In-Capillary Screening of Matrix Metalloproteinase Inhibitors by

Dec 13, 2010 - A capillary electrophoresis-based method with enzymatic reaction inside the capillary for the screening of matrix metalloproteinase (MM...
0 downloads 0 Views 470KB Size
Download PDF
Anal. Chem. 2011, 83, 425–430

In-Capillary Screening of Matrix Metalloproteinase Inhibitors by Electrophoretically Mediated Microanalysis with Fluorescence Detection Xin Hai, Xu Wang, Mohamed El-Attug, Erwin Adams, Jos Hoogmartens, and Ann Van Schepdael* Laboratory for Pharmaceutical Analysis, Faculty of Pharmaceutical Sciences, Katholieke Universiteit Leuven, Leuven, Belgium A capillary electrophoresis-based method with enzymatic reaction inside the capillary for the screening of matrix metalloproteinase (MMP) inhibitors has been developed. MMP-2 and MMP-9, which have been considered as promising targets for cancer therapy, were selected as the model enzymes. The hydrolysis of a fluorogenic substrate catalyzed by MMPs was determined by measuring the increase in fluorescence. For high-throughput screening, the short-end injection was employed. The enzyme, substrate containing inhibitors, and enzyme solutions were injected from the outlet of the capillary via the sandwich mode. They were mixed by alternating the potential at positive and negative polarities. Online hydrolysis, separation, and detection were achieved in 70 s with approximately 0.87 fmol of MMP required for each assay. Theapplicabilityofelectrophoreticallymediatedmicroanalysis (EMMA) with fluorescence detection to estimate the inhibitory mechanism and to determine the IC50 values was evaluated for two natural inhibitors, epigallocatechin gallate and oleic acid. A few other natural compounds such as resveratrol, quercetin, caffeic acid, glucosamine, and doxycycline were also screened to test their inhibitory potency. The results obtained were compared with those obtained by offline enzyme assay and confirm the effectiveness of the present method. A rapid, cost-effective, and fully automated method for MMP inhibitor screening is proposed. Matrix metalloproteinases (MMPs) are a multigene family of zinc-dependent endopeptidases that collectively degrade all components of the extracellular matrix. They play an important role in tissue remodeling associated with various physiological and pathological processes. MMP activity is regulated by endogenous tissue inhibitors of metalloproteinases. Deregulation of the balance between MMPs and their inhibitors may cause excessive degradation of extracellular matrix, allowing the invasion, metastasis, and angiogenesis of tumor cells. As a result, MMPs have become biomarkers and potential therapeutic targets in human cancer.1 Preclinical results of MMP suppression in cancer models were * Corresponding author: Laboratory for Pharmaceutical Analysis, K. U. Leuven, O&N2, Postbox 923, Herestraat 49, 3000 Leuven, Belgium. Tel: +32.16.323443. Fax: +32 0.16.323448. E-mail: ann.vanschepdael@ pharm.kuleuven.be. (1) Roy, R.; Yang, J.; Moses, M. A. J. Clin. Oncol. 2009, 27, 5287–5297. 10.1021/ac1027098  2011 American Chemical Society Published on Web 12/13/2010

