European Journal of Pharmaceutical Sciences 47 (2012) 15

Contents lists available at SciVerse ScienceDirect

European Journal of Pharmaceutical Sciences

journal homepage: www.elsevier.com/locate/ejps

A new pH/organic modier gradient RP HPLC method for convenient determination of lipophilicity and acidity of drugs as applied to established imidazoline agents
Pawe Wiczling a, Antoni Nasal a, ukasz Kubik a, Roman Kaliszan a,b,
a b

sk, Gen. J. Hallera 107, 80-416 Gdan sk, Poland Department of Biopharmaceutics and Pharmacodynamics, Medical University of Gdan , Poland Department of Biopharmacy, Collegium Medicum in Bydgoszcz Nicolaus Copernicus University in Torun

a r t i c l e

i n f o

a b s t r a c t
Convenient drug candidates testing methods for lipophilicity and acidity are highly requested in modern pharmaceutical research and development strategy. Reversed-phase high-performance liquid chromatography (RP HPLC) might be particularly useful for the determination of both pKa and the apparent (pH-dependent) octanolwater partition coefcient, applicable in high-throughput analysis of multicomponent mixtures. In this report the pH/organic modier gradient RP HPLC is presented as a means of simultaneous determination of acidity and lipophilicity of a series of 26 imidazoline-like drugs. The previously theoretically elaborated approach has been applied consisting in retention measurements in a series of methanol gradient runs differing in pH range and duration of the gradient. The simultaneously determined lipophilicity and dissociation constants have been demonstrated to correlate to the respective parameters form calculation chemistry. The proposed approach can be applied to compound mixtures, it requires only minute amounts of substances, and pKa values can be determined in the range 310 units and lipophilicity log P parameter in the range 07 units. 2012 Elsevier B.V. All rights reserved.

Article history: Received 18 January 2012 Received in revised form 25 April 2012 Accepted 26 April 2012 Available online 4 May 2012 Keywords: pH/organic modier Gradient RP HPLC Imidazoline-like drugs Lipophilicity Acidity

1. Introduction The pharmacokinetic properties of drugs depend strongly on lipophilicity. In the case of ionizable agents, the dissociation of the drug in aqueous compartments of a living system separated by lipid membranes is also of great importance for the processes determining pharmacokinetic phase of drug action (i.e. absorption, distribution, metabolism and elimination ADME). Therefore, knowledge of dissociation constant, pKa, lipophilicity parameters like log P logarithm of octanolwater partition coefcient and distribution coefcient, log D, of a drug (drug candidate) is of a primary concern in medicinal chemistry. That implies a need of procedures for measurement of lipophilicity parameters and pKa values that are reliable, but at the same time convenient, fast and which request relatively small amounts of a given, often rare and valuable, substance. The RP HPLC pH-gradient method previously elaborated by Kaliszan and coworkers (Kaliszan et al., 2001) is unique in that it allows simultaneous determination of the important biorelevant physicochemical properties of drugs
Corresponding author at: Department of Biopharmaceutics and Pharmacody sk, Gen. J. Hallera 107, 80-416 Gdan sk, Poland. namics, Medical University of Gdan Tel.: +48 58 3493260. E-mail addresses: roman.kaliszan@amg.gda.pl, roman.kaliszan@gumed.edu.pl (R. Kaliszan).
0928-0987/$ - see front matter 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.ejps.2012.04.021

