Vol. 181, No. 2, 1991

December 16, 1991 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

Pages 671-676

TIUNS-VlNBLASTINESECRETIONMEDIATEDBY

P-GLYCOPROTEIN Is INHIBITED BY FORSKOLIN DERIVATIVES

Janice Hunter, Barry H. Hint and Niiolas L. Simmons

Gastrointestinal Drug Delivery Research Centre, and Department of Physiological Sciences, University of Newcastle upon Tyne,

Medical School, Newcastle upon Tyne NE2 4HH, UK

Received October 28, 1991

Summary: [3HjVinblastine transport across MDCK (renal epithelial) cell layers has been character&d. The basal-to-apical FHlvinblastine flux (J& (at 10 nM) exceeded apical-to- basal flux by 19.6 fold. Net vinblastine secretion (JBA - J,,-B) was inhibited by verapamil(O.1 mM) primarily by a reduction in JB-*, consistent with net vinblastine secretion resulting from an inhibition of P-glycoprotein. 1,9-Dideoxy-forskolin and forskolin (0.1 mM) both resulted in significant inhibition of JBA and net vinblastine secretion of 64.3 f 3.1% and 29.1 f 4.8% respectively. 7~-deactyl-7/3-(~-N-methylpiperaxino)-butyryl-forskolin was ineffective. Half-maximal inhibition of vinblastine secretion by 1,9dideoxy-forskolin was observed at 65 PM. 1,9dideoxy-forskolin is unable to stimulate adenylate cyclase, suggesting that this forskolin derivative is a potentially important lead antagonist of P-glycoprotein for circumvention of pleiotropic drug resistance. (D 1991 Academic Press, Inc.

P-glycoprotein is a 170-180 kDa membrane glycoprotein associated with the phenomenon of

pleiotropic (multidrug) resistance (MDR) (reviewed in reference 1). Immunohistochemical

techniques have demonstrated the presence of P-glycoprotein in the apical regions of several

natural epithelia (2-5), especially in the gastrointestinal tract and kidney. Epithelial cell lines

capable of reforming intact epithelial layers when grown upon permeable matrices (e.g.,

MDCK, LLC-PK,), including those derived from human gastrointestinal tumours (T84, HCT-

8), have been shown to express P-glycoprotein and to mediate epithelial secretion (from basal

to apical cell surfaces) of P-glycoprotein substrates such as vinblastine (6-8). Retroviral

transfection of the Madin-Darby canine kidney (MDCK) epithelial cell line with mdrl cDNA

results in polarised expression of P-glycoprotein to the apical plasma membrane domain (9)

and an enhanced transepithelial secretion of vinblastine above intrinsic levels (6). We have

used the vinblastine secretion by native MDCK epithelia (6,7) as a direct functional measure

of P-glycoprotein to investigate its modulation by intrinsic stimuli and pharmacological

0006-291X/91 $1.50

671 Copyright 0 1991 by Academic Press, Inc.

All rights of reproduction in any fom reserved.

Vol. 181, No. 2, 1991 BIOCHEMtCAL AND BtOPHYSlCAL RESEARCH COMMUNlCATtONS

agents. Here we report that forskolin derivatives are effective inhibitors of P-glycoprotein

function. l,PDideoxy-forskolin has no effect upon stimulation of adenylate cyclase,

suggesting that such an otherwise inert biological molecule may be an important lead

compound in the development of specific inhibitors of P-glycoprotein function that will be

useful therapeutically.

Cell Culture MATERI[AIs AND METHODS

Strain 1 (60-75 serial passages) MDCK cells (10) were maintained in serial culture in Eagle’s Minimum Essential Medium supplemented with 2% v/v foetal calf serum, 8% v/v donor horse serum, kanamycin antibiotic (1 pg/ml), non-essential amino acids (1% v/v) and 2 mM glutamine. Confluent monolayers were subcultured every 7 days, by treatment with 0.05% trypsin and 0.2% EDTA in Ca”- and Mg*+-free phosphate-buffered saline (PBS). All cultures were incubated at 37°C in a humidified atmosphere of 5% C@/95% air.