so compelling that synthetic MMP inhibitors were brought into clinical testing. However, they eventually failed because of severe side effects caused by lack of selectivity.2 It was later discovered that some MMPs have been implicated in tumor progression (targets), but some MMPs also serve as host defense against tumors (antitargets). This dual role of MMPs in cancer explains in large part the failure of MMP inhibitors in clinical trials.3 It is therefore important to assess the inhibitory potency of test compounds against the specific MMP targets. Several procedures for assays of MMPs and their inhibitors have been described.4,5 Zymography is particularly useful to detect the activity of different MMP isoenzymes in biological samples.6 This technique is specific and highly sensitive, even more sensitive than fluorometric methods using either fluorescein isothiocyanate casein or fluorogenic peptides. However, it is not suitable for rapid screening of MMP inhibitors due to its complicated protocol.7 The fluorescence-based assays, such as fluorescent resonance energy transfer and fluorescence polarization, have become the most used techniques for screening MMP inhibitors.8,9 Routine procedures for performing such assays employ fluorogenic substrates, 96well microtiter plates, and compatible detection devices, with assay volume of 0.05-0.2 mL and analysis time of 30-60 min.10,11 They allow the continuous monitoring of activity and are thus suitable for mechanistic studies, but when they are used in highthroughput screening, the cost due to the consumption of samples is considerable. Moreover, being homogeneous methods without any isolation step, all fluorescent species in the sample will contribute to the total fluorescent signal. Chromatographic enzyme assay measures product formation by separating the reaction mixture into its components. It is usually done by offline incuba(2) Coussens, L. M.; Fingleton, B.; Matrisian, L. M. Science 2002, 295, 2387– 2392. (3) Overall, C. M.; Kleifeld, O. Nat. Rev. Cancer 2006, 6, 227–239. (4) Cheng, X. C.; Fang, H.; Xu, W. F. J. Enzym. Inhib. Med. Chem. 2008, 23, 154–167. (5) Lombard, C.; Saulnier, J.; Wallach, J. Biochimie 2005, 87, 265–272. (6) Snoek-van Beurden, P. A. M.; Von den Hoff, J. W. BioTechniques 2005, 38, 73–83. (7) Quesada, A. R.; Barbacid, M. M.; Mira, E.; Fernandez-Resa, P.; Marquez, G.; Aracil, M. Clin. Exp. Metastasis 1997, 15, 26–32. (8) Fields, G. B. Methods Mol. Biol. 2001, 151, 495–518. (9) Antczak, C.; Radu, C.; Diaballah, H. J. Biomol. Screening 2008, 13, 285– 294. (10) Peppard, J.; Pham, Q.; Clark, A.; Farley, D.; Sakane, Y.; Graves, R.; George, J.; Norey, C. Assay Drug Dev. Technol. 2003, 1, 425–433. (11) Marcotte, P. A.; Davidsen, S. K. In Current Protocols in Pharmacology; John Wiley & Sons, Inc.: New York, 2001; pp 3.7.1-3.7.13.

Analytical Chemistry, Vol. 83, No. 1, January 1, 2011

425

tion of enzyme and substrate and followed by HPLC analysis. Recently, an online chromatographic method using an immobilized enzyme reactor for screening of MMP inhibitors has been reported.12 MMP-9 was immobilized on the monolithic support and inserted into a HPLC system. Ellman’s reagent was added to the mobile phase to react with the hydrolyzed product in order to detect it with a UV detector. The online use of immobilized enzyme reactors typically reduced analysis time and cost but required 2 days to prepare. Capillary electrophoresis (CE) represents an attractive alternative for enzyme assays due to the low consumption of sample and reagent, high-efficiency separations, fast analysis time, and ability to employ several detection techniques.13,14 Besides being a powerful separation technique, CE offers possibilities for incapillary enzyme assays that combine the enzymatic reaction, separation, and detection all in one capillary. The technique to electrophoretically mix zones of reactants on the basis of their different electrophoretic mobilities under the influence of an electric field is commonly referred to as electrophoretically mediated microanalysis (EMMA). EMMA was first described by Bao and Regnier15 and has been successfully applied for studies of enzyme kinetics and enzyme inhibition.16-19 The advantages of EMMA are full automation and miniaturization since nanoliter samples can readily be injected and react. The analysis throughput can be dramatically improved by using multiplex CE.20 However, EMMA with UV detection may not be sensitive enough to analyze small amounts of product, due to the short optical path length and the nanoliter scale of the online reaction. Fluorescence detection is a very useful detection technique for EMMA enzyme assays,21 because it is able to detect extremely low concentrations of analytes in small sample volumes and it also minimizes interference from other components in the sample. Here, we describe a new EMMA method for screening of MMP inhibitors using fluorescence detection. MMP-2 and MMP-9 were selected as model enzymes because they degrade components of basement membranes and are believed to be crucial in tumor invasion. The method was developed based on the hydrolysis of a widely used fluorogenic substrate for MMP, (7-methoxycoumarin-4-yl)acetyl-Pro-Leu-Gly∼Leu-(3-[2,4-dinitrophenyl]-L-2,3-diaminopropionyl)-Ala-Arg-NH2 (Mca-Pro-Leu-Gly∼LeuDpa-Ala-Arg-NH2).22 This synthesized peptide contains a methoxycoumarin (Mca) fluorophore and a dinitrophenyl quencher located on opposite sides of the susceptible peptide bond (∼). Upon hydrolysis, the quencher and the fluorophore become (12) Ma, X.; Chan, E. C. Y. J. Chromatogr. B 2010, 878, 1777–1783. (13) Kostal, V.; Katzenmeyer, J.; Arriaga, E. A. Anal. Chem. 2008, 80, 4533– 4550. (14) Kraly, J.; Fazal, M. A.; Schoenherr, R. M.; Bonn, R.; Harwood, M. M.; Turner, E.; Jones, M.; Dovichi, N. J. Anal. Chem. 2006, 78, 4097–4110. (15) Bao, J. M.; Regnier, F. E. J. Chromatogr. A 1992, 608, 217–224. (16) Hai, X.; Konecny, J.; Zeisbergerova, M.; Adams, E.; Hoogmartens, J.; Van Schepdael, A. Electrophoresis 2008, 29, 3817–3824. (17) Hai, X.; Adams, E.; Hoogmartens, J.; Van Schepdael, A. Electrophoresis 2009, 30, 1248–1257. (18) Belenky, A.; Hughes, D.; Korneev, A.; Dunayevskiy, Y. J. Chromatogr. A 2004, 1053, 247–251. (19) Fan, Y.; Scriba, G. K. E. J. Pharm. Biomed. Anal. 2010, 53, 1076–1090. (20) Ma, L. J.; Gong, X. Y.; Yeung, E. S. Anal. Chem. 2000, 72, 3383–3387. (21) Regnier, F. E.; Patterson, D. H.; Harmon, B. J. TrAC, Trends Anal. Chem. 1995, 14, 177–181. (22) Knight, C. G.; Willenbrock, F.; Murphy, G. FEBS Lett. 1992, 296, 263– 266.

