- Email: [email protected]
1H NMR spectroscopic evidence of interaction between ibuprofen and lipoproteins in human blood plasma
ANALYTICAL BIOCHEMISTRY Analytical Biochemistry 324 (2004) 292–297 www.elsevier.com/locate/yabio
1
H NMR spectroscopic evidence of interaction between ibuprofen and lipoproteins in human blood plasma
Yunhuang Yang, Guoyun Bai, Xu Zhang, Chaohui Ye, and Maili Liu* State Key Laboratory of Magnetic Resonance and Atomic and Molecular Physics, Wuhan Institute of Physics and Mathematics, Chinese Academy of Sciences, Wuhan 430071, PeopleÕs Republic of China Received 5 September 2003
Abstract Recent studies have suggested that ibuprofen inhibits low-density lipoprotein oxidation in a high dose-dependent manner and is a promising drug for treatment of the conditions associated with atherosclerosis. In this article, we present the NMR spectroscopic evidence for the interaction between ibuprofen and phospholipids in lipoprotein particles in intact human plasma. Ibuprofen caused chemical shift upfield drifts for the protons of –Nþ (CH3 )3 moieties of phosphatidylcholine and sphingomyelin, olefinic chains (– CH@CH–, –CH@CHCH2 CH@CH–, –(CH2 )n CH2 CH@), and (CH2 )n and CH3 groups, from unsaturated lipids in lipoprotein particles. The ibuprofen may interact directly with the above-mentioned groups of phospholipids or induce structural changes in the lipoproteins. This may shed light on the mechanism by which the drug protects against oxidative modification of lipoproteins. Ó 2003 Elsevier Inc. All rights reserved. Keywords: Nuclear magnetic resonance; Ibuprofen; Lipoproteins; Unsaturated lipids; Phosphatidylcholine; Sphingomyelin
The lipoprotein particles in blood plasma or serum are classified by their size and density as very low density lipoprotein (VLDL),1 low-density lipoprotein (LDL), and high-density lipoprotein (HDL) [1]. They consist largely of a lipid core of nonpolar triacylglycerides and cholesteryl esters surrounded by more polar phospholipids, cholesterol, and apoproteins. VLDL and LDL, specifically oxidized LDL, are positively associated with coronary heart disease (CHD) and thus are considered atherogenic lipoproteins, whereas HDL is antiatherogenic because of its negative association with CHD [1,2]. Ibuprofen, 2-(4-isobutylphenyl)propionic acid, is a wellknown nonsteroidal anti-inflammatory drug. Recent studies have suggested that ibuprofen is a promising drug for treatment of the conditions associated with atherosclerosis, because of its novel properties [3–6].
*
Corresponding author. Fax: +86-27-8719-9291. E-mail address: [email protected] (M. Liu). 1 Abbreviations used: VLDL, very low density lipoprotien; LDL, low-density lipoprotien; HDL, high-density lipoproteins; CHD, coronary heart disease; PFG, pulsed field gradient; BP-LED, bipolar gradient longitudinal eddy-current delay; PC, phosphatidylcholine; SM, sphingomyelin. 0003-2697/$ - see front matter Ó 2003 Elsevier Inc. All rights reserved. doi:10.1016/j.ab.2003.10.002
These properties include: (1) inhibition of adhesion and transendothelial migration of leukocytes; (2) suppression of intracellular production of reactive oxygen species and oxidative modification of LDL; and (3) increment of HDL cholesterol levels and reduction of triglycerides. In addition, ibuprofen inhibited LDL oxidation in a high dose-dependent manner over the concentration range 0.1 to 2.0 mM (100 lg/mL LDL) [4]. However, the mechanism of the interaction between ibuprofen and lipoproteins in human blood remains unclear. Diffusion-based NMR techniques have been incorporated into a variety of methods and are widely used in pharmaceutical studies, such as screening potential bioactive ligands [7–14], studying ligand–receptor interaction [14–17], determining structures and orientations of bound ligands [15], and measuring the diffusion coefficients of small metabolites and proteins in blood plasma [17–21]. Molecular self-diffusion can cause attenuation of signal intensity in pulsed-field-gradient (PFG) NMR experiments [22,23]. At a suitable gradient strength, intensities of NMR peaks from the small molecules are reduced substantially due to their relatively fast self-diffusion and the remaining signals are
Y. Yang et al. / Analytical Biochemistry 324 (2004) 292–297
mainly from the larger molecules in complex biological samples, such as blood plasma [17–21]. This enables direct characterization of lipoproteins [19,20] and provides the possibility of studying the interaction between the proteins and drugs in blood plasma. In this article, a PFG NMR approach was used to study the interaction between ibuprofen and lipoproteins in human blood plasma without physical separation. We employed a high concentration of ibuprofen to saturate its binding sites on the serum albumin and to ensure sufficient ibuprofen to interact with the lipoproteins. Here the evidence of the interaction between ibuprofen and lipoproteins is presented.
Materials and methods Human blood was obtained from a healthy male volunteer and the plasma was prepared conventionally. The blood plasma was divided into four parts. One of the samples was used for the control. A certain amount of ibuprofen sodium salt (Sigma, Poole, Dorset, UK) was directly added into the other samples. The final concentrations of ibuprofen were 0, 5.0, 20.0, and 40.0 mM, respectively. Ten percent D2 O was added into the samples and used for the NMR magnetic field frequency lock. The samples were stored in a refrigerator (4 °C) before use. For convenience, the structure and numbering system of ibuprofen are given in Scheme 1. The NMR experiments were carried out on a Varian INOVA500 spectrometer operating at 1 H frequency of 500.12 MHz at 25 °C. The machine is equipped with a PFG accessory capable of delivering z-field gradients up to 60 G/cm. One-dimensional (1D) spectra were obtained conventionally with solvent suppression using W5 sequence [24]. The bipolar gradient longitudinal eddy-current delay (BP-LED) [25] method was used to obtained diffusion-weighted NMR spectra and to measure the diffusion coefficients. The solvent suppression pulse was incorporated into the BP-LED sequence [17– 19]. For diffusion coefficient measurement, the gradient strength was increased linearly from 10 to 55 G/m in 16 steps. The other experimental parameters were unchanged. Sixty-four transients were collected into 32,000 complex data points covering a spectral width of 8000 Hz. A cosine (0–p=2) window function was applied
Scheme 1. The structure and numbering system of ibuprofen.
