Secretory Transport of Methylprednisolone Possibly Mediated by P-Glycoprotein in Caco-2 Cells

March 2002 Notes Biol. Pharm. Bull. 25(3) 393—396 (2002) 393

a a ,a a a Ayako OKA, Masako ODA, Hiroshi SAITOH,* Akira NAKAYAMA, Masahiko TAKADA, and b Bruce J. AUNGST Department of Pharmaceutics, Faculty of Pharmaceutical Sciences,a Health Sciences University of Hokkaido, Kanazawa 1757, Ishikari-Tobetsu, Hokkaido 061–0293, Japan and DuPont Pharmaceuticals,b Wilmington, Delaware 19880–0400, U.S.A. Received June 20, 2001; accepted November 19, 2001

We recently reported that P-glycoprotein (MDR1) is capable of interfering with the absorption of methylprednisolone in the rat small intestine. This study was undertaken to examine the interaction between methylprednisolone and MDR1 using Caco-2 cells. The permeation of various steroid hormones (hydrocortisone, prednisolone, progesterone, b-estradiol, and testosterone) was compared. The basolateral-to-apical (secretory) permeation of methylprednisolone was more than 3-fold greater than the apical-to-basolateral (absorptive) permeation. When verapamil (0.1 mM), a potent modulator of MDR1, was added to both apical and basolateral sides of Caco-2 cells, the absorptive permeation of methylprednisolone was increased and its secretory permeation was decreased. As a result, the secretory-oriented manner of methylprednisolone permeation almost completely disappeared. Prednisolone and hydrocortisone exhibited weaker secretory-oriented movement than did methylprednisolone. The secretory-oriented permeation of prednisolone and hydrocortisone was also diminished by the addition of verapamil. There was no signiﬁcant directionality in progesterone permeation and the permeation of b- estradiol and testosterone tended to be absorptive. These results appear to suggest that methylprednisolone, prednisolone, and hydrocortisone interact with MDR1 as the substrates. In contrast, there was no evidence that MDR1 was capable of potently interfering with the absorption of the sex hormones tested in this study, supporting our previous ﬁndings in the rat. It was further found that apically-added verapamil demonstrated a modulating effect on MDR1 function even at 5 mM. Key words methylprednisolone; P-glycoprotein; Caco-2; secretory-oriented transport; steroid absorption

P-Glycoprotein (MDR1) is localized on the epithelial A human colorectal carcinoma cell line, Caco-2, has been membrane of various tissues including the intestine,1,2) and used to predict the intestinal absorption of various drugs and its physiological function may be to detoxify or protect the new chemical entities in humans for nearly a decade.11,12) Re- body from various drugs and cytotoxic compounds.3) Several cently, Caco-2 has been used increasingly as an experimental reports have described the interaction of various steroid hor- model for evaluating the interactions between MDR1 and mones with MDR1 as substrates or modulators.4—7) In a re- drugs, because this cell line expresses abundant MDR1 on its cent study, we reported that methylprednisolone disappeared apical membrane.13,14) This study was undertaken to elucidate from in situ rat intestinal loops in a region-dependent man- the permeation of methylprednisolone and several steroid ner; that is, its disappearance was greater in the upper small hormones across Caco-2 cell monolayers, with special focus intestine and relatively smaller in the lower small intestine.8) on their interaction with MDR1. It was also found that the addition of verapamil and cy- closporin A, potent modulators of MDR1, significantly in- MATERIALS AND METHODS creased the disappearance of methylprednisolone from the entire small intestine, though more markedly in the lower Materials 6a-Methylprednisolone, prednisolone, hydro- small intestine.8,9) A likely explanation was that methylpred- cortisone, b-estradiol, and testosterone were purchased from nisolone absorption was lowered through interaction with Sigma Chemical Co. (St. Louis, MO, U.S.A.). Progesterone MDR1, especially in the lower small intestine. It is known and verapamil hydrochloride were obtained from Wako Pure that MDR1 expression is rich in the lower small intestine.1) Chemical Ind. (Osaka, Japan). Dulbecco’s modified Eagle’s On the other hand, it was found that the steroid sex hormones medium (DMEM), fetal calf serum (FCS), and non-essential b-estradiol, progesterone, medroxyprogesterone and testos- amino acid solution were