FORMULATION - BIOMIMETIC LIPOPHILICITY | Liposome Water Partitioning

Alex Avdeef EDITOR’S NOTE: This chapter is from the book Absorption and Drug Development: Solubility, Permeability, and Charge State, 2nd Edition by Alex Avdeef, published in May by John Wiley & Sons (http://www.wiley.com/WileyCDA/WileyTitle/productCd-1118057457.html). For further information on the book, visit the author’s website: www.in-adme.com/book.html.
This chapter considers “biomimetic” lipophilicity, where partition coefficients of drugs are determined in the liposomes–water system. Unilamellar vesicles formed with phosphatidylcholine provide a lipophilicity scale, expressed as log P_MEM, which is different from that of octanol. The log P_MEM can be used as a component in property or biological activity prediction models. For ionizable molecules, the coefficients depend on pH and are called distribution or apparent partition coefficients, log D_MEM.
Given a wide range of pH, monoprotic molecules show a sigmoidal curve when log D_MEM is expressed as a function of pH. At the asymptotic top of such curves, log D_MEM is equal to the log P_MEM constant describing the liposome–water partition of the neutral species. At the asymptotic bottom of the sigmoidal curve, log D_MEM is equal to log P_^SIPMEM, the constant describing surface ion-pairing, charged drug paired with surface charge components in the bilayer. In the liposome system, charged species partitioning—association with the surface—is about 100 times greater than ion-pair partitioning in octanol. Consequently, the “diff 3–4” approximation in the octanol system becomes the “diff 1–2” approximation in the liposome system. The phospholipid-drug interaction discussed here serves as the foundation for the PAMPA model in Chapter 7. A database of log P_MEM and log P_^SIPMEMfor 114 molecules is listed at the end of the chapter.
The legacy octanol–water partition model has some shortcomings. Notably, it is not very “biological.” Given that liposomes—vesicles with internal aqueous compartments separated from the bulk solution by a phospholipid bilayer—are made of the main ingredients found in all biological membranes, a substantial effort has been made to characterize drug partitioning in the more biomimetic liposome–water system.^1-68

Tetrad of Equilibria and Surface Ion Pairing

Figure 5.1 shows a tetrad of equilibrium reactions related to the partitioning of a drug between an aqueous environment and that of the bilayer formed from phospholipids. (Only half of the bilayer is shown in Figure 5.1.) The subscript “MEM” designates the partitioning medium as that of a membrane vesicle formed from a phospholipid bilayer. Equations (4.1)–(4.4) apply.
The pK_a^MEM in Figure 5 .1 refers to the “membrane” pK_a. Its meaning is similar to that of pK_a^OCT: When the concentrations of the uncharged and the charged species in the membrane phase are equal, the aqueous pH at that point defines pK_a^MEM, which is described for a weak base as:
BH⁺_MEM B_MEM + H⁺ K_a^MEM = [B]_MEM[H⁺]/[BH⁺]_MEM

---

The salt dependency of constants discussed in Sections 4.2 and 4.3 also applies to the pK_a^MEM and log P^SIP_MEM constants in Figure 5.1. Although the surface ion-pair and membrane-pK_a are conditional constants, the dependence on solution counterion concentration differs from that of octanol.^{57, 66}

click for larger view

FIGURE 5.1. Phospholipid membrane–water tetrad equilibria. Only half of a bilayer is shown.

It is thought that when a charged drug migrates into the lipid environment of a liposome, the counterion that at first accompanies it may be exchanged with the zwitterionic phosphatidylcholine head groups, as suggested in Figure 5.1, while still maintaining local charge neutrality. As the nature of the ion pair may be different with liposome partitioning, the term “surface ion-pair” is used to denote it. The term diff log P_MEM will be used to designate the difference between the neutral species partitioning and the surface ion-pair partitioning [cf. Eq. (4.6)].

Data Sources

There are no convenient databases for liposome log P values. Most measured quantities need to be ferreted from original publications.^{1,2,5-11,67,69} The handbook edited by Cevc is a comprehensive collection of properties of phospholipids, including extensive compilations of structural data from X-ray crystallographic studies.⁴ Constituent-lipid distributions in various biological membranes have been reported.^4,12,57

Location of Drugs Partitioned into Bilayers

Based on the nuclear Overhauser effect in a ^31P{¹H} NMR study of egg phosphatidylcholine (eggPC) bilayers, Yeagle and colleagues concluded that the N-methyl hydrogen atoms were in proximity to phosphate oxygen atoms in neighboring phospholipids, suggesting that the surface of the bilayer was a “shell” of interlocking (intermolecular) electrostatic associations.²³ Added cholesterol bound below the polar head groups and did not interact with them directly. However, its presence indirectly broke up some of the surface structure, making the surface more polar and open to hydration.
Boulanger and colleagues studied the interactions of the local anesthetics procaine and tetracaine with egg PC multilamellar vesicles (MLV,52–650 mM), as a function of pH, using deuterium NMR as a structural probe.^44,45 They proposed a three-site model, similar to that in Figure 5.1, except that the membrane-bound species, both charged and uncharged, had two different locations, one a weakly bound surface site (predominantly occupied at pH 5.5) and the other a strongly bound deeper site (predominantly occupied at pH 9.5).

