Immobilized Artificial Membrane HPLC Derived Parameters vs PAMPA

Jul 5, 2016 - Although the first described PAMPA-BBB technique employed a PAMPA membrane made from 2% w/v lecithin dissolved in dodecane,(9) the data ...
0 downloads 4 Views 1MB Size
Subscriber access provided by University of Sussex Library

Article

Immobilized Artificial Membrane HPLC derived parameters vs PAMPA-BBB data in estimating in situ measured Blood-Brain Barrier permeation of drugs. Lucia Grumetto, Giacomo Russo, and Francesco Barbato Mol. Pharmaceutics, Just Accepted Manuscript • DOI: 10.1021/acs.molpharmaceut.6b00397 • Publication Date (Web): 05 Jul 2016 Downloaded from http://pubs.acs.org on July 7, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Molecular Pharmaceutics is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

1

Immobilized Artificial Membrane HPLC derived parameters vs PAMPA-BBB data in

2

estimating in situ measured Blood-Brain Barrier permeation of drugs. REFEREE VERSION

3 4

Lucia Grumetto, Giacomo Russo, Francesco Barbato*

5 6

Dipartimento di Farmacia, Università degli Studi di Napoli Federico II, Via D. Montesano, 49 I-80131

7

Naples, Italy

8 9 10

Keywords:

11

Immobilized Artificial Membrane;

12

Blood-Brain Barrier;

13

Parallel Artificial Membrane Permeability Assay;

14

in situ brain perfusion;

15

Quantitative Structure-Properties Relationships.

16 17 18 19

*To whom correspondence should be addressed:

20

Prof. Francesco Barbato

21

Dipartimento di Farmacia

22

Via D. Montesano, 49

23

I-80131 Naples (Italy)

24

Tel. +39 081 678628

25

Fax +39 081 678107

26

e-mail: [email protected]

27

ACS Paragon Plus Environment

1

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

28

Page 2 of 49

ABSTRACT

29 30

The affinity indexes for phospholipids (log kWIAM) for 42 compounds were measured by High

31

Performance Liquid Chromatography (HPLC) on two different phospholipid-based stationary phases

32

(Immobilized Artificial Membrane - IAM), i.e. IAM.PC.MG and IAM.PC.DD2. The polar/electrostatic

33

interaction forces between analytes and membrane phospholipids (∆log kWIAM) were calculated as the

34

differences between the experimental values of log kWIAM and those expected for isolipophilic neutral

35

compounds having Polar Surface Area (PSA) = 0. The values of passage through a porcine brain lipid

36

extract (PBLE) artificial membrane for 36 out of the 42 compounds considered, measured by so-called

37

PAMPA-BBB technique, were taken from the literature (P0PAMPA-BBB). The values of Blood-Brain

38

Barrier (BBB) passage measured in situ, P0in situ, for 38 out of the 42 compounds considered, taken from

39

the literature, represented the permeability of the neutral forms on “efflux minimized” rodent models.

40

The present work was aimed at verifying the soundness of ∆log kWIAM at describing the potential of

41

passage through BBB as compared to data achieved by PAMPA-BBB technique.

42

In a first instance, the values of log P0PAMPA-BBB (32 data points) were found significantly related to the

43

n-octanol lipophilicity values of the neutral forms (log PN) (r2 = 0.782) whereas no significant

44

relationship (r2 = 0.246) was found with lipophilicity values of the mixtures of ionized and neutral

45

forms existing at the experimental pH 7.4 (log D7.4) as well as with either log kWIAM or ∆log kWIAM

46

values. log P0PAMPA-BBB related moderately to log P0in situ values (r2 = 0.604). The latter did not relate

47

with either n-octanol lipophilicity indexes (log PN and log D7.4) or phospholipid affinity indexes (log

48

kWIAM). In contrast, significant inverse linear relationships were observed between log P0in situ (38 data

49

points) and ∆log kWIAM values for all the compounds but ibuprofen and chlorpromazine who behaved

50

as moderate outliers (r2 = 0.656 and r2 = 0.757 for values achieved on IAM.PC.MG and IAM.PC.DD2,

51

respectively). It should be remembered that BBB passage, parameterized by log BB values, i.e. the

ACS Paragon Plus Environment

2

Page 3 of 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

52

logarithm of the ratio brain concentration/ blood concentration, was previously found linearly related to

53

∆log kWIAM for bases, ampholytes, and neutral compounds whereas for acids the measure of

54

polar/electrostatic interactions required the correction for ionization giving a different parameter, ∆’log

55

kWIAM. However, the result of the present work are not in contrast with the results previously found

56

because log P0in

57

degree. Since log P0in situ refer to the “intrinsic permeability” of the analytes regardless their ionization

58

degree, no correction for ionization of ∆log kWIAM values was needed. Furthermore log P0in situ were

59

found roughly linearly related to log BB values (i.e. the logarithm of the ratio brain concentration/blood

60

concentration measured in vivo) for all the analytes but those predominantly present at the experimental

61

pH 7.4 as anions.

62

These results suggest that ∆logkWIAM parameter is effective at predicting log P0in situ values for all the

63

analytes and that ionization affects much more markedly BBB passage of acids (yielding anions) than

64

that of the other ionisable compounds.

65

These results suggest that, at least for the data set considered, ∆logkWIAM parameters are more effective

66

than log P0PAMPA-BBB at predicting log P0in

67

appears to affect differently, and much more markedly, BBB passage of acids (yielding anions) than

68

that of the other ionisable compounds.

situ

refer to the “intrinsic permeability” of the analytes regardless their ionization

situ

values for all the analytes. Furthermore, ionization

69 70

1 - Introduction

71

The Blood-Brain Barrier (BBB) is one of the most complex and extensively studied biological barriers

72

intended to protect the mammalian brain integrity against possible injurious substances. It is made of

73

endothelial cells, narrowly adherent to one other such as forming tight junctions, which restrict the

74

passage of solutes1,2. Drug transport is strongly limited by this peculiar biological structure to pure

75

transcellular diffusion, being the paracellular route, i.e. the passage of actives through the gaps

ACS Paragon Plus Environment

3

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 49

76

between the endothelial cells, completely hindered. Passive transcellular diffusion takes place for the

77

majority of drugs but active transport may also occasionally occur affecting their BBB passage.

78

Although recently debated3, one of the most used measures reflecting in vivo BBB penetration

79

capability of drugs is log BB which is defined as: log BB = log

80

C Brain C Blood

81

in which CBrain is the concentration of the analyte in the brain tissues, and CBlood is the concentration of

82

the analyte in the blood. However, the setting up of alternative in vitro methods aimed at appreciating

83

BBB penetration potential of drugs is more and more desirable, as log BB determination is time-

84

consuming and may raise ethical issues.

85

Several efforts have been devoted to the development of methods based on the employment of cultured

86

cell lines. However, astrocytes cell cultures4 are often difficult to grow whereas cultured Caco-2 cells

87

markedly differ from brain cell structures. Both methods are and this method is difficult to standardize,

88

preventing the comparison of data determined in different laboratories. Therefore, the methods based

89

on physico-chemical properties are increasingly attractive. Although they represent a simplification of

90

the more complex phenomena occurring in vivo, they can offer a rationale on the mechanisms driving

91

barrier passage, useful in the structural optimization of new drug candidates at the early stages of

92

pharmaceutical development.

93

In our previous studies we found that BBB permeation, expressed as log BB values, inversely related,

94

by highly significant inverse linear relationships, to the polar/electrostatic interaction forces occurring

95

between drugs and membrane phospholipids. These interactions were parameterized by ∆log kWIAM and

96

∆'log kWIAM values, i.e. the differences between experimental affinity values for phospholipids and the

97

expected affinity values based on n-octanol lipophilicity. The experimental affinity values for

98

phospholipids (log kWIAM) were measured by a HPLC method on phospholipid-based stationary phases

99

(IAM-HPLC). whereas n-octanol lipophilicity values referred to either the neutral form of the analytes ACS Paragon Plus Environment

4

Page 5 of 49

1 2 3 100 4 5 6 101 7 8 102 9 10 103 11 12 13 104 14 15 16 105 17 18 106 19 20 107 21 22 23 108 24 25 109 26 27 28 110 29 30 111 31 32 112 33 34 35 113 36 37 114 38 39 115 40 41 42 116 43 44 117 45 46 118 47 48 49 119 50 51 120 52 53 54 121 55 56 122 57 58 123 59 60

Molecular Pharmaceutics

(log PN), giving ∆log kWIAM, or the mixture of neutral and ionized forms existing at the physiological pH 7.4 (log D7.4), giving ∆'log kWIAM

5–7

. The expected phospholipid affinity values were calculated

taking into account either the lipophilicity of the neutral form of the analytes (log PN) or that of the mixture of neutral and ionized forms existing at the physiological pH 7.4 (log D7.4), yielding Δlog kWIAM and Δ'log kWIAM values, respectively5–7. However, the proposed model was based on a relatively small number of analytes and the study of possible relationships between “delta” values and data of BBB passage different than log BB would allow a further validation of the proposed model and a deeper understanding of the molecular mechanisms behind BBB uptake. Recently, the relationships between in situ measured BBB passive permeation data and in vitro passage data, achieved by the so called “PAMPA-BBB” technique, have been reported in the literature8. In that study the in situ BBB passive permeability was expressed as the values of P0in situ, i.e. in situ brain perfusion permeability values (on rat and mouse) selected from studies which used some sort of carriermediated transport inhibition (e.g. GF120918, PSC833, cyclosporin A, self-inhibition at high concentrations, mdr1a(−/−)/mrp1(− /−)/brcp - knockout rodent model), allowing that the in situ data can be assumed as free of efflux effects. It is important to underline that P0in situ values refer to the permeability of the neutral form of the analytes and represent the “intrinsic permeability” regardless any effect given by ionization. The in vitro data were achieved by PAMPA-BBB technique and expressed as P0PAMPA-BBB values. PAMPA (Parallel Artificial Membrane Permeability Assay) is a method based on the determination of the permeability of substances from a donor compartment, through a lipid-infused artificial membrane, into an acceptor compartment. Although the firstly described PAMPA-BBB technique employed a PAMPA membrane made from 2% w/v lecithin dissolved in dodecane9, the data considered in the study were achieved by an improved model consisting of a new PAMPA–BBB formulation based on ACS Paragon Plus Environment

5

Molecular Pharmaceutics

1 2 3 124 4 5 6 125 7 8 126 9 10 127 11 12 13 128 14 15 129 16 17 130 18 19 20 131 21 22 132 23 24 25 133 26 27 134 28 29 135 30 31 32 136 33 34 137 35 36 138 37 38 139 39 40 140 41 42 43 141 44 45 142 46 47 143 48 49 50 144 51 52 145 53 54 55 146 56 57 147 58 59 60

Page 6 of 49

10% w/v porcine brain lipid extract (PBLE), using a fivefold higher lipid concentration in a more viscous alkane solvent than dodecane and with thinner membranes8. The relationships between log P0in situ

and log P0PAMPA-BBB for 197 compounds became significant only after the classification of the

analytes into four predominant-charge groups (positive, negative, neutral, and zwitterionic) and after, for all groups, an Abraham solvation descriptor10 was added as a second term in the equations. In the present study 36 P0PAMPA-BBB and 38 P0in situ values were selected from the literature8,11. The aim was to investigate their possible relationships with various physico-chemical parameters, including log PN, log D7.4, log kWIAM, ∆log kWIAM, and ∆'log kWIAM. Finally, the effectiveness of the various physicochemical parameters at predicting P0in situ values was compared to that of P0PAMPA-BBB values. Indeed, PAMPA method, especially in its variation PAMPA-BBB, is regarded as one of the in vitro techniques mirroring rather closely the BBB passage taking place in vivo12. However, to the best of our knowledge its validation in terms of log BB predictivity was performed only on limited sets of compounds12,13, and the effectiveness of IAM derived indexes in surrogating PAMPA-BBB data has never been investigated.

