Figure 7.27 Membrane retentions with and without sink, 2% DOPC model.

expected to have a strong electrostatic component, as well as a hydrophobic component.

Furthermore, the membrane retentions of the lipophilic probe molecules are dramatically decreased in the presence of the sink condition in the acceptor wells, as shown in Fig. 7.27. All molecules show R < 35%, with progesterone and phenazo-pyridine showing the highest values, 34% and 26%, respectively.

The combination of increased Pe and decreased %R allowed the permeation time to be lowered to 4 h, in comparison to the originally specified time of 15 h [547,550], a considerable improvement for high-throughput applications. The quality of the measurements of the low-permeability molecules did not substantially improve with sink conditions or the reduced assay times. DOPC with Dodecylcarboxylic Acid under Sink Conditions

The free fatty acid model 5.1 shows dramatic differences in permeabilities over the neutral-charge model 1.1. For example, quinine, metoprolol, and primaquine are 10, 14, and 16 times more permeable, respectively, in the charged (0.6% wt/vol in dodecane) lipid system. The most remarkable enhancement is that of amiloride. In the DOPC system, no detectable amount of amiloride permeates; however, Pe is 1.6 x 10—6cm/s when 0.6% DA is added to the dodecane. It is thought that a very strong ion-pair complex forms between the positively-charged amiloride (Fig. 7.21) and the negative-charge dodecylcarboxylate group, through strong electrostatic and hydrogen bonding, perhaps forming an eight-membered ring =(=C=N+=H • • • =O=C=O • • • H=N=)=. Uncharged carboxylic acids are known to form dimeric units of a similar sort when dissolved in oil [538].

The increase of negative charge from 0.6% to 1.1% wt/vol in dodecane (modeling the expected increase between BBM and BBB lipid compositions; see Table 7.1) shows further increases to the permeabilities of the dramatically affected molecules, especially amiloride, which becomes effectively more permeable than piroxicam.

Most of the weak-acid probe molecules (ibuprofen, naproxen, ketoprofen, pirox-icam) show significant increase in permeabilities with models 5.1 and 6.1, compared to model 1.1. This is surprising, considering that most of the weak-acid probes are negatively charged themselves, and would be expected to be less permeable, due to electrostatic repulsions. Apparently, the increased membrane-water partitioning of weak acids in the two-component lipid models overcomes the expected negative charge repulsions between the ionized acids and the charged membrane components, and leads to increased permeability. Also, membrane surface negative charge is expected to lower the surface pH, thus increasing interfacial fu (Table 7.4), leading to higher permeabilities of ionizable acids [457]. DOPC with Phosphatidic Acid under Sink Conditions

The PA systems (models 6.1 and 7.1) seem to show some of the general patterns of changes seen above, but to a lesser extent. Amiloride permeates in its usual way (poorly). The weak-acid probes are more permeable in the PA models, compared to neutral DOPC, but to a lesser extent than in DA. As a predictor of GIT absorption, the phosphatidic acid system appears to be the best. (The rankings of all the investigated lipid systems are discussed in Section 7.8.3.) Figure 7.28a shows the effect of PA on the permeabilities of the weak-base probe molecules. Dramatic and systematic increases are seen in all the membrane-limited permeabilities (left side of the bar graph). When the permeabilities reach the UWL limit of model 1.1, then no substantial effects due to increasing amounts of PA are seen (right side of the bar graph). So, most of the charged bases are elevated to be nearly diffusion-limited in their permeabilities, when PA is part of the membrane constituents.

Figure 7.28b shows that membrane retention is very systematically increased for almost all of the weak bases. This is a general pattern for bases with any of the negatively charged membrane models, and is probably best explained by the increased electrostatic attractions between the drugs and the membranes. Still, all retentions are below 50%, due to the offsetting sink condition created in the acceptor wells.

Figure 7.28 (a) Permeability (with 2% DOPC +0.0 - 1.1% PA/sink in acceptor) and (b) membrane retentions as a function of phosphatidic acid in 2% DOPC/dodecane lipid membranes at pH 7.4 for a series of weak bases. DOPC with Phosphatidylglycerol under Sink Conditions

The PG models 9.1 and 10.1 show similar trends as indicated by PA, but the effects are somewhat muted. The increase in PG from 0.6% to 1.1% causes the permeabilities of weak bases to decrease and membrane retentions to increase, with many bases showing R > 60%. Many molecules were not detected in the acceptor compartments by UV spectrophotometry after 4 h permeation times (Table 7.7). These properties of the PG system make it less attractive for high-throughput applications than the other two-component models. DOPC with Negative Lipids without Sink

The two-component lipid models were also characterized in the absence of sink conditions (Table 7.8). Comparisons between models 7.0 (Table 7.7) and 1.0 (Table 7.5) suggest that negative charge in the absence of sink causes the permeabilities of many of the bases to decrease. Exceptions are quinine, prazosin, prima-quine, ranitidine, and especially metoprolol. The inclusion of 0.6% PA causes Pe of metoprolol to increase nearly 10-fold, to a value twice that of propranolol, a more lipophilic molecule than metoprolol (based on the octanol-water scale). Naproxen and ketoprofen become notably more permeable in the two-component system. Surprisingly, the neutral progesterone becomes significantly less permeable in this system.

With the noted exceptions above, the other negative-lipid combinations (Table 7.8) show consistently lower permeabilities compared to neutral DOPC. Surprisingly, the retentions are not concomitantly higher than in the neutral DOPC lipid.

7.7.4 Five-Component Anionic Lipid Model (Chugai Model)

The interesting five-component BBM model (11.0 in Table 7.3) proposed by Sugano et al. [561,562] was tested by us (Table 7.9). A small modification was made to the original composition: 1,7-octadiene was replaced by dodecane, due to safety concerns over the use of the octadiene in an unprotected laboratory setting [561]. The permeabilities in the dodecane-modified Chugai model were considerably lower than those shown in pION model 1.0 (and those reported by Sugano's group). This may be due to the lessened "fluidity" of the membrane mixture when the octadiene is replaced by dodecane. Retention is quite considerable in the modified Chugai model, with chlorpromazine and progesterone showing R 95% and 87%, respectively. As discussed later, the Sugano model actually has a good GIT absorption prediction property, about as good as that of model 7.1 (which contains only two lipid components).

The Chugai model was unstable in the presence of a sink-forming surfactant in the acceptor wells, and no further efforts were devoted to the untenable model 11.1. The 1% wt/vol cholesterol in dodecane may have interacted with the sink-forming micelles.

TABLE 7.9 Five-Component Anionic Lipid PAMPA Model (without Sink), pH 7.4"
0 0

Post a comment