Enhanced Microdialysis Relative Recovery of Inflammatory Cytokines

Anal. Chem. 2004, 76, 3777-3784

Enhanced Microdialysis Relative Recovery of Inflammatory Cytokines Using Antibody-Coated Microspheres Analyzed by Flow Cytometry Xiaoping Ao,† Timothy J. Sellati,‡ and Julie A. Stenken*,†

Department of Chemistry and Chemical Biology, 130 Cogswell Laboratories, Rensselaer Polytechnic Institute, 110 Eighth Street, Troy, New York 12180-3590, and Center for Immunology and Microbial Disease, Albany Medical College, 47 New Scotland Avenue, MC151, Albany, New York 12208-3479

Achieving high relative recovery (RR) of proteins during microdialysis sampling is difficult due to the diffusion limitations inherent to this sampling process. This often causes low microdialysis RR for proteins with molecular weight >10 kDa. A RR enhancement process for microdialysis sampling of proteins has been developed that can be readily used with flow cytometers. Multiplexed RR enhancement and detection of five different cytokines (TNF-r, IFN-γ, IL-2, IL-4, and IL-5) was achieved by including antibody-coated microspheres in the microdialysis perfusion fluid. Inclusion of these antibody-coated microspheres causes an increase in the analyte diffusive driving force across the dialysis membrane and a subsequent increase in the relative recovery. For the five cytokines, typical control and enhanced relative recoveries at a 1.0 µL/min flow rate were as follows (n ) 3): TNFr, 5.5 ( 0.6% and 60.4 ( 5.8%; IFN-γ, 2.6 ( 0.3% and 25.8 ( 2.3%; IL-5, 1.4 ( 0.2% and 4.9 ( 0.1%; IL-4, 10.9 ( 0.6% and 78.8 ( 8.0%; and IL-2, 4.1 ( 0.4% and 19.8 ( 2.5%. Using this approach, a four- to 12-fold enhancement of microdialysis RR was achieved for the five cytokines from a quiescent solution. The enhancement varies among the five cytokines and may be due to different diffusive and antibody binding properties. TNF-r exhibited the highest RR enhancement, while IL-5 exhibited the lowest. Experimental parameters that affect the enhancement, such as flow rate, sample collection volume, and bead density, were studied. Microdialysis sampling is a technique that has been widely used for neuroscience, pharmacokinetic, and biotechnology applications.1,2,3 Microdialysis sampling is a diffusion-based separation process that requires analytes to diffuse through a semipermeable membrane and to be carried to the outlet via a flowing perfusion fluid. The primary analyte diffusion path through the * Corresponding author. Phone: (518) 276-2045. Fax: (518) 276-4887. E-mail: [email protected]. † Rensselaer Polytechnic Institute. ‡ Albany Medical College. (1) Bourne, J. A. Clin. Exp. Pharmacol. Physiol. 2003, 30, 16-24. (2) de la Pena, A.; Liu, P.; Derendorf, H. Adv. Drug Delivery Rev. 2000, 45, 189-216. (3) Torto, N.; Laurell, T.; Gorton, L.; Marko-Varga, G. Anal. Chim. Acta 1999, 379, 281-305. 10.1021/ac035536s CCC: $27.50 Published on Web 05/18/2004

© 2004 American Chemical Society

membrane is via the water-filled pores. Hydrophilic solutes with low molecular weight or small size will freely diffuse through the water-filled pores. Proteins are generally excluded from the dialysate because they either do not pass through the pores or diffuse so slowly that low concentrations are collected in the dialysate. Microdialysis sampling has been used primarily for in vivo sampling procedures in neuroscience for neurotransmitter collection where in situ removal of proteins from the sample was highly advantageous for the analytical methodology employed, for example, prior to liquid chromatographic separations. As microdialysis sampling became a standard tool in neuroscience, researchers in other scientific disciplines became interested in applying the technique to a wide variety of different analytes.4 In particular, in vivo microdialysis for peptides and proteins became of considerable interest.5 However, detection of peptides and proteins is highly challenging using microdialysis sampling, because their larger size and, thus, smaller aqueous diffusion coefficients cause mass transport limitations through the probe, resulting in a lower recovery. This difficulty is better appreciated when considering the mathematics that describe microdialysis extraction efficiency, shown in eq 1.6

Ed )

