Existence of a p-Glycoprotein Drug Efflux Pump in
`Cultured Rabbit Conjunctival Epithelial Cells
`
`Pratt/e Sci/m,"3 Johnny]. Yang,‘ and VincentH. L. Leel’2
`
`PURPOSE. To determine whether a p-glyeoprotein (P-gp) drug efflux pump exists in cultured rabbit
`conjunctival epithelial cells (RCEs) to restrict the absorption of cyclosporin A (CSA) and other
`lipophilic drugs such as verapamil and dexamethasone.
`
`Msmons. The anti—ng monoclonal antibody (mAb) C219 was used in western blot analysis to
`reveal the presence of P—gp in freshly isolated and cultured RCEs. Bidirectional transport of tritiated
`CSA, verapamil, and dexamethasone (0.5 or 5.0 ptM) across cultured RCEs was evaluated in the
`absence and presence of P-gp inhibitors and an external mAb to P-gp (4133).
`
`RESULTS. Western blot analysis of lysates of freshly isolated and cultured RCEs with C219 mAb
`revealed a 170-kDa membrane protein band. At 0.5 nM CSA, the basal-to-apical (ba) apparent
`permeability coefficient (Paw; that is, efflux) was 9.5 times higher than that in the apical-to-basal
`direction (that is, influx). At 5 jiM, this ratio was halved: Net CSA secretion was blocked completely
`at 4°C. Verapamil (100 MM), progesterone (100 MM), and 4E3 mAb (5 jig/ml) increased CSA influx
`three times, while reducing efflux by 50% to 70%. Verapamil and progesterone inhibited CSA efflux
`in a concentration-dependent manner. In all cases, net secretory CSA flux was markedly reduced.
`
`The Pam, for verapamil (0.5 MM) and dexamethasone (0.5 nM) in the ba direction was 3.4 and LG
`times, respectively, which was higher than that in the opposite direction. The 4E3 mAb reduced net
`verapamil secretion by 65%.
`
`CONCLUSIONS. There may exist a P-gp—mediated drug efflux pump on the apical aspect of the
`rabbit conjunctiva to restrict the absorption of cyclosporinA and other lipophilic drugs. (Invest
`Ophthalmol Vis Sci. 1998;39:1221-1226)
`
`he p-glycoprotein (P-gp) drug efflux pump is a 170-kDa
`membrane protein (gp170) encoded by the MDR1 gene.
`First identified in nmltidrugresistant tumor cells,l it was
`subsequently found to exist in diverse mammalian epithelia,
`including the retina and iris,2 small intestine,“4 liver,“ kid-
`n ey,
`“ and blood— brain endotheliumf"5 In these locations,
`gp170 possibly serves a protective role against the entry of
`foreign chemicals, including cyclosporin A (CSA)("7 and pacli—
`taxel.“ In Caco-Z cell nionolayers,("9"” cultured bovine brain
`capillary endothelial cells,7"' human adenocarcinoma cell
`lines (I'lC'I'—8 and T84),'2 and Chinese hamster ovaty cells,”
`P-gp has been shown to reduce the transepithelial permeation
`of CSA,“7 vinblastine,’I2 vincristine,ll and the lipophilic pep-
`tides, AcI’he(NMeI’he)2NH2,l0 and N—acetyl-leucyl-leucyl-nor-
`leucinal.” In all instances, the apical-tobasolateral (ab) trans-
`port (influx) of each drug is lower than its basolateral-to-apical
`(ba) transport (efflux). Moreover, influx and efflux are subject
`
`
`
`From the Departments of 'Pharmaceutical Sciences and 2Ophthal-
`mology, University of Southern California, Los Angeles.
`Supported in part by National Institutes of Health grant EYIO/iZI,
`the United States l’harmacopeia Fellowship, and the Gavin S. Herbert
`Professorship.
`Submitted for publication November I9, 1997; revised February
`'10, I998; accepted February 24, I998.
`Proprietary interest category: N.