2 - Materials and methods All samples were obtained from commercial source Sigma-Aldrich (Milan, Italy), TCI (Milan, Italy) and Acros-Organics (Rodano, Italy) as reported in Table 1. All chemicals were of HPLC grade and used without further purification. 2.1 Chromatographic system LC-10AD liquid chromatographic apparatus (Shimadzu Corporation, Kyoto, Japan); SPD-10AV UV detector (Shimadzu), set at λ of maximum absorbance for each compound; 7725 Rheodyne injection valve (fitted with a 20 µl loop). Data processing: Cromatoplus software for personal computer (Shimadzu).

ACS Paragon Plus Environment

6

Page 7 of 49

1 2 3 148 4 5 6 149 7 8 150 9 10 11 151 12 13 152 14 15 153 16 17 18 154 19 20 155 21 22 156 23 24 25 157 26 27 158 28 29 159 30 31 32 160 33 34 161 35 36 37 162 38 39 163 40 41 42 164 43 44 165 45 46 166 47 48 49 167 50 51 168 52 53 54 169 55 56 170 57 58 171 59 60

Molecular Pharmaceutics

Analytical HPLC columns: - IAM.PC.MG (4.6 mm x 150 mm; 12 µm, 300Å; Regis Chemical Company, Morton Grove, IL); - IAM.PC.DD2 (4.6 mm x 100 mm, 10 µm, 300Å; Regis Chemical Company, Morton Grove, IL). 2.2 Chromatographic conditions The analyses were performed at room temperature with 0.1 M phosphate buffer at pH 7.0 in mixture with acetonitrile at various percentages. The flow rate was selected according to the retention time of each analyte (1.0, 2.0, and 3.0 mL/min). Sample preparation: each analyte was dissolved in the mobile phase or in methanol to ca. 10-4 M concentration. Chromatographic retention data are reported as log k (the logarithm of the retention factor), calculated by the expression: log k = log [(tr – t0)/t0] where tr and t0 are the retention times of the drug and a non-retained compound (acetone), respectively. Direct measurements of log k values in fully aqueous mobile phases (log kWIAM) were only possible for the compounds eluting within 20 min, whereas for the solutes requiring the addition of acetonitrile to the eluent, the log kWIAM values were calculated by an extrapolation method14: log k values were determined at four different acetonitrile percentages (ϕ) in the mobile phases (from 10 to 30% v/v) and the intercept values of the linear relationships between log k and ϕ values, found for all compounds in the range of eluent composition examined (r2 ≥ 0.99), were assumed as log kWIAM values. In particular, for basic and neutral analytes having log PN values higher than 2.5, the analyses were carried out employing 30, 25, 20, and 15% v/v acetonitrile percentages (φ). For highly lipophilic (log PN ≥ 2.5) acidic compounds and for all the other analytes, whose log PN values span the range 2.5-1.0 (2.5 > log PN ≥ 1.0) the employed acetonitrile percentages (φ) were 25, 20, 15 and 10% v/v; fully aqueous mobile phase (φ = 0) was used for analyzing the most polar compounds (log PN < 1.0). All reported log k values are the average of at least three measurements; for each log k value the 95% confidence interval associated with each value never exceeded 0.04. To avoid that the experimental ACS Paragon Plus Environment

7

Molecular Pharmaceutics

1 2 3 172 4 5 6 173 7 8 174 9 10 175 11 12 13 176 14 15 177 16 17 178 18 19 20 179 21 22 180 23 24 25 181 26 27 182 28 29 183 30 31 32 184 33 34 185 35 36 37 186 38 39 187 40 41 188 42 43 44 189 45 46 190 47 48 191 49 50 51 192 52 53 193 54 55 194 56 57 58 195 59 60

Page 8 of 49

measurements were affected by retention changes due to column ageing, the retention times of five test compounds (amlodipine, p-nitroaniline, toluene, isradipine, and ketoprofen) were checked weekly. No correction was done to the experimental retention values since no retention value of test compounds changed more than 4% during the study. 2.3 Lipophilic parameters: log PN values, i.e. partition coefficients n-octanol/aqueous phase of the neutral form of analytes, were either from the literature15–26 or calculated by the program ALOGPS version 2.127. The noctanol/aqueous buffer at pH 7.4 distribution coefficients (log D7.4) were taken from the literature16,28 or calculated according to the following equations: log D7.4 = log P – log (1 + 107.4 – pKa)

(for acids)

log D7.4 = log P – log (1 + 10 pKa – 7.4)

(for bases)

2.4 In vitro and in situ data P0PAMPA-BBB and P0in situ values were taken from the literature8,11. For the selection of in situ perfusion log P0BBB values, the following criteria were applied. If log P0BBB data on different animal models (rat, mouse, dog, cat and guinea pig) were available, the data achieved on rat were preferred as they were more numerous. For Verapamil and Diltiazem, for which log P0BBB data were not available on rat, data achieved on mouse were used instead. If more than one data achieved on rat were available, an average of the values was performed. For actively transported analytes, measures of BBB passive diffusion were selected by (i) eliminating the values referred to wild-type animal models, (ii) selecting the measures achieved either after the pharmacological inhibition of the active transport mechanisms obtained by administering cyclosporine A (Diphenhydramine, Pyrilamine and Loratadine) or determined on mdr1a(−/−)/mrp1(− /−)/brcp - knockout rat model (iii) preferring the measures achieved on non-saturable models as transport saturation is generally referred to some sort of active transport mechanism involvement. Since for indomethacin three log P0BBB values, based on data measured at

ACS Paragon Plus Environment

8

Page 9 of 49

1 2 3 196 4 5 6 197 7 8 198 9 10 11 199 12 13 200 14 15 201 16 17 18 202 19 20 203 21 22 204 23 24 25 205 26 27 206 28 29 207 30 31 32 208 33 34 209 35 36 37 210 38 39 211 40 41 212 42 43 44 213 45 46 214 47 48 215 49 50 51 216 52 53 217 54 55 218 56 57 58 219 59 60

Molecular Pharmaceutics

different pH values (5.5, 6.5, and 7.4), were available, the one extrapolated from measures at pH 7.4, i.e. the closest to the physiological and experimental pH, was selected. 2.5 Statistical analysis Linear regression analysis was performed by a commercially available statistical package for personal computer observing the requirements of significant regression analysis.

3.0 Results and discussion log P0PAMPA-BBB values were taken into account for a set of 36 analytes consisting of 23 bases, 8 acids, 2 ampholytes, and 3 neutral compounds (series 1). log P0in situ values were taken into account for a set of 38 analytes consisting of 24 bases, 6 acids, 4 ampholytes, and 4 neutral compounds (series 2). Table 1 summarizes the values of pKa, log PN, log D7.4, the logarithms of the chromatographic retention coefficients measured on IAM.PC.MG (log kWIAM.MG) and IAM.PC.DD2 (log kWIAM.DD2) stationary phases, and the suppliers of the substances for all the analytes. As can be seen, 32 compounds belong to both groups. The chemical structures of the analytes taken into account are reported in Scheme 1. As it is evident from Table 1, the analytes taken into account span a remarkably wide range of log PN values (0.02 – 5.19), covering more than five log PN units. The selection of the analytes was performed with the aim of covering and mirroring as faithfully as possible the chemical diversity of the marketed drugs and both CNS active (e.g. Buspirone, Chlopromazine, and Diazepam) and CNS inactive drugs (e.g. Cimetidine and Thiourea) were taken into account. It is interesting to note that log kwIAM.DD2 values cover a wider range than log kwIAM.MG values (5.0 vs 4.4 units). This evidence would suggest a higher selectivity of IAM.PC.DD2 in comparison with IAM.PC.MG column. Table 2 summarizes the values of ∆log kWIAM.MG, ∆log kWIAM.DD2, ∆'log kWIAM.MG and ∆'log kWIAM.DD2 values, as well as the values of log P0PAMPA-BBB, log P0in situ, and log BB. >SCHEME 1 ABOUT HERE
TABLE 1 and TABLE 2 ABOUT HERE< For the set of compounds taken into account in the present work, the chromatographic retention data achieved on the two different IAM columns, which support the same phosphatidylcholine analogues but differ from each other in the end-capping, were found strongly collinear (r2 = 0.914; F1,

40

=

426,71). This, as already previously verified for other sets of compounds5–7,29,30,32, once again suggests that IAM retention is based on the interactions with phospholipids with secondary interaction mechanisms playing an only minor role. Indeed, using two different phospholipid based stationary phases was aimed at confirming that the experimental retention data were not noticeably affected by any secondary retention mechanism. This can be reasonably assumed if the retention scales obtained on the two columns are collinear. The calculation of ∆log kWIAM and ∆'log kWIAM was based on the evidence that log kWIAM values of structurally non-related neutral compounds having Polar Surface Area (PSA) = 0 relate unambiguously with n-octanol lipophilicity values according to the following equations29: log kWIAM.MG = 0.854 (± 0.047) log P - 0.976 (± 0.156) n = 17

r2 = 0.957

s = 0.214

F1, 15 = 331,35

log kWIAM.DD2 = 1.039 (± 0.051) log P - 1.311 (± 0.169) n = 17

r2 = 0.965

s = 0.232

F1, 15 = 417.54

(1) F1, 15 α,0.001 = 16.59 (2) F1, 15 α,0.001 = 16.59

The values of ∆log kWIAM.MG, ∆'log kWIAM.MG, ∆log kWIAM.DD2, and ∆'log kWIAM.DD2 were determined according to the procedure described in our previous work29. They are the differences between the log kWIAM values observed experimentally and those calculated by equations (1) and (2) taking into account either log PN, to obtain ∆log kWIAM, or log D7.4, to obtain ∆'log kWIAM values.

3.1 Relationships between log P0PAMPA-BBB and physico-chemical parameters

ACS Paragon Plus Environment

10

Page 11 of 49

1 2 3 244 4 5 6 245 7 8 246 9 10 247 11 12 13 248 14 15 249 16 17 250 18 19 20 251 21 22 252 23 24 25 253 26 27 254 28 29 255 30 31 32 256 33 34 257 35 36 258 37 38 39 259 40 41 260 42 43 44 261 45 46 262 47 48 263 49 50 51 264 52 53 265 54 55 266 56 57 58 267 59 60

Molecular Pharmaceutics

A significant linear relationship was found between log P0PAMPA-BBB and log PN values (Fig. 1 and Eq. (3)). log P0PAMPA-BBB = 0.939 (±0.085) log PN – 6.210 (±0.276) r2 = 0.782

n = 36

s = 0.765

(3)

F1,34 = 121.63 F1, 34 α,0.001 = 12.90

>FIGURE 1 ABOUT HERE< No relationship was found with log D7.4 values (r2 = 0.246 F1,34 = 11.14). The relationships with phospholipid affinity data, log kWIAM.MG and log kWIAM.DD2 (Eq. (4) and Eq. (5)), as well as those with ∆log kWIAM.MG and ∆log kWIAM.DD2 (Eq. (6) and Eq. (7)), were poorer. Furthermore, no improvement of the relationships was observed by taking into account ∆'log kWIAM.MG and ∆'log kWIAM.DD2 for acids (data not shown). log P0PAMPA-BBB = 1.193 (±0.210) log kWIAM.MG – 5.654 (±0.425) n = 36

r2 = 0.487

s = 1.172

F1,34 = 32.29

F1, 34 α,0.001 = 12.90

log P0PAMPA-BBB = 1.192 (±0.192) log kWIAM.DD2 – 5.936 (±0.433) n = 36

r2 = 0.531

s = 1.120

F1,34 = 38.56

r2 = 0.441

s = 1.223

F1,34 = 26.86

r2 = 0.536

s = 1.114

F1,34 = 39.40

(6)

F1, 34 α,0.001 = 12.90

log P0PAMPA-BBB = -1.283 (±0.204) ∆log kWIAM.DD2 – 3.048 (±0.200) n = 36

(5)

F1, 34 α,0.001 = 12.90

log P0PAMPA-BBB = -1.460 (±0.282) ∆log kWIAM.MG – 3.045 (±0.223) n = 36

(4)

(7)

F1, 34 α,0.001 = 12.90

The above reported relationships suggest that, at least for the set of compounds considered, the data of passage measured by PAMPA-BBB substantially reflect n-octanol/water partition coefficients of the

