An Inexpensive, In-House-Made, Microdialysis Device for Measuring Drug–Protein Binding

Abstract

An inexpensive, in-house made microdialysis device is described that is suitable for measuring the binding of small molecules including drug candidates to serum proteins or other macromolecules. The device is based on the standard equilibrium dialysis method to measure the fraction of low molecular weight compound bound to proteins. It is constructed from a standard polypropylene 96-well plate, dialysis tubing, and low viscosity epoxy resin. The device can be readily prepared for a small fraction of the cost of a commercial, multichamber microdialysis device. Drug–protein binding results are provided, which validates the device.

For the development of new drugs, steps have been taken in the past several decades to create high-throughput methods for traditional ADME (absorption, distribution, metabolism, and excretion) testing.¹ These ADME tools are significant in estimating and understanding the pharmacokinetics and overall therapeutic action of a drug compound in preparation for more costly in vivo studies.^2,3 An important ADME parameter is the concentration of free drug in solution, i.e., that which is not bound to macromolecules including serum proteins. It is the free concentration of compound that dictates its biological potency on the drug-target as well as the rate of drug clearance by metabolic and excretion processes.^1,3

One method to measure plasma protein binding (PPB) of drug candidates to proteins is equilibrium dialysis.^1,3 Protein solution in buffer or a biological fluid such as serum or tissue homogenate is placed in a chamber on one side of a dialysis membrane that allows permeability of low molecular weight compounds but not that of macromolecules including proteins. Drug candidate is added to the protein-containing chamber while the chamber with buffer only is in contact with the other side of the dialysis membrane. The device is agitated at a constant temperature, typically 37 °C, for several hours to allow the concentration of drug candidate in solution (not bound to proteins) on both sides of the membrane to become equal (equilibrium). If the drug candidate binds to proteins, the total drug candidate concentration on the protein side of the membrane (bound + free) is higher than that on the other side (free only). The fraction of protein-bound and protein-free drug candidate is obtained by measuring these concentrations using a suitable analytical technique such as liquid chromatography–tandem mass spectrometry to measure the amount of drug candidate in an aliquot taken from both chambers of the dialysis device.

It is often desirable to carry out several equilibrium dialysis experiments in parallel and with a minimal amount of liquid so as to conserve drug candidate and biological fluid such as serum. Multichamber, equilibrium dialysis devices are available commercially for a cost of approximately $5–10 per microdialysis chamber (Table 1). Some of the more popular devices include Pierce Biotechnology’s Rapid Equilibrium Dialysis (RED) device (ThermoFisher Scientific, Waltham, MA), the Harvard Apparatus (Harvard Bioscience, Holliston, MA), and HTDialysis, LLC’s HTD 96 (HTDialysis, LLC, Gales Ferry, CT).^1,4,5 These devices provide a high-throughput option for scientists seeking small volume dialysis methods, but they come at a high price.

Table 1. Cost per Dialysis Well of the DIYM versus Commercially Available Devices.

dialysis device	cost per device	cost per dialysis well	dialysis wells per device
DIYM	$44.16	$0.46	96
RED devicea	$445.00	$9.27	48
Harvard Apparatusb	$446.00	$4.65	96
HTD96c	$4,149.00	$43.22	96

^aThe Rapid Equilibrium Dialysis (RED) device Single-Use Plate with Inserts (ThermoFisher Scientific, Waltham, MA).

^bThe Harvard Apparatus 96-well DispoEquilibrium Dialyzer Single Use Plate (Harvard Bioscience, Holliston, MA).

^cThe HTD96b Complete Unit is reusable (HTDialysis, LLC, Gales Ferry, CT).