426

Analytical Chemistry, Vol. 83, No. 1, January 1, 2011

physically separated and the fluorescence of the Mca group can be measured to determine the rate of the reaction. Nowadays, several natural compounds have shown great potential for innovative cancer therapy because they act as selective MMP inhibitors, namely, (-)-epigallocatechin gallate (EGCG), oleic acid (OA), resveratrol, quercetin, caffeic acid, glucosamine sulfate, and doxycycline.23,24 These known MMP inhibitors were therefore utilized in this study to evaluate the effectiveness of the developed method. EXPERIMENTAL SECTION Chemicals. Human recombinant MMP-2 and MMP-9 were purchased from Calbiochem (Merck, Darmstadt, Germany). Fluorogenic peptide substrate Mca-Pro-Leu-Gly∼Leu-Dpa-Ala-ArgNH2 was from R&D Systems (Minneapolis, MN). EGCG, OA, caffeic acid, resveratrol, quercetin, glucosamine sulfate, and sodium fluorescein were purchased from Sigma (St. Louis, MO). Doxycycline hyclate was obtained from the European Pharmacopoeia (Strasbourg, France). Tris(hydroxymethyl)aminomethane (Tris) and Tris-HCl were purchased from AppliChem (Darmstadt, Germany). Dimethyl sulfoxide (DMSO) was from Merck (Darmstadt, Germany). Ethylenediaminetetraacetic acid (EDTA) was purchased from Acros Organics (Geel, Belgium). All solutions were prepared with Milli-Q water (Millipore, Milford, MA). The buffer solutions were filtered through 0.2 µm regenerated cellulose (RC) filters (Whatman, Dassel, Germany). Instrumentation. A HP3DCE instrument (Agilent, Waldbronn, Germany) was used for all electrophoretic experiments. It was equipped with a standard HP cassette containing a fused-silica capillary of 60 cm total length (16.5 cm from the outlet to fluorescence detection window) with 75 µm inner diameter. The capillary was thermostated at 25 °C and coupled to an Argos 250B fluorescence detector (Flux Instruments, Basel, Switzerland) as external detector. The optical principle of the fluorescence detector is illustrated in Figure 1A. After being filtered, the light from a mercury-xenon lamp is led to the capillary by an optical fiber and focused onto the detection window by a ball lens. The fluorescence light emitted by the analyte is trapped inside the capillary and guided along the capillary by total internal reflection. The fluorescence light is optically decoupled from the capillary by use of a quartz cone, leading to a second filter and a photomultiplier by a liquid-core light guide. The excitation wavelength was selected with a broad-band filter (240-400 nm), and the emitted light was measured at >418 nm by use of a cutoff filter. Data acquisition and analysis were performed with Agilent Chemstation software (Hewlett-Packard, Waldbronn, Germany). Solution Preparation and Data Analysis. The background electrolyte (BGE) was 50 mM Tris-HCl adjusted to pH 7.5 with 50 mM Tris. The following sample solutions were used for both online and offline methods. To estimate the inhibitory mechanism and kinetic constants (Km), a stock solution of 300 µM fluorogenic substrate was prepared and diluted to 30, 60, 120, 180, and 240 µM. For IC50 determination, a stock solution of 12 mM (23) Li, N. G.; Shi, Z. H.; Tang, Y. P.; Duan, J. A. Curr. Med. Chem. 2009, 16, 3805–3827. (24) Mannello, F. Recent Pat. Anti-Cancer Drug Discovery. 2006, 1, 91–103.