(

)

Coutlet - Cinlet -1 ) 1 - exp Csample,∞ - Cinlet Qd(Rd + Rm + Rs)

(1)

The extraction efficiency is bidirectional and relates the dialysate outlet (Coutlet) and inlet (Cinlet) concentrations to the sample concentration at a distance far away from the probe (Csample,∞). When Cinlet has an analyte concentration of zero, the term relative recovery (RR), rather than extraction efficiency, is commonly used. The extraction efficiency is also related to the dialysis flow rate (Qd) and the mass transport resistances among the three different regions through which an analyte must pass: dialysate (Rd), membrane (Rm), and sample (Rs). All three resistance terms incorporate the analyte diffusion coefficient for transport across each of the respective regions. For a protein with a diffusion coefficient that is much less than that of a small hydrophilic (4) Hansen, D. K.; Davies, M. I.; Lunte, S. M.; Lunte, C. E. J. Pharm. Sci. 1999, 88, 14-27. (5) Kendrick, K. M. Methods Enzymol. 1989, 168, 182-205. (6) Bungay, P. M.; Morrison, P. F.; Dedrick, R. L. Life Sci. 1990, 46, 105-119.

Analytical Chemistry, Vol. 76, No. 13, July 1, 2004 3777

molecule, the mass transport resistances are greatly increased, causing an expected concomitant decrease in RR. Typical microdialysis RR values for 10 kDa or larger proteins across 100kDa molecular weight cutoff (MWCO) membranes range between 1 and 5% at flow rates of 0.5 and 1.0 µL/min.7 An additional challenge to microdialysis sampling of proteins as well as peptides is the often low concentration of these analytes in vivo. This combination of low analyte concentration and low RR requires either very sensitive methods8 or methods which can preconcentrate the sample prior to analysis.9 Furthermore, many proteins and peptides are inherently hydrophobic and tend to adsorb to the polymeric materials used for the membrane and outlet tubing. These nonspecific adsorption problems can be overcome by addition of bovine serum albumin (BSA) to the perfusion fluid.10 Cytokines are regulatory proteins (8-80 kDa) produced by different cells during inflammatory responses.11,12 Cytokines act via a complex network of interactions and are vitally important to many physiological functions. Cytokines are potent molecules that are active at picomolar to femtomolar concentrations. Each cytokine can modulate the synthesis or the actions of the others in a tightly regulated and complex network. Consequently, the makeup of the cytokine milieu is often of greater importance than a single cytokine concentration. The production of these potent molecules is transient, many lasting just a few hours to several days. Cytokines principally act in an autocrine and paracrine fashion and, thus, are highly localized within a microenvironment. The imbalances in the production of cytokines, particularly those affecting immunoregulation, can have profound effects and are implicated in numerous disease states.13 Thus, sensitive and selective analytical tools that can be applied to the detection and quantification of multiple cytokines are obvious prerequisites for the study of these molecules.13 Because of the tremendous biomedical relevance of cytokines to various human diseases, there has been a great interest in applying in vivo microdialysis sampling techniques to collect these proteins. Commercially available microdialysis probes with 100kDa MWCO are commonly used for in vivo microdialysis of cytokines;14-16 however, the RR across these membranes is quite low, for example, 3% at 1.0 µL/min for a 10-mm probe.15 In addition to this commercially available membrane, an in-house 3000-kDa MWCO microdialysis probe has been described to recover interleukin-1β (IL-1β), IL-6, and nerve growth factor from injured (7) Kjellstro ¨m, S.; Appels, N.; Ohlrogge, M.; Laurell, T.; Marko-Varga, G. Chromatographia 1999, 50, 539-546. (8) Freed, A. L.; Cooper, J. D.; Davies, M. I.; Lunte, S. M. J. Neurosci. Methods 2001, 109, 23-29. (9) Haskins, W. E.; Wang, Z.; Watson, C. J.; Rostand, R. R.; Witowski, S. R.; Powell, D. H.; Kennedy, R. T. Anal. Chem. 2001, 73, 5005-5014. (10) Trickler, W. J.; Miller, D. W. J. Pharm. Sci. 2003, 92, 1419-1427. (11) Dinarello, C. A. Chest 2000, 118, 503-508. (12) Gesualdo, L.; Pertosa, G.; Grandaliano, G.; Schena, F. P. Nephrol. Dial. Transplant 1998, 13, 1622-1626. (13) Thorpe, R.; Wadhwa, M.; Bird, C. R.; Mire-Sluis, A. R. Blood Rev. 1992, 6, 133-148. (14) Mayberry, A. J.; Klitzman, B.; Mathy, J. A.; Levin, L. S.; Phillips, T. M.; Brown, S. A. Surg. Forum 1998, 49, 616-617. (15) Sjo ¨gren, F.; Svensson, C.; Anderson, C. Br. J. Dermatol. 2002, 146, 375382. (16) Riese, J.; Boecker, S.; Hohenberger, W.; Klein, P.; Haupt, W. Surg. Infect. 2003, 4, 11-15.

