The importance of the pH to mimic the physiological conditions has been introduced in the section on "Physiological Relevance". It has been demonstrated that media buffered to a pH varying from 5 to 8 were compatible with the Caco-2 cell monolayer (Palm et al., 1999; Neuhoff et al., 2003, 2005). PAMPA models have even been shown to be compatible with a wider range of pH, i.e., from 4 to 10 (Bermejo etal., 2004).
Yamashita and coworkers studied the effect of medium pH on the transport rate of several passively and actively transported drugs in Caco-2 cells (including antipyrine, theophylline, hydrochlorothiazide, atenolol, terbutaline, nadolol, salicylic acid, furosemide and cephalexin). Transport was studied in the absence (apical and basolateral medium buffered at pH 7.4) or presence (apical medium buffered at pH 6.0, basolateral medium buffered at pH 7.4) of a pH-gradient. The observed differences in apparent permeability were attributed to both difference in drug partitioning and modification of carrier-mediated transport (Yamashita et al., 1997). A similar study was run in the PAMPA model (Sugano et al., 2001). The permeation of 30 model compounds, expected to be passively tran-scellularly transported, was assessed as a function of pH. For both models, it was concluded that a better prediction of the fraction absorbed in humans would be obtained under pH-gradient conditions. Although useful for standard screening tests, the value of this set-up can be questioned when performing mechanistic studies in the Caco-2 system. For instance, when performing bi-directional transport studies to explore the interaction with a carrier, the pH of the donor and receiver solvents should be similar in order to prevent bias created by the experimental conditions. Recently, the impact of a pH-gradient when performing bi-directional transport experiments of weak bases (Neuhoff et al., 2003) and weak acids (Neuhoff et al., 2005) has been discussed. The use of a system in which different pH values are maintained at the apical and basolateral side will lead to different concentrations of uncharged drug species, resulting in asymmetry in bi-directional transport and, for weak bases, a 'false' efflux component (Ungell et al., 2002; Neuhoff, 2005). This asymmetry in transport rate occurs independently of active transport mechanisms; it is therefore not advisable to use an efflux ratio obtained in a pH-gradient system as an indication of the involvement of an active transport system (Neuhoff et al., 2003; Volpe, 2004). Similarly, when transport of weak acids is studied, a pH gradient over the membrane will create a false asymmetry over the membrane which is not associated with interaction with transporters and can be interpreted as false uptake (Neuhoff et al., 2005). However, many acids are taken up actively via H+-dependent systems (e.g. MCT carrier system) and, thus, when studying weak acids, two different pH systems have to be used: one with and another without a gradient in order to rule out an inappropriate interpretation (Neuhoff et al., 2005). A compromise pH of 7.0 at both sides of the monolayer (Yamashita et al., 1997) has been suggested when performing bi-directional mechanistic studies. However, when assessing pH-dependent carrier systems, the pH-gradient remains key for mechanistic studies [e.g. human peptide transporter hPEPT1, MCT, PAT1(imino) and thiamine transporter] (Steffansen et al. 2004). A possible pH-dependence should also be considered in studies of drug/drug-interactions involving P-glycoprotein or proton dependent uptake transporters (Ungell et al., 2002; Neuhoff et al.,
2003; Neuhoff et al., 2005). For these reasons, we suggest to work under pH-gradient conditions for the absorptive ranking of compounds. When performing more mechanistic polarity studies, we suggest to select the same pH on both sides during the first screening phase and elaborate on the differential pH conditions when the compounds are moving forward through the later development stages (lower throughput).
Sodium taurocholate, sodium cholate, sodium taurodeoxycholate, sodium tau-rodihydrofusidate and other bile salts have been studied for their effect on the epithelial integrity of cell monolayers and on the transport of model compounds in Caco-2 cells (Anderberg et al., 1992; Lo and Huang, 2000). As described previously and summarized in Table 3, different bile salts were suggested as solubility enhancers (Lo and Huang, 2000, Meaney and O'Driscoll, 2000; Taub et al., 2001). More complex formulations containing bile salts as fasted state simulated intestinal fluid (FaSSIF) have also been evaluated in the Caco-2 model. It was demonstrated that this buffer was compatible with the Caco-2 monolayer for a period of at least two hours, without affecting the transport of theophyllin (passive diffusion) and phenylalanine (active absorptive transport). However, a concentration-dependent P-gp inhibitory effect of sodium taurocholate (present in FaSSIF) when assessing cyclosporin A transport was demonstrated (Ingels et al., 2002). Although the impact of using bile salts in Caco-2 experiments is not yet fully understood, we could expect that the use of bile salts included in the apical solvent could increase the solubilization of poorly water-soluble drugs in the Caco-2 cell culture model. In addition, it could also improve the physiological relevance of the model (Figure 3). Although we are convinced that the use of bile salts could become the preferred option in future especially for 'ranking' purposes, several aspects of this approach remain to be investigated, including the exact concentration and the type of bile salt(s) that should be used. In addition, the high cost of bile salts could also preclude the use of such buffers in high-throughput screening.