(drug candidates), which are available in minute amounts only, without prior isolation of an individual component of complex mixtures, which are presently often provided by combinatorial and other modern synthetic chemistry methods. That is impossible to achieve by means of standard acidimetric and spectrophotometric methods. In this work we managed to get necessary RP HPLC retention data from a single chromatographic run for 7, 8 analytes in a mixed sample when applying classical UV detection, which requires a distinct separation of the components. Currently available universal detection techniques, like mass spectrometry, allow to enlarge number of compounds analyzed in individual chromatographic runs. That opens perspectives for our approach to enable a preselective high-throughput screening (HTS) of drug candidates from large numbers of compounds obtained by modern organic synthesis methods. It is well known that imidazolines exert diverse pharmacological effects on cardiovascular system (Szabo, 2002). Clonidine-like drugs of rst generation reduce blood pressure by stimulation of central a2-adrenoceptors. Second generation antihypertensives, such as rilmenidine and moxonidine, attenuate sedative (a2-adrenoceptor-mediated) effects acting mainly at I1-imidazoline receptors, localized in the rostral ventrolateral medulla (RVLM) region of the brain (Bousquet and Feldman, 1999; Farsang, 2001; Head and Burke, 2000). Since the presence of imidazoline receptors of

P. Wiczling et al. / European Journal of Pharmaceutical Sciences 47 (2012) 15

Table 1 Pump program ensuring constant values of pH during linear organic modier gradient.% Buffer D can be calculated by: % Buffer D = 100%% MeOH% Buffer C. % MeOH 5 24 43 61 80 s w pH % Buffer C 77 66 53 37 20 3.40 66 57 46 34 18 4. 15 56 49 41 30 17 4.89 48 42 35 26 15 5.64 40 35 29 23 13 6.39 32 29 24 19 11 7.13 25 22 19 15 9 7.88 18 16 14 11 7 8.63 11 10 9 7 4 9.37 4 4 4 2 0 10.12

I1 and I2 type in peripheral tissues (heart, kidney, adrenal medulla) has been claimed, many studies aimed at elucidation of regulatory role of these receptors were performed (Bousquet et al., 2000; ElAyoubi et al., 2004; Evans and Haynes, 1995; Molderings, 1997; Radwanska et al., 2009; Szabo and Urban, 1995). To relate bioactivity differences to the structure of the agents it is necessary to have some comparable data on their physicochemical properties, in particular, those which are presumed to affect drug distribution within the body compartments. Such data are lacking for imidazoline-like drugs. The report deals with a series of well known 26 imidazoline-like drugs and pharmacological reference agents, for which no data on lipophilicity and acidity have so far been available in literature. Hence, the data provided by us in this paper should be of interest to medicinal chemists and molecular pharmacologists interested in the agents exerting their activity through a-adrenoceptors and I1/I2 imidazoline receptors. 2. HPLC equipment Experiments were done using a Merck-Hitachi LaChrome (Darmstadt, Germany-San Jose, CA, USA) apparatus equipped with a diode array detector, autosampler and thermostat. Chromatographic data were collected using a D-7000 HPLC System Manager, version 3.1 (Merck-Hitachi). A new XBridge C-18 column, 150 4.6 mm I.D., particle size 5 lm (Waters Corporation, Milford, MA, USA) was used. 1% urea was injected to determine the column dead volume, Vo, which was 1.28 ml. The system dwell volume Vd

equaled 1.74 ml. The extra column volume equaled 0.59 ml. It served to nd the extra column time that has been subtracted from all the retention times prior to any calculation. The chromatographic measurements were done at 25 C and at 0.5 ml/min ow rate. Prior to each measurement the column was equilibrated by passing six column volumes of the mobile phase. All the reagents and the analytes employed were of a highest commercially available quality. 3. pH measurements The details of proper pH measurements in the organic/water mobile phase are discussed by Ross (Roses, 2004). Briey, the pH of the mobile phase was measured after mixing the aqueous buffer and the organic modier. The electrode system was calibrated with the usual aqueous standard (w s pH scale). The pH of the buffers was measured at 25 C. The measurements were done with an HI 9017 pH-meter (Hanna Instruments, Bedfordshire, UK). An universal buffer was used to control the pH of linear methanol gradient. The base buffer solution was formed using three compounds: citric acid (CIT), tris(hydroxymethyl)-aminomethane (TRIS), glycine (GLY), each at a concentration of 0.008 M. Buffer of pH = 10.50 (buffer C) (corresponding to aqueous solutions) was made by adding the necessary volume of 3 M KOH to the base solution. Buffer of pH = 2.50 (buffer D), was made by adding the necessary volume of 1 M HCl. The mobile phases contained buffers D and C in different proportions and methanol as the organic modier (solvent B). The pH measurements were done with an HI 9017