Figure 1. (A) Optical principle of the Argos 250B fluorescence detector: (1) injection portion, outlet of the capillary; (2) excitation light; (3) quartz cone; (4) emission light; (5) light fiber to the photomultiplier. (B) Schematic illustration of EMMA with short-end injection: (1) inject plugs of enzyme, substrate (containing inhibitor), enzyme solutions, and background electrolyte (BGE); (2) mix substrate and inhibitor with enzyme by switching the polarity; (3) separate the mixture and detect the product.

EGCG was made to produce solutions with concentrations of 3.6 mM, 1.2 mM, 360 µM, 120 µM, 36 µM, 12 µM, and 1.2 µM. For the inhibitor screening experiment, stock solutions of each inhibitor were prepared at 3.6 mM. OA, caffeic acid, quercetin, and resveratrol were first dissolved in DMSO. The final concentration of DMSO never exceeded 1% (v/v). Unless stated otherwise, the stock solution of 120 nM for MMP and 6 µM for internal standard (IS) fluorescein was used throughout. All solutions were made to the desired concentrations with 5 mM Tris buffer (pH 7.5). The percentage of inhibition was calculated according to the equation % inhibition ) [1 - (x/blank)] × 100, where x represents the relative peak area (the corrected area of the product peak divided by the corrected area of the IS peak) of product determined at a given concentration of inhibitor, and the blank is the relative peak area determined without inhibitor being present. The corrected peak area is defined as the peak area divided by the migration time. Prism 5.0 (GraphPad software, San Diego, CA) was used for curve-fitting and Km and IC50 calculations. EMMA with Short-End Injection (Online). A schematic illustration of EMMA with short-end injection for the screening of MMP inhibitors is shown in Figure 1B. The capillary was first filled with the separation buffer. The enzyme, substrate, and inhibitor solutions were hydrodynamically injected into the outlet part of the capillary, that is, enzyme solution (-50 mbar for 3 s), substrate with inhibitor solution (-50 mbar for 3 s), enzyme solution (-50 mbar for 3 s), and a plug of BGE (-50 mbar for 5 s) to guard against loss to the outlet vial. After each injection step, the outlet end of the capillary was dipped into water to prevent sample carryover (0.02 min). Following the injection

