Abstract
In cell culture systems with aqueous buffers, concentration-response curves to lipophilic inhibitors are difficult to establish because plateau effects (Imax) are often not reached because of limited drug solubility. Consequently, the inhibitory potency of a compound will not be definable using IC50 values (concentration exerting 50% of Imax). Since alternative potency measures f2 values, the concentrations required to double baseline signals have been proposed. Using both methods, we reevaluated the concentration-response curves of calcein assays with 78 compounds in three different cell culture systems and found a close correlation between both methods (rs = 0.93–0.99, p ≤ 0.0028). These findings suggest that f2 values are a valuable alternative to define rank orders of highly lipophilic inhibitors as a basis for the prediction of pharmacological interaction properties in clinical settings. Although it was only tested for inhibition of P-glycoprotein, it seems likely that this method may be transferred to other assays with other proteins.
The assessment of the concentration-response relationship is a key element in the pharmacologic characterization of inhibitors of important targets like the ATP-binding cassette (ABC) transporter P-glycoprotein (P-gp; MDR1/ABCB1). The inhibitory potency of a compound is normally expressed as IC50 (concentration leading to half-maximal inhibition) and is typically derived from the Hill equation or its derivatives, which describe concentration-response curves using four parameters: background (Imin), maximum effect (Imax), slope (s), and IC50, the concentration leading to 50% of Imax.
Obviously, the accurate determination of IC50 values depends on the reliable description of Imin, s, and Imax, with the latter being the most challenging, since maximum effects can often not be reached because of limited solubility or cytotoxic effects. We have previously shown that the poor solubility is frequently neglected and that numerous papers report having tested inhibition at drug concentrations beyond the maximum solubility of the inhibitor in the respective buffers (Weiss et al., 2002).
In this study we have evaluated an alternative method to assess the inhibitory potency of compounds interacting with P-gp, the substrates of which are typically highly lipophilic and typically cannot be tested up to maximal effects in vitro because of their low solubility in the aqueous solutions used in cell culture systems. P-gp inhibition was investigated with the well established and widely used calcein assay in three different cell systems expressing moderate to high amounts of this efflux transporter: primary porcine brain capillary endothelial cells (pBCECs), L-MDR1 cells with overexpression of human P-gp, and the murine leukemic cell line P388/dx overexpressing murine P-gp.
Materials and Methods
Materials. Culture media, fetal calf serum, medium supplements, antibiotics, and Hanks' balanced salt solution were purchased from Invitrogen (Karlsruhe, Germany), dimethyl sulfoxide (DMSO) from Sigma-Aldrich (Taufkirchen, Germany), calcein acetoxymethyl ester from MoBiTec (Göttingen, Germany), and 96-well microtiter plates from NUNC GmbH & Co. KG (Wiesbaden, Germany). Drugs were obtained from Sigma-Aldrich (Taufkirchen, Germany) or from the corresponding manufacturer.
LLC-PK1 and L-MDR1 Cells. As a model for human P-gp we used L-MDR1 cells, a cell line generated by transfection of the porcine kidney epithelial cell line LLC-PK1 with the human MDR1 gene (Schinkel et al., 1996) and the parental cell line LLC-PK1 (available from The American Type Culture Collection, Manassas, VA) as a control. The L-MDR1 cell line was kindly provided by Dr. A. H. Schinkel (The Netherlands Cancer Institute, Amsterdam, The Netherlands). The cells were cultured and seeded as described previously (Weiss et al., 2003a).
P388 and P388/dx Cells. As an alternative model for P-gp we used the murine monocytic leukemia cell line P388 and the corresponding doxorubicin-resistant cell line P388/dx over-expressing mdr1a/1b (Boesch et al., 1991). Both cell lines were kindly provided by Dr. Dario Ballinari (Pharmacia and Upjohn, Milano, Italy). The cells were cultured and seeded as described previously (Fröhlich et al., 2004).
pBCECs. Isolation, culturing, and seeding of pBCECs expressing porcine pgp1A was essentially based on the method described by Audus et al. (1996) with minor alterations (Weiss et al., 2003a).