Table 2 Parameters obtained by nonlinear curve tting of Eq. (1) to the experimental data.%CV denotes coefcient of variation. Name 2-BFI Antazoline Benazoline BU 239 BU224 Cirazoline Clonidine Detomidine Efaroxan Guanabenz Harman Idazoxan Lofexidine Medetomidine Moxonidine Naphazoline Oxymetazoline Phentolamine RS 45041-190 RX 821002 Tetrahydrozoline Tiamenidine UK14304 Xylazine Xylometazoline Yohimbine log kw,I (%CV) 1.41 2.71 1.85 1.08 1.45 2.09 1.06 2.12 1.67 2.26 1.77 1.17 1.72 2.49 0.83 1.79 2.78 2.67 1.92 1.30 1.49 1.19 0.85 1.91 3.09 2.36 (1.67) (3.23) (1.74) (2.70) (1.56) (1.81) (1.59) (2.91) (1.53) (2.15) (2.97) (2.64) (4.67) (2.15) (2.26) (2.53) (1.40) (2.80) (1.11) (1.31) (1.78) (2.07) (2.05) (5.44) (2.60) (3.50) S (%CV) 4.80 5.35 5.18 5.14 5.25 4.73 4.46 4.53 4.70 4.95 4.90 4.49 3.96 5.00 5.44 4.84 4.96 5.78 5.22 4.24 4.91 4.82 6.52 5.77 5.07 5.17 (4.08) (4.93) (3.36) (7.54) (3.60) (3.33) (4.97) (5.32) (3.32) (3.68) (6.04) (7.69) (10.59) (3.48) (7.54) (5.11) (2.14) (4.23) (2.09) (3.56) (4.15) (5.82) (5.90) (10.00) (3.77) (5.80) log kw,N (%CV) 2.33 (2.56) 3.87 (9.21) 2.72 (4.00) 2.32 (4.61) 2.70 (2.49) 3.05 (2.59) 1.99 (2.65) 2.95 (3.52) 2.81 (2.62) 2.83 (2.47) 3.16 (4.22) 2.14 (4.39) 2.18 (6.08) 3.42 (2.50) 1.20 (2.94) >1.79 >2.78 >2.67 >1.92 2.09 (1.91) >1.49 2.16 (3.89) 1.53 (2.99) 3.67 (8.16) >3.09 3.43 (4.22) pKa,chrom (%CV) 9.38 (0.51) 10.86 (5.97) 10.35 (1.41) 8.13 (0.80) 9.29 (0.43) 9.94 (0.66) 8.71 (0.49) 7.07 (0.78) 10.02 (0.61) 8.67 (0.67) 7.21 (0.79) 9.04 (0.80) 9.28 (2.36) 7.08 (0.49) 7.84 (1.22) >10.50 >10.50 >10.50 >10.50 8.05 (0.42) >10.50 9.87 (0.84) 7.75 (0.67) 9.48 (1.51) >10.50 7.94 (0.72) log PN (ACD) 0.69 4.39 1.79 0.47 0.84 3.19 1.41 2.75 2.99 2.57 3.26 1.81 3.59 3.10 1.14 3.88 4.52 3.60 0.34 1.86 3.31 1.08 0.97 2.37 5.26 2.20 pKa (ACD) 9.23 10.29 9.87 8.77 9.87 9.84 8.10 7.14 9.79 8.17 8.62 9.51 9.80 6.65 6.47 10.27 10.61 10.31 9.18 9.51 10.42 9.47 7.20 7.67 10.62 8.44 log PN (lit) pKa (lit) 10.13 1.34

