0.00 0.02 0.04 0.06 0.08 0.10 Volume (mL) 0.5 M KOH

Figure 7.53 Universal buffer for robotic pH adjustment.

Antonenko et al. [540] considered pH gradients forming in the UWL under bulk solution iso-pH conditions. They elegantly expanded on the buffer effect model and made it more general by considering multicomponent buffer mixtures. Direct measurements of the pH gradients (using wire-coated micro-pH electrodes) near the membrane-water interface were described.

A 10 mM ionic strength universal buffer mixture, consisting of Good zwitterio-nic buffers, [174] and other components (but free of phosphate and boric acid), is used in the pION apparatus [116,556]. The 5-pKa mixture produces a linear response to the addition of base titrant in the pH 3-10 interval, as indicated in Fig. 7.53. The robotic system uses the universal buffer solution for all applications, automatically adjusting the pH with the addition of a standardized KOH solution. The robotic system uses a built-in titrator to standardize the pH mapping operation.

7.7.13 Effects of Stirring

Stirring the permeation cell solution increases the effective permeability, by decreasing the thickness of the UWL (Section Since the PAMPA sandwich (Fig. 7.9) has no airgaps in the bottom wells, and since the solution volumes are small (200-400 mL), the use of rotary-motion platforms to stir the plate is not very effective. Avdeef et al. [556] described the effects of stirring up to speeds of 600 rpm, and noted that the stirring efficiency is about 4 times greater along the periphery of the plate compared to the center locations. This is demonstrated in Fig. 7.54 by 96-replicate verapamil permeability measurements in a plate stirred at 500 rpm. The use of individual magnetic stir bars in each bottom well is a more effective way to stir the solutions. This is currently being developed at pION.

Figure 7.53 Universal buffer for robotic pH adjustment.

Figure 7.54 Effect of stirring: verapamil permeabilities (in units of 10 6 cm/s) in 96 replicates, orbital shaker at 500 rpm.

7.7.14 Errors in PAMPA: Intraplate and Interplate Reproducibility

Figure 7.55 shows a plot of over 2000 2%DOPC/dodecane Pe measurements (10~6 cm/s units), each representing at least three intra-plate replicates, vs. the estimated standard deviations, s(Pe). Over 200 different drug-like compounds were measured. The %CV (coefficient of variation 100 x s(Pe )/Pe) is about 10% near Pe 10 x 10~6cm/s, and slightly increases for higher values of permeability, but rapidly increases for Pe < 0.1 x 10~6 cm/s, as shown in Table 7.21. These statistics accurately reflect the errors that should be expected in general. For some molecules, such as caffeine and metoprolol, %CV has been typically about 3-6%.

The errors mentioned above represent the reproducibility obtained on the same microtiter plate when the sample molecule is assayed in several different wells. When the reproducibility of Pe measurement is assessed on the basis of assays performed at different times over a long period of time, more systematic sources of errors show up, and the reproducibility can be about 2-3 times worse. Figure 7.56 shows reproducibility of standard compounds taken over a period of about 12 months. Carbamazepine show a long-term reproducibility error of —15%. The other compounds show somewhat higher errors.

Considering that PAMPA is a high-throughput screening method, the errors are low enough to encourage the use of the method to study mechanistic properties, as our group at pION has done since 1997.

0 0

Post a comment