Here we describe the construction of a simple microdialysis device (do-it-yourself microdialysis device, DIYM) and show its validity in measuring the binding of drug candidates to serum protein (Figure 1). Construction of the DIYM costs less than $1 per microdialysis chamber (Table 1) and begins with a 96-well, polypropylene storage plate.⁶ Into each square well of the plate is deposited a low viscosity epoxy resin (made by mixing a bis-epoxide with a diamine hardener). A short piece of flat dialysis tubing is inserted into each well so that one of the open ends of tubing is fully submerged in the epoxy resin. The other end of the tubing protrudes above the surface of the epoxy resin and extends to just below the top of the well. The epoxy resin is allowed to cure at room temperature overnight. After curing, the wells are rinsed with nanopure water to allow the dialysis tubing to swell and to remove any impurities that are not trapped in the cured resin. In order to add structure to the dialysis tubing, a section cut from a 250 μL pipet tip is inserted into the opening of the tubing (see Figure 1). In this Technology Note, the volume inside the dialysis tubing is considered the donor chamber and is filled with protein solution in buffer (for example, diluted serum). The space between the well walls and the outside of the dialysis tubing is considered the acceptor chamber and is filled with buffer without protein. The cured epoxy serves as a water-impermeable barrier between chambers.⁸ The open end of the 3 cm length of dialysis tubing protrudes above the top of the acceptor chamber solution. The clearance is sufficient to prevent donor material from leaking into the acceptor chamber. Thus, the only movement of solvent and small molecular weight solute molecules between the donor and acceptor chambers must be by diffusion through the dialysis membrane. It is not necessary to seal the top of the dialysis tubing under experimental conditions described in this Technology Note.

Figure 1.

(A) An individual 2.2 mL square well in a 96-well storage plate is prepared for the DIYM. (B) Low-viscosity epoxy resin (0.5 mL) is transferred to the bottom of the well. (C) A section is cut from a 250 μL pipet tip. This segment is inserted into the soaked dialysis tubing to hold it open during dialysis. (D) A 3 cm strip of 12–14 kDa dialysis tubing is inserted into the square well so that the bottom opening is submerged in uncured epoxy. Once the epoxy has cured, the tubing is soaked in nanopure water for 30 min, rinsed twice with nanopure water, and stored in buffer until use.

High and low molecular weight dye compounds, dextran (∼2 million Daltons) and methyl orange, respectively, were used in initial experiments to validate the DIYM. Aliquots of buffer on both sides of the dialysis tubing were periodically sampled and submitted to spectrophotometry to measure the concentration of dyes. After a minimum of 8 h of incubation with agitation of the 96-well reservoir at 37 °C, the concentration of methyl orange was found to be the same in both chambers. Dextran was not found on the acceptor side of the membrane showing that the epoxy formed a molecularly tight barrier between the two compartments.

Next, the DIYM was used to investigate protein binding of eight drug-like compounds to serum protein. The compounds were chosen to represent a variety of structures and a range of serum binding. Dextromethorphan (DXM), diclofenac (DCF), mefloquine (MFQ), methotrexate (MTX), paclitaxel (PCL), progesterone (PRG), propranolol (PRO), and testosterone (TST) were added to the donor chamber at 5 μM concentration in 50% plasma solutions and dialyzed for 8 h.

In the initial stages of method development, equal volumes of donor and acceptor solutions were charged to the dialysis chambers (donor solutions containing plasma, acceptor solutions composed of only buffer). A volume shift between the acceptor and donor chambers was seen and was attributed to osmolarity differences between the chambers. Results from this initial stage of method development were corrected by applying a factor representing the change in the donor chamber’s plasma/buffer ratio from t = 0 to equilibrium. In order to prevent the volume shift, the acceptor chamber volume was reduced, resulting in no volume change. Dialysis of propranolol and mefloquine was retested under these conditions, and there was no change in the final protein binding data, supporting the use of a corrective factor in the initial calculations. For all further testing, the acceptor chamber volume was reduced to avoid volume shift.

Reproducibility of the DIYM was investigated using propranolol, a compound with moderate PPB, which is commonly used as an experimental control in the literature. The protein binding of propranolol was determined for n = 12 simultaneous replications using the DIYM device. It was found that the PPB for these 12 replications was 86.3 ± 1.0%. This result indicates that there is low variability between dialysis cells in the DIYM.