7.84 2.88 1.7

9.28 7.36 10.35

10.51 1.18 4.52 2.73 10.20 7.70

P. Wiczling et al. / European Journal of Pharmaceutical Sciences 47 (2012) 15

3
RX 821002

pH-meter (Hanna Instruments, Bedfordshire, UK) at 25 C. The relationship between w s pH (corresponding to organic-water solutions), methanol content and various buffer compositions was experimentally determined to ensure constant pH during organic modier gradient. Large number of solvent combinations was mixed by Merck-Hitachi LaChrome system and subsequently s w pH was measured. Next, the ve-step pump program was found that gave constant pH despite increase in methanol content (Table 1).
Retention Time, min

Moxonidine 80 40 0 Naphazoline 80 40 0 Detomidine 80 40 0 BU 239 80 40 0 RS 45041-190 80 40 0 Xylometazoline 80 40 0 Clonidine 80 40 0

Guanabenz

Tiamenidine

4. Tested analytes We analyzed 26 analytes listed in Table 2. All the samples were prepared in concentrations of 0.2 mg/ml. Bretylium tosylate was the test compound for the assessment of the silanol activity as a function of pH. The theoretical values of log P and pKa were calculated using the ACD/PhysChem Suite software (Advanced Chemistry Development Inc). The literature values of log P and pKa were available for selected analytes only and were extracted from ACD/PhysChem Suite and Physical/Chemical Property (Howard and Meylan, 1999) database. 5. Experimental design Retention times of tested analytes were measured in two series of wide organic modier gradients ranging from 0.05 to 0.8. The series differed in gradient duration of 32 min (series I) and 88 min (series II). Each series comprised chromatographic runs at 10 different pH values of mobile phase as listed in Table 1. The experimental data for each analyte are given in Fig. 1. The data was tted to the previously proposed model (Wiczling and Kaliszan, 2010; Wiczling et al., 2008):

Idazoxan

Cirazoline

UK14304

BU224

10 pH

10

Xylazine 80 40 0 Phentolamine 80 40 0 Retention Time, min

Efaroxan

Antazoline

Z
0

t R t 0 t e

10Sut 1 10pHpKa; chrom dt 1 t0 10log kw;I 10log kw;N 10pHpKa;chrom

Oxymetazoline 80 40 0 Tetrahydrozoline 80 40 0 Benazoline 80 40 0 Harman 80 40 0 4 6 8 10 pH 4 6

2-BFI

where tR is analyte retention time, t0 is the column hold-up time, te is the extra column time, log kw,I and log kw,N denote the chromatographic measure of lipophilicity for ionized and neutral form of an analyte and corresponds to the retention factor in neat water as an eluent, S is the slope of the relationship between logarithm of retention factor and organic modier content, and pKa,chrom denotes the apparent pKa value of an analyte. A nonlinear data-tting problem was solved by a least-square algorithm implemented in the Matlab 2010b (Mathworks Inc., Natick, MA, USA) lsqcurvet function. The dependence between retention time and pH Eq. (1) is sigmoidal. For such relationship a measurements in the region of pH covering full dissociation step are required to precisely estimate pKa value and log kw of the corresponding analyte form. In our data set few analytes had pKa larger than 10.5 (the maximal value of pH used in our experiments). Still the estimation of those parameters was possible, however, resulted in large coefcients of variation (%CV). As rule of thumb we decided to include only those pKa and log kw values for which %CV was smaller than 10%. 6. Silanol activity of the column The properties of a chromatographic column can change when chromatographic runs are conducted at different pH. For the purpose of pKa and log kw determination this pH-dependent changes, especially with regard to the presence of other retention mechanism than partitioning, need to be as small as possible. The probe used to examine the inuence of pH on the retention of the fully ionized form of an analyte was bretylium tosylate. The ionization of either the bretylium cation or tosylate anion does not change with pH, which makes them a useful marker of ion-exchange properties

Lofexidine

Yohimbine

Medetomidine

10

Fig. 1. The experimental (points) and predicted retention times (lines) versus pH for tested analytes observed during 20 chromatographic run. The chromatographic data was collected at 10 different pH of mobile phase for the 32 (s) and 88 min (h) duration of organic modier gradient. The predictions were obtained using Eq. (1).