`3Present address: Schering-Plough Research Institute, Kenilworth,
`NJ 07033.
`Reprint. requests: Vincent H. L. Lee, University of Southern Cali-
`fornia, School of Pharmacy, Department of Pharmaceutical Sciences,
`I985 Zonal Avenue, Los Angeles, CA 90033.
`
`Investigative Ophthalmology S: Visual Science, june 1998, Vol. 39, No. 7
`Copyright © Association for Research in Vision and Ophthalmology
`
`to modulation by other lipophilic substrates. For instance,
`Augustijns et al.“ reported that chlorpromazine (100 MM) and
`progesterone (100 MM) increased CSA (0.5 MM) influx across
`Caco-Z cell monolayers three times, while they reduced efflux
`by 75%. In addition, Tsuji et al.7 observed that P-gp inhibitors
`and MRK16 (an external monoclonal antibody [mAb] to P-gp)
`increased CSA uptake into primary cultured bovine brain cap-
`illary endothelial cells by 150% to 300%.
`The conjunctiva is the conduit for topically applied drugs
`to the posterior segment of the eye.'4 This tissue has been
`shown to be permeable to drugs of diverse physicochemical
`characteristics, including peptides.'5"H The main objective of
`the present study was to determine whether a P—gp drug efflux
`pump exists in rabbit conjunctiva that may restrict
`the
`conjunctival absorption of CSA,
`an undecapeptide P—gp
`substrate 6'7 that is being tested clinically for the treatment of
`uveitis '9 and tested experimentally for the treatment of aller-
`gic conjunctivitis.20 Two other drugs of comparable lipophi-
`Iicity also were studied: verapamil, a competitive substrate Ol:
`P-gpz'; and dexamethasone, a physiologic substrate of P-gp.22
`Cultured rabbit conjunctival epithelial cells (RCEs) on a per-
`meable support 2324 were used instead of the intact tissue to
`avoid complications caused by any P-gp that may also exist in
`the vascular endothelia in the intact tissue. Bidirectional trans-
`
`port of CSA was studied at 0.5 and 5.0 MM to bracket the
`threshold drug concentration (1 nM) previously identified in
`Caco-2 cells for CSA.6 Verapamil,9 progesterone,“ and a murine
`mAb that recognizes an external epitope of P-gp, namely,
`4E3,25 were used as P-gp inhibitors. A different mAb, C219,
`which binds to a cytoplasmic epitope near the adenosine
`
`1221
`
`ALCON 2259
`Apotex Inc. v. Alcon Pharmaceuticals, Ltd.
`Case |PR2013-00012
`
`

`

`1222
`
`Saha et a1.
`
`IOVS, June 1998, Vol. 39, No. 7
`
`triphosphate binding sites of gp170,26 was used in western
`blot analysis to further demonstrate the existence of P-gp in the
`conjunctiva. This antibody has been used extensively for that
`purpose in other tissues.27
`
`MATERIALS AND METHODS
`
`Materials
`
`3H—CSA (specific activity, 7.5 Ci/mmol) was obtained from Am-
`ersham Co. (Downers Grove, IL). D-[3H]mannitol (specific ac-
`tivity, 26.4 Ci/mmol), 3‘H-Verapamil (specific activity, 66.3 Ci/
`mmol), and 5H-dexamethasone (specific activity, 43.9 Ci/
`mmol) were obtained from Dupont NEN (Boston, MA). 36Cl
`was obtained from Amersham (Arlington Heights, IL). Unla-
`beled CSA was supplied by Sandoz Pharmaceuticals (Basel,
`Switzerland). Unlabeled verapamil, progesterone, dexametha-
`some, and D-mannitol were purchased from Sigma Chemical (St.