3778

Analytical Chemistry, Vol. 76, No. 13, July 1, 2004

human brain in vivo.17 The 3000-kDa membrane was capable of achieving in vitro RR of 27.8, 45.2, and 22.2% for IL-1β, IL-6, and nerve growth factor, respectively, at 1.0 µL/min. Other caveats must be considered when microdialysis sampling is applied to protein collection. Microdialysis sampling with large MWCO membranes is often plagued by membrane ultrafiltration that can cause significant fluid loss. To overcome these fluid losses, the osmotic pressure difference between the outside sample space and the dialysate has to be compensated for by adding either dextran-70 or additional protein, such as BSA.10,18 In addition to the problems with fluid loss, the most common analytical method for detecting cytokines is enzyme-linked immunosorbent assay (ELISA).19,20 The limitation of a standard ELISA method is that only a single cytokine can be analyzed per well. Furthermore, typically 100 µL of sample is required for the analysis, which at a moderate microdialysis sampling flow rate of 1 µL/min would require nearly a 2-h sampling time. Thus, when using microdialysis sampling for in vivo protein sampling, there is a severe tradeoff between temporal resolution and the need for high RR. The ability to measure multiple cytokines simultaneously is important in a variety of physiological conditions, because the concentration fluctuations of one cytokine often induce changes in other “networked” cytokines. Flow cytometry coupled with antibody-coated fluorophore-containing microspheres (∼5 to 10 µm o.d.) has been widely applied to multiplexed cytokine analyses.21,22 The bead-based multiplexed ELISAs that are coupled with flow cytometry are highly sensitive and provide a convenient means to simultaneously quantify multiple cytokines in low-volume (e50-µL) samples.23,24,25 In this paper, we demonstrate the use of bead-based flow cytometry with microdialysis sampling. Cytometric beads (7 µm) coated with antibodies specific to different cytokines (Table 1) were included in the microdialysis perfusion fluid to enhance cytokine RR, as shown in Figure 1. The inclusion of the antibodycoated beads increases the diffusive mass transport driving force through the membrane because of the antigen/antibody reaction that occurs inside the microdialysis probe. Increasing mass transport by combining diffusion with a chemical reaction is a well-known concept in chemical engineering.26 This increase in the mass transport driving force causes the microdialysis RR to increase. (17) Winter, C. D.; Iannotti, F.; Pringle, A. K.; Trikkas, C.; Clough, G. F.; Church, M. K. J. Neurosci. Methods 2002, 119, 45-50. (18) Rosdahl, H.; Ungerstedt, U.; Henriksson, J. Acta Physiol. Scand. 1997, 159, 261-262. (19) Meager, A. In Cytokine Cell Biology; Balkwill, F., Ed.; Oxford University Press: Oxford, UK.; 2000; pp 193-206. (20) Punnonen, R.; Punnonen, J. In Methods in Molecular Medicine; Humana Press: Totowa, NJ, 2001; Vol.60, pp 285-291. (21) Vignali, D. A. A. J. Immunol. Methods 2000, 243, 243-245. (22) Kettman, J. R.; Davies, T.; Chandler, D.; Oliver, K. G.; Fulton, R. J. Cytometry 1998, 33, 234-243. (23) Oliver, K. G.; Kettman, J. R.; Fulton, R. J. Clin. Chem. 1998, 44, 20572060. (24) Chen, R.; Lowe, L.; Wilson, J. D.; Crowther, E.; Tzeggai, K.; Bishop, J. E.; Varro, R. Clin. Chem. 1999, 45, 1693-1694. (25) Kellar, K. L.; Kalwar, R. R.; Dubois, K. A.; Crouse, D.; Chafin, W. D.; Kane, B-E. Cytometry 2001, 45, 27-36. (26) Cussler, E. L. Diffusion: Mass Transfer in Fluid Systems; Cambridge University Press: Cambridge, 1984.