In order to compare the results of the DIYM, the same compounds were investigated in ThermoFisher’s RED device using single-use inserts with an exclusion limit of 8.0 kDa and a reuseable base-plate (Table 2). Dialyses were run in triplicate for each compound on each device, and standard deviations are given in Table 2.

Table 2. PPB for Each Compound Was Investigated within the Same Laboratory Environment for the DIYM Device and ThermoFisher’s RED Device^a.

	PPB, n = 3 (%)
compd	DIYM	RED	lit. PPB (%)
DXM	66.8 ± 6.9	66.7 ± 3.3	659
DCF	98.0 ± 0.1	90.4 ± 1.3	99.513
MFQ	98.9 ± 0.1	100.0 ± 0.0	>987
MTX	54.0 ± 5.1	44.8 ± 10.8	50.411
PCL	94.2 ± 0.9	92.9 ± 1.7	9514
PRG	97.0 ± 0.4	94.6 ± 0.9	9815
PRO	82.5 ± 1.9	79.7 ± 1.5	80–821
TST	93.3 ± 0.5	94.7 ± 0.1	9810

^aLiterature PPB values may represent PPB in 100% human plasma. Thus, the exact values may vary from the findings of our study.

Compared to the RED, the DIYM, due to its 96-well design, is compatible with high-throughput systems but is not optimized for highly automated screening or commercialization. Rather, the DIYM is intended to provide researchers with limited budgets a noncommercialized and therefore lower-cost option compatible with multichannel pipetting devices for simultaneous dialysis of multiple samples. The simple design and availability of materials for the DIYM make it a lower-cost option with comparable performance to other multichambered devices.

Experimental Procedures

Materials

Pooled female mouse plasma (Lithium Heparin as a coagulant) was received from BioReclamationIVT (Westbury, NY). Test compounds (dextromethorphan, diclofenac, mefloquine, methotrexate, paclitaxel, progesterone, propranolol, and testosterone) were received from Sigma-Aldrich (St. Louis, MO). Spectra/Por brand regenerated cellulose dialysis tubing (MWCO 12–14 kDa, flat width 10 mm, part no. 132697) was received from Spectrum Laboratories (Rancho Dominiquez, CA). Aldon Corp SE dialysis tubing (MWCO 14 kDa, flat width 10 mm, cat. no. 470163-426) was received from VWR International (Radnor, PA). Abgene storage plates (2.2 mL square-well, polypropylene, 96-well plate, cat. no. AB0932B) were purchased from ThermoScientific (Rockwood, TN). The Rapid Equilibrium Dialysis (RED) Device Inserts, 8K MWCO, and Reusable Base Plate were also purchased from ThermoFisher Scientific. Resin Research Project 21 System Epoxy 2000 and 2100 Fast Hardener (item no. G02–0135) were purchased from Fiberglass Supply (Burlington, WA). Rainin LTS 250 μL pipet tips (item no. 17005092) were purchased from Rainin (Oakland, CA). Axyseal polyester sealing film with acrylic adhesive (cat. no. PCR-SP) was received from PlateMax. Anhydrous potassium phosphate monobasic, anhydrous sodium phosphate dibasic, sodium chloride, and methyl orange were purchased from Fisher (Hampton, NH). High molecular weight dextran was received from Pharmacia (Uppsala, Sweden).

Instrumentation

The DIYM was constructed from the 96-well Abgene storage plate, Spectra/Por regenerated cellulose dialysis tubing, Resin Research epoxy system, and the 250 μL Rainin LTS pipet tips. Total concentrations for initial testing were determined on a Lambda 3A UV/vis Spectrophotometer (PerkinElmer, Waltham, MA). Total concentrations were determined for later testing via mass spectrometry using a Quattro Micro triple-quadrupole mass spectrometer (Waters, Milford, MA). Samples were separated using a binary HPLC pump 1525μ and 2777C sample manager (Waters, Milford, MA) connected to a Zorbax C18, 3.5 μm, 2.1 × 100 mm column (Agilent, Santa Clara, CA). Mobile phases were used in a gradient program, with an aqueous phase (1% acetic acid in 5% acetonitrile/94% water) and an organic phase (1% acetic acid in 99% acetonitrile).