The partition of lipophilic drugs into lipid phases is often thought to be entropy-driven, a “hydrophobic” effect. Bäuerle and Seelig studied the thermodynamics of amlodipine and nimodipine binding.

Membrane partition coefficients (D_MEM) were estimated for both sites. Westman and colleagues further elaborated the model by applying the Gouy–Chapman theory.46 When a positively charged drug partitions into the bilayer, a Cl− is likely bound to the surface, to maintain local charge neutrality. They found unexpected low values of diff log P_MEM of 0.77 for tetracaine and 1.64 for procaine (cf. Section 4.6), much smaller than the value expected in octanol–water partitioning. Kelusky and Smith, also using deuterium NMR, proposed that there was an electrostatic bond formed at pH 5.5 between the protonated drug and the phosphate groups, )=P–O…+H₃N–(, and a hydrogen bond formed between the aminobenzene proton and the acyl carbonyl oxygen.⁴⁷ At pH 9.5, the electrostatic bond breaks as the secondary amine moves deeper into the interior of the bilayer; however, the aminobenzene H-bond, )=CO…H₂N–(, continues to be an anchoring point.
Bäuerle and Seelig studied the structural aspects of amlodipine (weak base, primary amine pK_a 9.24 [2]) and nimodipine (nonionizable) binding to phospholipid bilayers, using NMR, microcalorimetry, and zeta–potential measurements.¹⁹ They were able to see evidence of interactions of amlodipine with the cis double bond in the acyl chains. They saw no clear evidence for )=P–O−…+H₃N–( electrostatic interactions.
Herbette and co-workers studied the structures of drugs bound to liposomes using a low-angle X-ray diffraction technique.^49-52,70 Although the structural details were coarse, it was apparent that different drugs position in different locations of the bilayer. For example, amlodipine is charged when it partitions into a bilayer at physiological pH: The aromatic dihydropyridine ring is buried in the vicinity of the carbonyl groups of the acyl chains, while the –NH⁺₃ endpoints toward the aqueous phase, with the positive charge located near the phosphate negatively charged oxygen atoms.^50-52 A much more lipophilic molecule, amiodarone (weak base with pK_a 10.24; Table 3.14), positioned itself closer to the center of the hydrocarbon interior.⁴⁹

Thermodynamics of Partitioning: Entropy or Enthalpy Driven?

TABLE 5.1. Energy of Transfer (kJ·mol−1) into Lipid Phase for 4-Methylphenol

Davis and colleagues studied the thermodynamics of the partitioning process of substituted phenols and anisoles in octanol, cyclohexane, and dimyristoylphosphatidylcholine at 22° C (below the gel-liquid transition temperature of DMPC).¹⁸ Table 5.1 shows the results for 4-methylphenol.
The phenol partitioned into the lipid phases in the order DMPC > octanol > cyclohexane, as indicated by ΔG_tr. That is, the free energy of transfer into DMPC was greater than into octanol or cyclohexane. Partitioning was generally entropy-driven, but the enthalpy and entropy parts of the free energy of transfer differed greatly among the three lipid systems (Table 5.1).
Octanol was the only lipid to have an exothermic heat of transfer (negative enthalpy), due to H-bond stabilization of the transferred solute, not found in cyclohexane. Although ΔH_tr in the DMPC system is a high positive number (endothermic), not favoring partitioning into the lipid phase, the entropy increase (+114.1 J·mol−1) was even greater, more than enough to offset the enthalpy destabilization, to end up an entropy-driven process. The large ΔH_tr and TΔS_tr terms in the DMPC system are due to the disruption of the ordered gel structure, found below the transition temperature.
The partition of lipophilic drugs into lipid phases is often thought to be entropy-driven, a “hydrophobic” effect. Bäuerle and Seelig studied the thermodynamics of amlodipine and nimodipine binding to phospholipid bilayers (above the transition temperature) using highly sensitive microcalorimetry.¹⁹ The partitioning of the drugs into the lipid bilayer was enthalpy-driven, with ΔH_tr − 38.5 kJ·mol−1 bound amlodipine. The entropy of transfer is negative, contrary to the usual interpretation of the “hydrophobic” effect. Thomas and Seelig also found the partitioning of the calcium antagonist, flunarizine (a weak base), to be predominantly enthalpy- driven, with ΔH_tr− 22.1 kJ·mol−1, again at odds with the established ideas of entropy-driven partitioning of drugs.²¹ The same surprise was found for the partitioning of paclitaxel.²² So, these observations appear to suggest that these drugs partition into membrane phases because they are lipophilic and not because they are hydrophobic.