Table 1. Physical Properties of Cytokines Studieda,b cytokine

MW (kDa)

tumor necrosis factor-R interferon-γ interleukin-5 interleukin-4 interleukin-2

17.3 15.9 13.1 13.6 17.2

a

active protein (kDa)

dimensions (Å) a, b, c

angles (°) R, β, γ

trimer

51.9

95, 95,117

90, 90, 90

dimer dimer monomer monomer

31.8 26.2 13.6 17.2

43, 80, 82 122, 36, 56 92, 92, 46 56, 40, 34

90, 90, 90 90, 99, 90 90, 90, 90 90, 109, 93

conformation

Cytokine 3-D structure information is from the protein database bank (http://www.rcsb.org/pdb/). b See ref 34.

Figure 1. Schematic representation of conventional microdialysis sampling without trapping and a microdialysis scheme with antibody-coated beads included in the perfusion fluid. The bead darkness represents the fluorescence intensity of single bead population: TNF-R beads have the dimmest fluorescence intensity, and IL-2 beads have the brightest fluorescence intensity.

EXPERIMENTAL SECTION Chemicals. A Mouse Th1/Th2 Cytokine Cytometric Bead Array (CBA) kit from BD Biosciences (San Diego, CA) was used for all experiments. These kits contain five sets of polymeric beads (7-µm) each with distinct fluorescence intensities when suspended in solution. Beads with the same fluorescence intensity are coated with a specific antibody for each cytokine (tumor necrosis factoralpha [TNF-R], interferon-gamma [IFN-γ], interleukin-2 [IL-2], IL4, and IL-5). Phycoerythrin (PE) conjugated detection antibodies also are included in the kit. The kit contains proprietary mixtures of assay diluent and wash buffer. The wash buffer contains a proprietary concentration of phosphate-buffered saline, serum proteins, detergent, and sodium azide. All buffer dilutions were made using distilled deionized water (Nanopure). The cytokine standards, cytokine antibody-coated beads, and detection reagent were stored at 4 °C prior to use. Microdialysis and Sample Preparation. A microdialysis syringe pump with a Bee syringe pump controller (Bioanalytical System Inc., West Lafayette, IN) was used with a 1000-series gastight syringe (Hamilton, Reno, NV). Commercially available CMA/20 10-mm polyethersulfone (PES) microdialysis probes with a 100-kDa MWCO (CMA Microdialysis, North Chelmsford, MA) were used for all experiments. According to the manufacturer, the external diameter of the internal cannula is 350 µm, the membrane internal diameter is 420 µm, and the membrane external diameter is 500 µm. Microdialysis samples were collected at room temperature, which was usually 25.5 °C and was measured daily.

Solutions containing either 1250 or 2500 pg/mL of each cytokine were prepared in the assay diluent solution provided by the manufacturer. A 1-mL portion of the cytokine sample was placed into a 1.5-mL microcentrifuge tube. The microdialysis probe was immersed into this quiescent solution. Microdialysis control experiments were performed using the kit’s wash buffer as the perfusion fluid. The probe was perfused at flow rates ranging from 0.5 to 4.0 µL/min, and at least 30 µL of sample was collected. Collected samples were stored at 4 °C prior to sample analysis. From this sample, 25 µL was used and added to a solution containing 25 µL of mixed cytokine bead solution and 25 µL of PE detection reagent for 2 h, as per the manufacturer’s instructions. For microdialysis experiments that included the antibodycoated beads in the perfusion fluid, the mixed cytokine beads were diluted with different volumes of the wash buffer (v:v) and placed into the microdialysis syringe. Depending upon the dilution factor of mixed cytokine beads, either 50 or 100 µL of dialysate was collected for subsequent dilution (v:v) to ratios of 1:1 and 1:4. The syringe was agitated at 7-min (1:1 dilution) or 10-min time intervals (1:4 dilution) to keep the beads in suspension during microdialysis sampling. All samples were stored at 4 °C prior to sample analysis. Dialysates (50 µL) containing 1:1 (v:v) dilutions of beads and wash buffer were directly analyzed. For the 1:4 dilution, 100 µL of the bead-containing dialysate was centrifuged at 200g for 5 min, and 50 µL was removed from the supernatant. The remaining 50µL solution containing the cytokine beads was incubated with 25 Analytical Chemistry, Vol. 76, No. 13, July 1, 2004