ACS Paragon Plus Environment

11

Molecular Pharmaceutics

1 2 3 268 4 5 6 269 7 8 270 9 10 271 11 12 13 272 14 15 273 16 17 274 18 19 20 275 21 22 276 23 24 25 277 26 27 278 28 29 279 30 31 32 280 33 34 281 35 36 282 37 38 39 283 40 41 284 42 43 44 285 45 46 286 47 48 287 49 50 51 288 52 53 289 54 55 290 56 57 58 59 60

Page 12 of 49

neutral forms of the analytes, log PN. Since PAMPA barrier consists of porcine brain lipid extract (PBLE), which is better mirrored by phospholipids than n-octanol, it could appear surprising that phospholipophilicity indexes (log kWIAM) are ineffective at describing the passage through them. However, it was already reported in the literature5–7 that the passage through phospholipids is not described by log kWIAM values, i.e. the affinity indexes for them, but by “delta” values accounting for polar/electrostatic interactions. On the other hand, the fact that even “delta” values poorly relate with passage data through PBLE may indicate that the latter involve different phenomena with respect to in vivo BBB passage. 3.2 Relationships between log P0in situ, log P0PAMPA-BBB and physico-chemical parameters P0in situ are “efflux minimized” in situ brain perfusion permeability values of the neutral forms of the analytes. log P0in situ values for 197 analytes were demonstrated to relate to log P0PAMPA-BBB values8 but only after they were divided into four predominant-charge groups (at pH 7.4), i.e. bases, acids, neutral and zwitterions. Furthermore, the achievement of significant relationships required the inclusion in the equations of at least one Abraham linear free energy relation (LFER) solvation descriptor10, i.e. α (H bond acidity) and/or β (H - bond basicity). Abraham solvation descriptors are: α : H - bond acidity; β : H - bond basicity; π : polarity/polarizability due to solute–solvent interactions between bond dipoles and induced dipoles; R (dm3 mol-1 /10): excess molar refraction, which models dispersion force interaction arising from π and n electrons of the solute; Vx (dm3 mol-1 /100): McGowan molar volume of the solute.

ACS Paragon Plus Environment

12

Page 13 of 49

1 2 3 291 4 5 6 292 7 8 9 293 10 11 12 294 13 14 295 15 16 296 17 18 19 297 20 21 298 22 23 299 24 25 26 300 27 28 301 29 30 31 302 32 33 303 34 35 304 36 37 38 305 39 40 306 41 42 43 307 44 45 308 46 47 309 48 49 50 310 51 52 311 53 54 312 55 56 57 313 58 59 60

Molecular Pharmaceutics

The best relation equations between log P0in situ and log P0PAMPA-BBB found by the authors8 included the term α, for the bases, and the term (α + β), for acids and neutral compounds. For zwitterions, log P0in situ values related to the sole term (α - β) and no dependence on log P0PAMPA-BBB values was found so that the authors concluded that “ampholytes seem to be in a class by themselves, minimally dependent on lipophilicity”. In the present work 38 P0in situ values were taken from the literature11 and for 32 of such analytes the values of P0PAMPA-BBB were also reported8. The log P0in situ values for these compounds moderately related linearly to log P0PAMPA-BBB values (Fig. 2 and Eq. (8)). log P0in situ = 0.809 (±0.120) log P0PAMPA-BBB - 0.274 (±0.443) n = 32

r2 = 0.604

s = 0.989

F1,30 = 45.74

(8)

F1, 30 α,0.001 = 13.29

>FIGURE 2 ABOUT HERE< Although log P0PAMPA-BBB relate quite well to log PN values, the latter were less effective at describing in situ permeability (Fig. 3 and Eq. (9)) whereas the relationships between log P0in situ and either log kWIAM.MG or log kWIAM.DD2 values were not significant (data not shown). log P0in situ = 0.692 (±0.112) log PN – 4.970 (±0.347) n = 38

r2 = 0.517

s = 1.089

F1,36 = 38.57

(9)

F1, 36 α,0.001 = 12.61

As expected, no relationship was observed with log D7.4 values, being P0in situ permeability values referred to the neutral forms of the analytes (data not shown). >FIGURE 3 ABOUT HERE< In contrast, significant linear relationships were found between log P0in situ and both ∆log kWIAM.MG and ∆log kWIAM.DD2, but only after the exclusion of ibuprofen and chlorpromazine which deviated from the imaginary regression line generated by the other analytes (Fig. 4 and Eqs. (10) and (11)). Ibuprofen and chlorpromazine are the analytes having the lowest ∆log kWIAM values; although their behavior may

ACS Paragon Plus Environment

13

Molecular Pharmaceutics

1 2 3 314 4 5 6 315 7 8 316 9 10 317 11 12 13 318 14 15 319 16 17 320 18 19 20 321 21 22 322 23 24 25 323 26 27 324 28 29 325 30 31 32 326 33 34 327 35 36 328 37 38 39 329 40 41 330 42 43 44 331 45 46 332 47 48 333 49 50 51 334 52 53 335 54 55 336 56 57 58 337 59 60

Page 14 of 49

suggest a parabolic dependence of log P0in situ on ∆log kWIAM, this hypothesis should be verified on more than two data points. However, it is interesting to note that chlorpromazine behaved similarly in a previous work reporting the relationships between ∆log kWIAM and log BB values5. log P0in situ = -2.077 (±0.258) ∆log kWIAM.MG – 2.225 (±0.195) n = 36

r2 = 0.656

s = 0.907

F1,34 = 53.34

F1, 34 α,0.001 = 12.90

log P0in situ = -1.818 (±0.176) ∆log kWIAM.DD2 – 2.266 (±0.157) n = 36

r2 = 0.757

s = 0.762

(10)

(11)

F1,34 = 105.98 F1, 34 α,0.001 = 12.90

>FIGURE 4 ABOUT HERE< By replacing ∆log kWIAM values with ∆'log kWIAM values, either for all the analytes or for acids only, the relationships with log P0in situ were not significant (data not shown). These results confirm the soundness of ∆log kWIAM to predict BBB passage, in spite of the fact that ∆log kWIAM values, and not ∆'log kWIAM, must be used for also acidic analytes. Indeed, this may appear in contrast with the results of our previous studies, in which log BB, assumed as a measure of BBB passage, related to ∆'log kWIAM of acidic analytes and not with their ∆log kWIAM values5–7. However, log P0in situ values express the “intrinsic permeability” of the analytes, regardless their ionization degree, and these values would seem to greatly overestimate the actual capability to cross BBB of the acids that are extensively ionized at the physiological pH 7.4. To support this hypothesis, we took into account the log BB values of 34 analytes8 among those considered in the present study. It is important to underline that they should be regarded as rough estimates of the actual values, being from different sources and including 14 out of 34 values calculated in silico. Nevertheless, as can be seen in Fig. 5, log BB relate linearly with log P0in situ values for all the compounds, including the negligibly ionized acids phenytoin and theophylline, with the exception of the compounds which are present predominantly as anions at pH 7.4 (flurbiprofen, indomethacin,

ACS Paragon Plus Environment

14

Page 15 of 49

1 2 3 338 4 5 6 339 7 8 340 9 10 341 11 12 13 342 14 15 343 16 17 344 18 19 20 345 21 22 346 23 24 25 347 26 27 348 28 29 30 349 31 32 350 33 34 351 35 36 37 352 38 39 353 40 41 354 42 43 44 355 45 46 356 47 48 357 49 50 51 358 52 53 359 54 55 360 56 57 58 361 59 60

Molecular Pharmaceutics

naproxen, ibuprofen, chlorambucil). These compounds strongly deviate from the regression line generated by the other analytes (bases, neutral, ampholytes, and negligibly ionized acids) and show a trend similar to that of the latter only if log P0in situ are corrected for ionization (log Pcin situ) according to the following expression8: log Pcin situ = log P0in situ – (1 + log 10(pH-pKa)) >FIGURE 5 ABOUT HERE< Therefore, log P0in situ values reflect the actual BBB permeation potential for all the analytes but the analytes extensively ionized as anions. This suggests that, while ionization affects BBB penetration of both bases and ampholytes in a roughly constant degree, it impacts BBB penetration of acids in a quite different, and more dramatic, extent.

4 - Conclusion The present study confirms the soundness of ∆log kWIAM parameters to predict BBB passage of drugs. For prediction of logP0in situ values they demonstrated superior not only to the n-octanol lipophilicity parameters (log PN and log D7.4) but also to PAMPA-BBB data. In this study ∆log kWIAM values were able to describe in situ passage data for acidic compounds, too. However, this is not in contrast with the results of our previous studies6,7 in which the BBB passage of acidic compounds was better described by ∆'log kWIAM. Indeed, in those studies BBB passage was expressed by log BB values, accounting for the ionization degree of the acidic analytes, whereas in the present study logP0in

situ

values were

considered, which account for the permeation potential partitioning of the neutral forms regardless the relative abundance of the species existing at the experimental pH. Although our previous studies demonstrated the existence of an inverse linear dependence between membrane passage data and “delta” parameters, this study sheds new light on the different role played by ionization in the BBB passage of drugs. The results of the present study emphasize the different role

ACS Paragon Plus Environment

15

Molecular Pharmaceutics

1 2 3 362 4 5 6 363 7 8 364 9 10 365 11 12 13 366 14 15 367 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 49

played by ionization in the BBB passage of drugs. In fact, for bases and neutral compounds, log P0in situ appear as just scaled up log BB values, since a direct linear relationship, albeit weak, can be observed between them. In contrast, for compounds present predominantly as anions at the experimental pH, such relationship is no longer observed being their log BB not related to log P0in situ but to logPc in situ, i.e. BBB permeation values corrected for ionization. These results confirm that ionization affects more markedly BBB passage of acids than that of the other analytes.

ACS Paragon Plus Environment

16

Page 17 of 49

1 2 3 4 5 6 368 7 369 8 370 9 371 10 372 11 12 373 13 374 14 375 15 376 16 377 17 378 18 19 379 20 380 21 381 22 382 23 383 24 384 25 26 385 27 386 28 387 29 388 30 389 31 32 390 33 391 34 392 35 393 36 394 37 38 395 39 396 40 397 41 398 42 399 43 44 400 45 401 46 402 47 403 48 404 49 405 50 51 406 52 407 53 408 54 409 55 410 56 57 411 58 412 59 413 60

Molecular Pharmaceutics

REFERENCES (1)

(2) (3) (4)

(5)

(6)

(7)

(8)

(9)

(10) (11) (12) (13)

(14) (15)

(16) (17) (18) (19)