of the column. As reported earlier by (Neue et al., 2001), the knowledge about bretylium tosylate retention can be used for a complete and quantitative understanding of the retention behavior of ionizable compounds as a function of the mobile phase pH. 7. Results The experimental and model predicted retention times versus pH proles are shown in Fig. 1. These plots demonstrate that the

P. Wiczling et al. / European Journal of Pharmaceutical Sciences 47 (2012) 15

Bretylium cation 30 tg = 32 min Retention Time, min 25 tg = 88 min

20

15

10

7 Tosylate anion

10

11

25 tg = 32 min Retention Time, min tg = 88 min 20

15

10

7 pH

10

11

Fig. 2. The dependence of the retention time of bretylium cation and tosylate anion on pH.

5 log kw,N (Experimental) 4 3 2 1 0 log kw,N = 0.42*log PN+ 1.7 R = 0.41

2

not cover a full dissociation step, it was not possible to precisely determine pKa value of analyte and consequently a log kw of the neutral form of the analyte. The dependence of gradient retention time on pH for bretylium tosylate is illustrated on Fig. 2. The retention time of bretylium cation sigmoidally decreases with increasing pH. The difference between the retention time of the bretylium cation at pH 3.40 and pH 10.12 is 0.4 min for 32 min and 0.93 min for 88 min gradient duration. For tosylate anion a similar in magnitude increase has been observed. Since those changes are very small compared to the usual difference in retention time of ionized and non-ionized form of analyte (about 5 for 32 and 10 min for 88 min gradient duration) they were neglected. It considerably simplied the model used to describe the chromatographic data. Fig. 3 presents the correlations between the experimental and theoretical values of lipophilicity and dissociation constant. The experimental pKa,chrom values were moderately correlated with theoretical values. The observed differences between pKa,chrom and theoretical values of pKa were most likely caused by the approximate nature of both methods. Similarly, the chromatographic lipophilicity parameter log kw was moderately related to the theoretical log P values. Certainly one must feel disappointed by such limited correlations. The question arises then which data are more reliable: chromatographically determined or theoretically calculated ones? Evidently, for the six references log PN date from literature, given in Table 2, the correlation with chromatographic lipophilicity parameter, log kw,N is better (log PN = 0.38log kw,N + 2.06; R2 = 0.78) than with calculated log PN (ACD) (R2 = 0.42). Similarly, for the ve references pKa date from literature, given in Table 2, the correlation with chromatographic pKa,chrom is better (R2 = 0.92) than with calculated pKa(ACD) (R2 = 0.71). Because in our experience the ACD/PhysChem suite has performed best of a number of the calculation chemistry programs tested, we presume that specic chemical properties of imidazoline derivatives studied here have been poorly accounted for by the existing molecular modeling softwares. The more useful than appear to be out chromatographic method. 8. Conclusions A previously elaborated (Kaliszan et al., 2001; Wiczling et al., 2006, 2004) original pH/organic modier gradient RP HPLC method of determination of dissociation constants and lipophilicity parameters has been applied to a series of widely used imidazoline drugs. It was demonstrated that the method well applies for a preliminary evaluation of the pharmacokinetically important lipophilic and acidic properties of drug and drug candidate analytes, especially when present in mixtures and/or available in non-pure form. The approach here developed can facilitate a high-throughput preselection of drugs candidates of a rationally predicted bioavailability.