Stock Solutions. Stock solutions of test compounds were prepared, strictly following the manufacturer's instructions. Only very few compounds were soluble in aqua bidest. All others were dissolved in DMSO. The DMSO concentration in the assays never exceeded 1% (v/v), a concentration which was found not to influence the results of the assay.
Calcein Uptake Assay. The calcein assay was used to assess P-gp inhibition. Because the transport capacity of P-gp is inversely proportional to the accumulation of intracellular calcein fluorescence, inhibition of P-gp leads to an increase in intracellular calcein fluorescence. The assay was conducted and validated as described previously (Weiss et al., 2003a; Fröhlich et al., 2004). Each experiment was performed at least in duplicate (if no inhibition was observed) or in triplicate on different days.
Statistical Analysis. For calculation of the inhibitor effects, a nonlinear four-parameter fit was used (GraFit version 4; Erithacus Software, Middlesex, UK) according to the sigmoidal Imax model, with the formula specified in eq. 1 (Hill equation). The f1.5, f2, f3, and f4 values (concentration needed to increase baseline fluorescence by a factor of 1.5, 2, 3, and 4, respectively) were derived from the corresponding concentration-response curve (Fig. 1) as published previously for the f2 value (Weiss et al., 2003).
Correlations were assessed by Spearman rank correlation and characterized by the corresponding correlation coefficient rS (GraphPad Prism, version 4.0; GraphPad Software Inc., San Diego, CA). A p value of ≤0.05 was considered significant.
Results and Discussion
A total of 78 compounds were tested in the calcein assay in the three different cell systems (50 compounds in P388/dx cells, 67 in L-MDR1 cells, and 48 in pBCECs). Twenty-seven compounds were investigated in all three cell systems, 35 in two systems, and 16 in only one cell system. Among them were highly potent P-gp inhibitors like LY335979 (zosuquidar), SDZ-PSC833 (valspodar), and GG918 (elacridar), drugs used for HIV therapy, fungicides, newer antidepressants (Weiss et al., 2003a), progestins (Fröhlich et al., 2004), fibrates (Ehrhardt et al., 2004), amphetamines (Ketabi-Kiyanvash et al., 2003), antiepileptic drugs (Weiss et al., 2003b), and kava-kava extracts and kavalactones (Weiss et al., 2005). Of all these compounds, 21 revealed no P-gp inhibition up to the highest soluble concentration (10 in P388/dx cells, 19 in L-MDR1 cells, and 11 in pBCECs). Plateau effects were only reached by 15 compounds in P388/dx cells, 6 compounds in L-MDR1 cells, and 9 compounds in pBCECs, a prerequisite for a valid calculation of IC50 (Tables 1, 2, 3). For all other compounds (44 in P388/dx cells, 42 in L-MDR1 cells, and 28 in pBCECs) with obvious P-gp-inhibitory effects, evaluation of the inhibitory potency based on IC50 values would be inappropriate because they did not reach Imax. Indeed, the common practice to calculate IC50 values from truncated curves not covering the whole range of the dose-response relationship (e.g., Tiberghien and Loor, 1996; Ji and Morris, 2004; Zhou et al., 2005) bears a significant potential of over- or underestimation of a compound's potency. To illustrate this fact, we have calculated IC50 with truncated concentration-response curves (using only values below Imax) for the 15 compounds reaching plateau effects in P388/dx cells. Figure 2 demonstrates that such calculations will result in IC50 values up to 2 orders of magnitude over the “true” IC50 calculated with the complete concentration-response curves. Moreover, when IC50 values were calculated from truncated concentration-response curves, variation coefficients of the test series were up to 171%, illustrating the extent of inaccuracy by this mode of calculation. In addition, the ranking order of the compounds changed substantially from A-B-C-D-E-F-G-H-I-J-K-L-M-N-O to B-A-C-E-F-K-G-J-I-M-L-D-H-O-N.
Provided that the maximal intracellular calcein fluorescence was identical for all compounds, it would be possible to calculate IC50 by fixing Imax to a certain value. However, as Fig. 3 demonstrates, Imax values may differ between different inhibitors by half an order of magnitude, and indeed, when IC50 values were calculated with Imax values fixed to 35, they differed up to 5-fold (Fig. 2) and the potency ranking order changed to A-B-C-H-D-F-E-G-I-M-L-K-N-J-O. Moreover, when IC50 values were calculated with fixed Imax, variation coefficients reached up to 157%, emphasizing that this mode of calculation is also inadequate.
