Lipophilic and electrostatic forces encoded in IAM-HPLC indexes of basic drugs: Their role in membrane partition and their relationships with BBB passage data

Lipophilic and electrostatic forces encoded in IAM-HPLC indexes of basic drugs: Their role in membrane partition and their relationships with BBB passage data

European Journal of Pharmaceutical Sciences 45 (2012) 685–692 Contents lists available at SciVerse ScienceDirect European Journal of Pharmaceutical ...

339KB Sizes 0 Downloads 16 Views

European Journal of Pharmaceutical Sciences 45 (2012) 685–692

Contents lists available at SciVerse ScienceDirect

European Journal of Pharmaceutical Sciences journal homepage: www.elsevier.com/locate/ejps

Lipophilic and electrostatic forces encoded in IAM-HPLC indexes of basic drugs: Their role in membrane partition and their relationships with BBB passage data Lucia Grumetto, Carmen Carpentiero, Francesco Barbato ⇑ Dipartimento di Chimica Farmaceutica e Tossicologica, Università degli Studi di Napoli Federico II, Via D. Montesano, 49, I-80131 Naples, Italy

a r t i c l e

i n f o

Article history: Received 12 October 2011 Received in revised form 22 December 2011 Accepted 18 January 2012 Available online 28 January 2012 Keywords: Phospholipids Lipophilicity Membrane partition IAM-HPLC BBB passage

a b s t r a c t IAM

The membrane phospholipid affinity data, log kw , for 14 basic drugs spanning a wide lipophilicity range were measured by HPLC on two different phospholipid stationary phases, i.e. IAM.PC.MG and IAM.PC.DD2. These data related weakly with log PN values, the n-octanol/water partition coefficients of the neutral forms; poorer relationships were found with log D7.0 values, the n-octanol/water partition coefficients of the mixtures of neutral and ionized forms at pH 7.0. The lack of collinearity confirms that, differently from partition in n-octanol/water, partition in phospholipids encodes not only lipophilic/hydrophobic intermolecular recognition forces but also ionic bonds, due to electrostatic interactions between electrically charged species and phospholipids, according to the ‘‘pH-piston hypothesis’’. This component of interacIAM IAM tion was parameterized by D log kw values; they are the differences between the log kw values experiIAM mentally measured and the values expected for neutral isolipophilic compounds. D log kw values of the various analytes changed almost linearly from positive to negative values at increasing lipophilicity. This behavior is consistent with an interaction mechanism with membrane phospholipids including two intermolecular interaction forces: (i) lipophilic/hydrophobic interactions, which decrease on ionization proportionally to the lipophilicity of the neutral forms, and (ii) electrostatic interactions, which increase on ionization and are quite constant for all the analytes at a given ionization degree. Since BBB passage of the considered compounds is supposed to be based on passive mechanisms, we investigated the possible relationships between log BB values, i.e. the logarithms of the ratio between brain and blood concentrations, and three physico-chemical parameters, i.e. (i) log PN (lipophilic interaction of the neutral form), IAM IAM (ii) log kw (global interaction with phospholipids), and (iii) D log kw (electrostatic component of interacIAM tion with phospholipids). The results suggest that the electrostatic interactions encoded in log kw values might act as trapping forces in a phospholipid barrier. Actually, we observed an inverse linear dependence IAM IAM of log BB on D log kw values, but only for the compounds showing positive D log kw values. We conclude that the driving force for BBB passage is the lipophilicity of the neutral forms, log PN, and not the lipophilicity actually displayed at the experimental pH, log D7.0. Indeed, the latter does not adequately take into account the role played by protonation in the analyte/membrane interactions because protonation, although hindering membrane passage, can either reduce or enhance partition in phospholipids, depending on analyte lipophilicity. Ó 2012 Elsevier B.V. All rights reserved.

1. Introduction One important factor of bioavailability is the lipophilicity of a drug that is related to the classical octanol/water partition coefficient, log Poct (Leo et al., 1971). Determination of this thermodynamically defined parameter is a tedious, substance and time consuming procedure. Some efforts to replace this procedure are the usage of chromatographic retention data of reversed phase HPLC. In the case of fast equilibrium kinetics these data are closely related to log Poct values. Since the alkyl chains of a reversed phase ⇑ Corresponding author. Tel.: +39 081 678627; fax: +39 081 678630. E-mail address: [email protected] (F. Barbato). 0928-0987/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.ejps.2012.01.008

don’t resemble the real penetration behavior of a cell membrane, the introduction of phosphatidylcholine phases (Immobilized Artificial Membrane – IAM) should result in a more realistic model (Yang et al.,1996). It has been reported in the literature that, in the case of charged or partly charged molecules, IAM-HPLC data, IAM expressed as log kw , encode not only the balance of hydrophobic and polar intermolecular interactions but also ionic bond contribution (Barbato, 2006; Taillardat-Bertschinger et al., 2003). Depending on the pH value of the solution, ionisable drugs exist as a mixture at various percentages of neutral and ionized forms. Partition of ionized forms in an organic phase is very low and generally negligible as compared to the partition of the neutral ones. Therefore, partition of ionisable compounds in n-octanol is expressed as