3779

µL of PE detection reagent for 2 h while protected from light. After incubation, 1 mL of wash buffer was added, and the samples were vortexed and centrifuged at 200g for 5 min. After careful aspiration of the supernatant from each test tube, 300 µL of wash buffer was added to resuspend the beads prior to flow cytometric analysis. Cytokine concentrations in the sample medium were determined in duplicate via splitting of a 60-µL aliquot taken prior to and after in vitro microdialysis sampling. The average cytokine concentration between these duplicate samples was taken as the actual cytokine concentration in the sample medium for the purposes of microdialysis RR calibration. All the microdialysis experiments (both control and those containing antibody-coated beads in the perfusate) were performed in triplicate at each flow rate using one probe. Ultrafiltration Determination. The extent of ultrafiltration through the microdialysis probes was determined by delivering the different bead dilutions through the probe at 1.0 µL/min for 30 min and then directly weighing the collected dialysate on an Ohaus Analytical Plus electronic balance sensitive to 0.00001 g (Ohaus Corporation, Florham Park, NY). The density for the different bead dilutions was determined to be the same. Flow Cytometry. A FACScan flow cytometer (BD Immunocytometry Systems, San Jose, CA) equipped with a 488-nm laser capable of detecting and distinguishing fluorescence emissions at 576 and 670 nm was used for multiplexed cytokine analysis. BD CellQuest data acquisition software that included the BD CBA software was used. The instrument setup was carried out according to BD Pharmingen CBA instructions. Cytokine standards included one negative control (0 pg/mL cytokines) and additional standards ranging in concentration between 20 and 5000 pg/mL. Mass Action Control Experiments. The extent of enhancement due to overall mass collected was determined by adding different volumes of a 156 pg/mL cytokine standard to a set volume of bead solution. To 25 µL of mixed cytokine bead solution, 25, 100, or 225 µL of the 156 pg/mL cytokine standard was added. These solutions were allowed to incubate at room temperature for 30 min. After 30 min, the solutions containing 100 and 225 µL of cytokine standard were centrifuged. From the two samples, 75 and 200 µL of the supernatant was removed, leaving 50 µL of beads and buffer. These samples then were prepared as the others according to the manufacturer’s directions, that is, addition of secondary antibody, etc. Cytokine Stability. Cytokine standards ranging between 20 and 5000 pg/mL were prepared and stored for 1, 4, and 5 days at 4 °C. On the fifth day, all the samples were measured by the flow cytometry assay. Standards in this range were determined to be stable for up to 5 days. RESULTS AND DISCUSSION Microdialysis Sampling of Proteins. Microdialysis sampling of proteins coupled with quantitation is difficult due to the small diffusion coefficients associated with proteins. Since diffusive mass transport has to occur through three separate regionsssample, membrane, and dialysatesthe additive properties of the mass transport resistances causes low RR for many proteins. This difficulty is illustrated in Figure 2 for the control dialysis samples of the five different cytokines collected in the range between 0.5 and 4 µL/min across the CMA 100-kDa MWCO PES microdialysis probe. The control RR at 0.5 µL/min for the five cytokines studied, 3780 Analytical Chemistry, Vol. 76, No. 13, July 1, 2004

Figure 2. Microdialysis control relative recovery changes of cytokines with the perfusion flow rates (without trapping). Cytokine sample medium concentration was prepared to be 2500 pg/mL from the cytokine standards. (2) IL-5, (b) IFN-γ, ([) IL-2, (9) TNF-R, and (1) IL-4; n ) 3 from each flow rate.