293
to the time domain data before Fourier transformation to enhance signal-to-noise ratio. The areas of the NMR peaks were used to derive the diffusion coefficients [19] using the tow-parameter equation AðbÞ ¼ Að0Þ expðbDÞ, where b ¼ ð2cdGÞ2 ðD d=3Þ, c is the 1 H gyromagnetic ratio, d (2 ms) and G (16 values) are the duration and strength of the gradient pulse, D (100 ms) is the effective diffusion time, AðbÞ and Að0Þ are the NMR signal areas in the presence and absence of the gradient pulses, and D is the diffusion coefficient. For COSY experiments, 2048 (F2) 256 (F1) complex data points were acquired using conventional pulse sequence and with 16 transients for each increment. The spectral width of 8000 Hz was used for both dimensions. The time domain data were weighted using a sine-bell window function and zero-filled to 2048 (F2) 512 (F1) before Fourier transformation. The chemical shift was referred to H1 (d5.23) of a-glucose.
Results and discussion Fig. 1 shows the 1D 1 H NMR spectra of intact human blood plasma in the absence (a) and presence (b–d) of ibuprofen. Tow regions (d3:30–d2:95, d1:45–d0:70) were enlarged to show the ibuprofen-induced chemical shift changes. The broad resonances of ibuprofen (labeled with prefix I) indicated that the drug has been bound to the serum proteins. Serum albumin is the major ibuprofen-binding protein in blood plasma [16,26,27]. Considering the albumin concentration is about 0.7 mM in human blood [27] and its large binding capacity for ibuprofen (33 binding sites, Kd of 1.4 mM) [26], the concentration of the non-albumin bound ibuprofen can be derived using the first-order-fast-reversible binding model [16,26]. The calculated values were 0.4, 3.5, and 18.5 mM for three experimental samples with analytical ibuprofen concentrations of 5.0, 20.0, and 40.0 mM, respectively. The presence of excess ibuprofen ensures that the drug interacts with the other proteins in the plasma samples. Our previous studies indicated that there was no NMR spectral change for human serum albumin (0.2 mM) in the ibuprofen concentration range 4.0 to 60 mM [26]. In contrast, significant chemical shift changes were observed for the resonances at d3.23 (Fig. 1). The resonances were assigned to –Nþ (CH3 )3 headgroups of phosphatidylcholine (PC) and sphingomyelin (SM) since they were the major phospholipids at the surface of lipoproteins [28,29]. It was noted that the ibuprofen-induced chemical shift changes were apparently ibuprofen concentration dependent. When 5.0 mM ibuprofen was added into the sample, –Nþ (CH3 )3 resonance shifted 2.0 Hz upfield (Fig. 1B). At an ibuprofen concentration of 20.0 mM, the resonance separated into two peaks and the corresponding upfield drifts were 5.5 and 20.0 Hz,
294
Y. Yang et al. / Analytical Biochemistry 324 (2004) 292–297
D C B A
Fig. 1. Single-pulse 1 H NMR spectra of intact human blood plasma in the absence (A) and presence of 5.0 mM (B), 20.0 mM (C), and 40.0 mM (D) ibuprofen. The tow regions (I and II) were enlarged to show the ibuprofen-induced chemical shift and line shape changes. Assignments of the major peaks are labeled: PC and SM, Nþ (CH3 )3 of phosphatidylcholine (PC) and sphingomyelin (SM); I2, I3, I47, I56, I8, I9, and I10, ibuprofen; Lact, lactate; CH3 and (CH2 )n , methyl and methylene of lipoproteins.
respectively (Fig. 1C). The upfield changes were further increased to 21.5 and 63.5 Hz, respectively, when the ibuprofen concentration was increased to 40.0 mM (Fig. 1D). Since a larger proportion of PC headgroups, compared with SM, were tightly bound to apoB and a majority of SM headgroups were more mobile than PC [28,29], the SM headgroups were expected to have stronger interaction with ibuprofen than were PC headgroups. We therefore assigned the peak with a larger chemical shift change to SM headgroups and the peak with the smaller chemical shift change to PC. The negatively charged carboxyl group of ibuprofen may play an important role in interactions with positively charged headgroups of SM and PC. In addition, the resonances to the right of the CH3 and (CH2 )n bands of proteins were shifted in a lower-frequency direction (Fig. 1, enlarged region II). These facts revealed that ibuprofen interacted with the macromolecules, possibly lipoproteins, in blood plasma. To emphasize the ibuprofen-induced chemical shift changes for the macromolecules, we performed diffusion-weighted NMR experiments. In this case, resonances from small metabolites were suppressed due to their fast diffusion. The remaining peaks were mainly from the macromolecules in blood plasma. From the spectra obtained at the gradient strength of 16.2 G/cm (Fig. 2), we found that, in addition to the CH3 (L1) and (CH2 )n (L2) bands, the line shapes of the resonances for the allylic methylene protons (L4, –(CH2 )n CH2 CH@), the methylene protons between the olefinic groups (L6, –CH@CHCH2 CH@CH–), and the olefinic protons (L7,
–CH@CH–) were changed when ibuprofen was introduced. It was noted that the chemical shift changes of peaks L1, L2, L4, L6, and L7 were all at their low-frequency side. These regions had been attributed to HDL [19,30–32]. These observations were undoubtedly confirmed by two-dimensional correlation (COSY) experiments (Fig. 3). Comparing the COSY cross-peaks in the absence (Fig. 3A) and presence (Fig. 3B) of ibuprofen (40.0 mM), a small proportion of resonances from the L1, L2, L4, L6, and L7 bands drifted upfield and separated from the main peaks, respectively. However, the resonances of L3 (–CH2 CH2 CO–) and L5 (–CH2 CO–) in the lipoproteins were less affected. Fig. 4 shows diffusion-weighted spectra obtained at the higher gradient strength of 52.5 G/cm. The resonances of PC and SM headgroups and the upfield-drifted regions of L1, L2, L4, L6, and L7 were fully attenuated. The remaining resonances were mainly from larger lipoprotein particles, such as LDL and VLDL. The results suggest that the protons that experienced chemical shift changes were from relatively small and rapidly diffusing lipoprotein particles. This was further improved by diffusion coefficient measurements. At an ibuprofen concentration of 20.0 mM, the diffusion coefficients were 3.55 1011 and 3.62 1011 m2 s1 for PC and SM, respectively, as determined from the peaks of the headgroups. These values were close to the diffusion coefficient of HDL (3.45 1011 m2 s1 ), but they were about three times larger than those of LDL and VLDL (1:15 0:98 1011 m2 s1 ), as measured in the same experiment [19].