`Louis, MO). C219 and 4133 murine mAbs were obtained from
`Signet Laboratories (Dedham, MA). Cell culture media and
`supplies were obtained from Gibco (Grand Island, NY). PC-l
`medium (a serum-free, low-protein, defined medium) was pur-
`chased from Hycor Biomedical
`(Portland, ME). Collagen-
`treated Transwells (Transwell—COL inserts with 0.45pm pore
`size, 12-mm diameter, 1.13 cm2) and cluster plates (tissue
`culture-treated multiwell plates for holding the inserts) were
`obtained from Costar (Cambridge, MA). Male Dutch-belted
`pigmented rabbits, weighing 3.0 to 3.5 kg, were purchased
`from American Rabbitry (Los Angeles, CA). The studies using
`rabbits described in this report conformed to the Guiding
`Principles in the Care and Use of Animals (Department of
`Health, Education, and Welfare, National Institutes of Health;
`Publication 80 —23), and all animals were treated in accordance
`with the the ARVO Statement for the Use of Animals in Oph-
`thalmic and Vision Research.
`
`Primary Culture of Rabbit Conjunctival Epithelial
`Cell Layers
`
`The procedure for the routine generation of tight RCE layers
`was reported elsewhere.” Briefly, epithelial cells were isolated
`by treating the excised conjunctiva with 0.2% protease for 90
`minutes. Isolated cells were declumped with DNase I and
`centrifuged. The cell pellet was washed, filtered through a
`40-pin cell strainer, and centrifuged. The final cell pellet was
`resuspended in defined PC-l growth medium supplemented
`with 1% fetal bovine serum, 2 mM uglutamine, 100 U/ml
`penicillin-streptomycin, 0.5% gentamicin, and 0.4% fungizone.
`The cells were plated at a density of 1.5 X 106 cells/cm2
`on Costar 12-mm Transwell-COL inserts. After 48 hours, adher-
`ent cells were washed with a defined PC-l medium containing
`0.1% fetal bovine serum, and they were fed daily with serum-
`free medium from day 3 onward. The transepithelial electrical
`resistance (in kilo-ohms per square centimeter) and potential
`difference (apical negative; in millivolts) of the cultures were
`monitored with an epithelial voltohmeter (EVOM; World Pre-
`cision Instruments, Sarasota, FL) and were corrected for back-
`ground transepithelial electrical resistance and potential differ-
`ence contributed by the blank filter and culture medium.
`Confluent cultured RCEs (days 6—8) with transepithelial elec-
`trical resistance 2 1.0 kQ/cm2 were used for transport studies.
`
`Western Blot Analysis
`
`Western blot analysis was performed with the C219 murine
`mAb using the procedure modified from that of Muller et a1.23
`Freshly isolated and confluent (day 8) cultured RCEs grown on
`Transwells underwent lysis for 45 minutes in ice-cold phos-
`phate-buffered saline containing 3% sodium dodecyl sulfate
`and protease inhibitors (83 MM antipain, 73 MM pepstatin A,
`and 0.1 mM leupeptin). The cell lysate then was centrifuged at
`12,000g for 15 minutes, and the supernatant was used for
`protein assay. Total cell protein was measured by a protein
`assay (DC; Bio—Rad, Hercules, CA), using bovine serum albu-
`min as a standard. Cell proteins (10 Mg) underwent electro-
`phoresis on 8% sodium dodecyl sulfate—polyacrylamide gel and
`subsequently were electrotransferred to a nitrocellulose mem-
`brane (Amersham). An immunoblot procedure based on the
`enhanced chemiluminescence method was performed by ex-
`posing the protein side of the nitrocellulose membrane to the
`enhanced chemiluminescence method detection reagent (Am-
`ersham) for 1 minute, followed by immediate exposure to x-ray
`film.
`
`Bidirectional Transport of P-Glycoprotein
`Substrates
`
`Before each experiment, cultured cells grown on Transwell
`inserts were washed with modified Ringer's solution (pH 7.4)
`and were allowed to equilibrate at 37°C. This solution (300
`mOsm) comprised 116 mM NaCl, 5.6 mM KCl, 0.8 mM
`NaHZPOm 25 mM NaHCOS, 1.8 mM CaCl2.2H20, 0.8 mM
`MgSO4, 5.5 mM D-glucose, and 15 mM HEPES. Hypotonic
`modified Ringer’s solution (200 mOsm) was made by reducing
`the NaCl concentration to 66 mM.29 After incubation at 37°C,
`the transepithelial electrical resistance was measured to gauge
`the integrity of the cell layers. The same protocol was used for
`the 4°C incubations.
