. Specialist providers of bioanalytical mass spectrometry and immunochemistry capabilities have long recognized the need for dependable evaluation of PB.

The latest innovations offered by specialist service providers involve the application of radioactively labeled compounds to improve the accuracy of the equilibrium dialysis assay using an RED device (Thermo Fisher Scientific; Waltham, Mass).^1,2 This may not be suitable for discovery studies because radioactively labeled test compounds are not available for most early stage discovery compounds. In addition, impurities in the radiolabeled compounds may preclude determining the PB for highly (>99%) bound drugs.

Experiment
The Thermo Scientific RED dialysis system utilizes disposable tube inserts and a single-use base plate. Each insert is made of two side-by-side chambers separated by a vertical cylinder of dialysis membrane with a molecular weight cut-off of 8000. Pooled K3 EDTA human plasma (Bioreclamation; Liverpool, N.Y.) was adjusted to a pH value of 7.4 by addition of sodium phosphate salt and phosphoric acid to generate buffered blank plasma. An aqueous buffer at pH value of 7.⁴ was also prepared by addition of sodium phosphate salt and phosphoric acid to water (Sigma-Aldrich; St. Louis, Mo.).

click to enlarge Table 3: A special example that verapamil and a test compound have the same free fraction 10%.

Stock standards at a concentration of 20 mg/mL in acetonitrile for verapamil, chloropromazine, diazepam, doxepin, and trazodone (Sigma-Aldrich) were prepared. Diluting with buffered blank plasma, a cocktail of multi-drug reinforced plasma containing these five compounds was prepared. Concentrations in plasma were 100, 250, 250, 200, and 250 ng/mL for verapamil, chloropromazine, diazepam, doxepin, and trazodone respectively.

An Agilent 1100 (Agilent Technologies; Santa Clara, Calif.) - API Sciex 4000 LC-MS/MS system (MDS Sciex, Concord, Canada) was applied to determine the peak area ratios of each compound over the reference compound verapamil. The LC method was a gradient reverse phase method and the mass spectrometer applied MS/MS methods for the six compounds. A C18 Synergi 50x2mm column was used. Peak area ratios were used to calculate the PB value for each compound.

This assay offers two unique advantages: it does not require accurate aliquot sample collection before extraction and it does not need to reach equilibrium. In order to experimentally evaluate these two advantages, a specially designed experiment was conducted.

Analysis
To measure the PB value of a test compound, the test compound was spiked into the cocktail plasma. Aliquots of 300 µL of aqueous buffer and 500 µL of the cocktail plasma were filled into the inner and out-side chambers of the RED insert, respectively. The RED dialysis plate with the inserts was kept at 37°C.

In order to prove that dialysis time does not significantly affect the final result, after 1.25 or 2.5 hours, partial samples were collected from the two chambers and mixed with drug-free plasma or buffer according to conditions listed in Table 1. An aliquot of 100 µL of sample from buffer chamber was mixed with 10 µL of drug-free plasma before extraction. An aliquot of five or 10 µL of sample from plasma chamber was mixed with 100 µL of drug-free buffer before extraction. The samples were prepared in triplicate (#1 to #3).

Results and Discussion
The PB data of the test and four QC compounds are listed in Table 2. It was confirmed that accurate aliquot sample collection before extraction is not required because collecting 5 or 10 µL from the plasma chamber did not significantly affect the final results for all of the five compounds. It also indicated that dialysis for 1.25 or 2.5 hours did not make a significant difference on the final result. The system required at least four hours to reach equilibrium.2,7 It was noted that peak areas of buffer chamber samples after 2.5-hours were significantly larger than those after 1.25-hours. These two advantages would not have been achieved using the traditional equilibrium dialysis assay.

The data process method to generate the findings presented in Table 2 is illustrated by a special and a general example. In the special example (Table 3), all of the peak area ratios from the plasma chamber and buffer chamber at time points of one and two hours were identical (1/2). The plasma chamber was larger than the buffer chamber and the concentration change in the plasma chamber during dialysis was omitted. The fact that the peak area ratios in the two chambers were identical resulted in the PB values and the free fractions of the test compound and the verapamil reference compound being the same. This means that any compound can be used to replace verapamil as a reference compound.

In the general example (Table 4), peak area ratios from the buffer chamber at time points of 1 and 2 hours were identical (1/2) but they were different from that in the plasma chamber (1/4). The determined free fraction of the test compound was 5%, resulting in 95% PB value. Just like in the special example, the plasma chamber was larger than the buffer chamber and the concentration change in the plasma chamber during dialysis was omitted.

Conclusion
A high-throughput dialysis drug-protein binding assay has been successfully developed using the Thermo Scientific RED dialysis chambers and a human plasma containing one or more test compound(s), verapamil as a reference compound, and chloropromazine, diazepam, doxepin, and trazodone as four QCs. Experimental results have demonstrated that the assay offers important benefits including only peak area ratios are needed for data processing, the method does not require accurate aliquot sample collection during sample preparation and a significant amount of time is saved as the assay is independent of the dialysis reaching equilibrium. However, for full confirmation of the suitability and efficiency of the assay in drug-protein binding applications, further research is required.

About the Author
Dr. Dahai Dong is a Senior Scientist in the Discovery Services group at Taylor Technology. Mike Koleto oversees daily operations of the Discovery Services group at Taylor Technology. Dr. Rohan Thakur is an Associate Director and responsible for Discovery Services at Taylor Technology. Dr. Richard LeLacheur is the Vice President of Taylor Technology. Dr. Christopher Kemper is a Business Director at Taylor Technology.

References
1. Qian Mark G, et al. High Throughput Plasma Protein Binding Assay Using Rapid Equilibrium Dialysis (RED) - Faster PPB Assay Results Using a 2 Hour Protocol, Millennium Pharmaceuticals; Posted by vendor of RED at http://www.piercenet.com.
2. Nasreen Malik, et al. Measurement of Plasma Protein Binding: A Comparison of Conventional, 96-Well Plate and Rapid Equilibrium Dialysis (RED) Methodologies; Covance Laboratories Ltd., Harrogate, UK; Posted by vendor of RED at http://www.piercent.com.
3. Hardman JG, Limbird LE. Goodman and Gilman’s The Pharmacological Basis of Therapeutics, 9th ed. McGraw-Hill New York 1996.
4. Edeki T, Dillon-Moore B, He N, Am J. Therapeu, 1996;Sep 3(9),611-15.
5. Virtanen R, Lisalo E, Irjala K. Acta Pharmacol Toxicol (Copenh), 1982;Aug 51(2), 59-64.
6. Abernethy DR, Todd EL. J Clin Psychopharmacol, 1980, Feb 6(1),8-12.
7. Protein binding value of Verapamil determined by RED device vendor and listed in the manual of the product.

This article was published in Drug Discovery & Development magazine: Vol. 13, No. 4, May 2010, pp. 18-20.