Y. Yang et al. / Analytical Biochemistry 324 (2004) 292–297
295
D
C
B
A
Fig. 2. Diffusion-weighted 1 H NMR spectra of intact human blood plasma in the absence (A) and presence of 5.0 mM (B), 20.0 mM (C), and 40.0 mM (D) ibuprofen at a gradient strength of 16.2 G/cm. PC and SM2 Nþ (CH3 )3 of phosphatidylcholine (PC) and sphingomyelin (SM); L1 , –CH3 ; L2 , (CH2 )n ; L3, CH2 CH2 COO; L4, (CH2 )n CH2 CH@; L5, –CH2 CH2 COOC–; L6: –CH@CHCH2 CH@CH–; L7: –HC@CH–CH2 –.
Fig. 3. High-field region of two-dimensional COSY spectra of intact human blood plasma in the absence (A) and presence of 40.0 mM (B) ibuprofen. Diffusion-weighted one-dimensional spectra were plotted as projections.
It is interesting to note that the resonances that experiened relatively large chemical shift changes were mostly from unsaturated lipids in lipoproteins. It can be seen from the COSY spectrum (Fig. 3) that L6 and L4 were coupled, respectively, to different olefinic protons of L7: L4 was coupled to L2, and L2 to L1. The results revealed that the environment of the unsaturated phospholipids, including –Nþ (CH3 )3 headgroups, in the
lipoproteins had been changed by the introduced ibuprofen. There were two possible mechanisms for the interactions. One possibility is direct interaction of ibuprofen with phospholipids. In this case, the negatively charged carboxyl group of ibuprofen may play an important role in interactions with the positively charged trimethyl headgroups of SM and PC. The isobutyl and phenyl groups of ibuprofen may interact
296
Y. Yang et al. / Analytical Biochemistry 324 (2004) 292–297
D
C
B
A
Fig. 4. Diffusion-weighted 1 H NMR spectra of intact human blood plasma in the absence (A) and presence of 5.0 mM (B), 20.0 mM (C), and 40.0 mM (D) ibuprofen at a gradient strength of 52.5 G/cm.
with the olefinic and acyl regions of the lipids. The other possible mechanism is that ibuprofen binding induces structural changes in the lipoproteins. It is known that one of the oxidizing targets for lipoproteins is the olefinic group of unsaturated lipids, and oxidative modification of lipoproteins is associated with atherosclerosis [1,2]. Although we observed ibuprofen-induced environmental changes for unsaturated phospholipids, we cannot conclude that such changes relate to inhibition of lipoprotein oxidation [4]. In addition, the intensities of the lactate resonances at d1.33 and d4.12 are moderately increased because the protein-bound lactate is replaced by the introduced ibuprofen. The resonances of the protein-bound lactate are so broad that they became NMR-invisible [33,34].
Conclusions In summary, we have demonstrated here that proton NMR spectroscopy can provide direct evidence of the interaction of drugs (ibuprofen in this case) and lipoproteins in blood plasma without any physical separation. The approach is effective and easily feasible. We found that ibuprofen can cause environmental changes for the unsaturated phospholipids in the lipoproteins. The negatively charged carboxyl group of ibuprofen may interact with the positively charged þ N(CH3 )3 moiety of PC and SM, while the phenyl and isobutyl groups of the ibuprofen may take part in the hydrophobic interaction with the olefinic chains of the lipids. In addition, ibuprofen-induced structural changes in lipoproteins cannot be excluded. This may shed light on the mechanism by which this drug protects against oxidative modification of lipoproteins.
Acknowledgments This work was supported by the grants from National Natural Science Foundation of China and the Multidisciplinary Research Program of the Chinese Academy of Sciences. We are grateful to Ms. Xiqun Shi for her assistance in the preparation of blood plasma samples. We thank Professor J.C. Lindon, Imperial College, UK, for valuable discussion.
References [1] M.S. Brown, J.L. Goldstein, A receptor-mediated pathway for cholesterol homeostasis, Science 232 (1986) 34–47. [2] D. Steinberg, S. Parthasarathy, T.E. Carew, J.C. Khoo, J.L. Witztum, Beyond cholesterol: modifications of low-density lipoprotein that increase its atherogenicity, N. Engl. J. Med. 320 (1989) 915–924. [3] D. Zapolska-Downar, M. Naruszewicz, A pleiotropic antiatherogenic action of ibuprofen, Med. Sci. Monit. 7 (2001) 837– 841. [4] D. Zapolsa-Downar, M. Naruszewicz, A. Zapolska-Downar, M. Markiewski, H. Bukowska, B. Millo, Ibuprofen inhibits adhesiveness of monocytes to endothelium and reduces cellular oxidative stress in smokers and non-smokers, Eur. J. Clin. Invest. 30 (2000) 1002–1010. [5] A. Rayyes, B. Ahren, C.H. Floren, Enhancement of low density lipoprotein catabolism by non-steroidal anti-inflammatory drugs in cultured HepG2 cells, Eur. J. Pharmacol. 372 (1999) 311–318. [6] D. Zapolska-Downar, A. Zapolska-Downar, H. Bukowska, H. Galka, M. Naruszewicz, Ibuprofen protects low density lipoproteins against oxidative modification, Life Sci. 65 (1999) 2289–2303. [7] J. Weigelt, M. Dongen, J. Uppenberg, J. Schultz, M. Wikstr€ om, Site-selective screening by NMR spectroscopy with labeled amino acid pairs, J. Am. Chem. Soc. 124 (2002) 2446–2447. [8] B.J. Stockmana, C. Dalvit, NMR screening techniques in drug discovery and drug design, Prog. Nucl. Magn. Reson. Spectrosc. 41 (2002) 187–231. [9] L.M. Lin, M.J. Shapiro, J.R. Wareing, Screening mixtures by affinity NMR, J. Org. Chem. 62 (1997) 8930–8931.