Van Bree, J. B.; De Boer, A. G.; Danhof, M.; Breimer, D. D. Drug Transport across the Blood-Brain Barrier. I. Anatomical and Physiological Aspects. Pharm. Weekbl. Sci. Ed. 1992, 14 (5), 305–310. Keaney, J.; Campbell, M. The Dynamic Blood-Brain Barrier. FEBS J. 2015, 282 (21), 4067– 4079. Bickel, U. How to Measure Drug Transport across the Blood-Brain Barrier. NeuroRx 2005, 2 (1), 15–26. Lundquist, S.; Renftel, M.; Brillault, J.; Fenart, L.; Cecchelli, R.; Dehouck, M.-P. Prediction of Drug Transport through the Blood-Brain Barrier in Vivo: A Comparison between Two in Vitro Cell Models. Pharm. Res. 2002, 19 (7), 976–981. Grumetto, L.; Carpentiero, C.; Barbato, F. Lipophilic and Electrostatic Forces Encoded in IAMHPLC Indexes of Basic Drugs: Their Role in Membrane Partition and Their Relationships with BBB Passage Data. Eur. J. Pharm. Sci. 2012, 45 (5), 685–692. Grumetto, L.; Carpentiero, C.; Di Vaio, P.; Frecentese, F.; Barbato, F. Lipophilic and Polar Interaction Forces between Acidic Drugs and Membrane Phospholipids Encoded in IAM-HPLC Indexes: Their Role in Membrane Partition and Relationships with BBB Permeation Data. J. Pharm. Biomed. Anal. 2013, 75, 165–172. Grumetto, L.; Russo, G.; Barbato, F. Indexes of Polar Interactions between Ionizable Drugs and Membrane Phospholipids Measured by IAM-HPLC: Their Relationships with Data of BloodBrain Barrier Passage. Eur. J. Pharm. Sci. 2014, 65, 139–146. Tsinman, O.; Tsinman, K.; Sun, N.; Avdeef, A. Physicochemical Selectivity of the BBB Microenvironment Governing Passive Diffusion - Matching with a Porcine Brain Lipid Extract Artificial Membrane Permeability Model. Pharm. Res. 2011, 28 (2), 337–363. Di, L.; Kerns, E. H.; Fan, K.; McConnell, O. J.; Carter, G. T. High Throughput Artificial Membrane Permeability Assay for Blood-Brain Barrier. Eur. J. Med. Chem. 2003, 38 (3), 223– 232. Abraham, M. H. Scales of Solute Hydrogen-Bonding - Their Construction and Application to Physicochemical and Biochemical Processes. Chem. Soc. Rev. 1993, 22 (2), 73–83. Avdeef, A. Absorption and Drug Development; John Wiley & Sons, Inc.: Hoboken, NJ, USA, 2012; pp 625–663. Müller, J.; Esso, K.; Dargó, G.; Könczöl, Á.; Balogh, G. T. Tuning the Predictive Capacity of the PAMPA-BBB Model. Eur. J. Pharm. Sci. 2015, 79, 53–60. Bicker, J.; Alves, G.; Fortuna, A.; Soares-Da-Silva, P.; Falcão, A. A New PAMPA Model Using an in-House Brain Lipid Extract for Screening the Blood-Brain Barrier Permeability of Drug Candidates. Int. J. Pharm. 2016, 501 (1-2), 102–111. Braumann, T.; Weber, G.; Grimme, L. H. Quantitative Structure—activity Relationships for Herbicides. J. Chromatogr. A 1983, 261, 329–343. Wishart, D. S.; Knox, C.; Guo, A. C.; Shrivastava, S.; Hassanali, M.; Stothard, P.; Chang, Z.; Woolsey, J. DrugBank: A Comprehensive Resource for in Silico Drug Discovery and Exploration. Nucleic Acids Res. 2006, 34 (Database issue), D668–D672. Avdeef, A. Absorption and Drug Development; John Wiley & Sons, Inc.: Hoboken, NJ, USA, 2012; pp 201–209. Lombardo, F.; Shalaeva, M. Y.; Tupper, K. A.; Gao, F.; Abraham, M. H. ElogP(oct): A Tool for Lipophilicity Determination in Drug Discovery. J. Med. Chem. 2000, 43 (15), 2922–2928. Biobyte Clog P. Claremont, CA. Gambaro, V.; Argo, A.; Cippitelli, M.; Dell’Acqua, L.; Far??, F.; Froldi, R.; Guerrini, K.; Roda, ACS Paragon Plus Environment

17

Molecular Pharmaceutics

1 2 3 414 4 415 5 6 416 7 417 8 418 9 419 10 420 11 12 421 13 422 14 423 15 424 16 425 17 426 18 19 427 20 428 21 429 22 430 23 431 24 432 25 26 433 27 434 28 435 29 436 30 437 31 32 438 33 439 34 440 35 441 36 442 37 38 443 39 444 40 445 41 446 42 447 43 44 448 45 449 46 450 47 451 48 452 49 453 50 51 454 52 455 53 456 54 55 56 57 58 59 60

(20) (21)

(22)

(23)

(24)

(25)

(26)

(27)

(28)

(29)

(30)

(31)

(32)

Page 18 of 49

G.; Rusconi, C.; Procaccianti, P. Unexpected Variation of the Codeine/morphine Ratio Following Fatal Heroin Overdose. J. Anal. Toxicol. 2014, 38 (5), 289–294. Seydel, J. K.; Wiese, M. Drug-Membrane Interactions and Pharmacodynamics; 2003; Vol. 4. Barbato, F.; La Rotonda, M. I.; Quaglia, F. Interactions of Nonsteroidal Antiinflammatory Drugs with Phospholipids: Comparison between Octanol/buffer Partition Coefficients and Chromatographic Indexes on Immobilized Artificial Membranes. J. Pharm. Sci. 1997, 86 (2), 225–229. Van de Waterbeemd, H.; Testa, B. Drug Bioavailability; van de Waterbeemd, H., Testa, B., Eds.; Methods and Principles in Medicinal Chemistry; Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim, Germany, 2008; Vol. 40. Barbato, F.; La Rotonda, M. I.; Quaglia, F. Chromatographic Indexes on Immobilized Artificial Membranes for Local Anesthetics: Relationships with Activity Data on Closed Sodium Channels. Pharm. Res. 1997, 14 (12), 1699–1705. Barbato, F.; Caliendo, G.; La Rotonda, M. I.; Morrica, P.; Silipo, C.; Vittoria, A. Relationships between Octanol-Water Partition Data, Chromatographic Indices and Their Dependence on pH in a Set of Beta-Adrenoceptor Blocking Agents. Farmaco 1990, 45 (6), 647–663. Aravagiri, M.; Yuwiler, A.; Marder, S. R. Distribution after Repeated Oral Administration of Different Dose Levels of Risperidone and 9-Hydroxy-Risperidone in the Brain and Other Tissues of Rat. Psychopharmacology (Berl). 1998, 139 (4), 356–363. Govers, H.; Ruepert, C.; Stevens, T.; van Leeuwen, C. J. Experimental Determination and Prediction of Partition Coefficients of Thioureas and Their Toxicity to. Chemosphere 1986, 15 (4), 383–393. Tetko, I. V; Gasteiger, J.; Todeschini, R.; Mauri, A.; Livingstone, D.; Ertl, P.; Palyulin, V. A.; Radchenko, E. V; Zefirov, N. S.; Makarenko, A. S.; Tanchuk, V. Y.; Prokopenko, V. V. Virtual Computational Chemistry Laboratory--Design and Description. J. Comput. Aided. Mol. Des. 2005, 19 (6), 453–463. Avdeef, A.; Barrett, D. A.; Shaw, P. N.; Knaggs, R. D.; Davis, S. S. Octanol-, Chloroform-, and Propylene Glycol Dipelargonat-Water Partitioning of Morphine-6-Glucuronide and Other Related Opiates. J. Med. Chem. 1996, 39 (22), 4377–4381. Grumetto, L.; Russo, G.; Barbato, F. Polar Interactions Drug/phospholipids Estimated by IAMHPLC vs Cultured Cell Line Passage Data: Their Relationships and Comparison of Their Effectiveness in Predicting Drug Human Intestinal Absorption. Int. J. Pharm. 2016, 500 (1-2), 275–290. Grumetto, L.; Russo, G.; Barbato, F. Relationships between Human Intestinal Absorption and Polar Interactions Drug/phospholipids Estimated by IAM-HPLC. Int. J. Pharm. 2015, 489 (1-2), 186–194. Barbato, F.; Grumetto, L.; Carpentiero, C.; Rocco, A.; Fanali, S. Capillary Electrochromatography as a New Tool to Assess Drug Affinity for Membrane Phospholipids. J. Pharm. Biomed. Anal. 2011, 54 (5), 893–899. Tsopelas, F.; Malaki, N.; Vallianatou, T.; Chrysanthakopoulos, M.; Vrakas, D.; OchsenkuhnPetropoulou, M.; Tsantili-Kakoulidou, A. Insight into the Retention Mechanism on Immobilized Artificial Membrane Chromatography Using Two Stationary Phases. J. Chromatogr. A 2015, 1396, 25–33.

ACS Paragon Plus Environment

18

Page 19 of 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

1

Immobilized Artificial Membrane HPLC derived parameters vs PAMPA-BBB data in

2

estimating in situ measured Blood-Brain Barrier permeation of drugs.

3 4

Lucia Grumetto, Giacomo Russo, Francesco Barbato*

5 6

Dipartimento di Farmacia, Università degli Studi di Napoli Federico II, Via D. Montesano, 49 I-80131

7

Naples, Italy

8 9 10

Keywords:

11

Immobilized Artificial Membrane;

12

Blood-Brain Barrier;

13

Parallel Artificial Membrane Permeability Assay;

14

in situ brain perfusion;

15

Quantitative Structure-Properties Relationships.

16 17 18 19

*To whom correspondence should be addressed:

20

Prof. Francesco Barbato

21

Dipartimento di Farmacia

22

Via D. Montesano, 49

23

I-80131 Naples (Italy)

24

Tel. +39 081 678628

25

Fax +39 081 678107

26

e-mail: [email protected]

27

ACS Paragon Plus Environment

19

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

28

Page 20 of 49

ABSTRACT

29 30

The affinity indexes for phospholipids (log kWIAM) for 42 compounds were measured by High

31

Performance Liquid Chromatography (HPLC) on two different phospholipid-based stationary phases

32

(Immobilized Artificial Membrane - IAM), i.e. IAM.PC.MG and IAM.PC.DD2. The polar/electrostatic

33

interaction forces between analytes and membrane phospholipids (∆log kWIAM) were calculated as the

34

differences between the experimental values of log kWIAM and those expected for isolipophilic neutral

35

compounds having Polar Surface Area (PSA) = 0. The values of passage through a porcine brain lipid

36

extract (PBLE) artificial membrane for 36 out of the 42 compounds considered, measured by so-called

37

PAMPA-BBB technique, were taken from the literature (P0PAMPA-BBB). The values of Blood-Brain

38

Barrier (BBB) passage measured in situ, P0in situ, for 38 out of the 42 compounds considered, taken from

39

the literature, represented the permeability of the neutral forms on “efflux minimized” rodent models.

40

The present work was aimed at verifying the soundness of ∆log kWIAM at describing the potential of

41

passage through BBB as compared to data achieved by PAMPA-BBB technique.

42

In a first instance, the values of log P0PAMPA-BBB (32 data points) were found significantly related to the

43

n-octanol lipophilicity values of the neutral forms (log PN) (r2 = 0.782) whereas no significant

44

relationship (r2 = 0.246) was found with lipophilicity values of the mixtures of ionized and neutral

45

forms existing at the experimental pH 7.4 (log D7.4) as well as with either log kWIAM or ∆log kWIAM

46

values. log P0PAMPA-BBB related moderately to log P0in situ values (r2 = 0.604). The latter did not relate

47

with either n-octanol lipophilicity indexes (log PN and log D7.4) or phospholipid affinity indexes (log

48

kWIAM). In contrast, significant inverse linear relationships were observed between log P0in situ (38 data

49

points) and ∆log kWIAM values for all the compounds but ibuprofen and chlorpromazine who behaved

50

as moderate outliers (r2 = 0.656 and r2 = 0.757 for values achieved on IAM.PC.MG and IAM.PC.DD2,

51

respectively). Since log P0in

situ

refer to the “intrinsic permeability” of the analytes regardless their

ACS Paragon Plus Environment

20

Page 21 of 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

52

ionization degree, no correction for ionization of ∆log kWIAM values was needed. Furthermore log P0in

53

situ

54

concentration/blood concentration measured in vivo) for all the analytes but those predominantly

55

present at the experimental pH 7.4 as anions.

56

These results suggest that, at least for the data set considered, ∆logkWIAM parameters are more effective

57

than log P0PAMPA-BBB at predicting log P0in

58

appears to affect differently, and much more markedly, BBB passage of acids (yielding anions) than

59

that of the other ionisable compounds.

were found roughly linearly related to log BB values (i.e. the logarithm of the ratio brain

situ

values for all the analytes. Furthermore, ionization

60 61

1 - Introduction

62

The Blood-Brain Barrier (BBB) is one of the most complex and extensively studied biological barriers

63

intended to protect the mammalian brain integrity against possible injurious substances. It is made of

64

endothelial cells, narrowly adherent to one other such as forming tight junctions, which restrict the

65

passage of solutes1,2. Drug transport is strongly limited by this peculiar biological structure to pure

66

transcellular diffusion, being the paracellular route, i.e. the passage of actives through the gaps

67

between the endothelial cells, completely hindered. Passive transcellular diffusion takes place for the

68

majority of drugs but active transport may also occasionally occur affecting their BBB passage.