obtained from Life Technologies terone were absorbed very rapidly and almost completely (Grand Island, NY, U.S.A.). Other chemicals were of the from the entire small intestine, implying that MDR1 did not highest grade available. interfere with the absorption of these hormones in this Cell Culture Caco-2 cells (passage #40) were purchased organ.9) These results supported previous reports which de- from the Riken Cell Bank (Tsukuba, Japan) and were kept scribed that progesterone and medroxyprogesterone were not frozen in aliquots in liquid nitrogen. At passage #42—44, the transported by MDR1 on multidrug-resistant cancer cells.7,10) cells were seeded on 60-mm plastic culture dishes coated Progesterone and progesterone-like compounds are regarded with rat tail collagen type I (Becton Dickinson, Bedford, MD, as potent modulators rather than substrates of MDR1.7) How- U.S.A.) at a density of 1ϫ104 cells/cm2 in 5 ml of DMEM ever, our previous results from the rat small intestine are not with 5% FCS. They were grown at 37 °C in an atmosphere of always direct evidence that MDR1 is able to carry methyl- 5% CO2 in air and 90% relative humidity. The culturing prednisolone more efficiently compared with other steroid medium was replaced every 3—4 d after inoculation. Subcul- hormones. turing was carried out every 6 d using 0.02% EDTA and

0.25% trypsin. Caco-2 cells (passage #45—55) were seeded Table 1. Apical-to-basolateral and Basolateral-to-apical Permeation of Six at a density of 1ϫ105 cells/cm2 on polycarbonate filter inserts Steroid Hormones and Vinblastine across Caco-2 Cell Monolayers 2 (Becton Dickinson) with a surface area of 4.2 cm . For per- P (ϫ10Ϫ6, cm sϪ1) meation experiments, Caco-2 cell monolayers, with transep- app Ratio B to A/A to B ithelial electrical resistance over 400 W ·cm2, were used on A to B B to A 20—22 d (mostly on 21 d) after seeding. Ϯ Ϯ Transport Study across Caco-2 Cell Monolayers Dul- Methylprednisolone (50 m M) 3.1 0.1 10.4 0.2** 3.4 ϩVerapamil (0.1mM) 6.4Ϯ0.3 7.5Ϯ0.4 1.2 becco’s phosphate-buffered saline (D’s PBS, 137mM NaCl, Methylprednisolone (100 m M) 3.3Ϯ0.2 5.2Ϯ0.1* 1.6 3mM KCl, 8 mM Na2HPO4, 1.5 mM KH2PO4, 1 mM CaCl2, Prednisolone (50 m M) 3.6Ϯ0.1 7.0Ϯ0.1* 1.9 ϩ Ϯ Ϯ 0.5 mM MgCl2, and 5 mMD-glucose, pH 7.4) was used as an Verapamil (0.1mM) 3.6 0.1 3.6 0.1 1.0 experimental medium. The apical-to-basolateral and basolat- Hydrocortisone (50 m M) 3.1Ϯ0.3 4.8Ϯ0.3 1.6 ϩ Ϯ Ϯ eral-to-apical permeation of test compounds across Caco-2 Verapamil (0.1mM) 3.3 0.2 3.5 0.4 1.1 Progesterone (50 m M) 3.7Ϯ0.2 4.5Ϯ0.1 1.2 cell monolayers was determined at 37 °C. Caco-2 cells were b-Estradiol (25 m M) 4.8Ϯ0.3 2.1Ϯ0.3** 0.4 rinsed three times with 2 ml of D’s PBS at pH 7.4. Caco-2 Testosterone (50 m M) 5.2Ϯ0.2 1.5Ϯ0.1** 0.3 cells were preincubated for 10 min at 37 °C. The permeation experiment was initiated by adding D’s PBS including a test Vinblastine (50 m M) 1.3Ϯ0.1 6.9Ϯ0.2** 5.5 compound either to the apical or basolateral side of cell Caco-2 cells were incubated at 37 °C for 60 min with test compounds added either to monolayers that were then incubated for 60 min. The medium the apical or basolateral side of the cell monolayers. Verapamil was added to both the apical and basolateral sides. Data represents the meanϮS.E. (nϭ3—6). * pϽ0.01, volume of the apical and basolateral side was 1.5 and 2.2 ml, Ͻ ** p 0.001 significantly different from apical-to-basolateral permeation. Papp, apparent respectively. At stated times, aliquots of the receiver medium permeation coefficent; A to B, apical-to-basolateral permeation; B to A, basolateral-to- were taken to measure the amount that had permeated Caco- apical permeation. 2 cell monolayers. The cumulative amount of test compounds permeated in both directions was calculated and plotted vs. time. The slope calculated was corrected for membrane surface area and donor concentration to obtain an apparent permeation coefficient (Papp). Analytical Methods Determination of steroid hormones tested was carried out by HPLC as described previously.8,9) Protein Assay A Bio-Rad Protein Assay Kit (Bio-Rad Laboratories, Richmond, CA, U.S.A.) was used to determine the protein contents of the cell monolayers, which were dis- solved in 1 M NaOH, with bovine serum albumin used as a standard. Statistical Analysis Statistical significance was evaluated by Tukey’s multiple comparison procedure (ANOVA) and Student’s t-test. The level of significance was pϽ0.05.