686

L. Grumetto et al. / European Journal of Pharmaceutical Sciences 45 (2012) 685–692

log D values; they are the weighted average of the log P values of the various forms existing in solution, decrease at increasing ionization degree, and are lower than log PN, i.e. the partition of neutral forms. In contrast, partition of ionisable compounds in phosphatidylcholine phases suggests that the presence of an electric charge on the analyte, mainly a positive one, does not hinder its interaction with charged phospholipids but, in some cases, can even promote the interaction, according to the ‘‘pH-piston hypothesis’’ (Avdeef et al., 1998). Although a monolayer of phosphatidylcholine covalently bound to a propilamino–silica core (IAM) is an approximation of a biomembrane phospholipid bilayer, it is a better model than n-octanol to mimic drug/biomembrane interactions (Ong et al., 1996). In fact, it not only supports charged moieties but is also an anisotropic phase, having polar and apolar moieties ordered in the three-dimensional space, so offering different interaction capability depending on the side they are approached by the analyte. Nevertheless, it should be pointed out that partition measured in membrane-like systems may not always reflect transmembrane permeation, as solutes may associate with the membrane interface without entering the membrane interior (Mälkiä et al., 2004). In other words, it is not clear whether ionic bonds analyte/phospholipids, i.e. the so-called ‘‘extra-lipophilic’’ contribution to the global partition observed, act as a driving force for membrane permeation or as a trapping mechanism, i.e. promoting partition but hindering passage. Although the passive passage of a drug across a biological barrier can occur through paracellular or transcellular pathway, paracellular pathway can only occur for small, usually hydrophilic, solutes and transcellular pathway is the main passive permeation mechanism for the large majority of drugs (Krämer et al., 2001; Abbott et al., 2010). With the exclusion of compounds absorbed by active transport mechanisms, transcellular pathway is the only mechanism which can be hypothesized for the passage of particular barriers such as Blood–Brain Barrier (BBB). It is a selective barrier with the endothelium forming a much tighter interface than peripheral endothelia, because the gaps between capillary endothelial cells in most part of the brain are sealed by tight junctions and thus have a severely limited permeability (van Bree et al., 1992). To date an effective model to predict BBB permeation based on the physicochemical properties of the molecules is lacking. For small sets of molecules, the passage has been found not only directly related to lipophilicity but also inversely related to Dlog P (Abraham et al., 1994; Young et al., 1988), i.e. the difference between n-octanol/ water and alkane/water partition coefficients. It is a physicochemical meaningful lipophilicity-derived parameter which mainly expresses H-bond donor acidity (El Tayar et al., 1991). Analogously, the excess of polar component to the global partition in phospholipids, mainly ionic bonds, can be quantitatively expressed by IAM IAM D log kw parameter, i.e. the difference between the log kw value observed for an analyte on an IAM-HPLC column and the value observed for an unionizable isolipophilic compound. For a series IAM of structurally diverse drugs, D log kw values were found to inversely correlate with transdermal transport (Barbato et al., 1998). In this work we considered 14 basic drugs, namely atenolol, chlorpromazine, cimetidine, clonidine, haloperidol, imipramine, mianserin, midazolam, morphine, nicotine, promazine, promethazine, propranolol, and ranitidine (scheme 1). We calculated their phospholipid affinity data at 100% aqueous eluent at pH 7.0 by measuring their retention on two different HPLC phospholipid stationary phases (namely, IAM.PC.MG and IAM.PC.DD2) at various experimental conditions. The aim of this study has been to highlight similarities and dissimilarities between partition behavior of the considered bases in n-octanol and in phospholipids in order to elucidate the role

played by lipophilic/hydrophobic and electrostatic forces in the interaction. Moreover, the investigated drugs are supposed to cross BBB by passive transport and their in vivo capability to cross BBB is reported in the literature as log BB values (Platts et al., 2001). log BB is the logarithm of the ratio between brain and blood concentration of the drug:

log BB ¼ log ð½Dbrain =½Dblood Þ We investigated the relationships between log BB values and the various partition indexes to compare their possible effectiveIAM ness in describing BBB passage and to verify whether D log kw values can offer a contribution to explain the passage mechanism. 2. Materials and methods 2.1. Materials All samples were obtained from commercial source. All chemicals were of HPLC grade and used without further purification. 2.2. Chromatographic system A Shimadzu liquid chromatographic apparatus (LC-10AD), equipped with a 7725 Rheodyne injection valve (fitted with a 20 ll loop) and a SPD-10AV UV detector (Shimadzu), set at k of maximum absorbance for each compound, was used. The chromatograms were recorded and processed by Cromatoplus software for personal computer (Shimadzu). The stainless steel columns were: – IAM.PC.MG (4.6  150 mm; Regis Chemical Company, Morton Grove, IL); – IAM.PC.DD2 (4.6  100 mm; Regis Chemical Company, Morton Grove, IL). 2.3. Chromatographic conditions Eluent: 0.1 M phosphate buffer at pH 7.0 and acetonitrile at various percentages; the flow rate was 1.0, 2.0, and 3.0 mL/min, depending on the analyte retention time. The chromatography was carried out at room temperature. Samples were dissolved in the mobile phase or in methanol (ca. 104 M) and 20 lL sample was injected. Chromatographic retention data are expressed by the logarithm of the retention factor, log k, defined as log k = log [(tr – t0)/t0] where tr and t0 are the retention times of the drug and a non-retained compound (citric acid), respectively. log k values were determined with completely aqueous eluents for all the compounds eluting within 20 min. For solutes requiring the addition of acetonitrile in the eluent, the log k values relative to IAM 100% aqueous eluent (log kw ) were calculated by performing a polycratic method of extrapolation (Braumann et al., 1983): each of them was eluted with at least four different mobile phases containing acetonitrile in percentages (u) ranging from 10% to 30% (v/v). Linear relationships between log k and u values were found for all compounds in the range of eluent composition examined (r2 P 0.99). All values of log k are the average of at least three measurements; the 95% confidence interval associated with each value never exceeded 0.04 for each log k value. Possible occurrence of retention changes due to column aging was monitored by checking the retention times of five test compounds (amlodipine, p-nitroaniline, toluene, isradipine, and ketoprofen). During the study no retention value of test compounds changed more than 4% and no correction was done to the retention values experimentally determined for the analytes.