For experimental purposes MDCK cells were grown as epithelial layers by high- density seeding (106 cellslcm2) onto permeable filter matrices (Anocel125mm culture inserts, 4 cm* growth area). Culture inserts were cultured in &-well plates for 2-3 days. The formation of functional epithelial layers was monitored visually and by the development of a significant transepithelial resistance (IQ, as measured using a WPI Evometer fitted with “chopstick” electrodes to allow transepithelial current passage and potential sensing (11). Cell monolayers were used when the transepithelial resistance exceeded 1 kQ *cm*. Values of resistance for filter inserts alone, - 300 fl l cm2, are subtracted from all data.

Measurement of bidirectional transepithelial13H/vinblasrim sulphate flures

Measurements of transepithelial solute flux were made essentially as described by Hunter et al. (7). Functional epithelial layers in filter cups were washed with 2 x 3 ml serum-free medium and placed into fresh 6-well plates containing 3 ml serum-free medium (basal solution), a further 3 ml serum-free medium was then pipetted into the upper chamber (apical solution) of the filter cup. Transepithelial resistance was measured following 10 min incubation of the cells at 37”C, as described above.

The medium on either the apical or basal side of the monolayers was then removed and replaced with 3 ml serum-free medium containing 10 nM [3avinblastine sulphate as tracer, and that on the contralateral side replaced with 3 ml serum-free medium containing 10 nM vinblastine sulphate, in the presence or absence of varying concentrations of verapamil or forskolin derivatives, followed by incubation at 37°C. In order to measure the bidirectional fluxes of vinblastine sulphate (JxeB, flux from apical to basal solutions, and JBA, flux from basal to apical solutions), 100~1 samples of medium from each side of the monolayer were taken at regular intervals, [3H+activities in these samples were determined by liquid scintillation counting. Each incubation was performed at least in triplicate. On completion of the flux experiments epithelial integrity was determined by measurement of transepithelial resistance.

Measurement of adenylate cyclase activities in MDCK cell homogenates

Adenylate cyclase was measured in trip&ate on homogenates of MDCK cells (12) in a buffer containing 80 mM Tris-maleate, 4 mM MgSO,, 0.2 mM EGTA, 1 mM isobutylmethyl- xanthine, 1 mM ATP, 0.1 mM GTP, 10 mM creatine phosphate and 10 pg cm-3 creatine kinase. Incubations were terminated by HCl and after neutralisation, samples were assayed for their CAMP content using a competitive protein binding assay (13).

672

Vol. 181, No. 2, 1991 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

Data are expressed as mean f SEM of n replicates (within experiment) or N experiments (between experiment). Tests of significance of differences between mean values were made using a two-tailed Student’s I test, where appropriate. Inhibition constants for verapamil inhibition of vinblastine flux were determined by non-linear regression with the method of least-squares fitting for a logistic sigmoid (P.Fit, Biosoft, Cambridge UK)

Materials

[3H]vinblastine sulphate was obtained from Amersham International (Little Chalfont, Bucks, UK). All tissue culture media and reagents (Gibco BRL) and tissue culture plastics and Anocell tissue culture inserts (Nunc) were supplied by Life Technologies Ltd. (Paisley, Scotland). Forskolin, 1,9-dideoxy-forskolin (DiD-forskolin) and 7/3-deactyl-7/3-(7-N- methylpipcrazino)-butyryl-forskolin dihydrochloride (DMPB-forskolin) were obtained from Calbiochem (Novabiochem, Nottingham, UK). All other chemicals were obtained from Sigma Chemical Co. (Poole, Dorset) or BDH Chemicals Ltd. (Poole, Dorset).

RESULTS

Strain 1 MDCK epithelia have a high transepithelial electrical resistance (3354 + 214

@I= 18) Q *cm’> typical of right epithelia. With tracer concentrations of vinblastine (10 nM)

a marked asymmetry in the bidirectional transepithelial fluxes of vinblastine is observed (Fig.

1); the basal-to-apical flux JB,, (0.253 + 0.014 pmol crnm2 he1 (N=15)) being considerably

: E 2.0 - 0 -

0 . g 1.6-

s .