sequence, a positive potential of 5 kV for 5 s was initiated across the capillary to electrophoretically mix the substrate and inhibitor with enzyme. Then the polarity was reversed to -5 kV for 5 s to move the reactant plugs apart. This two-step process was repeated one more time. After the mixing step, a voltage of -30 kV was applied to separate the fluorescent product from other components. Between runs, the capillary was rinsed with DMSO and BGE, each for 1.5 min. For the inhibition experiments, a 60 µL mixture containing 10 µL of inhibitor solution, 10 µL of substrate solution, 10 µL of IS, and 30 µL of Tris buffer (5 mM, pH 7.5) was prepared in one mini-CE vial, while 20 µL of MMP solution (20 nM) was kept in another vial. Offline CE Assay. The offline incubation was performed as previously described.25 MMP (120 nM), inhibitor (0-12 mM), and Tris buffer (50 mM, pH 7.5) and IS solutions, 10 µL of each, were mixed and preincubated for 15 min at 25 °C. The hydrolysis was initiated by adding 10 µL of the substrate solution (60 µM) to the mixture. The reaction was allowed to proceed at 25 °C for 20 min and then was stopped by adding 10 µL of EDTA (1 mM). The offline incubation mixture was injected from the outlet side at -50 mbar for 3 s and separated by -30 kV with 50 mM Tris buffer (pH 7.5) as BGE. RESULTS AND DISCUSSION This part looks into the setup needed to investigate MMP inhibition in a fast, selective, and automated way. Fluorescence detection was applied in conjunction with an internally quenched substrate. In view of the analysis speed and the automation aspect, an EMMA system was developed combined with short-end injection and rapid polarity switching. This development work is discussed in the following section. In the subsequent sections, the use of this system is described to explore the inhibition of MMP. EMMA with Fluorescence Detection. It has been reported that MMP displays optimal performance at pH 7.5.8 Therefore, 50 mM Tris buffer (pH 7.5) was used for both the in-capillary enzyme reaction and electrophoretic separation. In this case, MMP cleaves the substrate into a negatively charged Mca-Pro-Leu-GlyOH and a positively charged Leu-Dpa-Ala-Arg-NH2, allowing them to be easily separated from one another due to the significantly different electrophoretic mobilities. This hydrolysis also relieves the quenching of the fluorophore in the intact substrate and increases the fluorescence, which was detected by a lamp-based fluorescence detector. It was observed that using filters instead of a monochromator results in a greater light intensity, thereby improving the instrument sensitivity. A broad-band filter (240-400 nm) was therefore used for excitation, and the emission light was measured by a cutoff filter (>418 nm). EMMA with fluorescence detection developed for in vitro characterization of MMP inhibitors is shown in Figure 1B. Sandwich EMMA was used as an alternative to the classical plug-plug mode, in which a plug of fluorogenic substrate containing inhibitor is injected between plugs of MMP solution. This establishes conditions under which inhibitors, regardless of charge, will mix with the enzyme when the mixing potential is applied. The dual plug arrangement also allows for diffusional (25) Troeberg, L.; Nagase, H. In Current Protocols in Protein Science; Wiley: New York, 2003; pp 21.16.1-21.16.9.

Analytical Chemistry, Vol. 83, No. 1, January 1, 2011

427

Figure 2. Typical electropherograms of MMP-2 online hydrolysis by EMMA with (A) short-end injection and (B) normal-end injection. EMMA conditions: fused-silica capillary, 75 µm i.d. × 60 cm (16.5 cm from outlet to detection window); BGE, 50 mM Tris buffer (pH 7.5) Injections: MMP-2 (-50 mbar, 3 s), substrate (-50 mbar, 3 s), MMP-2 (-50 mbar, 3 s), BGE (-50 mbar, 5 s). Mixing: -5 kV, 5 s; +5 kV, 5 s (two iterations). Fluorescence detection: excitation, 240-400 nm (broad-band filter); emission, > 418 nm (cutoff filter). Separation voltage: - 30 kV. Concentrations: MMP-2, 10 nM; fluorogenic substrate (S), 20 µM; IS, 1 µM. Note: The negative settings used in the short-end injection were reversed to positive in the normalend injection for the injection pressure and applied potential. P ) product.

mixing of any neutral inhibitors that may be present in the sample. A series of positive and negative potentials were applied to the capillary after the final BGE plug. This is referred to as rapid polarity switching (RPS), which has been reported for increasing the efficiency of in-capillary EMMA reactions.26 By alternating the polarities, the backward and forward movements are analogous to mechanical shaking. This improves the quality of mixing by increasing the convection of reactants. After the mixing step, a separation voltage was immediately applied to separate the product and remaining components. For rapid analysis, short-end injection can be employed.27-29 The effective separation length of the capillary was decreased from 43.5 to 16.5 cm by injecting components from the outlet of the capillary. Figure 2A shows a typical electropherogram of EMMA with fluorescence detection for MMP online hydrolysis using short-end injection. The intact substrate has low intrinsic fluorescence and migrates before the fluorescent product. The other product, a peptide, and the enzyme were not detected because they are not fluorescent. An IS, fluorescein was introduced in the substrate solution with the aim to control variations in the injection. The separation was achieved in 70 s. Compared with normal-end injection at the inlet of the capillary (Figure 2B), shortend injection provided 3-4-fold reduction in the analysis time as well as 1.5-2 times increase of the sensitivity. The effect of RPS on the product formation was investigated by increasing the alternation of potentials at both polarities, each (26) Sanders, B. D.; Slotcavage, R. L.; Scheerbaum, D. L.; Kochansky, C. J.; Strein, T. G. Anal. Chem. 2005, 77, 2332–2337. (27) Altria, K. D.; Kelly, M. A.; Clark, B. J. Chromatographia 1996, 43, 153– 158. (28) Van Dyck, S.; Vissers, S.; Van Schepdael, A.; Hoogmartens, J. J. Chromatogr. A 2003, 986, 303–311. (29) Nemec, T.; Glatz, Z. J. Chromatogr. A 2007, 1155, 206–213.