Y. Yang et al. / Analytical Biochemistry 324 (2004) 292–297 [10] P.J. Hajduk, E.T. Olejniczak, S.W. Fesik, One-dimensional relaxation- and diffusion-edited NMR methods for screening compounds that bind to macromolecules, J. Am. Chem. Soc. 119 (1997) 12257–12261. [11] K.F. Morris, P. Stilbs, C.S. Johnson Jr., Analysis of mixtures based on molecular-size and hydrophobicity by means of diffusion-ordered 2D NMR, Anal. Chem. 66 (1994) 211–215. [12] A.D. Chen, M.J. Shapiro, NOE pumping: 2. A high-throughput method to determine compounds with binding affinity to macromolecules by NMR, J. Am. Chem. Soc. 122 (2000) 414–415. [13] K. Bleicher, L.M. Lin, M.J. Shapiro, J.R. Wareing, Diffusion edited NMR: screening compound mixtures by affinity NMR to detect binding ligands to vancomycin, J. Org. Chem. 63 (1998) 8486–8490. [14] J.L. Yan, A.D. Kline, H.P. Mo, E.R. Zartler, M.J. Shapiro, Epitope mapping of ligand–receptor interactions by diffusion NMR, J. Am. Chem. Soc. 124 (2002) 9984–9985. [15] E.R. Zartler, J.L. Yan, H.P. Mo, A.D. Kline, M.J. Shapiro, 1D NMR methods in ligand–receptor interactions, Curr. Top. Med. Chem. 3 (2003) 25–37. [16] M.L. Liu, J.K. Nicholson, J.C. Lindon, Analysis of drug–protein binding using nuclear magnetic resonance based molecular diffusion measurements, Anal. Commun. 34 (1997) 225–228. [17] M.L. Liu, J.K. Nicholson, J.A. Parkinson, J.C. Lindon, Measurements of biomolecular diffusion coefficients in blood plasma using two-dimensional 1 H–1 H diffusion-edited total-correlation NMR spectroscopy, Anal. Chem. 69 (1997) 1504–1509. [18] M.L. Liu, J.K. Nicholson, J.C. Lindon, High-resolution diffusion and relaxation edited one- and two-dimensional 1 H NMR spectroscopy of biological fluids, Anal. Chem. 68 (1996) 3370– 3376. [19] M.L. Liu, H.R. Tang, J.K. Nicholson, J.C. Lindon, Use of 1 H NMR-determined diffusion coefficients to characterize lipoprotein fractions in human blood plasma, Magn. Reson. Chem. 40 (2002) S83–S88. [20] R.A. de Graaf, K.L. Behar, Quantitative 1 H NMR spectroscopy of blood plasma metabolites, Anal. Chem. 75 (2003) 2100–2104. [21] X. Zhang, C.G. Li, M.L. Liu, Determination of molecular selfdiffusion coefficient using multiple spin-echo NMR spectroscopy with removal of convection and background gradient artifacts, Anal. Chem. 73 (2001) 3528–3534.
297
[22] C.S. Johnson Jr., Diffusion ordered nuclear magnetic resonance spectroscopy: principles and applications, Prog. Nucl. Magn. Reson. Spectrosc. 34 (1999) 203–256. [23] P.J. Stilbs, Fourier transform pulsed-gradient spin-echo studies of molecular diffusion, Prog. Nucl. Magn. Reson. Spectrosc. 19 (1987) 1–45. [24] M.L. Liu, X.A. Mao, C.H. Ye, J.K. Nicholson, J.C. Lindon, Improved WATERGATE pulse sequences for solvent NMR resonance suppression, J. Magn. Reson. 132 (1998) 125–129. [25] D.H. Wu, A.D. Chen, C.S. Johnson Jr., An improved diffusionordered spectroscopy experiment incorporating bipolar-gradient pulses, J. Magn. Reson. A115 (1995) 260–264. [26] Y.F. Cui, G.Y. Bai, C.G. Li, C.H. Ye, M.L. Liu, Analysis of competitive binding of ligands to human serum albumin using NMR relaxation measurements, J. Pharmaceut. Biomed. Anal. (2004) in press. [27] T. Peters Jr. (Ed.), All About Albumin: Biochemistry, Genetics and Medical Applications, Academic Press, San Diego/London, 1996. [28] H.C. Murphy, S.P. Burns, J.J. White, J.D. Bell, R.A. Iles, Investigation of human low-density lipoprotein by 1 H nuclear magnetic resonance spectroscopy: mobility of phosphatidylcholine and sphingomyelin headgroups characterizes the surface layer, Biochemistry 39 (2000) 9763–9770. [29] H.C. Murphy, M.K. Ala-Korpela, J.J. White, A. Raoof, J.D. Bell, M.L. Barnard, S.P. Burns, R.A. Iles, Evidence for distinct behaviour of phosphatidylcholine and sphingomyelin at the low density lipoprotein surface, Biochem. Biophys. Res. Commun. 234 (1997) 733–737. [30] M. Ala-Korpela, 1 H NMR spectroscopy of human blood plasma, Prog. Nucl. Magn. Reson. Spectrosc. 27 (1995) 475–554. [31] J.A. Hamilton, J.D. Morrisett, Nuclear magnetic resonance studies of lipoproteins, Methods Enzymol. 128A (1986) 472–515. [32] J.K. Nicholson, P.J.D. Foxall, M. Spraul, R.D. Farrant, J.C. Lindon, 750 MHz 1 H and 1 H–13 C NMR spectroscopy of human blood plasma, Anal. Chem. 67 (1995) 793–811. [33] J.D. Bell, J.C.C. Brown, G. Kubal, P.J. Sadler, 1 H NMR-invisible lactate in blood plasma, FEBS J. 235 (1988) 81–86. [34] J.D. Bell, P.J. Sadler, A.F. Macled, P.R. Turner, A.L. Ville, 1 H NMR studies of human blood plasma: assignment of resonances for lipoproteins, FEBS J. 219 (1987) 239–243.