687

L. Grumetto et al. / European Journal of Pharmaceutical Sciences 45 (2012) 685–692 OH

O

NH2

CH3

N

O H3C

Cl

O

NH OH

F

Haloperidol NH

Atenolol NH

N H3C

N

Cl

H N

N

S

H3C

N

N H

N H

cimetidine

Cl

clonidine

CH3

N

N CH3 N

Cl

H3C

N CH3

S

chlorpromazine

imipramine HO

H3C

OH

O

N

N H N H3C

mianserin

morphine N

H 3C N

N

Cl

N F N H3C

midazolam

nicotine H3C

CH3

N

CH3

N CH3

CH3

N

N

S

S

promazine

promethazine

CH3 H3 C

NH

O

H3C

N

O

CH3

+

O

N

O

OH

-

NH S HN CH3

propranolol

ranitidine Scheme 1. Chemical structures of the compounds considered.

2.4. Lipophilic parameters log PN values, i.e. partition coefficients n-octanol/aqueous phase of the neutral form of analytes, were either experimentally determined or calculated (clog P) values as reported by the program ClogP for Windows version 2.0 (Biobyte Corp., Claremont, CA).

log D7.0 values, i.e. partition coefficients n-octanol/aqueous phase measured at pH 7.0 and relative to the mixture of neutral and ionized forms existing at this pH of the aqueous phase, were calculated according to the following equation

log D7:0 ¼ log P  logð1 þ 10pKa7:0 Þ

688

L. Grumetto et al. / European Journal of Pharmaceutical Sciences 45 (2012) 685–692

Log BB values were from the literature (Platts et al., 2001). 2.5. Statistical analysis A commercially available statistical package for personal computer was used for linear regression analysis. Requirements of significant regression analysis were observed. 3. Results and discussion IAM

The values of log kw were determined on two different phospholipid stationary phases (i.e. IAM.PC.MG and IAM.PC.DD2) which differ from each other in the end-capping of the residual amino groups of the silica–propylamine core. End-capping is performed with methyl glycolate on IAM.PC.MG and with C10 and C3 alkyl chains on IAM.PC.DD2. Therefore, retention on IAM.PC.MG may be affected by additional interactions with the hydroxy groups of the end-capping agent whereas analyses of strongly lipophilic analytes on IAM.PC.DD2 may be difficult due to their too long retention times, very often requiring the addition of organic IAM modifiers to the eluents. Table 1 summarizes the log kw values IAM:MG measured for the bases considered (log kw for IAM.PC.MG IAM:DD2 and log kw for IAM.PC.DD2), together with pKa, log P, log D7.0 values. IAM data measured on IAM.PC.MG and IAM.PC.DD2 are strongly interrelated (Eq. (1)), suggesting that they substantially reflect the interactions between analytes and phospholipids with secondary mechanisms only playing a minor role in the retention. IAM:DD2

IAM:MG

log kw

¼ 0:972ð0:077Þ log kw

n ¼ 14

2

r ¼ 0:930

IAM:MG

þ 0:385ð0:149Þ

s ¼ 0:261

ð1Þ

In this and the following equations, n denotes the number of molecules considered in the derivation of the regression equation, r is the correlation coefficient, and s is the standard error of the estimate. Numbers in parentheses represent the standard error of the regression coefficients. IAM:MG IAM:DD2 As can be seen in Fig 1, both log kw and log kw values N are roughly related to the values of log P , i.e. the lipophilicity values of the neutral species of the analytes in the n-octanol/water system. The relative relation equations are: IAM:MG

log kw

¼ 0:443ð0:064Þ log PN þ 0:528ð0:210Þ

n ¼ 14

r 2 ¼ 0:797

log

IAM:DD2 kw

n ¼ 14

s ¼ 0:443

ð2Þ

¼ 0:464ð0:054Þ log PN þ 0:808ð0:174Þ

r 2 ¼ 0:862

s ¼ 0:368

ð3Þ

Table 1 Logarithms of lipophilicity values in n-octanol, of acidity constants, and of chromatographic retention factors on IAM phases for the bases considered.