428

Analytical Chemistry, Vol. 83, No. 1, January 1, 2011

Figure 3. Effect of RPS on product formation during online hydrolysis by MMP without an in-capillary incubation period. The plugs were mixed with an alternating potential of 5 kV for 5 s at each polarity with increasing iterations. Zero point: no RPS, after injection of plugs the separation voltage was directly applied. Concentrations: MMP-2, 20 nM; MMP-9, 20 nM; fluorogenic substrate, 10 µM; IS, 1 µM. Other conditions are as in Figure 2.

for 5 kV and 5 s. No in-capillary incubation time was employed. The curve of relative peak areas against the RPS time span is shown in Figure 3. The time on the horizontal axis corresponds to the total duration of electrophoretic mixing by RPS. As can be seen, online hydrolysis takes place during the RPS. The increasing product amount for both MMP-2 and MMP-9 was linear within 20 s. The slope of the linear curve is the initial velocity of the reaction where less than 10% of the substrate has been converted to product. At low substrate depletion (i.e., initial velocity conditions), the factors that contribute to nonlinear progression curves for enzyme reactions do not influence the reaction. That means a steady state for the reaction system is ensured. Therefore, 20 s of RPS time was chosen to perform the reaction under initial velocity conditions and to have reasonable peak size for further inhibition studies. At point zero where no RPS was employed, the separation voltage was immediately applied after the injection of plugs. Some fluorescence product was formed during the passage of the plugs through each other, which indicates the fast kinetics of MMP hydrolysis. Unless stated otherwise, the MMP solution was kept at 20 nM and substrate concentration was 10 µM, so that approximately 0.87 fmol of MMP and 0.22 pmol of substrate are required for each assay. In addition, the influence of BGE concentration on MMP activity was tested as well. No significant effect on MMP activity was observed. The repeatability for migration time and relative peak areas was determined for a set of three replicates. The RSD values were lower than 2.3% for migration time and did not exceed 5.0% for relative peak areas. Evaluation of Inhibition Mechanism and Efficacy. EGCG, a polyphenol catechin isolated from green tea, was first tested by EMMA with fluorescence detection. The overlay electropherograms in Figure 4 illustrate the inhibition effect of EGCG at different concentrations on MMP-9 hydrolysis. As can be observed, the peak area of fluorescent product decreases with increasing EGCG concentrations.

Table 1. Km, IC50, and RSD Values for Relative Peak Area Obtained by Different Methods Km (µM)

EMMA offline CE Fa/Zb a

MMP-2

MMP-9

10.4 7.2 3.0634

6.8 5.2 2.4634/3.335

EGCG IC50 (µM) RSD (%) MMP-2 5.0 2.3

59.7 24.4 631/830

MMP-9 31.5 17.3 0.831/1330

RSD (%) 7.1 2.7

Fluorometric assay. b Zymography; no reference value is available.

Figure 4. Overlay electropherograms for EGCG inhibition of MMP-9 online hydrolysis. Concentrations: EGCG, varied from 0 to 2 mM; MMP-9, 20 nM; fluorogenic substrate, 10 µM; IS, 1 µM. EMMA conditions: fused-silica capillary, 75 µm i.d. × 60 cm (16.5 cm from outlet to detection window); BGE, 50 mM Tris buffer (pH 7.5). Injections: MMP-9 (-50 mbar, 3 s); substrate containing inhibitor (-50 mbar, 3 s); MMP-9 (-50 mbar, 3 s); BGE (- 50 mbar, 5 s). Mixing: -5 kV, 5 s; +5 kV, 5 s (two iterations). Fluorescence detection: excitation, 240-400 nm (broad-band filter); emission, > 418 nm (cutoff filter). Separation voltage: -30 kV.