1
H NMR spectroscopic evidence of interaction between ibuprofen and lipoproteins in human blood plasma
Yunhuang Yang, Guoyun Bai, Xu Zhang, Chaohui Ye, and Maili Liu* State Key Laboratory of Magnetic Resonance and Atomic and Molecular Physics, Wuhan Institute of Physics and Mathematics, Chinese Academy of Sciences, Wuhan 430071, PeopleÕs Republic of China Received 5 September 2003
Abstract Recent studies have suggested that ibuprofen inhibits low-density lipoprotein oxidation in a high dose-dependent manner and is a promising drug for treatment of the conditions associated with atherosclerosis. In this article, we present the NMR spectroscopic evidence for the interaction between ibuprofen and phospholipids in lipoprotein particles in intact human plasma. Ibuprofen caused chemical shift upfield drifts for the protons of –Nþ (CH3 )3 moieties of phosphatidylcholine and sphingomyelin, olefinic chains (– CH@CH–, –CH@CHCH2 CH@CH–, –(CH2 )n CH2 CH@), and (CH2 )n and CH3 groups, from unsaturated lipids in lipoprotein particles. The ibuprofen may interact directly with the above-mentioned groups of phospholipids or induce structural changes in the lipoproteins. This may shed light on the mechanism by which the drug protects against oxidative modification of lipoproteins. Ó 2003 Elsevier Inc. All rights reserved. Keywords: Nuclear magnetic resonance; Ibuprofen; Lipoproteins; Unsaturated lipids; Phosphatidylcholine; Sphingomyelin
The lipoprotein particles in blood plasma or serum are classified by their size and density as very low density lipoprotein (VLDL),1 low-density lipoprotein (LDL), and high-density lipoprotein (HDL) [1]. They consist largely of a lipid core of nonpolar triacylglycerides and cholesteryl esters surrounded by more polar phospholipids, cholesterol, and apoproteins. VLDL and LDL, specifically oxidized LDL, are positively associated with coronary heart disease (CHD) and thus are considered atherogenic lipoproteins, whereas HDL is antiatherogenic because of its negative association with CHD [1,2]. Ibuprofen, 2-(4-isobutylphenyl)propionic acid, is a wellknown nonsteroidal anti-inflammatory drug. Recent studies have suggested that ibuprofen is a promising drug for treatment of the conditions associated with atherosclerosis, because of its novel properties [3–6].
*
Corresponding author. Fax: +86-27-8719-9291. E-mail address: [email protected] (M. Liu). 1 Abbreviations used: VLDL, very low density lipoprotien; LDL, low-density lipoprotien; HDL, high-density lipoproteins; CHD, coronary heart disease; PFG, pulsed field gradient; BP-LED, bipolar gradient longitudinal eddy-current delay; PC, phosphatidylcholine; SM, sphingomyelin. 0003-2697/$ - see front matter Ó 2003 Elsevier Inc. All rights reserved. doi:10.1016/j.ab.2003.10.002
These properties include: (1) inhibition of adhesion and transendothelial migration of leukocytes; (2) suppression of intracellular production of reactive oxygen species and oxidative modification of LDL; and (3) increment of HDL cholesterol levels and reduction of triglycerides. In addition, ibuprofen inhibited LDL oxidation in a high dose-dependent manner over the concentration range 0.1 to 2.0 mM (100 lg/mL LDL) [4]. However, the mechanism of the interaction between ibuprofen and lipoproteins in human blood remains unclear. Diffusion-based NMR techniques have been incorporated into a variety of methods and are widely used in pharmaceutical studies, such as screening potential bioactive ligands [7–14], studying ligand–receptor interaction [14–17], determining structures and orientations of bound ligands [15], and measuring the diffusion coefficients of small metabolites and proteins in blood plasma [17–21]. Molecular self-diffusion can cause attenuation of signal intensity in pulsed-field-gradient (PFG) NMR experiments [22,23]. At a suitable gradient strength, intensities of NMR peaks from the small molecules are reduced substantially due to their relatively fast self-diffusion and the remaining signals are
Y. Yang et al. / Analytical Biochemistry 324 (2004) 292–297
mainly from the larger molecules in complex biological samples, such as blood plasma [17–21]. This enables direct characterization of lipoproteins [19,20] and provides the possibility of studying the interaction between the proteins and drugs in blood plasma. In this article, a PFG NMR approach was used to study the interaction between ibuprofen and lipoproteins in human blood plasma without physical separation. We employed a high concentration of ibuprofen to saturate its binding sites on the serum albumin and to ensure sufficient ibuprofen to interact with the lipoproteins. Here the evidence of the interaction between ibuprofen and lipoproteins is presented.
Materials and methods Human blood was obtained from a healthy male volunteer and the plasma was prepared conventionally. The blood plasma was divided into four parts. One of the samples was used for the control. A certain amount of ibuprofen sodium salt (Sigma, Poole, Dorset, UK) was directly added into the other samples. The final concentrations of ibuprofen were 0, 5.0, 20.0, and 40.0 mM, respectively. Ten percent D2 O was added into the samples and used for the NMR magnetic field frequency lock. The samples were stored in a refrigerator (4 °C) before use. For convenience, the structure and numbering system of ibuprofen are given in Scheme 1. The NMR experiments were carried out on a Varian INOVA500 spectrometer operating at 1 H frequency of 500.12 MHz at 25 °C. The machine is equipped with a PFG accessory capable of delivering z-field gradients up to 60 G/cm. One-dimensional (1D) spectra were obtained conventionally with solvent suppression using W5 sequence [24]. The bipolar gradient longitudinal eddy-current delay (BP-LED) [25] method was used to obtained diffusion-weighted NMR spectra and to measure the diffusion coefficients. The solvent suppression pulse was incorporated into the BP-LED sequence [17– 19]. For diffusion coefficient measurement, the gradient strength was increased linearly from 10 to 55 G/m in 16 steps. The other experimental parameters were unchanged. Sixty-four transients were collected into 32,000 complex data points covering a spectral width of 8000 Hz. A cosine (0–p=2) window function was applied
Scheme 1. The structure and numbering system of ibuprofen.