IL-2, IL-4, IL-5, IFN-γ, and TNF-R, ranged between ∼2 and 18%. These RR values are within the range of previously reported RR values (1-10%) among different cytokines for the commercially available 100-kDa CMA membrane.10,15,16,27 It is interesting to note that the obtained RR does not correlate with cytokine molecular weight. We initially thought the cytokine standards might only contain the monomeric form of each cytokine. Gel electrophoresis was used to confirm the molecular weight of each of the five cytokines in the standard, and cytokines such as TNF-R are, in fact, in their trimeric form (data not shown). Differences in RR between the cytokines may be due more to their tertiary structure than their molecular weight. Each cytokine has a different tertiary structure and a different hydrodynamic radius, as shown in Table 1. For the control studies, the 0.5 µL/min flow rate gives the highest RR values. Although lower perfusion fluid flow rates could be used, they would require significantly longer sample times. A flow rate of 0.5 µL/min requires a 1-h sampling period to obtain a sufficient volume for the CBA detection method. With this low perfusion flow rate, RR values of 126.2b 2.2 ( 0.2 28.3 ( 3.1 48.1 ( 10.6 1.2 ( 0.1 9.2 ( 0.5 18.1 ( 4.1 16.6 ( 2.2 103.2 ( 1.3 110.4 ( 7.5 5.8 ( 1.0 33.6 ( 1.8 30.3 ( 4.1

4.3 ( 1.0 90.8 ( 5.6 71.7 ( 0.8 1.3 ( 0.2 17.4 ( 1.7 14.8 ( 1.1 0.7 ( 0.1 5.1 ( 0.4 5.6 ( 0.2 8.5 ( 2.3 78.3 ( 4.8 93.2 ( 1.2 3.1 ( 0.7 17.8 ( 2.3 18.0 ( 0.7

1.5 ( 0.04 8.1 ( 2.4 22.0 ( 3.1 0.5 ( 0.1 2.1 ( 1.0 4.7 ( 1.2 0.3 ( 0.1 0.6 ( 0.2 1.3 ( 0.2 2.7 ( 0.2 12.4 ( 4.2 23.9 ( 1.3 1.0 ( 0.1 2.9 ( 0.4 3.7 ( 0.2

0.90 ( 0.10 4.8 ( 3.8 5.5 ( 0.4 0.3 ( 0.1 1.7 ( 1.9 0.8 ( 0.1 0.2 ( 0.1 0.5 ( 0.3 0.4 ( 0.1 1.6 ( 0.2 3.6 ( 1.4 6.8 ( 0.7 0.5 ( 0.1 1.9 ( 1.2 1.1 ( 0.1

a Cytokine sample medium concentration was prepared to be 2,500 pg/mL from the cytokine standards. b For TNF-R, the relative recovery should be much larger than 126.2% since the concentration was out of range (>5000 pg/mL) for 1:4 dilution at 0.5 µL/min.

large difference in the concentration between the interior of the dialysis probe and the external concentration causes a large shift in the overall concentration gradient as compared to control microdialysates. This, in turn, causes the significant enhancement in RR. Microdialysis RR is affected by the perfusion fluid flow rate. Higher RR is generally obtained at lower perfusion fluid flow rates.30 It would be expected that RR enhancement is dependent on flow rate as well as the dilution factor for the antibody-coated bead solution. Table 3 shows the microdialysis RR for a range of flow rates between 0.5 and 4.0 µL/min and for different bead dilution factors. The RR for control and bead-containing perfusion fluids shows an increase in RR as the perfusion fluid flow rate decreases. The largest enhancements in RR were observed for the 0.5 and 1.0 µL/min flow rates. This flow rate dependence is similar to that observed with cyclodextrin enhancement.31 As the perfusion fluid flow rate increases, the residence time in the membrane “window” decreases, and the wall of the membrane is swept at such a rate that the effective analyte concentration is nearly zero. Thus, having a chemical reaction in the fluid with such rapidly flowing fluid causes less enhancement than with more slowly moving fluids. With lower perfusion fluid flow rates, the cytokines have more time to interact with the antibody and more time for cytokine to diffuse through the microdialysis membrane. The cytokine RR enhancement is also a function of the bead density in the microdialysis perfusion fluid. Although higher RR was obtained for the 1:4 (v:v) bead dilution, as compared to the 1:1 (v:v), it is important to note that more volume had to be collected for the 1:4 dilution. For dilutions greater than 1:1, it is necessary to centrifuge and then resuspend the beads so that the volumes are consistent with those used for the standards. Although 25 µL of sample was collected for the controls, and 50 µL of sample for the 1:1 dilution, the 1:4 dilutions required at least 100 µL of sample to recover a sufficient number of beads to count in the flow cytometer. This means that at least four times the mass of cytokine had been collected, as compared to the controls. Even