69

Although recently debated3, one of the most used measures reflecting in vivo BBB penetration

70

capability of drugs is log BB which is defined as: log BB = log

71

C Brain C Blood

72

in which CBrain is the concentration of the analyte in the brain tissues, and CBlood is the concentration of

73

the analyte in the blood. However, the setting up of alternative in vitro methods aimed at appreciating

74

BBB penetration potential of drugs is more and more desirable, as log BB determination is time-

75

consuming and may raise ethical issues. ACS Paragon Plus Environment

21

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 49

76

Several efforts have been devoted to the development of methods based on the employment of cultured

77

cell lines. However, astrocytes cell cultures4 are often difficult to grow and this method is difficult to

78

standardize, preventing the comparison of data determined in different laboratories. Therefore, the

79

methods based on physico-chemical properties are increasingly attractive. Although they represent a

80

simplification of the more complex phenomena occurring in vivo, they can offer a rationale on the

81

mechanisms driving barrier passage, useful in the structural optimization of new drug candidates at the

82

early stages of pharmaceutical development.

83

In our previous studies we found that BBB permeation, expressed as log BB values, inversely related,

84

by highly significant inverse linear relationships, to the polar/electrostatic interaction forces occurring

85

between drugs and membrane phospholipids. These interactions were parameterized by ∆log kWIAM and

86

∆'log kWIAM values, i.e. the differences between experimental affinity values for phospholipids and the

87

expected affinity values based on n-octanol lipophilicity. The experimental affinity values for

88

phospholipids (log kWIAM) were measured by a HPLC method on phospholipid-based stationary phases

89

(IAM-HPLC). The expected phospholipid affinity values were calculated taking into account either the

90

lipophilicity of the neutral form of the analytes (log PN) or that of the mixture of neutral and ionized

91

forms existing at the physiological pH 7.4 (log D7.4), yielding Δlog kWIAM and Δ'log kWIAM values,

92

respectively5–7.

93

However, the proposed model was based on a relatively small number of analytes and the study of

94

possible relationships between “delta” values and data of BBB passage different than log BB would

95

allow a further validation of the proposed model and a deeper understanding of the molecular

96

mechanisms behind BBB uptake.

97

Recently, the relationships between in situ measured BBB passive permeation data and in vitro passage

98

data, achieved by the so called “PAMPA-BBB” technique, have been reported in the literature8. In that

99

study the in situ BBB passive permeability was expressed as the values of P0in situ, i.e. in situ brain ACS Paragon Plus Environment

22

Page 23 of 49

1 2 3 100 4 5 6 101 7 8 102 9 10 103 11 12 13 104 14 15 105 16 17 106 18 19 20 107 21 22 108 23 24 25 109 26 27 110 28 29 111 30 31 32 112 33 34 113 35 36 114 37 38 39 115 40 41 116 42 43 44 117 45 46 118 47 48 119 49 50 51 120 52 53 121 54 55 122 56 57 58 123 59 60

Molecular Pharmaceutics

perfusion permeability values (on rat and mouse) selected from studies which used some sort of carriermediated transport inhibition (e.g. GF120918, PSC833, cyclosporin A, self-inhibition at high concentrations, mdr1a(−/−)/mrp1(− /−)/brcp - knockout rodent model), allowing that the in situ data can be assumed as free of efflux effects. It is important to underline that P0in situ values refer to the permeability of the neutral form of the analytes and represent the “intrinsic permeability” regardless any effect given by ionization. The in vitro data were achieved by PAMPA-BBB technique and expressed as P0PAMPA-BBB values. PAMPA (Parallel Artificial Membrane Permeability Assay) is a method based on the determination of the permeability of substances from a donor compartment, through a lipid-infused artificial membrane, into an acceptor compartment. Although the firstly described PAMPA-BBB technique employed a PAMPA membrane made from 2% w/v lecithin dissolved in dodecane9, the data considered in the study were achieved by an improved model consisting of a new PAMPA–BBB formulation based on 10% w/v porcine brain lipid extract (PBLE), using a fivefold higher lipid concentration in a more viscous alkane solvent than dodecane and with thinner membranes8. The relationships between log P0in situ

and log P0PAMPA-BBB for 197 compounds became significant only after the classification of the

analytes into four predominant-charge groups (positive, negative, neutral, and zwitterionic) and after, for all groups, an Abraham solvation descriptor10 was added as a second term in the equations. In the present study 36 P0PAMPA-BBB and 38 P0in situ values were selected from the literature8,11. The aim was to investigate their possible relationships with various physico-chemical parameters, including log PN, log D7.4, log kWIAM, ∆log kWIAM, and ∆'log kWIAM. Finally, the effectiveness of the various physicochemical parameters at predicting P0in situ values was compared to that of P0PAMPA-BBB values. Indeed, PAMPA method, especially in its variation PAMPA-BBB, is regarded as one of the in vitro techniques mirroring rather closely the BBB passage taking place in vivo12. However, to the best of our knowledge its validation in terms of log BB predictivity was performed only on limited sets of compounds12,13, and

ACS Paragon Plus Environment

23

Molecular Pharmaceutics

1 2 3 124 4 5 6 125 7 8 126 9 127 10 11 12 128 13 14 129 15 16 17 130 18 19 131 20 21 132 22 23 24 133 25 26 134 27 28 29 135 30 31 136 32 33 137 34 35 36 138 37 38 139 39 40 41 140 42 43 141 44 45 142 46 47 48 143 49 50 144 51 52 53 145 54 55 146 56 57 147 58 59 60

Page 24 of 49

the effectiveness of IAM derived indexes in surrogating PAMPA-BBB data has never been investigated.

2 - Materials and methods All samples were obtained from Sigma-Aldrich (Milan, Italy), TCI (Milan, Italy) and Acros-Organics (Rodano, Italy) as reported in Table 1. All chemicals were of HPLC grade and used without further purification. 2.1 Chromatographic system LC-10AD liquid chromatographic apparatus (Shimadzu Corporation, Kyoto, Japan); SPD-10AV UV detector (Shimadzu), set at λ of maximum absorbance for each compound; 7725 Rheodyne injection valve (fitted with a 20 µl loop). Data processing: Cromatoplus software for personal computer (Shimadzu). Analytical HPLC columns: - IAM.PC.MG (4.6 mm x 150 mm; 12 µm, 300Å; Regis Chemical Company, Morton Grove, IL); - IAM.PC.DD2 (4.6 mm x 100 mm, 10 µm, 300Å; Regis Chemical Company, Morton Grove, IL). 2.2 Chromatographic conditions The analyses were performed at room temperature with 0.1 M phosphate buffer at pH 7.0 in mixture with acetonitrile at various percentages. The flow rate was selected according to the retention time of each analyte (1.0, 2.0, and 3.0 mL/min). Sample preparation: each analyte was dissolved in the mobile phase or in methanol to ca. 10-4 M concentration. Chromatographic retention data are reported as log k (the logarithm of the retention factor), calculated by the expression: log k = log [(tr – t0)/t0] where tr and t0 are the retention times of the drug and a non-retained compound (acetone), respectively. Direct measurements of log k values in fully aqueous mobile phases (log kWIAM) were only possible for the compounds eluting within 20 min,

ACS Paragon Plus Environment

24

Page 25 of 49

1 2 3 148 4 5 6 149 7 8 150 9 10 11 151 12 13 152 14 15 16 153 17 18 154 19 20 155 21 22 23 156 24 25 157 26 27 158 28 29 30 159 31 32 160 33 34 35 161 36 37 162 38 39 163 40 41 42 164 43 44 165 45 46 166 47 48 49 167 50 51 168 52 53 169 54 55 56 170 57 58 59 60

Molecular Pharmaceutics

whereas for the solutes requiring the addition of acetonitrile to the eluent, the log kWIAM values were calculated by an extrapolation method14: log k values were determined at four different acetonitrile percentages (ϕ) in the mobile phases (from 10 to 30% v/v) and the intercept values of the linear relationships between log k and ϕ values, found for all compounds in the range of eluent composition examined (r2 ≥ 0.99), were assumed as log kWIAM values. In particular, for basic and neutral analytes having log PN values higher than 2.5, the analyses were carried out employing 30, 25, 20, and 15% v/v acetonitrile percentages (φ). For highly lipophilic (log PN ≥ 2.5) acidic compounds and for all the other analytes, whose log PN values span the range 2.5-1.0 (2.5 > log PN ≥ 1.0) the employed acetonitrile percentages (φ) were 25, 20, 15 and 10% v/v; fully aqueous mobile phase (φ = 0) was used for analyzing the most polar compounds (log PN < 1.0). All reported log k values are the average of at least three measurements; for each log k value the 95% confidence interval associated with each value never exceeded 0.04. To avoid that the experimental measurements were affected by retention changes due to column ageing, the retention times of five test compounds (amlodipine, p-nitroaniline, toluene, isradipine, and ketoprofen) were checked weekly. No correction was done to the experimental retention values since no retention value of test compounds changed more than 4% during the study. 2.3 Lipophilic parameters: log PN values, i.e. partition coefficients n-octanol/aqueous phase of the neutral form of analytes, were either from the literature15–26 or calculated by the program ALOGPS version 2.127. The noctanol/aqueous buffer at pH 7.4 distribution coefficients (log D7.4) were taken from the literature16,28 or calculated according to the following equations: log D7.4 = log P – log (1 + 107.4 – pKa)

(for acids)

log D7.4 = log P – log (1 + 10 pKa – 7.4)

(for bases)

ACS Paragon Plus Environment

25

Molecular Pharmaceutics

1 2 3 171 4 5 6 172 7 8 173 9 10 174 11 12 13 175 14 15 176 16 17 177 18 19 20 178 21 22 179 23 24 25 180 26 27 181 28 29 182 30 31 32 183 33 34 184 35 36 185 37 38 39 186 40 41 187 42 43 44 188 45 46 189 47 48 190 49 50 51 191 52 53 192 54 55 193 56 57 58 194 59 60

Page 26 of 49

2.4 In vitro and in situ data P0PAMPA-BBB and P0in situ values were taken from the literature8,11. For the selection of in situ perfusion log P0BBB values, the following criteria were applied. If log P0BBB data on different animal models (rat, mouse, dog, cat and guinea pig) were available, the data achieved on rat were preferred as they were more numerous. For Verapamil and Diltiazem, for which log P0BBB data were not available on rat, data achieved on mouse were used instead. If more than one data achieved on rat were available, an average of the values was performed. For actively transported analytes, measures of BBB passive diffusion were selected by (i) eliminating the values referred to wild-type animal models, (ii) selecting the measures achieved either after the pharmacological inhibition of the active transport mechanisms obtained by administering cyclosporine A (Diphenhydramine, Pyrilamine and Loratadine) or determined on mdr1a(−/−)/mrp1(− /−)/brcp - knockout rat model (iii) preferring the measures achieved on non-saturable models as transport saturation is generally referred to some sort of active transport mechanism involvement. Since for indomethacin three log P0BBB values, based on data measured at different pH values (5.5, 6.5, and 7.4), were available, the one extrapolated from measures at pH 7.4, i.e. the closest to the physiological and experimental pH, was selected. 2.5 Statistical analysis Linear regression analysis was performed by a commercially available statistical package for personal computer observing the requirements of significant regression analysis.