$1.2-

a . r ;Epo.e -

2

go.4 - =

?i - a o- E >

0 1

T

0-A

DID-for8kolb

I ,

0 60 120 160 240 300

Time Iminl

-‘%0

0 ’ , “Tlllq~‘Ilnl~ 0 2 10-10 103 10-a [modulator1 10-7 103 M 10-5 10-4 0.001

Fr~uan 1. Transepithelial vinblastine sulphate fluxes. [31-Ijvinblastine flux was measured in the basolateral-*apical (B-A), apical-to-basolatera1 (A-B), and in the basolateml-to-apical direction in the presence 0.2 mM verapamil (Verqamil) or 10 CM DiD-forskolin @ii Forslcolin), added to both apical and basal bathing solutions. The lines illustrate the least squares best fit to the data, the slopes of which give the vinblastine flux rate. Data arc mean values f SE (n=3) from 5, 2 and 1 experiment! for control, vesapamil and DiD-forskolin respectively. Data for A-B fluxes plus veraPamiI and DiD-forskolin superimpose on control data and are omitted for clarity (see text).

Fmr.utn 2. Concentration-dependence of forskolin (B). DiD-forskolin (0) and veqxunil @,) inhibition of J&,, vinblastine flux in MDCK epithelial cell layers. Vinblastine flux rates are expressed as a percentage of the control flux in the absence of any modulator in the same experiment. Data arc the mean f SE of three observations per data point.

673

Vol. 181, No. 2, 1991 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

larger than flux in the reverse direction (JAea; 0.013 f 0.003 pmol crne2 h-’ (N= 15)). A net

secretion (0.24 pmol cmq h-l) of vinblastine thus occurs from basal-to-apical surfaces. Under

control conditions and in the presence of vinblastine the spontaneous transepithelial electrical

potential difference is low (l-3 mV, basal solution electropositive). The observed

asymmetries for vinblastine flux (JbA/JAm8 of 19.5 fold), therefore, cannot be accounted by

passive electrica? forces and must represent active transport by the epithelium. Net

vinblastine flux (Jad = JBA - J,& demonstrates saturation kinetics; K,,, = 2.8 f 0.6 (N=6)

PM, V, 35.9 f 1.93 (N=6) pmol cmS2 h-l.

Verapamil(O.1 mM) effectively abolishes the net secretion of vinblastine (Fig. 1) by

reducing J, to 0.045 f 0.001 (N=9) pmol crnm2 h-l (P<O.OOl vs control flux), that is

towards the levels of JAeB (0.023 f 0.009 pmol cm” h-r; P>O.O5 vs control data). The

inhibition of vinblastine secretory flux by verapamil is dose-dependent (Fig. 2), half-maximal

inhibition being observed at 3.2 f 0.3 PM (df=17).

Forskolin (10 PM in both apical and basal solutions) reduced vinblastine secretion,

by a reduction in JB,, to 0.216 + 0.017 (N=6) pmol cm” h-l, with no effect upon JAmB (0.015

f 0.003 pmol cm-’ h-‘) (see also Fig. 2). This inhibition showed variability from batch to

batch of MDCK layers; for 4 separate batches the mean inhibition of vinblastine secretion

with 10 FM forskolin was 31.2 f 8.4% (range 17-5556). DiD-forskolin at 10 FM gave a

pronounced inhibition of vinblastine secretion (Fig. 1); JBA was reduced to 0.145 f 0.02

(N=6) pmol cmm2 h-l (P< 0.01 vs controls) while JA.B was only slightly elevated to 0,025 f

0.016 (N=6) pmol cm3 h I. Again variation in the inhibition from batch to batch of MDCK

cell layers was observed (see Fig. 2). DMPB-forskolin, 10 PM, was without effect upon

vinblastine fluxes. Figure 2 illustrates the concentration-dependent effects of forskolin and

its derivative; DiD-forskolin was most potent, half-maximal inhibition of JBA being observed

at 65 * 32 PM (df=17). Forskolin at the maximal concentration used (0.3 mM) reduced

vinblastine flux by 35% in this experiment.

The inhibition of vinblastine flux by forskolin is unrelated to the ability of these

compounds to stimulate adenylate cyclase. At 10 PM forskolin increased adenylate cyclase

in MDCK cell homogenates from basal values of 464 f 41 to 5304 f 323 (n=3) pmol mg

protein-’ 15 min“, whereas DMPB-forskolin increased adenylate cyclase to 3317 f 167 pmol

mg protein” 15 min-‘, and DiD-forskolin did not increase adenylate cyclase above basal

values.