though the 1:4 dilution gave higher cytokine concentrations, these samples require longer collection times as well as additional manipulations during the centrifugation process. The cytokine RR enhancement was greatest for TNF-R, IFNγ, and IL-4. These high enhancements do not correlate with the control RR values for these cytokines. IL-4 consistently exhibited the highest control RR, followed by TNF-R and IL-2. Although it is expected that diffusion would play a role in the RR enhancement process, it is not entirely clear why the cytokines exhibit such unpredictable microdialysis control RR values. This may suggest that diffusion through the membrane pores is affected by membrane polymeric domain/protein interactions or a hindered diffusion effect, possibilities that are under investigation. Additionally, the kinetics of physically orienting and ultimately binding the cytokine to the antibody may play a role in the enhancement process, as well. One could envision that the large beads may roll along the inner lumen of the membrane fiber and “pluck” cytokine molecules out of the external milieu as they move through the membrane’s pores. In addition to diffusive arguments, the kinetic “on” rate for each cytokine/antibody pair is not known. What is known is that binding curves (standard curves) among all five cytokines are generally quite similar, thus suggesting that the equilibrium constants for the binding processes are similar among all five cytokines. A commonly neglected aspect of microdialysis sampling is that higher overall mass recoveries are often obtained at higher flow rates (maximal at 10 µL/min).32 Higher mass recovery at higher flow rates may be advantageous with the beads since the antibodycoated beads can be used as a preconcentration step. Figure 3a-e shows the mass recovery (picograms per minute) of five cytokines for different perfusion flow rates. The mass recovery is closely related to the antibody concentration (cytokine antibody-coated bead density) in the perfusion fluid. The mass recovery for 1:1 beads in the perfusion fluid was significantly higher than the mass recovery for 1:4 beads; however, the mass recovery for both 1:1 and 1:4 beads was significantly higher than the controls shown

(30) Johnson, R. D.; Justice, J. B., Jr. Brain Res. Bull. 1983, 10, 567-571. (31) Khramov, A. N.; Stenken, J. A. Analyst 1999, 24, 1027-1033.

(32) Scott, D. O.; Bell, M. A.; Lunte, C. E. J. Pharm. Biomed. Anal. 1989, 7, 1249-1259.

3782 Analytical Chemistry, Vol. 76, No. 13, July 1, 2004

Figure 3. (a-e) Mass recovery enhancement changes for five cytokines vs perfusion flow rates for different antibody-coated bead dilutions; n ) 3 for each flow rate. (2) 1:4, (b) 1:1, and (9) control. The sample concentration was 2500 pg/mL. The insets show the control sample mass recovery using the same axis label.