0.5

1.5

2.5 log PN (ACD)

3.5

4.5

11 pK a,chrom (Experimental) 10 9 8 7 6 5 4 3 3 4 5 6 7 pK a (ACD) 8 9 10 11 pK a,chrom = 0.72*pKa + 2.5 R = 0.57

2

Acknowledgements This work was supported by Grants No. NN405G30038 and No. NN405376037 from the Polish Ministry of Science and Higher Education. References
Bousquet, P., Bruban, V., Schann, S., Feldman, J., 2000. Imidazoline receptors: a challenge. Pharm. Acta Helv. 74, 205209. Bousquet, P., Feldman, J., 1999. Drugs acting on imidazoline receptors: a review of their pharmacology, their use in blood pressure control and their potential interest in cardioprotection. Drugs 58, 799812. El-Ayoubi, R., Menaouar, A., Gutkowska, J., Mukaddam-Daher, S., 2004. Imidazoline receptors but not alpha 2-adrenoceptors are regulated in spontaneously

Fig. 3. The correlations between the experimental and theoretical parameters of lipophilicity (a) and pKa (b).

proposed model predicts retention times very well. The parameter estimates obtained by the nonlinear curve ttings are listed in Table 2. Majority of parameters were estimated with high precision, as conrmed by small coefcients of variation (%CV). All 26 analytes were weak monoprotic bases as their retention time increased with increasing pH. For six analytes this increase was present at relatively small range of pH values. As the increase did

P. Wiczling et al. / European Journal of Pharmaceutical Sciences 47 (2012) 15 hypertensive rat heart by chronic moxonidine treatment. J. Pharmacol. Exp. Ther. 310, 446451. Evans, R.G., Haynes, J.M., 1995. Alpha 2 adrenoceptor- and imidazoline-preferring binding sites in the dog kidney. Ann. NY Acad. Sci. 763, 357360. Farsang, C., 2001. Moxonidine: clinical prole. J. Clin. Basic Cardiol. 4, 197200. Head, G.A., Burke, S.L., 2000. I1 imidazoline receptors in cardiovascular regulation: the place of rilmenidine. Am. J. Hypertens. 13, 89S98S. Howard, P., Meylan, W., 1999. Physical/Chemical Property Database (PHYSPROP); Syracuse Research Corp.; Environmental Science Center, North Syracuse, NY. Kaliszan, R., Haber, P., Baczek, T., Siluk, D., 2001. Gradient HPLC in the determination of drug lipophilicity and acidity. Pure Appl. Chem. 73, 14651475. Molderings, G.J., 1997. Imidazoline receptors: basic knowledge, recent advances and future prospects for therapy and diagnosis. Drugs Future 22, 757772. Neue, U.D., Phoebe, C.H., Tran, K., Cheng, Y.F., Lu, Z., 2001. Dependence of reversedphase retention of ionizable analytes on pH, concentration of organic solvent and silanol activity. J. Chromatogr. A 925, 4967. Radwanska, A., Dlugokecka, J., Wasilewski, R., Kaliszan, R., 2009. Testing conception of engagement of imidazoline receptors in imidazoline drugs effects on isolated rat heart atria. J. Physiol. Pharmacol. 60, 131142.

Roses, M., 2004. Determination of the pH of binary mobile phases for reversedphase liquid chromatography. J. Chromatogr. A 1037, 283298. Szabo, B., 2002. Imidazoline antihypertensive drugs: a critical review on their mechanism of action. Pharmacol. Ther. 93, 135. Szabo, B., Urban, R., 1995. Mechanism of sympathoinhibition by imidazolines. Ann. NY. Acad Sci. 763, 552565. Wiczling, P., Kaliszan, R., 2010. Retention time and peak width in the combined pH/ organic modier gradient high performance liquid chromatography. J. Chromatogr. A 1217, 33753381. Wiczling, P., Kawczak, P., Nasal, A., Kaliszan, R., 2006. Simultaneous determination of pKa and lipophilicity by gradient RP HPLC. Anal. Chem. 78, 239249. Wiczling, P., Markuszewski, M.J., Kaliszan, R., 2004. Determination of pKa by pH gradient reversed-phase HPLC. Anal. Chem. 76, 30693077. Wiczling, P., Waszczuk-Jankowska, M., Markuszewski, M.J., Kaliszan, R., 2008. The application of gradient reversed-phase high-performance liquid chromatography to the pK(a) and log k(w) determination of polyprotic analytes. J. Chromatogr. A 1214, 109114.