RESULTS AND DISCUSSION Fig. 1. Effects of Verapamil on Methylprednisolone Permeation across Permeation of Methylprednisolone across Caco-2 Cell Caco-2 Cell Monolayers Apical-to-basolateral (open symbols) and basolateral-to-apical (closed symbols) per- Monolayers Figure 1 shows the apical-to-basolateral (ab- meation of 50 m M of methylprednisolone was investigated in the presence (triangle) and sorptive) and basolateral-to-apical (secretory) permeation of absence (circle) of 100 m M of verapamil, which was added to both apical and basolateral media. Each point represents the meanϮS.E. (nϭ3—6). methylprednisolone (50 m M) across Caco-2 cell monolayers in the presence or absence of verapamil. In the absence of verapamil, methylprednisolone was much more permeable in losing its directionality in Caco-2 cell permeation. It is well the basolateral-to-apical direction than in the apical-to-baso- known that verapamil is capable of modulating MDR1 func- lateral direction, demonstrating a typical secretory-oriented tion. Therefore, it is very likely that this glycoprotein played transport in Caco-2 cells. The apparent permeation coeffi- significant role in the secretory permeation of methylpred- cients (Papp) in both directions are listed in Table 1. In this nisolone, supporting our previous findings that it signifi- study, Papp was not corrected for the paracelluar permeation. cantly interfered with methylprednisolone absorption in the The ratio of secretory to absorptive permeability (B to A/A rat small intestine.8) However, since several recent reports to B) was 3.4 at 50 m M. When verapamil was added to both have demonstrated that verapamil is also able to interfere apical and basolateral sides of Caco-2 cells at a concentration with the function of other efflux transporters like multidrug 15,16) of 0.1 mM, the apical-to-basolateral permeation of 50 m M resistance-associated proteins (MRPs), which is also pre- methylprednisolone significantly increased at each sampling sent on the Caco-2 cell membrane,17) the possibility that non- time (Fig. 1). The amount permeated at 60 min was more MDR1 type transporters are involved in the secretory perme- than 200% of the control (i.e., in the absence of verapamil). ation of methylprednisolone cannot be excluded. On the other hand, the basolateral-to-apical permeation of As shown in Table 1, when methylprednisolone perme- methylprednisolone decreased to ca. 70% of the control at ation was examined at 100 m M, Papp in the absorptive direc- 60 min in the presence of verapamil. As a result, methylpred- tion was increased and that in the opposite direction was de- nisolone permeation became almost equal in both directions, creased; the ratio of secretory to absorptive permeability thus March 2002 395

Table 2. Effects of Apically-added Verapamil on the Apical-to-Basolateral and a pH 7.4 buffer and found them to be 2.09, 1.64 and Permeation of Methylprednisolone across Caco-2 Cell Monolayers 1.55, respectively.8) Accordingly, our present results may be 7) Verapamil concentration Methylprednisolone permeation inconsistent with those of Barnes et al. Further information (m M) (% of control) on various steroid hormones is required to clarify the dis- crepancy. 