3.0 Results and discussion log P0PAMPA-BBB values were taken into account for a set of 36 analytes consisting of 23 bases, 8 acids, 2 ampholytes, and 3 neutral compounds (series 1). log P0in situ values were taken into account for a set of 38 analytes consisting of 24 bases, 6 acids, 4 ampholytes, and 4 neutral compounds (series 2). Table 1 summarizes the values of pKa, log PN, log D7.4, the logarithms of the chromatographic retention

ACS Paragon Plus Environment

26

Page 27 of 49

1 2 3 195 4 5 6 196 7 8 197 9 10 198 11 12 13 199 14 15 200 16 17 201 18 19 20 202 21 22 203 23 24 25 204 26 27 205 28 29 206 30 31 32 207 33 34 208 35 36 209 37 38 39 210 40 41 211 42 43 44 212 45 46 213 47 48 214 49 50 51 215 52 53 216 54 55 217 56 57 58 59 60

Molecular Pharmaceutics

coefficients measured on IAM.PC.MG (log kWIAM.MG) and IAM.PC.DD2 (log kWIAM.DD2) stationary phases, and the suppliers of the substances for all the analytes. As can be seen, 32 compounds belong to both groups. The chemical structures of the analytes taken into account are reported in Scheme 1. As it is evident from Table 1, the analytes taken into account span a remarkably wide range of log PN values (0.02 – 5.19), covering more than five log PN units. The selection of the analytes was performed with the aim of covering and mirroring as faithfully as possible the chemical diversity of the marketed drugs and both CNS active (e.g. Buspirone, Chlopromazine, and Diazepam) and CNS inactive drugs (e.g. Cimetidine and Thiourea) were taken into account. It is interesting to note that log kwIAM.DD2 values cover a wider range than log kwIAM.MG values (5.0 vs 4.4 units). This evidence would suggest a higher selectivity of IAM.PC.DD2 in comparison with IAM.PC.MG column. Table 2 summarizes the values of ∆log kWIAM.MG, ∆log kWIAM.DD2, ∆'log kWIAM.MG and ∆'log kWIAM.DD2 values, as well as the values of log P0PAMPA-BBB, log P0in situ, and log BB. >SCHEME 1 ABOUT HERE< >TABLE 1 and TABLE 2 ABOUT HERE< For the set of compounds taken into account in the present work, the chromatographic retention data achieved on the two different IAM columns, which support the same phosphatidylcholine analogues but differ from each other in the end-capping, were found strongly collinear (r2 = 0.914; F1,

40

=

426,71). This, as already previously verified for other sets of compounds5–7,29,30,32, once again suggests that IAM retention is based on the interactions with phospholipids with secondary interaction mechanisms playing an only minor role. Indeed, using two different phospholipid based stationary phases was aimed at confirming that the experimental retention data were not noticeably affected by any secondary retention mechanism. This can be reasonably assumed if the retention scales obtained on the two columns are collinear.

ACS Paragon Plus Environment

27

Molecular Pharmaceutics

1 2 3 218 4 5 6 219 7 8 220 9 10 221 11 12 13 222 14 15 223 16 17 224 18 19 20 225 21 22 226 23 24 25 227 26 27 228 28 29 229 30 31 32 230 33 34 231 35 36 232 37 38 39 233 40 41 234 42 43 44 235 45 46 236 47 48 237 49 50 51 238 52 53 239 54 55 240 56 57 58 241 59 60

Page 28 of 49

The calculation of ∆log kWIAM and ∆'log kWIAM was based on the evidence that log kWIAM values of structurally non-related neutral compounds having Polar Surface Area (PSA) = 0 relate unambiguously with n-octanol lipophilicity values according to the following equations29: log kWIAM.MG = 0.854 (± 0.047) log P - 0.976 (± 0.156) n = 17

r2 = 0.957

s = 0.214

F1, 15 = 331,35

log kWIAM.DD2 = 1.039 (± 0.051) log P - 1.311 (± 0.169) n = 17

r2 = 0.965

s = 0.232

F1, 15 = 417.54

(1) F1, 15 α,0.001 = 16.59 (2) F1, 15 α,0.001 = 16.59

The values of ∆log kWIAM.MG, ∆'log kWIAM.MG, ∆log kWIAM.DD2, and ∆'log kWIAM.DD2 were determined according to the procedure described in our previous work29. They are the differences between the log kWIAM values observed experimentally and those calculated by equations (1) and (2) taking into account either log PN, to obtain ∆log kWIAM, or log D7.4, to obtain ∆'log kWIAM values.

3.1 Relationships between log P0PAMPA-BBB and physico-chemical parameters A significant linear relationship was found between log P0PAMPA-BBB and log PN values (Fig. 1 and Eq. (3)). log P0PAMPA-BBB = 0.939 (±0.085) log PN – 6.210 (±0.276) n = 36

r2 = 0.782

s = 0.765

(3)

F1,34 = 121.63 F1, 34 α,0.001 = 12.90

>FIGURE 1 ABOUT HERE< No relationship was found with log D7.4 values (r2 = 0.246 F1,34 = 11.14). The relationships with phospholipid affinity data, log kWIAM.MG and log kWIAM.DD2 (Eq. (4) and Eq. (5)), as well as those with ∆log kWIAM.MG and ∆log kWIAM.DD2 (Eq. (6) and Eq. (7)), were poorer. Furthermore, no improvement of the relationships was observed by taking into account ∆'log kWIAM.MG and ∆'log kWIAM.DD2 for acids (data not shown).

ACS Paragon Plus Environment

28

Page 29 of 49

1 2 3 242 4 5 6 243 7 8 244 9 10 245 11 12 13 246 14 15 247 16 17 248 18 19 20 249 21 22 250 23 24 25 251 26 27 252 28 29 253 30 31 32 254 33 34 255 35 36 256 37 38 39 257 40 41 258 42 43 44 259 45 46 260 47 48 261 49 50 51 262 52 53 263 54 55 264 56 57 58 59 60

Molecular Pharmaceutics

log P0PAMPA-BBB = 1.193 (±0.210) log kWIAM.MG – 5.654 (±0.425) n = 36

r2 = 0.487

s = 1.172

F1,34 = 32.29

F1, 34 α,0.001 = 12.90

log P0PAMPA-BBB = 1.192 (±0.192) log kWIAM.DD2 – 5.936 (±0.433) n = 36

r2 = 0.531

s = 1.120

F1,34 = 38.56

r2 = 0.441

s = 1.223

F1,34 = 26.86

r2 = 0.536

s = 1.114

F1,34 = 39.40

(6)

F1, 34 α,0.001 = 12.90

log P0PAMPA-BBB = -1.283 (±0.204) ∆log kWIAM.DD2 – 3.048 (±0.200) n = 36

(5)

F1, 34 α,0.001 = 12.90

log P0PAMPA-BBB = -1.460 (±0.282) ∆log kWIAM.MG – 3.045 (±0.223) n = 36

(4)

(7)

F1, 34 α,0.001 = 12.90

The above reported relationships suggest that, at least for the set of compounds considered, the data of passage measured by PAMPA-BBB substantially reflect n-octanol/water partition coefficients of the neutral forms of the analytes, log PN. Since PAMPA barrier consists of porcine brain lipid extract (PBLE), which is better mirrored by phospholipids than n-octanol, it could appear surprising that phospholipophilicity indexes (log kWIAM) are ineffective at describing the passage through them. However, it was already reported in the literature5–7 that the passage through phospholipids is not described by log kWIAM values, i.e. the affinity indexes for them, but by “delta” values accounting for polar/electrostatic interactions. On the other hand, the fact that even “delta” values poorly relate with passage data through PBLE may indicate that the latter involve different phenomena with respect to in vivo BBB passage. 3.2 Relationships between log P0in situ, log P0PAMPA-BBB and physico-chemical parameters

ACS Paragon Plus Environment

29

Molecular Pharmaceutics

1 2 3 265 4 5 6 266 7 8 267 9 10 268 11 12 13 269 14 15 16 270 17 18 19 271 20 21 272 22 23 24 25 273 26 27 274 28 29 275 30 31 32 276 33 34 277 35 36 278 37 38 39 279 40 41 280 42 43 44 281 45 46 282 47 48 283 49 50 51 284 52 53 285 54 55 286 56 57 58 59 60

Page 30 of 49

P0in situ are “efflux minimized” in situ brain perfusion permeability values of the neutral forms of the analytes. log P0in situ values for 197 analytes were demonstrated to relate to log P0PAMPA-BBB values8 but only after they were divided into four predominant-charge groups (at pH 7.4), i.e. bases, acids, neutral and zwitterions. Furthermore, the achievement of significant relationships required the inclusion in the equations of at least one Abraham linear free energy relation (LFER) solvation descriptor10, i.e. α (H bond acidity) and/or β (H - bond basicity). The best relation equations between log P0in situ and log P0PAMPA-BBB found by the authors8 included the term α, for the bases, and the term (α + β), for acids and neutral compounds. For zwitterions, log P0in situ values related to the sole term (α - β) and no dependence on log P0PAMPA-BBB values was found so that the authors concluded that “ampholytes seem to be in a class by themselves, minimally dependent on lipophilicity”. In the present work 38 P0in situ values were taken from the literature11 and for 32 of such analytes the values of P0PAMPA-BBB were also reported8. The log P0in situ values for these compounds moderately related linearly to log P0PAMPA-BBB values (Fig. 2 and Eq. (8)). log P0in situ = 0.809 (±0.120) log P0PAMPA-BBB - 0.274 (±0.443) n = 32

r2 = 0.604

s = 0.989

F1,30 = 45.74

(8)

F1, 30 α,0.001 = 13.29

>FIGURE 2 ABOUT HERE< Although log P0PAMPA-BBB relate quite well to log PN values, the latter were less effective at describing in situ permeability (Fig. 3 and Eq. (9)) whereas the relationships between log P0in situ and either log kWIAM.MG or log kWIAM.DD2 values were not significant (data not shown). log P0in situ = 0.692 (±0.112) log PN – 4.970 (±0.347) n = 38

r2 = 0.517

s = 1.089

F1,36 = 38.57

(9)

F1, 36 α,0.001 = 12.61

ACS Paragon Plus Environment

30

Page 31 of 49

1 2 3 287 4 5 6 288 7 8 289 9 10 290 11 12 13 291 14 15 292 16 17 293 18 19 20 294 21 22 295 23 24 25 296 26 27 297 28 29 298 30 31 32 299 33 34 300 35 36 301 37 38 39 302 40 41 303 42 43 44 304 45 46 305 47 48 306 49 50 51 307 52 53 308 54 55 309 56 57 58 59 60

Molecular Pharmaceutics

As expected, no relationship was observed with log D7.4 values, being P0in situ permeability values referred to the neutral forms of the analytes (data not shown). >FIGURE 3 ABOUT HERE< In contrast, significant linear relationships were found between log P0in situ and both ∆log kWIAM.MG and ∆log kWIAM.DD2, but only after the exclusion of ibuprofen and chlorpromazine which deviated from the imaginary regression line generated by the other analytes (Fig. 4 and Eqs. (10) and (11)). Ibuprofen and chlorpromazine are the analytes having the lowest ∆log kWIAM values; although their behavior may suggest a parabolic dependence of log P0in situ on ∆log kWIAM, this hypothesis should be verified on more than two data points. However, it is interesting to note that chlorpromazine behaved similarly in a previous work reporting the relationships between ∆log kWIAM and log BB values5. log P0in situ = -2.077 (±0.258) ∆log kWIAM.MG – 2.225 (±0.195) n = 36

r2 = 0.656

s = 0.907

F1,34 = 53.34

F1, 34 α,0.001 = 12.90

log P0in situ = -1.818 (±0.176) ∆log kWIAM.DD2 – 2.266 (±0.157) n = 36

r2 = 0.757

s = 0.762

(10)

(11)

F1,34 = 105.98 F1, 34 α,0.001 = 12.90

>FIGURE 4 ABOUT HERE< By replacing ∆log kWIAM values with ∆'log kWIAM values, either for all the analytes or for acids only, the relationships with log P0in situ were not significant (data not shown). These results confirm the soundness of ∆log kWIAM to predict BBB passage, in spite of the fact that ∆log kWIAM values, and not ∆'log kWIAM, must be used for also acidic analytes. Indeed, this may appear in contrast with the results of our previous studies, in which log BB, assumed as a measure of BBB passage, related to ∆'log kWIAM of acidic analytes and not with their ∆log kWIAM values5–7. However, log P0in situ values express the “intrinsic permeability” of the analytes, regardless their ionization degree,

ACS Paragon Plus Environment

31

Molecular Pharmaceutics

1 2 3 310 4 5 6 311 7 8 312 9 10 313 11 12 13 314 14 15 315 16 17 316 18 19 20 317 21 22 318 23 24 25 319 26 27 320 28 29 321 30 31 32 322 33 34 323 35 36 324 37 38 39 325 40 41 326 42 43 44 327 45 46 328 47 48 329 49 50 51 330 52 53 331 54 55 332 56 57 58 333 59 60

Page 32 of 49

and these values would seem to greatly overestimate the actual capability to cross BBB of the acids that are extensively ionized at the physiological pH 7.4. To support this hypothesis, we took into account the log BB values of 34 analytes8 among those considered in the present study. It is important to underline that they should be regarded as rough estimates of the actual values, being from different sources and including 14 out of 34 values calculated in silico. Nevertheless, as can be seen in Fig. 5, log BB relate linearly with log P0in situ values for all the compounds, including the negligibly ionized acids phenytoin and theophylline, with the exception of the compounds which are present predominantly as anions at pH 7.4 (flurbiprofen, indomethacin, naproxen, ibuprofen, chlorambucil). These compounds strongly deviate from the regression line generated by the other analytes (bases, neutral, ampholytes, and negligibly ionized acids) and show a trend similar to that of the latter only if log P0in situ are corrected for ionization (log Pcin situ) according to the following expression8: log Pcin situ = log P0in situ – (1 + log 10(pH-pKa)) >FIGURE 5 ABOUT HERE< Therefore, log P0in situ values reflect the actual BBB permeation potential for all the analytes but the analytes extensively ionized as anions. This suggests that, while ionization affects BBB penetration of both bases and ampholytes in a roughly constant degree, it impacts BBB penetration of acids in a quite different, and more dramatic, extent.