Analytical Chemistry, Vol. 76, No. 13, July 1, 2004

3783

in the insets. There were some unexpected differences in overall mass recovery of individual cytokines. For control experiments, the mass recovery does not show a consistent trend. The minimal mass recovery was observed at a flow rate of 2.0 µL/min for all five cytokines. In most cases, the maximal mass recovery was obtained at 1.0 µL/min, except for TNF-R (maximal mass recovery at 0.5 µL/min). The mass recovery increased as the perfusion flow rate decreased for dialysates that contained the antibody-coated beads in the perfusion fluid. The mass recovery decreased when the perfusion flow rate decreased from 1.0 to 0.5 µL/min for all the cytokines except IFN-γ and IL-5. For IFN-γ and IL-5, the mass recovery kept increasing when flow rate changed from 1.0 to 0.5 µL/min with the 1:4 antibody-coated beads as trapping agents in the perfusion fluid. The observed maximums obtained for the mass recovery are consistent with other microdialysis protein relative recovery data. Data from ref 7 were converted to mass recovery, and for different proteins, including insulin (5.7 kDa), cytochrome c (12.4 kDa), ribonuclease A (13.7 kDa), and lysozyme (14.4 kDa), the mass recovery was at a maximum at 1.0 µL/min for the flow rates used, which were 1, 3, and 5 µL/min. This is in contrast with our own small-molecule data as well as small molecule data observed by others. Using our data published in ref 33 for a cyclodextrinmediated approach, we found that flow rates above 3.0 µL/min exhibit a plateau effect for mass recovery that is consistent with previously published reports.32 These differences between small and large molecule microdialysis mass recoveries clearly suggest that diffusion plays a vital role in microdialysis mass transport. Mass Action Control. As mentioned above, when the antibodycoated beads were included in the microdialysis perfusion fluid, a greater volume of sample was needed. For controls, 25 µL dialysate was collected, whereas 50 and 100 µL samples were collected for 1:1 and 1:4 antibody-coated bead dialysates, respectively. Since a higher volume of sample was used for the dilutions, it is possible that enhancement will be partially due to a preconcentration of the sample. To determine the volume and mass effect on the cytokine concentrations, different volumes of a known concentration were added to 25 µL of the cytokine beads. A 156 pg/mL cytokine standard was chosen, since complete preconcentration of a 250-µL sample to 25 µL would be expected to give a 1560 pg/mL concentration, a value within the detection range of the cytokine standard curve (20-5000 pg/mL). The mass action control data are shown in Figure 4. The concentration determined using the same bead volume (25 µL of mixed cytokine beads) varied from 2.8× (TNF-R) to 4.0× (IL-5) when the cytokine standard volume increased from 25 to 100 µL. More apparent is the change in cytokine concentration when the cytokine standard volume was increased from 25 to 225 µL. The concentration determined by the same bead volume varied from 3.7× (TNF-R) to 6.3× (IL-5). The highest microdialysis RR enhancement achieved by using 1:1-diluted cytokine beads as the trapping agent in the microdialysis perfusion fluid was for TNF-R (∼12-fold at flow rate 1.0 µL/min), and the lowest was observed for IL-5 (∼4-fold). TNF-R showed the highest microdialysis RR enhancement, while it exhibited the lowest increase in the mass action control experi(33) Ao, X.; Stenken, J. A. Analyst 2003, 128, 1143-1149. (34) Fitzgerald, K. A.; O’Neill, L. A. J.; Gearing, A. J. H.; Callard, R. E. The Cytokine FactsBook, 2nd ed.; Academic Press: New York, 2001.

3784 Analytical Chemistry, Vol. 76, No. 13, July 1, 2004

Figure 4. Mass action control experiments were performed using 156 pg/mL cytokine standard samples at different sample volumes (25, 100, and 225 µL) to incubate with 25 µL of mixed cytokine beads (n ) 3).

ment. IL-5 showed the highest increase during mass action control experiments but the lowest microdialysis RR enhancement. All these data suggest that no direct relationship currently exists between the volume effect (material mass) and the microdialysis RR enhancement of cytokines using cytokine beads as trapping agents. CONCLUSIONS The extremely low microdialysis RR of proteins that has been a shortcoming of the microdialysis sampling technique has been overcome in an in vitro system. Cytokine antibody-coated beads were used as trapping agents in the perfusion fluid to significantly enhance the in vitro microdialysis RR of cytokines. These dialysates can be readily quantified using a standard flow cytometer. Antibodies specific for a particular cytokine on the bead surface rapidly trap cytokine molecules as they traverse the microdialysis membrane as a function of high affinity between antigen and antibody. This trapping effect dramatically increases the concentration gradient and, thus, diffusion driving force within the system. In turn, the mass transport of cytokines through the polymeric dialysis membrane increases, causing an enhanced microdialysis RR of cytokines. Five cytokines, including TNF-R, IFN-γ, IL-5, IL-4, and IL-2, were used as target small proteins and 4-12-fold microdialysis RR enhancement was achieved. This work suggests that microdialysis sampling coupled with antibody-coated beads may be a way to detect low concentrations of these important proteins in vivo. ACKNOWLEDGMENT We gratefully acknowledge the support from NSF 99-84150 (J.A.S.), NIH EB001441 (J.A.S.), the RPI Office of Research SEED funding program (J.A.S.), Arthritis Foundation Investigator Award (T.J.S.), and an Infectious Diseases Society of America OrthoMcNeil Pharmaceutical Young Investigator Award (T.J.S.). We thank our colleague, Professor Dan Loegering, Albany Medical College, for helpful comments about the manuscript and the staff of the Albany Medical College Flow Cytometry Core Facility. Note Added after ASAP Posting. In the version posted on May 18, eq 1 was incorrect. The corrected version was posted on May 28, 2004. Received for review December 24, 2003. Accepted March 25, 2004. AC035536S