5 142.8Ϯ18.8 The basolateral-to-apical permeation of progesterone was Ϯ 10 168.3 20.1* slightly but not significantly greater than its apical-to-baso- 20 203.0Ϯ31.5* 40 257.5Ϯ13.7** lateral permeation. In the case of b-estradiol and testos- 60 272.0Ϯ12.1*** terone, on the other hand, apical-to-basolateral permeation 80 257.8Ϯ13.2*** was significantly greater than basolateral-to-apical perme- 100 278.3Ϯ14.1*** ation, suggesting that the transepithelial movements of b- estradiol and testosterone were absorptive-oriented. It seems Caco-2 cells were incubated for 60 min with 50 m M methylprednisolone and the indi- cated concentrations of verapamil added to the apical side of the monolayers. Data rep- that these results are associated with earlier reports showing resents the meanϮS.E. (nϭ3). * pϽ0.05, ** pϽ0.01, *** pϽ0.001, significantly differ- that the effect of MDR1 on the absorption of these three ent from the control. steroid hormones was small.9) However, it is still possible that MDR1-mediated secretory transport is fully covered by a shifted to 1.6. It is probable that this result may relate to a specialized absorption mechanism for b-estradiol and testos- concentration dependence of methylprednisolone transport. terone; further studies are now being conducted. For comparison, we evaluated the transport of vinblastine, Concentration Dependence of Apically-added Vera- a well-known MDR1 substrate,14) across Caco-2 cell mono- pamil Effect on Absorptive Permeation of Methylpred- layers in this study. The secretory permeation of vinblastine nisolone To date, information on the relationship between was 5.5-fold greater than its absorptive permeation (Table 1). verapamil concentration and its efficacy to inhibit the func- We did not confirm the degree of MDR1 expression on the tion of transporters is inadequate. Table 2 shows the effects apical membrane of Caco-2 cells used in this study, for ex- of verapamil, which was added to the apical side of cell ample, using a technique like immunohistochemical detec- monolayers over a concentration range between 5 and tion. However, the highly secretory-oriented transport of vin- 100 m M, on the apical-to-basolateral permeation of methyl- blastine could confirm that the present Caco-2 cells retained prednisolone. The Papp gradually increased with increase in enough MDR1 to evaluate the transport it mediated. Accord- verapamil concentration to 40 m M and then remained con- ing to these observations, it is likely that methylprednisolone stant. For comparison, the effect of 5 m M of apically-added is transported by MDR1 to a lesser extent than vinblastine. verapamil was examined on the basolateral-to-apical perme- Permeation of Various Steroid Hormones across Caco- ation of methylprednisolone, and it was found that secretory 2 Cell Monolayers The permeability of prednisolone, hy- permeation of methylprednisolone was decreased to ca. 60% drocortisone, progesterone, b-estradiol, and testosterone of that in the absence of verapamil (data not shown). The site across Caco-2 cells is also presented in Table 1. Permeation and mechanism of verapamil action have not been fully char- experiments were done at a concentration of 50 m M, except acterized. Very recently, we reported that the presence of a b-estradiol which was evaluated at 25 m M due to its limited substrate on the apical surface of Caco-2 cell membrane was solubility in D’s PBS. Prednisolone and hydrocortisone ex- necessary to initiate MDR1 transport of the same substrate in hibited greater basolateral-to-apical permeation than apical- the basolateral-to-apical direction and that 5 m M of verapamil to-basolateral permeation, consistent with the results of completely inhibited this process.18) The present results im- methylprednisolone (Table 1). The apical-to-basolateral per- plied that apically-added verapamil is capable of fully dis- meation of prednisolone and hydrocortisone was almost playing the effect at relatively low concentration. identical to methylprednisolone, whereas their basolateral-to- In conclusion, it has been suggested that methylpred- apical permeation