4 - Conclusion The present study confirms the soundness of ∆log kWIAM parameters to predict BBB passage of drugs. For prediction of logP0in situ values they demonstrated superior not only to the n-octanol lipophilicity parameters (log PN and log D7.4) but also to PAMPA-BBB data. In this study ∆log kWIAM values were able to describe in situ passage data for acidic compounds, too. However, this is not in contrast with the

ACS Paragon Plus Environment

32

Page 33 of 49

1 2 3 334 4 5 6 335 7 8 336 9 10 337 11 12 13 338 14 15 339 16 17 340 18 19 20 341 21 22 342 23 24 25 343 26 27 344 28 29 345 30 31 32 346 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

results of our previous studies6,7 in which the BBB passage of acidic compounds was better described by ∆'log kWIAM. Indeed, in those studies BBB passage was expressed by log BB values, accounting for the ionization degree of the acidic analytes, whereas in the present study logP0in

situ

values were

considered, which account for the permeation potential of the neutral forms regardless the relative abundance of the species existing at the experimental pH. Although our previous studies demonstrated the existence of an inverse linear dependence between membrane passage data and “delta” parameters, this study sheds new light on the different role played by ionization in the BBB passage of drugs. In fact, for bases and neutral compounds, log P0in situ appear as just scaled up log BB values, since a direct linear relationship, albeit weak, can be observed between them. In contrast, for compounds present predominantly as anions at the experimental pH, such relationship is no longer observed being their log BB not related to log P0in situ but to logPc in situ, i.e. BBB permeation values corrected for ionization. These results confirm that ionization affects more markedly BBB passage of acids than that of the other analytes.

ACS Paragon Plus Environment

33

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 34 of 49

REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46

(1)

(2) (3) (4)

(5)

(6)

(7)

(8)

(9)

(10) (11) (12) (13)

(14) (15)

(16) (17) (18) (19)

Van Bree, J. B.; De Boer, A. G.; Danhof, M.; Breimer, D. D. Drug Transport across the Blood-Brain Barrier. I. Anatomical and Physiological Aspects. Pharm. Weekbl. Sci. Ed. 1992, 14 (5), 305–310. Keaney, J.; Campbell, M. The Dynamic Blood-Brain Barrier. FEBS J. 2015, 282 (21), 4067– 4079. Bickel, U. How to Measure Drug Transport across the Blood-Brain Barrier. NeuroRx 2005, 2 (1), 15–26. Lundquist, S.; Renftel, M.; Brillault, J.; Fenart, L.; Cecchelli, R.; Dehouck, M.-P. Prediction of Drug Transport through the Blood-Brain Barrier in Vivo: A Comparison between Two in Vitro Cell Models. Pharm. Res. 2002, 19 (7), 976–981. Grumetto, L.; Carpentiero, C.; Barbato, F. Lipophilic and Electrostatic Forces Encoded in IAMHPLC Indexes of Basic Drugs: Their Role in Membrane Partition and Their Relationships with BBB Passage Data. Eur. J. Pharm. Sci. 2012, 45 (5), 685–692. Grumetto, L.; Carpentiero, C.; Di Vaio, P.; Frecentese, F.; Barbato, F. Lipophilic and Polar Interaction Forces between Acidic Drugs and Membrane Phospholipids Encoded in IAM-HPLC Indexes: Their Role in Membrane Partition and Relationships with BBB Permeation Data. J. Pharm. Biomed. Anal. 2013, 75, 165–172. Grumetto, L.; Russo, G.; Barbato, F. Indexes of Polar Interactions between Ionizable Drugs and Membrane Phospholipids Measured by IAM-HPLC: Their Relationships with Data of BloodBrain Barrier Passage. Eur. J. Pharm. Sci. 2014, 65, 139–146. Tsinman, O.; Tsinman, K.; Sun, N.; Avdeef, A. Physicochemical Selectivity of the BBB Microenvironment Governing Passive Diffusion - Matching with a Porcine Brain Lipid Extract Artificial Membrane Permeability Model. Pharm. Res. 2011, 28 (2), 337–363. Di, L.; Kerns, E. H.; Fan, K.; McConnell, O. J.; Carter, G. T. High Throughput Artificial Membrane Permeability Assay for Blood-Brain Barrier. Eur. J. Med. Chem. 2003, 38 (3), 223– 232. Abraham, M. H. Scales of Solute Hydrogen-Bonding - Their Construction and Application to Physicochemical and Biochemical Processes. Chem. Soc. Rev. 1993, 22 (2), 73–83. Avdeef, A. Absorption and Drug Development; John Wiley & Sons, Inc.: Hoboken, NJ, USA, 2012; pp 625–663. Müller, J.; Esso, K.; Dargó, G.; Könczöl, Á.; Balogh, G. T. Tuning the Predictive Capacity of the PAMPA-BBB Model. Eur. J. Pharm. Sci. 2015, 79, 53–60. Bicker, J.; Alves, G.; Fortuna, A.; Soares-Da-Silva, P.; Falcão, A. A New PAMPA Model Using an in-House Brain Lipid Extract for Screening the Blood-Brain Barrier Permeability of Drug Candidates. Int. J. Pharm. 2016, 501 (1-2), 102–111. Braumann, T.; Weber, G.; Grimme, L. H. Quantitative Structure—activity Relationships for Herbicides. J. Chromatogr. A 1983, 261, 329–343. Wishart, D. S.; Knox, C.; Guo, A. C.; Shrivastava, S.; Hassanali, M.; Stothard, P.; Chang, Z.; Woolsey, J. DrugBank: A Comprehensive Resource for in Silico Drug Discovery and Exploration. Nucleic Acids Res. 2006, 34 (Database issue), D668–D672. Avdeef, A. Absorption and Drug Development; John Wiley & Sons, Inc.: Hoboken, NJ, USA, 2012; pp 201–209. Lombardo, F.; Shalaeva, M. Y.; Tupper, K. A.; Gao, F.; Abraham, M. H. ElogP(oct): A Tool for Lipophilicity Determination in Drug Discovery. J. Med. Chem. 2000, 43 (15), 2922–2928. Biobyte Clog P. Claremont, CA. Gambaro, V.; Argo, A.; Cippitelli, M.; Dell’Acqua, L.; Far??, F.; Froldi, R.; Guerrini, K.; Roda, ACS Paragon Plus Environment

34

Page 35 of 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90

(20) (21)

(22)

(23)

(24)

(25)

(26)

(27)

(28)

(29)

(30)

(31)

(32)

Molecular Pharmaceutics

G.; Rusconi, C.; Procaccianti, P. Unexpected Variation of the Codeine/morphine Ratio Following Fatal Heroin Overdose. J. Anal. Toxicol. 2014, 38 (5), 289–294. Seydel, J. K.; Wiese, M. Drug-Membrane Interactions and Pharmacodynamics; 2003; Vol. 4. Barbato, F.; La Rotonda, M. I.; Quaglia, F. Interactions of Nonsteroidal Antiinflammatory Drugs with Phospholipids: Comparison between Octanol/buffer Partition Coefficients and Chromatographic Indexes on Immobilized Artificial Membranes. J. Pharm. Sci. 1997, 86 (2), 225–229. Van de Waterbeemd, H.; Testa, B. Drug Bioavailability; van de Waterbeemd, H., Testa, B., Eds.; Methods and Principles in Medicinal Chemistry; Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim, Germany, 2008; Vol. 40. Barbato, F.; La Rotonda, M. I.; Quaglia, F. Chromatographic Indexes on Immobilized Artificial Membranes for Local Anesthetics: Relationships with Activity Data on Closed Sodium Channels. Pharm. Res. 1997, 14 (12), 1699–1705. Barbato, F.; Caliendo, G.; La Rotonda, M. I.; Morrica, P.; Silipo, C.; Vittoria, A. Relationships between Octanol-Water Partition Data, Chromatographic Indices and Their Dependence on pH in a Set of Beta-Adrenoceptor Blocking Agents. Farmaco 1990, 45 (6), 647–663. Aravagiri, M.; Yuwiler, A.; Marder, S. R. Distribution after Repeated Oral Administration of Different Dose Levels of Risperidone and 9-Hydroxy-Risperidone in the Brain and Other Tissues of Rat. Psychopharmacology (Berl). 1998, 139 (4), 356–363. Govers, H.; Ruepert, C.; Stevens, T.; van Leeuwen, C. J. Experimental Determination and Prediction of Partition Coefficients of Thioureas and Their Toxicity to. Chemosphere 1986, 15 (4), 383–393. Tetko, I. V; Gasteiger, J.; Todeschini, R.; Mauri, A.; Livingstone, D.; Ertl, P.; Palyulin, V. A.; Radchenko, E. V; Zefirov, N. S.; Makarenko, A. S.; Tanchuk, V. Y.; Prokopenko, V. V. Virtual Computational Chemistry Laboratory--Design and Description. J. Comput. Aided. Mol. Des. 2005, 19 (6), 453–463. Avdeef, A.; Barrett, D. A.; Shaw, P. N.; Knaggs, R. D.; Davis, S. S. Octanol-, Chloroform-, and Propylene Glycol Dipelargonat-Water Partitioning of Morphine-6-Glucuronide and Other Related Opiates. J. Med. Chem. 1996, 39 (22), 4377–4381. Grumetto, L.; Russo, G.; Barbato, F. Polar Interactions Drug/phospholipids Estimated by IAMHPLC vs Cultured Cell Line Passage Data: Their Relationships and Comparison of Their Effectiveness in Predicting Drug Human Intestinal Absorption. Int. J. Pharm. 2016, 500 (1-2), 275–290. Grumetto, L.; Russo, G.; Barbato, F. Relationships between Human Intestinal Absorption and Polar Interactions Drug/phospholipids Estimated by IAM-HPLC. Int. J. Pharm. 2015, 489 (1-2), 186–194. Barbato, F.; Grumetto, L.; Carpentiero, C.; Rocco, A.; Fanali, S. Capillary Electrochromatography as a New Tool to Assess Drug Affinity for Membrane Phospholipids. J. Pharm. Biomed. Anal. 2011, 54 (5), 893–899. Tsopelas, F.; Malaki, N.; Vallianatou, T.; Chrysanthakopoulos, M.; Vrakas, D.; OchsenkuhnPetropoulou, M.; Tsantili-Kakoulidou, A. Insight into the Retention Mechanism on Immobilized Artificial Membrane Chromatography Using Two Stationary Phases. J. Chromatogr. A 2015, 1396, 25–33.