`
`Transport of 3H-CSA in the Apical-to-Basolateral Di-
`rection. CSA transport was performed based on the proce-
`dures of Augustijns et al.6 and Hosoya et al.,30 which were
`designed to minimize the adsorption of CSA to Transwells,
`thereby maximizing its recovery (approximately 90%). Sili-
`conized pipet tips were used in all experiments.
`Transport was initiated by adding a solution containing
`3H—CSA (0.3 uCi) and unlabeled CSA to the apical compartment
`(0.4 ml) of Transwell
`inserts bathed with 1.5 ml modified
`Ringer’s solution in the basolateral compartment, followed by
`incubation at 37°C in a 5% CO2 atmosphere. The final concen-
`tration of CSA in the apical compartment was either 0.5 MM or
`5 [.LM. At predetermined times during a 3-hour period, the
`inserts were transferred to new wells containing 1.5 ml fresh
`modified Ringer‘s solution. Twenty microliters of 1 mM unla-
`beled CSA then was added to the receiver fluid to minimize
`
`3’H—CSA adsorption to the well. After 1 hour of incubation at
`37°C, the receiver solution was removed and placed (along
`with the tip) in a scintillation vial (first sampling). To maximize
`the recovery of the approximately 20% 3H-CSA remaining that
`was adsorbed into the receiver well, 2 ml of modified Ringer’s
`solution containing 10 uM cold CSA in 10% dimethylsulfoxide
`was added to the receiver wells for incubation overnight be-
`fore collection (along with the tip) in another scintillation vial
`(second sampling). The amount of 3H-CSA transported (sum of
`both samplings) was measured in a liquid scintillation spec-
`trometer (Beckman, Fullerton, CA).
`
`

`

`IOVS, June 1998, Vol. 59, No. 7
`
`Conjunctival Drug Efflux Pump
`
`1223
`
`Transport of all-GSA in the Baselateralto-Apical Di-
`rection. aH—CSA (0.3 [LCD and unlabeled CSA were added
`separately to the basolateral compartment of Transwells to
`attain a final drug concentration of 0.5 ,LLM or 5 MM in 1.5 ml,
`and they were mixed with gentle agitation. Inserts containing
`0.4 ml modified Ringer‘s solution in the apical Compartment
`then were transferred to the basolateral donor compartment
`containing radiolabeled CSA to initiate transport, as described
`above. In time-course studies, 0.4 ml modified Ringer’s solution
`on the apical side containing transported (ISA was removed
`and replaced with fresh modified Ringer’s solution. The apical
`samples (along with the tip) were placed in scintillation vials
`for counting in a liquid scintillation spectrometer.
`Transport in the Presence of Inhibitors and 4133
`Monoclonal Antibody. CSA (0.5 uM) transport in the pres—
`ence of li-gp inhibitors, verapamil, and progesterone (S, 10, 50,
`and 100 ,ulvl) was carried out essentially as described earlier
`with minor modifications. verapamil and progesterone stock
`solutions were prepared in modified Ringer‘s solution contain—
`ing 1% dimethylsulfoxidefa The same concentration of P—gp
`inhibitor was present in the apical and basolateral compart—
`ments atall times during transport. In the case of 4B} mAb (5
`pig/ml), it was present only in the apical compartment.
`Transport of Verapamil, Dexamethasone and o-Man-
`nitol. The same protocol described for CSA was used to cval~
`uate the transport: of verapamil (0.5 ,uM), dexamethasone (0.5
`nM; a transcellular transport marker“), and omannitol (1 mM;
`:1 paracellular marker”). Verapamil transport also was evaluw
`ated in the presence of the 4E3 math to further confirm its
`interaction with the P—gp efflux pump.