293
to the time domain data before Fourier transformation to enhance signal-to-noise ratio. The areas of the NMR peaks were used to derive the diffusion coefficients [19] using the tow-parameter equation AðbÞ ¼ Að0Þ expðbDÞ, where b ¼ ð2cdGÞ2 ðD d=3Þ, c is the 1 H gyromagnetic ratio, d (2 ms) and G (16 values) are the duration and strength of the gradient pulse, D (100 ms) is the effective diffusion time, AðbÞ and Að0Þ are the NMR signal areas in the presence and absence of the gradient pulses, and D is the diffusion coefficient. For COSY experiments, 2048 (F2) 256 (F1) complex data points were acquired using conventional pulse sequence and with 16 transients for each increment. The spectral width of 8000 Hz was used for both dimensions. The time domain data were weighted using a sine-bell window function and zero-filled to 2048 (F2) 512 (F1) before Fourier transformation. The chemical shift was referred to H1 (d5.23) of a-glucose.
Results and discussion Fig. 1 shows the 1D 1 H NMR spectra of intact human blood plasma in the absence (a) and presence (b–d) of ibuprofen. Tow regions (d3:30–d2:95, d1:45–d0:70) were enlarged to show the ibuprofen-induced chemical shift changes. The broad resonances of ibuprofen (labeled with prefix I) indicated that the drug has been bound to the serum proteins. Serum albumin is the major ibuprofen-binding protein in blood plasma [16,26,27]. Considering the albumin concentration is about 0.7 mM in human blood [27] and its large binding capacity for ibuprofen (33 binding sites, Kd of 1.4 mM) [26], the concentration of the non-albumin bound ibuprofen can be derived using the first-order-fast-reversible binding model [16,26]. The calculated values were 0.4, 3.5, and 18.5 mM for three experimental samples with analytical ibuprofen concentrations of 5.0, 20.0, and 40.0 mM, respectively. The presence of excess ibuprofen ensures that the drug interacts with the other proteins in the plasma samples. Our previous studies indicated that there was no NMR spectral change for human serum albumin (0.2 mM) in the ibuprofen concentration range 4.0 to 60 mM [26]. In contrast, significant chemical shift changes were observed for the resonances at d3.23 (Fig. 1). The resonances were assigned to –Nþ (CH3 )3 headgroups of phosphatidylcholine (PC) and sphingomyelin (SM) since they were the major phospholipids at the surface of lipoproteins [28,29]. It was noted that the ibuprofen-induced chemical shift changes were apparently ibuprofen concentration dependent. When 5.0 mM ibuprofen was added into the sample, –Nþ (CH3 )3 resonance shifted 2.0 Hz upfield (Fig. 1B). At an ibuprofen concentration of 20.0 mM, the resonance separated into two peaks and the corresponding upfield drifts were 5.5 and 20.0 Hz,
294
Y. Yang et al. / Analytical Biochemistry 324 (2004) 292–297
D C B A
Fig. 1. Single-pulse 1 H NMR spectra of intact human blood plasma in the absence (A) and presence of 5.0 mM (B), 20.0 mM (C), and 40.0 mM (D) ibuprofen. The tow regions (I and II) were enlarged to show the ibuprofen-induced chemical shift and line shape changes. Assignments of the major peaks are labeled: PC and SM, Nþ (CH3 )3 of phosphatidylcholine (PC) and sphingomyelin (SM); I2, I3, I47, I56, I8, I9, and I10, ibuprofen; Lact, lactate; CH3 and (CH2 )n , methyl and methylene of lipoproteins.
respectively (Fig. 1C). The upfield changes were further increased to 21.5 and 63.5 Hz, respectively, when the ibuprofen concentration was increased to 40.0 mM (Fig. 1D). Since a larger proportion of PC headgroups, compared with SM, were tightly bound to apoB and a majority of SM headgroups were more mobile than PC [28,29], the SM headgroups were expected to have stronger interaction with ibuprofen than were PC headgroups. We therefore assigned the peak with a larger chemical shift change to SM headgroups and the peak with the smaller chemical shift change to PC. The negatively charged carboxyl group of ibuprofen may play an important role in interactions with positively charged headgroups of SM and PC. In addition, the resonances to the right of the CH3 and (CH2 )n bands of proteins were shifted in a lower-frequency direction (Fig. 1, enlarged region II). These facts revealed that ibuprofen interacted with the macromolecules, possibly lipoproteins, in blood plasma. To emphasize the ibuprofen-induced chemical shift changes for the macromolecules, we performed diffusion-weighted NMR experiments. In this case, resonances from small metabolites were suppressed due to their fast diffusion. The remaining peaks were mainly from the macromolecules in blood plasma. From the spectra obtained at the gradient strength of 16.2 G/cm (Fig. 2), we found that, in addition to the CH3 (L1) and (CH2 )n (L2) bands, the line shapes of the resonances for the allylic methylene protons (L4, –(CH2 )n CH2 CH@), the methylene protons between the olefinic groups (L6, –CH@CHCH2 CH@CH–), and the olefinic protons (L7,
–CH@CH–) were changed when ibuprofen was introduced. It was noted that the chemical shift changes of peaks L1, L2, L4, L6, and L7 were all at their low-frequency side. These regions had been attributed to HDL [19,30–32]. These observations were undoubtedly confirmed by two-dimensional correlation (COSY) experiments (Fig. 3). Comparing the COSY cross-peaks in the absence (Fig. 3A) and presence (Fig. 3B) of ibuprofen (40.0 mM), a small proportion of resonances from the L1, L2, L4, L6, and L7 bands drifted upfield and separated from the main peaks, respectively. However, the resonances of L3 (–CH2 CH2 CO–) and L5 (–CH2 CO–) in the lipoproteins were less affected. Fig. 4 shows diffusion-weighted spectra obtained at the higher gradient strength of 52.5 G/cm. The resonances of PC and SM headgroups and the upfield-drifted regions of L1, L2, L4, L6, and L7 were fully attenuated. The remaining resonances were mainly from larger lipoprotein particles, such as LDL and VLDL. The results suggest that the protons that experienced chemical shift changes were from relatively small and rapidly diffusing lipoprotein particles. This was further improved by diffusion coefficient measurements. At an ibuprofen concentration of 20.0 mM, the diffusion coefficients were 3.55 1011 and 3.62 1011 m2 s1 for PC and SM, respectively, as determined from the peaks of the headgroups. These values were close to the diffusion coefficient of HDL (3.45 1011 m2 s1 ), but they were about three times larger than those of LDL and VLDL (1:15 0:98 1011 m2 s1 ), as measured in the same experiment [19].