ACS Paragon Plus Environment

35

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

Aminopyrine

Caffeine

Codeine

Amitriptyline

Carbamazepine

Diazepam

Antipyrine

Page 36 of 49

Buspirone

Chlorambucil

Cimetidine

Diltiazem

Diphenhydramine

ACS Paragon Plus Environment

Page 37 of 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

Molecular Pharmaceutics

Domperidone

Fluphenazine

Hydroxyzine

Doxorubicin

Flurbiprofen

Fluvastatin

Ibuprofen

Haloperidol

Indomethacin

ACS Paragon Plus Environment

Fluoxetine

Hydrocortisone

Lidocaine

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

Loratadine

Naproxen

Propranolol

Methadone

Page 38 of 49

Metoclopramide

Nicotinamide

Morphine

Phenytoin

Pyrilamine

ACS Paragon Plus Environment

Progesterone

Quinine

Risperidone

Page 39 of 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

Molecular Pharmaceutics

Temazepam

Theobromine

Theophylline

Verapamil

Scheme 1. Chemical structures of the compounds considered.

ACS Paragon Plus Environment

Thiourea

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 40 of 49

Table 1 – pKa values, logarithms of lipophilicity values in n-octanol and of chromatographic retention factors on IAM stationary phases for the compounds considered. Series 2 1, 2 1, 2 1, 2 1, 2 1, 2 2 1, 2 1, 2 1, 2 1, 2 1, 2 1, 2 1, 2 1, 2 1, 2 1, 2 1, 2 1 1 1, 2 1 1, 2 1, 2 1, 2 2 1, 2 1, 2 1, 2 1, 2 2 1, 2 1, 2 1, 2 1, 2 1, 2 1, 2 2 1 1, 2 2 1, 2

Compound Aminopyrine Amitriptyline Antipyrine Buspirone Caffeine Carbamazepine Chlorambucil Chlorpromazine Cimetidine Codeine Diazepam Diltiazem Diphenydramine Domperidone Doxorubicin Flurbiprofen Fluoxetine Fluphenazine Fluvastatin Haloperidol Hydrocortisone Hydroxyzine Ibuprofen Indomethacin Lidocaine Loratadine Methadone Metoclopramide Morphine Naproxen Nicotinamide Phenytoin Progesterone Propranolol Pyrilamine Quintine Risperidone Temazepam Teobromine Theophylline Thiourea Verapamil

logPN

pKa 4.50 9.49 0.65 7.57 0.52 4.82/4.62 9.41 6.80 7.99 3.40 8.94 8.76 9.00 8.00/9.93 4.18 9.96 7.84 4.56 8.29 6.62 4.41 3.96 7.90 4.27 9.05 9.08 9.48/9.25 4.15 3.54 8.28 9.50 9.12 8.57 7.81 11.66/1.58 9.90 8.60 8.90

15

1.00 4.9215 0.5616 2.6315 0.0715 2.1917 3.4116 5.1918 0.4015 1.3919 2.9920 3.4118 3.1816 3.9015 1.9716 4.1621 4.5016 4.3615 4.3022 4.3015 1.6115 3.5516 4.1316 4.2715 2.4823 4.8027 3.9315 2.7216 0.8916 3.2416 0.3715 2.4715 3.8718 3.2824 3.2715 3.4415 3.0425 2.1915 0.7815 0.0215 1.0826 3.7915

logD7.4 0.99 2.8016 0.56 2.3916 -0.07 2.19 0.6116 2.89 0.19 0.2216 2.99 2.02 1.80 2.29 -0.3316 0.91 2.2816 4.33 1.1716 3.57 1.61 3.48 1.4416 0.68 1.53 4.80 2.26 1.03 -0.0728 0.0916 -0.37 2.42 3.87 0.48 1.80 2.19 1.66 2.19 -0.78 -0.04 -1.08 1.88

logkWIAM.MG 29

0.536 2.8817 0.5996 1.742 0.12829 1.0397 1.2887 1.7995 0.6335 0.8557 1.73129 2.12129 2.21929 2.7907 2.223 1.870 3.181 3.5887 2.21030 2.6705 1.550 2.9087 0.9726 2.39021 1.112 3.354 2.646 1.199 0.7675 1.26021 0.351 1.7876 2.7696 1.8215 2.1097 2.313 2.1897 2.190 -0.1566 -0.13029 -0.817 2.89229

logkWIAM.DD2 29

0.573 3.1227 0.39329 1.986 0.11629 1.7177 1.8977 2.2255 1.0485 1.2907 2.1986 2.78029 2.17029 3.2137 1.764 1.950 3.522 3.9577 2.84330 2.78030 1.660 2.9657 1.1706 2.0806 1.650 3.623 2.828 1.902 1.1805 1.33930 -0.179 1.7896 3.3176 2.4805 1.8937 2.810 2.0287 1.69731 -0.0886 0.10029 -1.081 3.08530

Chemical character B B B B B N B/A B B B B B B B B/A A B B A B N B A A B B B B B/A A B A N B B B B B/A A A N B

Supplier Sigma Sigma Sigma Sigma Sigma Sigma Sigma Sigma Sigma Sigma Sigma TCI TCI TCI TCI ACROS TCI Sigma Acros TCI TCI Sigma ACROS ACROS ACROS TCI Sigma TCI Sigma ACROS ACROS ACROS ACROS ACROS TCI TCI TCI Sigma ACROS ACROS ACROS ACROS

Series 1: compounds for relationships with logP0PAMPA-BBB values. Series 2: compounds for relationships with log P0in situ values. A = acid; B = base; N = neutral; B/A = ampholyte. For the ampholytes the two pKa values reported refer to the acidic and basic functions, respectively.

ACS Paragon Plus Environment

Page 41 of 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

Molecular Pharmaceutics

Table 2 – Values of the differences between observed and expected logarithms of retention factors on IAM.PC.MG and IAM.PC.DD2 stationary phases (∆log kWIAM.MG, ∆'log kWIAM.MG, ∆log kWIAM.DD2, and ∆'log kWIAM.DD2, respectively), of logarithms of PAMPA-BBB permeation data (logP0PAMPA-BBB), of in situ BBB permeation data (log P0in situ), and of log BB. Compound Aminopyrine Amitriptyline Antipyrine Buspirone Caffeine Carbamazepine Chlorambucil Chlorpromazine Cimetidine Codeine Diazepam Diltiazem Diphenydramine Domperidone Doxorubicin Flurbiprofen Fluoxetine Fluphenazine Fluvastatin Haloperidol Hydrocortisone Hydroxyzine Ibuprofen Indomethacin Lidocaine Loratadine Methadone Metoclopramide Morphine Naproxen Nicotinamide Phenytoin Progesterone Propranolol

∆log kWIAM.MG

∆log kWIAM.DD2

∆'log kWIAM.MG

∆'log kWIAM.DD2

logP0PAMPA-BBB

log P0in situ

log BB

0.658 -0.345 1.097 0.472 1.164 0.145 -0.648 -1.657 1.267 0.644 0.154 0.185 0.479 0.435 1.517 -0.707 0.314 0.841 -0.486 -0.026 1.151 0.852 -1.579 -0.281 -0.030 0.231 0.266 -0.148 0.983 -0.531 1.643 0.654 0.440 -0.004

0.845 -0.679 1.122 0.564 1.500 0.753 -0.335 -1.856 1.943 1.157 0.402 0.548 0.177 0.472 1.028 -1.061 0.158 0.738 -0.314 -0.377 1.298 0.588 -1.810 -1.046 0.384 -0.053 0.056 0.387 1.566 -0.716 1.516 0.534 0.607 0.383

0.667 1.466 1.097 0.677 1.164 0.145 1.743 0.307 1.447 1.643 0.154 1.372 1.658 1.810 3.481 2.069 2.210 0.866 2.187 0.597 1.151 0.912 0.718 2.785 0.781 0.231 1.692 1.295 1.803 2.159 1.643 0.696 0.440 2.387

0.855 1.524 1.122 0.814 1.500 0.753 2.574 0.533 2.162 2.372 0.402 1.992 1.611 2.145 3.418 2.316 2.464 0.769 2.938 0.382 1.298 0.660 0.985 2.684 1.371 -0.053 1.791 2.142 2.564 2.556 1.516 0.586 0.607 3.292

-1.27 -6.14 -3.85 -5.92 -4.54 -1.46 -6.40 -3.68 -3.83 -3.18 -2.64 -3.36 -4.23 -2.35 -1.39 -2.36 -3.56 -2.06 -5.17 -3.72 -2.64 -2.67 -3.65 -2.18 -1.11 -4.47 -2.63 -4.34 -3.58 -1.93

-3.30 -1.48 -3.98 -2.53 -3.85 -3.74 -0.80 -1.33 -5.92 -3.80 -3.35 -2.81 -1.90 -4.45 -5.55 -0.58 -1.11 -3.35 -5.85 -1.22 -1.06 -3.24 -3.48 -2.02 -2.86 -5.43 -0.77 -4.88 -4.09 -3.74 -1.26

1.30a -0.10 0.40a -0.06 0.00 -1.70 1.06a -1.42 0.55 0.52 0.30a 0.70a -0.80 -0.83 0.30a 0.50 1.51 -0.90a -0.18 -1.26 0.10a 0.90a -0.70a -0.16 0.10a -0.14 0.20a 0.64a

ACS Paragon Plus Environment

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

Pyrilamine Quinine Risperidone Temazepam Theobromine Theophylline Thiourea Verapamil

0.292 0.351 0.569 1.296 1.486 0.863 1.081 0.631

-0.194 0.547 0.180 0.733 2.033 1.432 1.352 0.458

1.548 1.419 1.747 1.296 1.486 0.880 1.081 2.262

1.334 1.846 1.614 0.733 2.033 1.453 1.352 2.443

a : value calculated8

ACS Paragon Plus Environment

Page 42 of 49

-2.63 -2.99 -4.00 -8.00 -6.41 -2.03

-2.04 -3.45 -2.94 -3.35 -5.09 -5.45 -2.19

0.49 0.60a -0.02 -0.29 -0.70

Page 43 of 49

0.0 -1.0 -2.0 logP0PAMPA-BBB

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

-3.0 -4.0 -5.0 -6.0 -7.0 -8.0 -9.0 -1.0

0.0

1.0

2.0

3.0

4.0

log PN

ACS Paragon Plus Environment

5.0

6.0

Molecular Pharmaceutics

0.0 -1.0 -2.0 logP0in situ

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 44 of 49

-3.0 -4.0 -5.0 -6.0 -7.0 -7.0

-6.0

-5.0

-4.0

-3.0

-2.0

logP0PAMPA-BBB

ACS Paragon Plus Environment

-1.0

0.0

Page 45 of 49

0.0 -1.0 -2.0 logP 0in situ

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

-3.0 -4.0 -5.0 -6.0 -7.0 -2.0

-1.0

0.0

1.0

2.0

3.0

logPN

ACS Paragon Plus Environment

4.0

5.0

6.0

Molecular Pharmaceutics

A 0.0 -1.0

logP0in situ

-2.0

ibuprofen chlorpromazine

-3.0 -4.0 -5.0 -6.0 -7.0 -2.0

-1.5

-1.0

-0.5

0.0 ∆ logkW

0.5

1.0

1.5

2.0

0.5

1.0

1.5

2.0

IAM.MG

B 0.0 -1.0 -2.0 logP0in situ

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 46 of 49

ibuprofen chlorpromazine

-3.0 -4.0 -5.0 -6.0 -7.0 -2.0

-1.5

-1.0

-0.5

0.0 ∆ logkW

IAM.DD2

ACS Paragon Plus Environment

Page 47 of 49

2.0 1.5 1.0 0.5 log BB

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

Flurbiprofen Naproxen Ibuprofen

0.0 -0.5 -1.0

Indomethacin

-1.5 Chlorambucil

-2.0 -7.0

-6.0

-5.0

-4.0

-3.0

-2.0

-1.0

logP0in situ

♦ compounds; ■ acids (log P0in situ); × acids (log Pcin situ).

ACS Paragon Plus Environment

0.0

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 48 of 49

Captions to the figures: Figure 1. Relationship between log P0PAMPA-BBB and log PN values Figure 2. Relationship between log P0in situ and log P0PAMPA-BBB values. Figure 3. Relationship between log P0in situ and log PN values. Figure 4. Relationships between log P0in situ and ∆log kwIAM values. Figure 5. Relationship between log BB and logP0in situ values. For acids both log P0in situ (■) and log Pcin situ (×) values are reported.

ACS Paragon Plus Environment

Page 49 of 49

Molecular Pharmaceutics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Paragon Plus Environment