Proteins or Micellar Additives for the Creation of Sink Conditions In order to preserve sink conditions in the static in vitro models, it has been proposed to include additives in the receiver compartment that are able to decrease the free drug concentration (Figure 5). Among all the proposed alternatives, the ideal option should be suitable for high-throughput screening and not increase the workload associated with the Caco-2 transport assay. The inclusion of serum albumin (used as such or as present in culture medium) in the receiver compartment has been demonstrated to modify the transport properties of drugs (Mathieu et al., 1999; Walgren and Walle, 1999; Aungst et al., 2000; Yamashita et al., 2000; Krishna et al., 2001; Deferme et al., 2002; Saha and Kou 2002; Demirbas and Stavchansky, 2003; Neuhoff, 2005). The presence of albumin in the basolateral compartment can promote drug partitioning from the cell monolayer into the basolateral compartment; in addition, it can also prevent
the non-specific binding of the drug compounds to the plastic material. Inclusion of 4% bovine serum albumin (BSA) in the basolateral compartment has been proposed to be most relevant for in vivo conditions (Aungst et al., 2000, Saha and Kou, 2002; Demirbas and Stavchansky, 2003; Neuhoff, 2005). The inclusion of proteins in the receiver compartment can, however, have important consequences. It was for instance shown that the inclusion of BSA in the receiver compartment may affect the BCS permeability ranking of highly lipophilic new chemical entities (Saha and Kou, 2002). For protein-bound compounds, it was shown that failure to consider plasma binding could result in an overestima-tion of the basolateral to apical flux and so to a misleading net flux calculation. Under classic test conditions, furosemide and verapamil were shown to have a net apical secretion ratio of 4.2 and 1.3, respectively (Chung et al., 2001). In presence of human plasma in the basolateral compartment, the increase of the AP-to-BL and decrease of the BL-to-AP transport resulted in a reduction of the net apical secretion ratio, suggesting a much less significant efflux (Neuhoff, 2005). However, taking into account the low fraction of unbound drug in the BL compartment (e.g. low concentration of drug applied), the efficiency of the efflux transporter may increase, resulting in lower AP-to-BL and higher BL-to-AP transport (Neuhoff, 2005). The use of more physiologically relevant conditions is expected to be more in agreement with the in vivo absorption of many drugs and may partly explain why the efflux transporter in the intestine apparently has no or very little effect on limiting the in vivo oral absorption of many drugs (Chung and Chiou, 1999).
The studies mentioned so far all support the utilization ofserum proteins at the basolateral side to increase the quality of the model. However, to our knowledge, no in vitro/in vivo correlation studies have been performed to assess the impact of the absence or presence of albumin. An important disadvantage of using protein containing media is the additional procedure needed for sample preparation, which is an important drawback for high-throughput screening assays. In addition, if transport studies are conducted in diffusion chambers, where stirring is provided by gas-bubbling, the surfactant property of BSA gives rise to an undesirable frothing and foaming effect (Saha and Kou, 2002). To overcome these issues, efforts have been made to find alternatives, so that sink conditions could be maintained without using materials which could interfere with the analytical method used. When assessing the transport of the antiviral compound UC-781 in Caco-2 cells, Deferme et al. demonstrated that vitamin E TPGS included in the basolateral compartment enhanced the transport rate of the compound (Deferme et al., 2002) while no additional sample preparation was needed. The use of analysis-friendly additives requires that their compatibility with the absorption model needs to be thoroughly studied. It is known that some of these additives can cause conformational changes in membrane bound proteins and that they may also alter membrane fluidity. Both effects may change drug permeation (Nerurkar et al., 1996; Rege et al., 2002). However, by only including these surfactants in the basolateral compartment, one might expect that the interaction with apically localized (efflux) carriers is limited (Deferme et al., 2002). A few studies indicate that non-ionic surfactants not only bind, adhere and incorporate into the lipid membrane (causing changes in membrane fluidity), but also cross the lipid bilayer into the cell interior and cause ATP depletion (Batrakova et al. 2001, Kabanov et al. 2003). This mechanism has been proposed to partly explain the P-gp inhibition by surfactants in chemotherapy (Kabanov et al., 2003). At present, no systematic studies comparing the different options and evaluating the impact of such additives on the in vitro/in vivo correlation, have been performed.
Also in the PAMPAmodel, alternatives to BSAhave been explored to improve the biorelevance of the model. To overcome the adsorption and/or absence of sink conditions, different additives that do not require an additional step of sample preparation as compared with the addition of albumin, have been proposed. Recently, the use of Double-Sink PAMPA (DS-PAMPA) was proposed as biorelevant alternative to the classic PAMPA methodology (Avdeef, 2003). In DS-PAMPA, a non-specific binding agent (lipophilic sink) was included in the receiver compartment to create sink conditions.
Was this article helpful?
If you suffer with asthma, you will no doubt be familiar with the uncomfortable sensations as your bronchial tubes begin to narrow and your muscles around them start to tighten. A sticky mucus known as phlegm begins to produce and increase within your bronchial tubes and you begin to wheeze, cough and struggle to breathe.