Y. Yang et al. / Analytical Biochemistry 324 (2004) 292–297
295
D
C
B
A
Fig. 2. Diffusion-weighted 1 H NMR spectra of intact human blood plasma in the absence (A) and presence of 5.0 mM (B), 20.0 mM (C), and 40.0 mM (D) ibuprofen at a gradient strength of 16.2 G/cm. PC and SM2 Nþ (CH3 )3 of phosphatidylcholine (PC) and sphingomyelin (SM); L1 , –CH3 ; L2 , (CH2 )n ; L3, CH2 CH2 COO; L4, (CH2 )n CH2 CH@; L5, –CH2 CH2 COOC–; L6: –CH@CHCH2 CH@CH–; L7: –HC@CH–CH2 –.
Fig. 3. High-field region of two-dimensional COSY spectra of intact human blood plasma in the absence (A) and presence of 40.0 mM (B) ibuprofen. Diffusion-weighted one-dimensional spectra were plotted as projections.
It is interesting to note that the resonances that experiened relatively large chemical shift changes were mostly from unsaturated lipids in lipoproteins. It can be seen from the COSY spectrum (Fig. 3) that L6 and L4 were coupled, respectively, to different olefinic protons of L7: L4 was coupled to L2, and L2 to L1. The results revealed that the environment of the unsaturated phospholipids, including –Nþ (CH3 )3 headgroups, in the
lipoproteins had been changed by the introduced ibuprofen. There were two possible mechanisms for the interactions. One possibility is direct interaction of ibuprofen with phospholipids. In this case, the negatively charged carboxyl group of ibuprofen may play an important role in interactions with the positively charged trimethyl headgroups of SM and PC. The isobutyl and phenyl groups of ibuprofen may interact
296
Y. Yang et al. / Analytical Biochemistry 324 (2004) 292–297
D
C
B
A
Fig. 4. Diffusion-weighted 1 H NMR spectra of intact human blood plasma in the absence (A) and presence of 5.0 mM (B), 20.0 mM (C), and 40.0 mM (D) ibuprofen at a gradient strength of 52.5 G/cm.
with the olefinic and acyl regions of the lipids. The other possible mechanism is that ibuprofen binding induces structural changes in the lipoproteins. It is known that one of the oxidizing targets for lipoproteins is the olefinic group of unsaturated lipids, and oxidative modification of lipoproteins is associated with atherosclerosis [1,2]. Although we observed ibuprofen-induced environmental changes for unsaturated phospholipids, we cannot conclude that such changes relate to inhibition of lipoprotein oxidation [4]. In addition, the intensities of the lactate resonances at d1.33 and d4.12 are moderately increased because the protein-bound lactate is replaced by the introduced ibuprofen. The resonances of the protein-bound lactate are so broad that they became NMR-invisible [33,34].
Conclusions In summary, we have demonstrated here that proton NMR spectroscopy can provide direct evidence of the interaction of drugs (ibuprofen in this case) and lipoproteins in blood plasma without any physical separation. The approach is effective and easily feasible. We found that ibuprofen can cause environmental changes for the unsaturated phospholipids in the lipoproteins. The negatively charged carboxyl group of ibuprofen may interact with the positively charged þ N(CH3 )3 moiety of PC and SM, while the phenyl and isobutyl groups of the ibuprofen may take part in the hydrophobic interaction with the olefinic chains of the lipids. In addition, ibuprofen-induced structural changes in lipoproteins cannot be excluded. This may shed light on the mechanism by which this drug protects against oxidative modification of lipoproteins.
Acknowledgments This work was supported by the grants from National Natural Science Foundation of China and the Multidisciplinary Research Program of the Chinese Academy of Sciences. We are grateful to Ms. Xiqun Shi for her assistance in the preparation of blood plasma samples. We thank Professor J.C. Lindon, Imperial College, UK, for valuable discussion.
References [1] M.S. Brown, J.L. Goldstein, A receptor-mediated pathway for cholesterol homeostasis, Science 232 (1986) 34–47. [2] D. Steinberg, S. Parthasarathy, T.E. Carew, J.C. Khoo, J.L. Witztum, Beyond cholesterol: modifications of low-density lipoprotein that increase its atherogenicity, N. Engl. J. Med. 320 (1989) 915–924. [3] D. Zapolska-Downar, M. Naruszewicz, A pleiotropic antiatherogenic action of ibuprofen, Med. Sci. Monit. 7 (2001) 837– 841. [4] D. Zapolsa-Downar, M. Naruszewicz, A. Zapolska-Downar, M. Markiewski, H. Bukowska, B. Millo, Ibuprofen inhibits adhesiveness of monocytes to endothelium and reduces cellular oxidative stress in smokers and non-smokers, Eur. J. Clin. Invest. 30 (2000) 1002–1010. [5] A. Rayyes, B. Ahren, C.H. Floren, Enhancement of low density lipoprotein catabolism by non-steroidal anti-inflammatory drugs in cultured HepG2 cells, Eur. J. Pharmacol. 372 (1999) 311–318. [6] D. Zapolska-Downar, A. Zapolska-Downar, H. Bukowska, H. Galka, M. Naruszewicz, Ibuprofen protects low density lipoproteins against oxidative modification, Life Sci. 65 (1999) 2289–2303. [7] J. Weigelt, M. Dongen, J. Uppenberg, J. Schultz, M. Wikstr€ om, Site-selective screening by NMR spectroscopy with labeled amino acid pairs, J. Am. Chem. Soc. 124 (2002) 2446–2447. [8] B.J. Stockmana, C. Dalvit, NMR screening techniques in drug discovery and drug design, Prog. Nucl. Magn. Reson. Spectrosc. 41 (2002) 187–231. [9] L.M. Lin, M.J. Shapiro, J.R. Wareing, Screening mixtures by affinity NMR, J. Org. Chem. 62 (1997) 8930–8931.