was relatively smaller. Although the direc- nisolone is possibly a more potent substrate of MDR1 than tionality in their permeation was almost gone in the presence either prednisolone or hydrocortisone. The magnitude of the of verapamil, the modulator of MDR1 mainly affected the se- interaction with MDR1 is closely related to the extent of the cretory permeation of the two hormones but not their absorp- absorption in the small intestine. Poor response to the tive permeation. These results might imply that the secre- methylprednisolone therapy, possibly due to incomplete ab- tory-oriented transport of prednisolone and hydrocortisone sorption of the glucocorticoid, has been reported clinically.19) was also related to MDR1, but their interaction with MDR1 Our present results propose that methylprednisolone is less might be weaker than methylprednisolone. The order of the absorbable in a patient who has over-expression of MDR1 in ratio of secretory to absorptive permeability was methylpred- the gastrointestinal tract. nisoloneϾprednisoloneϾhydrocortisone. Earlier, we demonstrated that the absorption of these three steroid hormones REFERENCES from the rat intestine was in the order of hydrocortisoneϾ 8) prednisoloneϾmethylprednisolone. Therefore, it can be said 1) Fojo A. T., Ueda K., Slamon D. J., Poplack D. G., Gottesman M. M., that their absorption is significantly dependent on the extent Pastan I., Proc. Natl. Acad. Sci. U.S.A., 84, 265—269 (1987). of their interaction with MDR1. Barnes et al.7) previously re- 2) Thiebaut F. T., Tsuruo T., Hamada H., Gottesmann M. M., Pastan I., ported that hydrophilic steroid hormones were more suitable Willingham M. C., Proc. Natl. Acad. Sci. U.S.A., 84, 7735—7738 (1987). substrates for MDR1 in multidrug-resistant cancer cells. 3) Tanigawara Y., Ther. Drug Monit., 22, 137—140 (2000). We compared the log partition coefficients of methylpred- 4) Ueda K., Okamura N., Hirai M., Tanigawara Y., Saeki T., Kioka N., nisolone, prednisolone and hydrocortisone between n-octanol Komano T., Hori R., J. Biol. Chem., 267, 24248—24252 (1992). 396 Vol. 25, No. 3

5) Ueda K., Saeki T., Hirai M., Tanigawara Y., Tanaka K., Okamura M., 880—885 (1991). Yasuhara M., Hori R., Inui K., Komano T., Jpn. J. Physiol., 44, S67— 13) Hunter J., Jepson M. A., Tsuruo T., Simmons N. L., Hirst B. H., J. S71 (1994). Biol. Chem., 268, 14991—14997 (1993). 6) Rao U. S., Fine R. L., Scarborough G. A., Biochem. Pharmacol., 48, 14) Wils P., Phung-Ba V., Warnery A., Lechardeur D., Raeissi S., Hidalgo 287—292 (1994). I. J., Scherman D., Biochem. Pharmacol., 48, 1528—1530 (1994). 7) Barnes K. M., Dickstein B., Cutler G. B., Fojo T., Bates S. E., Bio- 15) Gaj C. L., Anyanwutaku I., Chang Y. H., Cheng Y. C., Biochem. Phar- chemistry, 35, 4820—4827 (1996). macol., 55, 1199—1211 (1998). 8) Saitoh H., Hatakeyama M., Eguchi O., Oda M., Takada M., J. Pharm. 16) Grech K. V., Davey R. A., Davey M. W., Biochem. Pharmacol., 55, Sci., 87, 73—75 (1998). 1283—1289 (1998). 9) Nakayama A., Eguchi O., Hatakeyama M., Saitoh H., Takada M., Biol. 17) Gutmann H., Fricker G., Torok M., Michael S., Beglinger C., Drewe J., Pharm. Bull., 22, 535—538 (1999). Pharm. Res., 16, 402—407 (1999). 10) Yang C. P., Cohen D., Greenberger L. M., Hsu S. I., Horwitz S. B., J. 18) Nakayama A., Saitoh H., Oda M., Takada M., Aungst B. J., Eur. J. Biol. Chem., 265, 10282—10288 (1990). Pharm. Sci., 11, 317—324 (2000). 11) Hilgers A. R., Conradi R. A., Burton R. S., Pharm. Res., 7, 902—910 19) Hill M. R., Szeﬂer S. J., Ball B. D., Bartoszek M., Brenner A. M., (1990). Clin. Pharmacol. Ther., 48, 390—398 (1990). 12) Artursson P., Karlsson J., Biochem. Biophys, Res. Commun., 175,