DISCUSSION

The present study has confumed the functional expression of P-glycoprotein in native

(untransfected) epithelial layers of MDCK cells. The observed net transepithelial secretion

of vinblastine sulphate by native MDCK epithelial layers, which is inhibited by verapamil,

674

Vol. 181, No. 2, 1991 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

provides an in vitro test system by which endogenous and exogenous regulation of P-

glycoprotein function may be conveniently monitored. Forskolin and DiD-forskolin are

shown to inhibit P-glycoprotein mediated flux. The mechanism of this inhibition is unlikely

to correlate with stimulation of adenylate cyclase since DiD-forskolin is an ineffective activator of adenylate cyclase. Conversely, DMPB-forskolin stimulates adenylate cyclase

activity in MDCK cells, but is unable to modify vinblastine flux.

The inhibition of P-glycoprotein mediated vinblastine secretion by forskolin and DiD-

forskolin is likely to correlate with chemosensitisation of drug-resistant cells by these

compounds, since agents which reverse multi-drug resistance are associated with inhibition

of vinblastine transport in isolated membrane vesicles (14). Further studies will be aimed

at correlating the effects of forskolin and derivatives upon membrane transport, intracellular

drug accumulation and chemosensitisation. A chemosensitising action of forskolin and DiD-

forskolin for doxorubicin-resistance in murine sarcoma S 180 variants has been reported (15).

The present data also point to the apparent diversity of chemical agents that are

capable of modulating P-glycoprotein function (1); recently intrinsic multidrug resistance in

Chinese hamster ovary cells has been reversed by amiloride analogues (16). An advantage I

of DiD-forskolin may reside in a lack of a specific pharmacology unrelated to P-glycoprotein

inhibition; verapamil, calmodulin antagonists, steroids and amiloride are (clinical) entities

whose use in therapeutic concentrations (as chemosensitisers) is limited by their primary

actions (1) (e.g., as Ca-channel or Na-channel blockers).

ACKNOWLEDGMENT

We wish to thank the North of England Cancer Research Campaign for generous financial support for this work.

REFERENCES

[l] Ford, J.M. and Hait, W.N. (1990) Pharmacol. Rev. 42, 155-199 [2] Cordon-Cardo, C., O’Brien, J.P., Boccia, J., Cassals, D., Bertino, J.R. and Melamed,

M.R. (1990) J. Histochem. Cytochem. 38, 1277-1287 [33 Sugawara, I., Kataoka, I., Morishita, Y., Hamada, H., Tsuruo, T., Hoyama, S. and

Mori, S. (1988) Cancer Res. 48, 19261929 [4] Thiebaut, F., Tsuruo, T., Hamada, H., Gottesman, M.M., Pastan, I. and Willingham,

M.C. (1987) Proc. Natl Acad. Sci. USA 84, 77357738 [5] Thiebaut, F., Tsuruo, T., Hamada, H., Gottesman, M.M., Pastan, I. and Willingham,

M.C. (1989) J. Histochem. Cytochem. 37, 159-164 [6] Horio, M., Chin, K.V., Currier, S.J., Goldenberg, S., Williams, C., Pastan, I.,

Gottesman, M.M., and Handler, J.S. (1988) J. Biol. Chem. 264, 14880-14884 [7] Horio, M., Pastan, I., Got&man, M.M. and Handler J.S. (1990) Biochim. Biophys. Acfa

1027, 116-122 [8] Hunter, J., Hirst, B.H. and Simmons, N.L. (1991) Br. J. Cancer 64, 437-444

675

Vol. 181, No. 2, 1991 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

[9] Pastan, I., Gottesman, M.M., Kazumitsu, U., Lovelace, E., Rutherford, A.V., and Willingham, M.C. (1988) Proc. Narl. Acad. Sci. USA g&4486-4490

[lo] Barker, G. and Simmons, N.L. (1981) Qu. J. Exp. Physiol. 66, 61-72 [ 1 l] Simmons, N.L. (1990) Mehods in Enzymology 191,426-436 [12] Griffiths, N.M and Simmons, N.L. (1987) J. Physiol. 387, 1-17 [13] Brown, B.L., Albano, J.D., Ekins, R.P., Sgherzi, A.M. and Tampion, W. (1971)

Biochem. J. 121,561-562 [14] Horio, M., Lovelace, E., Pastan, I. and Gottesman, M.M. (1991) B&him. Biophys.

Acta 1061, 106-110 [15] Wadler, S. and Wiemik, P.H. (1988) Cancer Res. 48, 539-543 [16] Epand, R.F., Epand, R-M., Gupta, R.S and Cragoe, E.J. (1991) Br. J. Cancer63,247-

251

676