Y. Yang et al. / Analytical Biochemistry 324 (2004) 292–297 [10] P.J. Hajduk, E.T. Olejniczak, S.W. Fesik, One-dimensional relaxation- and diffusion-edited NMR methods for screening compounds that bind to macromolecules, J. Am. Chem. Soc. 119 (1997) 12257–12261. [11] K.F. Morris, P. Stilbs, C.S. Johnson Jr., Analysis of mixtures based on molecular-size and hydrophobicity by means of diffusion-ordered 2D NMR, Anal. Chem. 66 (1994) 211–215. [12] A.D. Chen, M.J. Shapiro, NOE pumping: 2. A high-throughput method to determine compounds with binding affinity to macromolecules by NMR, J. Am. Chem. Soc. 122 (2000) 414–415. [13] K. Bleicher, L.M. Lin, M.J. Shapiro, J.R. Wareing, Diffusion edited NMR: screening compound mixtures by affinity NMR to detect binding ligands to vancomycin, J. Org. Chem. 63 (1998) 8486–8490. [14] J.L. Yan, A.D. Kline, H.P. Mo, E.R. Zartler, M.J. Shapiro, Epitope mapping of ligand–receptor interactions by diffusion NMR, J. Am. Chem. Soc. 124 (2002) 9984–9985. [15] E.R. Zartler, J.L. Yan, H.P. Mo, A.D. Kline, M.J. Shapiro, 1D NMR methods in ligand–receptor interactions, Curr. Top. Med. Chem. 3 (2003) 25–37. [16] M.L. Liu, J.K. Nicholson, J.C. Lindon, Analysis of drug–protein binding using nuclear magnetic resonance based molecular diffusion measurements, Anal. Commun. 34 (1997) 225–228. [17] M.L. Liu, J.K. Nicholson, J.A. Parkinson, J.C. Lindon, Measurements of biomolecular diffusion coefficients in blood plasma using two-dimensional 1 H–1 H diffusion-edited total-correlation NMR spectroscopy, Anal. Chem. 69 (1997) 1504–1509. [18] M.L. Liu, J.K. Nicholson, J.C. Lindon, High-resolution diffusion and relaxation edited one- and two-dimensional 1 H NMR spectroscopy of biological fluids, Anal. Chem. 68 (1996) 3370– 3376. [19] M.L. Liu, H.R. Tang, J.K. Nicholson, J.C. Lindon, Use of 1 H NMR-determined diffusion coefficients to characterize lipoprotein fractions in human blood plasma, Magn. Reson. Chem. 40 (2002) S83–S88. [20] R.A. de Graaf, K.L. Behar, Quantitative 1 H NMR spectroscopy of blood plasma metabolites, Anal. Chem. 75 (2003) 2100–2104. [21] X. Zhang, C.G. Li, M.L. Liu, Determination of molecular selfdiffusion coefficient using multiple spin-echo NMR spectroscopy with removal of convection and background gradient artifacts, Anal. Chem. 73 (2001) 3528–3534.
297
[22] C.S. Johnson Jr., Diffusion ordered nuclear magnetic resonance spectroscopy: principles and applications, Prog. Nucl. Magn. Reson. Spectrosc. 34 (1999) 203–256. [23] P.J. Stilbs, Fourier transform pulsed-gradient spin-echo studies of molecular diffusion, Prog. Nucl. Magn. Reson. Spectrosc. 19 (1987) 1–45. [24] M.L. Liu, X.A. Mao, C.H. Ye, J.K. Nicholson, J.C. Lindon, Improved WATERGATE pulse sequences for solvent NMR resonance suppression, J. Magn. Reson. 132 (1998) 125–129. [25] D.H. Wu, A.D. Chen, C.S. Johnson Jr., An improved diffusionordered spectroscopy experiment incorporating bipolar-gradient pulses, J. Magn. Reson. A115 (1995) 260–264. [26] Y.F. Cui, G.Y. Bai, C.G. Li, C.H. Ye, M.L. Liu, Analysis of competitive binding of ligands to human serum albumin using NMR relaxation measurements, J. Pharmaceut. Biomed. Anal. (2004) in press. [27] T. Peters Jr. (Ed.), All About Albumin: Biochemistry, Genetics and Medical Applications, Academic Press, San Diego/London, 1996. [28] H.C. Murphy, S.P. Burns, J.J. White, J.D. Bell, R.A. Iles, Investigation of human low-density lipoprotein by 1 H nuclear magnetic resonance spectroscopy: mobility of phosphatidylcholine and sphingomyelin headgroups characterizes the surface layer, Biochemistry 39 (2000) 9763–9770. [29] H.C. Murphy, M.K. Ala-Korpela, J.J. White, A. Raoof, J.D. Bell, M.L. Barnard, S.P. Burns, R.A. Iles, Evidence for distinct behaviour of phosphatidylcholine and sphingomyelin at the low density lipoprotein surface, Biochem. Biophys. Res. Commun. 234 (1997) 733–737. [30] M. Ala-Korpela, 1 H NMR spectroscopy of human blood plasma, Prog. Nucl. Magn. Reson. Spectrosc. 27 (1995) 475–554. [31] J.A. Hamilton, J.D. Morrisett, Nuclear magnetic resonance studies of lipoproteins, Methods Enzymol. 128A (1986) 472–515. [32] J.K. Nicholson, P.J.D. Foxall, M. Spraul, R.D. Farrant, J.C. Lindon, 750 MHz 1 H and 1 H–13 C NMR spectroscopy of human blood plasma, Anal. Chem. 67 (1995) 793–811. [33] J.D. Bell, J.C.C. Brown, G. Kubal, P.J. Sadler, 1 H NMR-invisible lactate in blood plasma, FEBS J. 235 (1988) 81–86. [34] J.D. Bell, P.J. Sadler, A.F. Macled, P.R. Turner, A.L. Ville, 1 H NMR studies of human blood plasma: assignment of resonances for lipoproteins, FEBS J. 219 (1987) 239–243.