*

Compound

log P

pKa

log D7.0

log kw

log kw

Atenolol Ranitidine Cimetidine Morphine Nicotine Clonidine Propranolol Haloperidol Midazolam Mianserin Promazine Imipramine Promethazine Chlorpromazine

0.16 0.27 0.40 0.76 1.17 1.57 2.98 3.23 3.27 4.41* 4.55 4.80 4.81 5.19

9.43 8.36 6.80 8.25 8.00 8.02 9.50 8.04 6.03 8.26 9.43 9.49 8.98 9.41

2.44 1.15 0.19 0.52 0.13 0.25 0.48 1.91 3.23 3.13 2.15 2.30 2.71 2.89

0.458 0.834 0.633 0.767 0.844 0.948 1.821 2.670 2.302 3.003 2.462 3.064 2.432 1.799

0.765 0.812 1.048 1.180 1.184 1.316 2.480 2.780 2.505 3.131 3.260 3.008 3.075 2.225

log P value calculated.

IAM:MG

As already reported in the literature (Amato et al., 2000; Barbato IAM et al., 1997, 2004, 2005), log kw values of bases do not relate well to log PN values because positively charged species can partition in phospholipids as strongly as, or even more strongly than, the corresponding neutral species, depending on the analyte chemical structure. According to the ‘‘pH piston hypothesis’’ (Avdeef et al., 1998), the partition behavior of bases in phospholipids can be rationalized taking into account the occurrence of electrostatic interactions between cations and the negatively charged phosphate residues of phospholipids. It is evident that electrically charged species can only partition poorly in neutral partition phases, such as n-octanol, as compared to the corresponding neutral ones. Since the presence of an electric charge on a basic analyte affects its partition in phospholipids in a different, sometimes opposite, way with respect to partition in n-octanol, the relationships IAM IAM log kw vs log D7.0 are expected to be poorer than those log kw N 7.0 vs log P , despite log D values are the partition coefficients in nIAM octanol measured at the same pH (7.0) of log kw values. Actually, IAM 7.0 the relationships between log kw and log D values of the bases considered were weaker than those with log PN (r2 = 0.742 and IAM:MG IAM:DD2 r2 = 0.750 for log kw and log kw ; respectively). In our previous studies (Taillardat-Bertschinger et al., 2002; Barbato et al., 2007) we found that, in the log P range 1.0–4.8, a single relationship exists between log P values and IAM retention data for neutral compounds, even if structurally non-related. The relationships are expressed by the following equations (Eqs. (4) and (5)):

IAM:DD2

log kw

¼ 0:792ð0:038Þ log P  0:732ð0:105Þ

n ¼ 36

r2 ¼ 0:926

log

IAM:DD2 kw

n ¼ 36

s ¼ 0:248

ð4Þ

¼ 0:934ð0:038Þ log P  0:883ð0:104Þ

r2 ¼ 0:946

s ¼ 0:246

ð5Þ IAM kw

N

A comparison between the plots log vs log P for the bases considered in the present work and the plots of neutral compounds is reported in Fig. 2. As a matter of the fact, the retention of neutral compounds on both IAM.PC.MG and IAM.PC.DD2 depends in a very similar extent on the same intermolecular recognition forces as partition in n-octanol (Taillardat-Bertschinger et al., 2002). Therefore, we assumed that the distances observed between the points of the ionized compounds under investigation and the line for neutral compounds were generated by the ‘‘extra-lipophilic’’ interactions (mainly ionic bonds) encoded in the overall interaction analyte/phospholipids. IAM They can be expressed as D log kw ; i.e. the difference between IAM the logarithm of the chromatographic retention factor (log kw ) measured for each analyte and the value expected for a hypothetical isolipophilic neutral compound, i.e. the value calculated by either equation 4 or 5, for IAM.PC.MG and IAM.PC.DD2 phase, respectively. IAM The values of D log kw of the compounds considered on both IAM:MG IAM.PC.MG and IAM.PC.DD2 phases (D log kw and D log IAM:DD2 kw ; respectively) are listed in Table 2 where the corresponding values of log BB are also reported. As can be seen, atenolol, cimetidine, clonidine, haloperidol, midazolam, morphine, nicotine, and ranitidine show significantly IAM positive D log kw values on both IAM phases, suggesting that their phospholipid affinity data include appreciable ‘‘extra-lipophilic’’ contributions. Propranolol also shows a significantly positive value, IAM but on IAM.PC.DD2 stationary phase only. Very small D log kw values were found for: (i) mianserin, on both IAM phases, (ii) imipramine and propranolol, on IAM.PC.MG phase only, and (iii) promazine, on IAM.PC.DD2 phase only. These data are in agreement with our previous observations (Amato et al., 2000; Barbato et al., 1996, 1997, 2004, 2005) and suggest that most considered bases

689

L. Grumetto et al. / European Journal of Pharmaceutical Sciences 45 (2012) 685–692

B

3.5

3.0

3.0

2.5

2.5

log k W IAM.DD2

logkW IAM.MG

A 3.5 2.0 1.5 1.0 0.5 1.0

2.0

3.0 log PN

4.0

IAM:MG

IAM:DD2

(A) or log kw

2.0 1.5 1.0 0.5 0.0 0.0

1.0

2.0

3.0 log PN bases

IAM:MG

Fig. 2. Relationships between either log kw compounds.