Having faced this problem when P-gp-inhibitory properties of antidepressants were assessed (Weiss et al., 2003a), we have established the f2 method. The f2 value corresponds to the concentration of the compound leading to a 2-fold increase in the baseline calcein fluorescence (Fig. 1). Whereas f2 is, obviously, rather independent of Imax, its values are only a useful alternative if they are closely correlated with IC50 values. Table 4 demonstrates the excellent correlation between f2 and IC50 in P388/dx cells, L-MDR1 cells, and pBCECs for all compounds that reached plateau effects in the calcein assay, and Fig. 4 exemplifies this correlation in P388/dx cells. In all cell lines, which differ in the expression level and the species origin of P-gp, the relationship was very close and highly significant. The same applies to the correlation between IC50 and f1.5 and f3 but not to f4, where the correlation is the weakest (Table 4). Figure 5 demonstrates that the ranking order between IC50, f2, and f3 only differed slightly, and switches mostly occurred between compounds with IC50 values lying closely together. However, f2 was favored over f1.5 or f3, because 1) in contrast to f1.5, with the common slopes observed, f2 always lies in the log-linear steep part of the concentration-response curve (similar to IC50), and 2) in contrast to f3 only, for a few of the compounds tested in the calcein assay, the poor solubility prevented the quantification of the f2 value (Table 5). Only for very weak inhibitors, which do not lead to a 2-fold increase in basal fluorescence up to the highest soluble concentration, this method is not suitable. However, such minor effects seem negligible because these compounds do not bear a clinically meaningful interaction potential.
The following limitations merit discussion. First, differences in rank order between the two methods are more likely if compounds with substantially different Imax values are compared. However, as illustrated in Fig. 4, it is very unlikely that potent inhibitors are misclassified as weak and that their interaction potential remains undiscovered.
Second, in this paper, we have only assessed a series of P-gp inhibitors, and we have, thus, not proven universal validity of the proposed approach. It seems, however, rather likely that this method may also readily be transferred to other assays evaluating the inhibition of other transporters, provided the slopes of the concentration-response curves are not too variable.
Finally, f2 values may depend on assay conditions and the cell system used, and these values are thus not readily comparable to absolute f2 values generated in other test systems. This is, however, also true for full concentration-response curves and figures directly derived from it, such as IC50. Irrespective of the method used, it is therefore critical to have appropriate controls to which the results may be related. Moreover, the clinical relevance of the observed potencies in vitro always depends on the concentrations encountered in vivo at the site of the transporter and on other factors influencing the interaction with the transporter, such as the presence of inhibitors or inducers. Indeed, this represents a general problem of data assessed in vitro and is not restricted to f2 or IC50 values.
In conclusion, in the frequent case where inhibitor potencies cannot be assessed with full concentration-response relationships because of restricted solubility of the (often lipophilic) test compounds, calculation of IC50 values as a gold standard is inaccurate and will cause misinterpretation of relative potencies. In these instances, the computation of f2 values (the concentration doubling baseline effects) is a robust and reliable alternative, which even provides correct rank orders if Imax values largely differ between different test compounds.
Footnotes
-
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
-
doi:10.1124/dmd.105.007377.
-
ABBREVIATIONS: ABC, ATP-binding cassette; f1.5, f2, f3, and f4, concentration needed to increase baseline fluorescence by a factor of 1.5, 2, 3, and 4, respectively; rS, Spearman rank correlation coefficient; DMSO, dimethyl sulfoxide; LY335979, zosuquidar; GG918, elacridar; IC50, concentration leading to 50% of Imax; Imax, maximum effect; Imin, minimum effect; pBCEC, porcine brain capillary endothelial cell; P-gp, P-glycoprotein; SDZ-PSC833, valspodar.
- Received September 14, 2005.
- Accepted November 2, 2005.
- The American Society for Pharmacology and Experimental Therapeutics