DIYM Method

For each dialysis well, a 3 cm section was cut from the Spectra/Por Dialysis Tubing. Epoxy mixture was prepared by mixing 0.45 g of 2100 Fast hardener to 1.0 g of 2000 Epoxy and stirring well with a wooden stick or spatula for 1 min, scraping the sides of the vessel to ensure complete incorporation. The Abgene plate was prepared by transferring 0.5 mL of uncured epoxy mixture to each well. The viscous epoxy was transferred to each well with a 1000 μL manual pipet, using a disposable pipet tip with a trimmed tip. Before the epoxy cured, a 3 cm length of dialysis tubing was inserted diagonally into the well so that the bottom end was completely encased in the epoxy resin. The epoxy was allowed to cure overnight prior to moving forward with testing.

It was noted that there was some diameter variability in the Spectra/Por dialysis tubing across lot numbers. Aldon Corp SE tubing was found to be an acceptable substitute for the Spectra/Por tubing as an option with decreased diameter variability.

Once the epoxy resin was cured, the dialysis membrane was soaked for 30 min in nanopure water to remove the humectant glycerin. The membrane was rinsed twice with nanopure water, then stored in PBS solution (10× Phosphate Buffered Saline, pH 7.4, 0.1 M phosphate, 0.154 M NaCl) until the assay was ready to begin. Immediately before charging donor solutions, a segment of the 250 μL Rainin pipet tip (which had been cut from the middle section of the pipet tip using a razor blade) was inserted into each dialysis tube and was pushed to the bottom of the dialysis tubing to hold it open.

For every sample, each dialysis tubing donor chamber was charged with 300 μL of donor solution (50% plasma). Each acceptor chamber was charged with 250 μL of phosphate buffered saline to avoid volume shift. Samples were covered with a CO₂-permeable membrane, and equilibrium dialysis was carried out at 37 °C and 5% CO₂ in order to control the postdialysis plasma pH by maintaining a consistent CO₂ content in solution.¹² The device was agitated using an orbital shaker set to 150 rpm. Initial experiments using methyl orange as an analyte showed that equilibrium is reached by 8 h, but equilibration was run between 8 to 12 h with consistent results. Experiments with dextran-2000 showed no leaking between acceptor and donor chambers under these conditions.

RED Method

A method for the use of the RED apparatus is provided on the ThermoFisher Web site. The inserts are ready for use without presoaking, and each donor chamber was charged with 300 μL of donor solution (50% plasma). Each acceptor chamber was charged with 550 μL of phosphate buffered saline. (These values differ from those listed on the Web site method, which requires 500 μL of buffer be used with 300 μL of sample. Our procedure differed from this method based on information provided in correspondence with ThermoScientific’s Protein Biology Technical Support (case 2-3452279952)). Following this step, the device was covered with a CO₂-permeable membrane and incubated for 6 h.

Calculations

After reaching equilibrium, the concentration of analyte present in the donor and acceptor chambers was measured via mass spectrometry. These values were used to determine the percent fraction unbound of drug (%FU) using the following equation:¹

The value %FU, found in the above equation, was corrected for plasma dilution and for volume shift (if applicable). The percent bound values reported in this text were calculated by subtracting the %FU from 100%.

Glossary

ABBREVIATIONS

DIYM

do-it-yourself microdialysis device

ADME

absorption, distribution, metabolism, and excretion

EDTA

ethylenediaminetetraacetic acid

PBS

phosphate buffered saline

FU

fraction unbound

Supporting Information Available

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsmedchemlett.7b00316.

• Full experimental details, in vitro biological methods and formulas, MS method, structures and peak shape for all compounds, tabulated experimental data, and DIYM device images (PDF)

Author Contributions

The manuscript was written through contributions of all authors.

This work was funded by grants from the National Institutes of Health (Grants AI070218 and AI097177).

The authors declare no competing financial interest.