For kinetic studies, two reference inhibitors, EGCG and OA, were selected as model compounds for MMP-9 and MMP-2, respectively. To estimate the inhibition mechanism and kinetic constant (Km), Michaelis-Menten plots were constructed by injecting different substrate solutions in the range 5-50 µM containing increased inhibitor concentrations 0, 20, and 60 µM. The corresponding Lineweaver-Burk plots for EGCG and OA are shown in Figure 5. The increase in y-intercept (decreased Vmax) and decrease in x-intercept (higher Km) with higher inhibitor concentrations indicates mixed-type inhibition for OA to inhibit MMP-2 (Figure 5A). In the case of EGCG, an increase in y-intercept (decreased Vmax) and an unvaried x-intercept (same Km) at increasing inhibitor concentrations shows that EGCG acts as a noncompetitive inhibitor for MMP-9 (Figure 5B). The patterns of inhibition found for EGCG and OA were consistent with the reported results.30-33 The determined Km values for MMP-2 and MMP-9 online hydrolysis are summarized in Table 1, together with the values reported in literature. EMMA with fluorescence detection was further applied to determine the inhibition efficacy of EGCG for MMP-2 and MMP-

Figure 6. Inhibition curves for EGCG inhibition of MMP-2 and MMP9. Concentrations: EGCG, 200 nM-2 mM; MMP, 20 nM; fluorogenic substrate, 10 µM; IS, 1 µM. Experimental conditions were as in Figure 4.

9. The reactions were performed with increasing EGCG concentrations (200 nM-2 mM) and a fixed substrate and enzyme concentration. The dose-response curves (Figure 6) were obtained by plotting the percentage of inhibition versus the logarithm of EGCG concentration. The inhibitor concentration yielding 50% inhibition is the IC50 value. The IC50 values found were 59.7 µM for MMP-2 and 31.5 µM for MMP-9. This result implies that

Figure 5. Lineweaver-Burk plots for (A) MMP-2 inhibited by OA and (B) MMP-9 inhibited by EGCG. Concentrations: inhibitors, (b) 0, (0) 20, and (2) 60 µM; fluorogenic substrate, 5-50 µM; MMP, 20 nM; IS, 1 µM. Experimental conditions were as in Figure 4. Analytical Chemistry, Vol. 83, No. 1, January 1, 2011