4.0

5.0

6.0

2.0

4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 0.0

1.0

2.0

neutral

IAM:DD2

(A) or log kw

3.0

3.0 log PN bases

4.0

5.0

6.0

4.0

5.0

6.0

neutral

(B) and log PN values for the 14 basic compounds considered in comparison to the plots of 36 neutral

Table 2 Values of the differences between observed and expected logarithms of retention IAM:MG factors on IAM.PC.MG and IAM.PC.DD2 stationary phases (D log kw and IAM:DD2 D log kw , respectively) and logarithms of the ratio brain concentration/blood concentration for the bases considered. Compound

D log kIAM:MG w

D log kIAM:DD2 w

log BB

Atenolol Ranitidine Cimetidine Morphine Nicotine Clonidine Propranolol Haloperidol Midazolam Mianserin Promazine Imipramine Promethazine Chlorpromazine

1.063 1.352 1.048 0.897 0.649 0.437 0.193 0.844 0.444 0.242 0.410 0.006 0.646 1.579

1.499 1.443 1.557 1.353 0.974 0.733 0.580 0.646 0.334 0.105 0.107 0.592 0.535 1.739

1.420 1.230 1.420 0.160 0.401 0.110 0.640 1.340 0.360 0.990 1.230 1.300 0.824 1.060

interact with phospholipids more strongly than, or at least as strongly as, an isolipophilic neutral compound. However, both chlorpromazine and promethazine show significantly negative IAM D log kw values on both IAM stationary phases; negative IAM D log kw values were also observed for promazine (on IAM.PC.MG phase only) and for imipramine (on IAM.PC.DD2 phase only). To the best of our knowledge, it is the first time that, under the reported experimental conditions, some basic compounds are found to interact with phospholipids weaker than isolipophilic neutral compounds. IAM The strength and sign, positive or negative, of the D log kw values are not related to the ionization degree of analytes, being all of them, but cimetidine and midazolam, more than 90% in their protonated form. In contrast, a weak dependence on the lipophilicity of the molecules is suggested by the pattern reported in Fig 3. As can be IAM seen, all poorly lipophilic compounds show positive D log kw values, whereas all strongly lipophilic compounds show negative valIAM ues. However, the linear regressions between D log kw and log PN (r2 = 0.788 and 0.865 for IAM.PC.MG and IAM.PC.DD2, respectively) IAM are not suitable for prediction of D log kw values from lipophilicity

B

1.5

IAM.DD2

1.0

2.0 1.5 1.0 0.5 0.0 -0.5 -1.0 -1.5 -2.0 0.0

Δ logk

0.5 0.0

-0.5 -1.0

-1.5 -2.0 0.0

1.0

(B) and log PN values for the 14 basic compounds considered.

B

3.0 2.5

IAM.MG

0.0 0.0

6.0

logk W IA M .D D 2

logkW IAM .M G

5.0

3.5

Δ logk

1.0

log PN

Fig. 1. Relationships between either log kw

A

1.5 0.5

0.0 0.0

A

2.0

1.0

2.0

3.0

4.0

5.0

6.0

N

log P N

IAM:MG

Fig. 3. Relationships between log P and either D log kw

IAM:DD2

(A) or D log kw

1.0

2.0

3.0 log PN

4.0

5.0

6.0

(B) values for the 14 basic compounds considered.

690

L. Grumetto et al. / European Journal of Pharmaceutical Sciences 45 (2012) 685–692

1.5

because the points are quite scattered and some compounds strongly deviate from an imaginary regression line. A case in point is chlorpromazine, i.e. the compound at highest log PN, which shows IAM on both IAM phases an unexpected high negative D log kw value. The observed behavior is consistent with a partition mechanism in phospholipids resulting from two concurrent interaction forces, i.e. (i) lipophilic/hydrophobic interactions and (ii) electrostatic interactions. It can be rationalized assuming that, at a given ionization degree, the first kind of interactions decrease proportionally to analyte lipophilicity (i.e. higher decrease, in absolute value, for more lipophilic compounds) whereas electrostatic interactions occur in a quite constant measure for all analytes. Therefore, for less lipophilic analytes (log PN 6 2) the net balance of the two opposite effects results in retention times higher than observed for isolipophilic neutral compounds, suggesting that the interactions are mainly supported by electrostatic forces. In the case of analytes at medium–high lipophilicity (log PN in the range 2–4.5) the decrease of lipophilicity is roughly compensated by the electrostatic interactions and the retention observed is close to that expected for isolipophilic neutral compounds. Finally, for strongly lipophilic analytes (log P P 4.5) the lipophilicity decrease caused by the presence of an electric charge is too high to be compensated by the electrostatic interactions and the observed retention is lower than the one expected for isolipophilic neutral compounds. Nevertheless, even in the case of highly lipophilic compounds the global interaction with phospholipids is stronger than expected on the basis of their partition in n-octanol at pH 7.0. Indeed, log D7.0 values, which are the partition values in n-octanol at the same pH of IAM measurements (i.e. at the same ionization degree of the analytes), underestimate the interactions actually occurring with phospholipids even for the comIAM pounds with negative D log kw values. As a matter of the fact, if we IAM:MG IAM:DD2 calculate log kw and log kw values of these analytes taking 7.0 into account their log D instead of log PN, we obtain values lower than those actually observed (Table 3). Since partition in n-octanol is uniquely based on lipophilic/ hydrophobic interactions and these forces decrease on ionization, log D7.0 values are systematically lower than the corresponding IAM log PN values, in contrast to log kw values which can be higher or lower than those expected for neutral forms. Being phospholipids

