Induction of P-glycoprotein expression by HIV protease inhibitors in cell culture

p-Glycoprotein is an ATP-driven drug efflux pump that can extrude a broad spectrum of substances from the cells that express it [1], and is now recognized as an important factor in drug sensitivity and disposition for many kinds of therapeutic agents [2,3]. Drugs and other substances that modulate P-glycoprotein activity have the potential to alter the pharmacokinetic and pharmacodynamic profiles of therapeutically administered P-glycoprotein substrates. Bioavailability, excretion, and penetration into sites such as the central nervous system and lymphocytes can be dependent on the level of P-glycoprotein activity [4,5]. The HIV protease inhibitors (PI) are substrates of P-glycoprotein (and in some cases, modulators or inhibitors) [6–12], and could be vulnerable to such changes, with resulting alterations in effectiveness. We report here that the HIV PI differentially induce the expression of P-glycoprotein in a transformed cell line. Saquinavir, ritonavir, indinavir, nelfanivir, and amprenavir were extracted into methanol solution from their commercial dosage forms, with purity verified by high performance liquid chromatography. Solvent was removed before use in cell cultures. Verapamil, ivermectin, and vinblastine were obtained from Sigma Chemical Co. (St Louis, MO, USA). C219 monoclonal antibody to human MDR1 (P-glycoprotein) was purchased from Signet Co. (Dedham, MA, USA). The human colon adenocarcinoma cell line (LS-180) [13] was obtained from ATCC (Manassas, VA, USA). Cell media consisted of minimal essential media and Earle's salts (GibcoBRL, Grand Island, NY, USA) supplemented with 10% fetal bovine serum, 0.1 mM non-essential amino acid solution (GibcoBRL), 1 mM pyruvic acid, 100 units/ml penicillin and 0.1 mg/ml streptomycin. Cells were incubated at 37°C in a humidified chamber with 5% carbon dioxide. The LS-180V cell line used in all experiments was selected by passage with increasing concentrations of vinblastine, and then maintained in 4 ng/ml vinblastine until the initiation of induction experiments [14]. This yielded a cell population homogenous in morphology and in detectable levels of P-glycoprotein. Dimethyl sulphoxide (DMSO; 0.5%) was included in the media to ensure drug dissolution. At 50% confluence, cells were incubated with saquinavir, ritonavir, indinavir, nelfanivir, and amprenavir, as well as verapamil (positive control) and ivermectin (negative control), at concentrations of 3 μM and 10 μM for 72 h (n = 3), at which time confluence was 70–80%. Cell lysates were spun down and the supernatant was collected. The total protein concentration was determined by a bicinchoninic acid protein assay (Pierce, Rockford, IL, USA) with bovine serum albumin as a standard. Sample protein was separated by sodium dodecyl sulphate–polyacrylamide gel electrophoresis on a 4–15% gradient polyacrylamide gel [15] and immunoblotted using monoclonal mouse anti-human MDR1 (C219), with sheep anti-mouse Ig-horseradish peroxidase as the secondary antibody. SuperSignal West Pico Chemiluminescent Substrate (Pierce) was used to activate the horseradish peroxidase signal. Blots were then exposed to radiographic film and the quantification of protein was completed via computer image analysis (Image Pro Plus; Media Cybernetics, Houston, TX, USA) [15], with quantitation based on comparison with a standard curve of an induced, concentrated LS-180V sample. Because the DMSO vehicle had a small inductive effect on P-glycoprotein expression compared with no treatment, DMSO-treated cells were established as the baseline. 3 μM verapamil produced a small increase in P-glycoprotein expression compared with DMSO-treated cells, but it increased P-glycoprotein expression 2.5-fold (± 0.41) at 10 μM (P < 0.05). Ivermectin-treated cells showed no increase in P-glycoprotein expression at either 3 μM or 10 μM. All HIV PI increased P-glycoprotein expression (P < 0.05) at 10 μM when compared with DMSO controls (Fig. 1). Ritonavir produced the greatest increase (4.2-fold ± 0.47), and indinavir the least (1.6-fold ± 0.10). Amprenavir and nelfinavir (3.5-fold increase), and to a lesser extent saquinavir (2.4-fold increase), moderately induced P-glycoprotein expression at 10 μM. Only ritonavir, amprenavir, and nelfinavir increased P-glycoprotein expression at 3 μM concentrations (2-fold).Fig. 1.: Induction of P-glycoprotein in LS-180V intestinal carcinoma cells. LS-180V cells were incubated with HIV protease inhibitors (PI) saquinavir (Saq), ritonavir (Rit), indinavir (Ind), nelfinavir (Nel), and amprenavir (Amp), and verapamil (Ver) and ivermectin (Ive). Immunoreactive protein was detected via Western blotting, loading 50 μg of each sample. (a) Representative blot, all drugs at 10 μM, and computer image analysis was used to quantify the immunoblots (pixel area * pixel density). Quantification was based on a calibration curve of induced/concentrated LS-180V standard. The bars (b) represent mean immunoreactive protein detected (± standard deviation) and are expressed as a percentage of dimethyl sulphoxide (DMSO) control. Significant differences from the DMSO control are indicated by *(P < 0.05) or (P < 0.01)as determined by Student's t-test.Ritonavir, nelfinavir and amprenavir are powerful inducers of P-glycoprotein expression in this cell culture system. To our knowledge, this inductive response to HIV PI exposure has not been previously reported for P-glycoprotein in any in-vitro or in-vivo system. The possibility that prolonged exposure to HIV PI could produce changes in drug disposition and activity at distinct critical sites via the similar induction of P-glycoprotein in vivo should be considered and explored. Michael D. Perloffa Lisa L. von Moltkeab Jeanne M. Faheya Johanna P. Dailyc David J. Greenblattab