429

the inhibitory effect of EGCG on MMP-2 was weaker than that on MMP-9. The obtained IC50 values as well as values reported in literature are presented in Table 1. The RSD values for triple analyses were lower than 2.7% for migration time and did not exceed 7.1% for relative peak areas. Quantitative Comparison of Online and Offline Methods. Comparison between the values determined by EMMA and other procedures was quite difficult, since IC50 values of 0.8 µM and 13 µM were reported for EGCG inhibition of MMP-9 by zymography.30,31 It was found that different substrates, enzyme sources, and assay conditions such as substrate concentration were employed in the literature. For this reason, offline CE assay was conducted to determine Km and IC50 values with the same enzyme, substrate, and inhibitor concentrations as for EMMA determination. The offline results are listed in Table 1. As can be seen, the Km values determined with the EMMA method were higher than those obtained from the offline CE assay. However, these differences were small. Unlike for the Km values, there is a 2-2.5-fold difference in the IC50 values, with the offline approach reflecting a more potent inhibitory effect. This may be due to the fact that the inhibitors were preincubated with the enzyme in the offline mode. However, in the EMMA method, the inhibitors were dissolved in the substrate plug and both are mixed with the enzyme at the same time. The differences in the IC50 values obtained between the online (flow equilibrium) and offline procedures (static equilibrium) were also observed in other studies.12,36-38 The repeatability of the EMMA method is relatively low with a high RSD value for the relative peak areas. This is probably due to the complex injection procedure of plugs in EMMA. For more accurate quantitative determination, offline CE assay is therefore preferred. Inhibitor Screening. Seven known MMP inhibitorssEGCG, OA, caffeic acid, quercetin, doxycycline, resveratrol, and glucosamineswere used to evaluate the screening potency of the present method. The effect of these inhibitors at 600 µM on both MMP-2 and MMP-9 was compared. The percentage of inhibition for these natural compounds is shown in Table 2. As can be seen, the effect of the inhibitors varies with different MMP isoenzymes. EGCG and OA appear to be the most potent inhibitors, while resveratrol is less effective than the others with activity only for MMP-2, but no activity was detectable for MMP-9. For MMP-2, the inhibition (30) Garbisa, S.; Sartor, L.; Biggin, S.; Salvato, B.; Benelli, R.; Albini, A. Cancer 2001, 91, 822–832. (31) Demeule, M.; Brossard, M.; Page, M.; Gingras, D.; Beliveau, R. Biochim. Biophys. Acta 2000, 1478, 51–60. (32) Berton, A.; Rigot, V.; Huet, E.; Decarme, M.; Eeckhout, Y.; Patthy, L.; Godeau, G.; Hornebeck, W.; Bellon, G.; Emonard, H. J. Biol. Chem. 2001, 276, 20458–20465. (33) Polette, M.; Huet, E.; Birembaut, P.; Maquart, F. X.; Hornebeck, W.; Emonard, H. Int. J. Cancer 1999, 80, 751–755. (34) Olson, M. W.; Gervasi, D. C.; Mobashery, S.; Fridman, R. J. Biol. Chem. 1997, 272, 29975–29983. (35) Ende, C.; Gebhardt, R. Planta Med. 2004, 70, 1006–1008. (36) Kim, H. S.; Wainer, I. W. Anal. Chem. 2006, 78, 7071–7077. (37) Martin-Biosca, Y.; Asensi-Bernardi, L.; Villanueva-Camanas, R. M.; Sagrado, S.; Medina-Hernandez, M. J. J. Sep. Sci. 2009, 32, 1748–1756. (38) Schuchert-Shi, A.; Hauser, P. C. Electrophoresis 2009, 30, 3442–3448.

430

Analytical Chemistry, Vol. 83, No. 1, January 1, 2011

Table 2. Percent Inhibition for Natural Compoundsa % inhibition

a

compound (600 µM)

MMP-2

MMP-9

EGCG OA caffeic acid quercetin doxycycline glucosamine sulfate resveratrol

70.1 83.6 53.7 39.3 21.2 13.4 9.5

87.0 69.1 58.2 34.8 25.7 16.3 0

Conditions were as in Figure 4.

potency ranking was OA > EGCG > caffeic acid > quercetin > doxycycline > glucosamine sulfate> resveratrol, while for MMP-9 the order was EGCG > OA > caffeic acid > quercetin > doxycycline > glucosamine sulfate. The inhibitory activities of these compounds were also determined by offline CE assay. The ranking orders obtained were consistent with the online results. This indicates that EMMA with fluorescence detection is well suited for initial screening of a large number of compounds for MMP inhibition. CONCLUSIONS EMMA with fluorescence detection has proved to be useful for screening and characterization of MMP inhibitors. The advantages over conventional methods are rapid analysis, full automation, and miniaturization. Fluorescence detection successfully improved the sensitivity of EMMA and allows working with even smaller sample amounts. Approximately 0.87 fmol of enzyme was required for each assay, whereas about 2 pmol of enzyme was used for an offline incubation (100 µL sample volume). EMMA integrates the assay procedures and enables system automation, and when it is combined with short-end injection, the screening time can be further reduced. The total analysis time for online hydrolysis, separation, and detection was 70 s as compared to 1 h for the traditional enzymatic assays. For high-throughput screening, the method can easily be transferred to capillary array electrophoresis, for instance, to search for more MMP inhibitors. This method was developed on a general MMP substrate, allowing it to be expanded to assess the activity and selectivity of inhibitors against other MMP isoenzymes. ACKNOWLEDGMENT We gratefully acknowledge Dr. Herbert Funke (Flux Instruments) and Marcel Schulze and Professor Dr. Detlev Belder (Institute of Analytical Chemistry, University of Leipzig) for their kind assistance with the Argos 250 B detector. This work was supported by the Flemish Fund for Scientific Research (Research project G.0681.08). X.H. and X.W. contributed equally to this work. M.N.A. thanks the Libyan General Peoples Committee (Ministry) for Education and Scientific Research for a scholarship. Received for review November 23, 2010. AC1027098

October

13,

2010.

Accepted