1.0

log BB

0.5 0.0 -0.5 -1.0 -1.5 -2.0 0.0

1.0

2.0

3.0 log P

4.0

5.0

6.0

N

Fig. 5. Relationship between log BB and log PN values for the 14 basic compounds considered.

a partition phase supporting much more polar moieties than n-octanol, log D7.0 values cannot be assumed as suitable parameters even to estimate the lipophilic/hydrophobic component of interaction with phospholipids remaining after analyte protonation. Therefore, to investigate about possible relationships between BBB passage and physico-chemical parameters, we decided to take into account only the following interaction indexes with biomemIAM branes: (i) log kw values, as interaction indexes for phospholipids including both lipophilic/hydrophobic and electrostatic intermolecular interaction forces, (ii) log PN values, as indexes taking into account only the lipophilic/hydrophobic based interaction potenIAM tial, and (iii) D log kw values, as measures of the excess of polarity (mainly due to electrostatic forces) in the interactions between analytes and phospholipids with respect to n-octanol partition. IAM The relationships between log BB and log kw values are reported in Fig 4. As can be seen, the plots show a trend of increasing log BB at IAM increasing log kw values but the points are quite scattered and IAM the equations describing the dependence of log BB on log kw values are less than satisfactory, both hypothesizing a linear (Eqs. (6) and (7)) and a parabolic relationship (Eqs. (8) and (9)).

Table 3 IAM:MG IAM:DD2 Comparison between the experimental affinity values for phospholipids and those calculated by log D7.0 on the two stationary IAM phases (log kw and log kw ) for the IAM compounds showing negative D log kw values. Compound

log D7.0

Chlorpromazine Promethazine Promazine Imipramine

2.89 2.71 2.15 2.30

IAM:DD2

log kw

Experimental

Calculated

Experimental

Calculated

1.799 2.432 2.462 –

1.557 1.414 0.971 –

2.225 3.075 – 3.008

1.816 1.648 – 1.265

B

1.5

1.5

1.0

1.0

0.5

0.5

0.0

log B B

log B B

A

IAM:MG

log kw

-0.5

0.0 -0.5

-1.0

-1.0

-1.5

-1.5

-2.0 0.0

0.5

1.0

1.5 2.0 2.5 log kW IAM.MG

3.0

IAM:MG

Fig. 4. Relationships between log BB and either log kw

3.5

-2.0 0.0

0.5

IAM:DD2

(A) or log kw

1.0

1.5 2.0 2.5 log kW IAM.DD2

3.0

3.5

(B) values for the 14 basic compounds considered.

691

L. Grumetto et al. / European Journal of Pharmaceutical Sciences 45 (2012) 685–692

B

1.5 1.0 0.5 0.0 -0.5 -1.0 -1.5 -2.0 -2.0 -1.5 -1.0 -0.5 0.0 ΔlogkW

0.5

IAM.MG

1.0

1.5

2.0

IAM:MG

IAM:MG

IAM:DD2

n ¼ 14 r ¼ 0:754

ð6Þ  1:580ð0:337Þ ð7Þ

n ¼ 14 r 2 ¼ 0:825

0.5

1.0

1.5

ð8Þ IAM:DD2 2

 0:530ð0:250Þðlog kw

2.0

IAM.DD2

(B) values for the 14 basic compounds considered.

is found for D log kw IAM:MG D log kw (r2 = 0.667).

values only (equation 11) but not for

log BB ¼ 0:694ð0:133ÞD log kw

 0:477ð0:122Þ

IAM:DD2 2

Þ þ 1:085ð0:177Þ

n ¼ 14 r2 ¼ 0:826 s ¼ 0:451

s ¼ 0:507 IAM:DD2

IAM:DD2

ðD log kw

 2:271ð0:617Þ log BB ¼ 3:029ð1:009Þlog kw  3:249ð0:841Þ

-1.0 -0.5 0.0

IAM:DD2

s ¼ 0:491

IAM:MG IAM:MG 2 log BB ¼ 2:528ð0:878Þlog kw  0:472ð0:249Þðlog kw Þ

n ¼ 14 r 2 ¼ 0:781

-1.5 -2.0 -2.0 -1.5

IAM:DD2

n ¼ 14 r ¼ 0:709 s ¼ 0:560 2

-0.5 -1.0

(A) or D log kw

 1:237ð0:319Þ

2

log BB ¼ 0:909ð0:150Þlog kw

0.5 0.0

ΔlogkW

Fig. 6. Relationships between log BB and either D log kw

log BB ¼ 0:888ð0:164Þlog kw

1.5 1.0

log BB

log BB

A

Þ

s ¼ 0:452

ð9Þ

ð11Þ

IAM:MG

IAM

For D log kw values, assuming that only positive D log kw values can affect BBB penetration and excluding the outlier haloperidol, a reasonable linear relationships is found with log BB values (Eq. (12)). IAM:MG

These results suggest that, in partial contrast to previously reported studies (Ducarme et al., 1998; Pehourcq et al., 2004), IAM log kw values do not describe adequately BBB permeation of the bases considered. This may occur because only one of the two components, lipophilic and electrostatic, involved in membrane phospholipid interaction is effective for membrane permeation. To verify this hypothesis we related log BB to log PN values, i.e. the partition coefficients of the neutral forms of the analytes, which take into account only lipophilic/hydrophobic interaction forces (Fig 5). As can be seen in Fig. 5, the plot suggests a parabolic dependence of BBB penetration on lipophilicity of the neutral forms of the bases considered; it is expressed by Eq. (10), slightly more significant than both Eqs. (8) and (9). However, the parabolic pattern was hypothesized on the basis of only 14 data points and should be confirmed on a larger set of data points to be extrapolated to other bases. log BB ¼ 1:229ð0:252Þlog PN  0:147ð0:047Þðlog P N Þ2  1:453ð0:240Þ n ¼ 14 r 2 ¼ 0:880 s ¼ 0:376

ð10Þ

It is interesting to note that a much poorer relationship would be found considering log D7.0 instead of log PN values (r2 = 0.685). Since IAM log kw values do not demonstrate more effective than log PN to describe BBB passage, we hypothesized that the electrostatic interaction component may act either as a noise on the driving forces promoting membrane permeation (lipophilicity) or as a ‘‘trapping’’ force in the membrane bilayer, hindering the analyte from transmembrane permeation. To elucidate this issue the relationships beIAM tween log BB and D log kw values were studied. The plots are reported in Fig. 6. IAM All analytes with negative D log kw values, on the left of the graphs, show high log BB values. The compounds are: chlorpromazine, imipramine, promazine, and promethazine on both stationIAM:DD2 ary IAM phases, as well as mianserin on the plot of D log kw values, only. Haloperidol is an exception because it shows a high IAM log BB value despite its appreciable positive D log kw values (0.844 and 0.646, on IAM.PC.MG and IAM.PC.DD2, respectively). IAM The compounds with positive D log kw values (i.e. atenolol, cimetidine, clonidine, midazolam, morphine, nicotine, propranolol, and ranitidine), located in the right side of the graphs, show IAM decreasing log BB values at increasing D log kw values. A reasonIAM able parabolic relationship between log BB and D log kw values

log BB ¼ 2:087ð0:356ÞD log kw n¼9

2

r ¼ 0:830

s ¼ 0:410

þ 1:275ð0:285Þ ð12Þ

These relationships indicate not only that analyte/phospholipid electrostatic interactions negatively affect BBB permeation but also that the effect is linearly related to the strength of these interacIAM tions as parameterized by D log kw values. Therefore, electrostatic interactions appear to act as ‘‘trapping’’ forces, hindering permeation through phospholipids although enhancing interaction. 4. Conclusion The present study supports a model of interaction between ionisable basic compounds and phospholipids based on the simultaneous occurrence of lipophilic and electrostatic forces. As the first time, the interactions between bases and phospholipids are found not only stronger or equal but also lower than those of neutral isolipophilic compounds. Thus, the influence of protonation on phospholipid partition of bases is complex and cannot be described by the log D parameter measured in n-octanol/water system. The relationships between the various physico-chemical parameters and log BB values suggest that the interaction forces with phospholipids govern partition and permeation in a different way. The only driving force for permeation is the lipophilicity of the neutral forms, log PN, whereas electrostatic attractions between cations and phospholipids promote the global interaction with phospholipids but not permeation where they act as ‘‘trapping’’ forces. According to this model, protonation should affect the membrane passage of basic compounds much less than expected on the basis of the ‘‘classical’’ partition in n-octanol; indeed, as regards to phospholipid partition, the loss of lipophilicity on protonation is, at least partly, compensated by the rise of electrostatic interactions whereas in the microenvironment of membrane thickness it is lipophilicity of the neutral forms to govern the passage. From a practical point of view, this study indicates that it is the lipophilicity of the neutral forms (log PN) to be taken into account to evaluate the BBB penetration potential of a given compound and not the ‘‘apparent’’ lipophilicity (log DpH) estimated in the noctanol partition system at a given pH value. In the design of prodrugs aimed at improve solubility (and dissolution rate) of active compounds, the reported observations suggest that the addition of a basic moiety to a molecule should improve, following its

692

L. Grumetto et al. / European Journal of Pharmaceutical Sciences 45 (2012) 685–692

ionization, its water solubility without noticeable loss of its membrane affinity and permeation capability.

References Abbott, N.J., Patabendige, A.A., Dolmen, D.E., Yusof, S.R., Begley, D.J., 2010. Structure and function in the blood–brain barrier. Neurobiol. Dis. 37 (1), 13–25. Abraham, M.H., Chadha, H.S., Mitchell, R.C., 1994. Hydrogen bonding. 33. Factors that influence the distribution of solutes between blood and brain. J. Pharm. Sci. 83, 1257–1268. Avdeef, A., Box, K.J., Comer, J.E.A., Hibbert, C., Tam, K.Y., 1998. PH-metric log P 10. Determination of liposomal membrane–water partition coefficients of ionizable drugs. Pharm. Res. 15, 209–215. Amato, M., Barbato, F., Morrica, P., Quaglia, F., La Rotonda, M.I., 2000. Interactions between amines and phospholipids: a chromatographic study on immobilized artificial membrane (IAM) stationary phase at various pH values. Helv. Chim. Acta 83, 2836–2847. Barbato, F., La Rotonda, M.I., Quaglia, F., 1996. Chromatographic indices determined on an immobilized artificial membrane (IAM) column as descriptors of lipophilic and polar interactions of 4-phenyldihydropyridine calcium-channel blockers with biomembranes. Eur. J. Med. Chem. 31, 311–318. Barbato, F., La Rotonda, M.I., Quaglia, F., 1997. Chromatographic indexes on immobilized artificial membranes for local anesthetics: relationships with activity data on closed sodium channels. Pharm. Res. 14, 1699–1705. Barbato, F., Cappello, B., Miro, A., La Rotonda, M.I., Quaglia, F., 1998. Chromatographic indexes on immobilized artificial membranes for the prediction of transdermal transport of drugs. Il Farmaco. 53, 655–661. Barbato, F., di Martino, G., Grumetto, L., La Rotonda, M.I., 2004. Prediction of drug-membrane interactions by IAM-HPLC: effects of different phospholipid stationary phases on the partition of bases. Eur. J. Pharm. Sci. 22, 261–269. Barbato, F., di Martino, G., Grumetto, L., La Rotonda, M.I., 2005. Can protonated betablockers interact with biomembranes stronger than neutral isolipophilic compounds? A chromatographic study on three different phospholipid stationary phases (IAM-HPLC). Eur. J. Pharm. Sci. 25, 379–386. Barbato, F., 2006. The use of immobilized artificial membrane (IAM) chromatography for determination of lipophilicity. Curr. Comput.-Aided Drug Des. 02 (4), 341–352. Barbato, F., Cirocco, V., Grumetto, L., La Rotonda, M.I., 2007. Comparison between immobilized artificial membrane (IAM) HPLC data and lipophilicity in n-octanol for quinolone antibacterial agents. Eur. J. Pharm. Sci. 31, 288–297.

Braumann, Th., Weber, G., Gromme, L.H., 1983. Quantitative structure–activity relationships for herbicides. Reversed phase liquid chromatographic retention parameters, log kw versus liquid–liquid partition coefficients as a model of the hydrophobicity of phenylureas, S-triazines and phenoxycarbonic acid derivatives. J. Chromatogr. 261, 329–343. Ducarme, A., Neuwels, M., Goldstein, S., Massingham, R., 1998. IAM retention and blood brain barrier penetration. Eur. J. Med. Chem. 33, 215–223. El Tayar, N., Tsai, R.S., Testa, B., Carrupt, P.-A., Leo, A., 1991. Partitioning of solutes in different solvent systems: the contribution of hydrogen bonding capacity and polarity. J. Pharm. Sci. 80 (6), 590–598. Krämer, S.D., Abbott, N.J., Begley, D.J., 2001. Biological models to study blood-brain barrier permeation. In: Testa, B., van de Waterbeemd, H., Folkers, G., Guy, R.H. (Eds.), Pharmacokinetic Optimization in Drug Research: Biological, Physicochemical and Computational Strategies. Wiley-VHCA, Zurich, pp. 127–153. Leo, A.J., Hansch, C., Elkins, D., 1971. Partition coefficients and their use. Chem. Rev. 71, 525–616. Mälkiä, A., Murtomäki, L., Urtti, A., Kontturi, K., 2004. Drug permeation in biomembranes: in vitro and in silico prediction and influence of physicochemical properties. Eur. J. Pharm. Sci. 23, 13–47. Ong, S., Liu, H., Pidgeon, C., 1996. Immobilized-Artificial-Membrane Chromatography: measurements of membrane partition coefficient and predicting drug membrane permeability. J. Chromatogr. A 728, 113–128. Pehourcq, F., Matoga, M., Bannwarth, B., 2004. Diffusion of arylproprionate nonsteroidal anti-inflammatory drugs into the cerebrospinal fluid: a quantitative structure-activity relationship approach. Fundam. Clin. Pharmacol. 18 (1), 65–70. Platts, J.A., Abraham, M.H., Zhao, Y.H., Hersey, A., Ijaz, L., Butina, D., 2001. Correlation and prediction of a large blood–brain distribution data set – an LFER study. Eur. J. Med. Chem. 36 (9), 719–730. Taillardat-Bertschinger, A., Barbato, F., Quercia, M.T., Carrupt, P.A., Reist, M., La Rotonda, M.I., Testa, B., 2002. Structural properties governing retention mechanisms on immobilized artificial membrane (IAM) HPLC columns. Helv. Chim. Acta 85, 519–532. Taillardat-Bertschinger, A., Carrupt, P.A., Barbato, F., Testa, B., 2003. Immobilized artificial membrane HPLC in drug research. J. Med. Chem. 46, 655–665. van Bree, J.B., de Beer, A.G., Danhof, M., Breiner, D.D., 1992. Drug transport across the blood–brain barrier I. Anatomical and physiological aspects. Pharm. Weekbl. Sci. 14, 305–310. Yang, C.Y., Cai, S.J., Liu, H., Pidgeon, C., 1996. Immobilized artificial membranes: screens for drug–membrane interactions. Adv. Drug Deliv. Rev. 23, 229–256. Young, R.C., Mitchell, R.C., Brown, T.H., Ganellin, C.R., Griffits, R., Jones, M., Rana, K.K., Saunders, D., et al., 1988. Development of a new physicochemical model for brain penetration and its application to the design of centrally acting H2 receptor histamine antagonists. J. Med. Chem. 31 (3), 656–671.