`
`Data Analysis
`
`The steady-State flux was estimated from the slope of the linear
`portion of a plot of a cumulative amount (moles) of the drug or
`36C! appearing in the receiver fluid versus time. The apparent
`permeability coefficient (flaw; centimeters per second) was
`calculated from the observed flux (rigid: in moles per second)
`normalized against the nominal surface area of the filter mem—
`brane (A; 1.13 cm2) and the initial drug concentration (Co in
`moles per milliliter) in the donor fluid, according to the equa—
`tion: l3app : (dQ/dt)/(A*C0). Statistical significance was tested
`by two-tailed Student’s I—test or one-way analysis of variance
`(Fisher PLSD) using a statistical software package (Starview II;
`Ahacus Concepts, Berkeley, CA). Statistical significance was set
`at P < 0.35.
`
`RESULTS
`
`Western Blot Analysis
`
`As shown in Figure 1, a distinct major immunoreactive band at
`approximately 170 kDa was evident in cultured (lanes 2 and 3)
`and freshly isolated RCEs (lane 4). There exist numerous other
`minor bands, which Suggests that the C219 mAb might recog-
`nize more than one epitope or might signify the presence of
`truncated P-gp fractions in the RCE. The absence of any posi-
`tive band in lane 1 (cultured RCEs) and lane 5 (freshly isolated
`RCES)
`in the presence of a mouse Ingn isotype—matchcd
`negative control antibody served as a negative control.
`
`MW. (ltDa)
`
`1
`
`9
`
`3
`
`5
`
`‘
`
`203 —D-
`
`1 18
`
`I
`
`-- W
`“filmW
`“rm-ma w
`.....mat
`
`‘_ P-g p
`,
`-
`( “' l70de)
`
`s
`
`.....
`
`51—p-
`
`34—r
`
`.3”;
`
`FIGURE 1. Western blot analysis of pnglycoprotein (P-gp) ex-
`pression in cultured conjunctiva] epithelial cells. The numbers
`to the left of the blot indicate the positions of molecular mass
`markers that underwent electrophoresis in parallel (expressed
`in kilodaltons). The proteins used as molecular mass standards
`were myosin (205 kDa), B—galactosidase (11,8 kDa), ovalbumin
`(51 kDa), and carhmiic anhydrase ('54 kDa) (presrainetl sodium
`dodecyl sulfate—polyacrylamide gel electrophoresis standards,
`broad range). Lane 1, mouse lngfl isotype—matched negative
`control in cultured conjunctiva] epithelial cells (RCES); lanes 2
`and 3, mouse C219 monoclonal antibody to P—gp in cultured
`RCES; {ans 4, mouse (3219 monoclonal antibody to P-gp in
`freshly isolated R035; and tone 5, mouse lgGZ,
`isowpe—
`matched negative control in freshly isolated REES.
`
`Directionality and Temperature Dependency of
`Conjunctival Cyclosporin A Transport
`
`The time course of 0.5 nM CSA transport in either direction
`across cultured RCE cell layers, in the presence and absence of
`100 ,uM verapamil, is shown in Figure 2. CSA (0.5 ,uM) flux
`exhibited marked directional asymmetry. The bit Pam, was 9.5
`times higher than in the ab direction (P < 0.01; see Table 1).
`By contrast,
`the bidirectional transport of Elli-mannitol was
`symmetric. Its ba Papp was 1 X 10—7 cm/sec, and its ab Papp
`was 0.9 X 10—7 cm/sec. When the CSA concentration was
`
`raised to 5 pm, the ratio of ha and ab PM, was reduced to 4.6
`(Table 1). This was attributed principally to a 70% increase in
`
`the (ISA ab Pam When the temperature was lowered to 4°C,
`the ab Papp was increased three times, while the ba Pm, was
`reduced three times (Table 1), which resulted in no net secre-
`tion.
`
`Effect of P—Glycoprotein on Conjunctival
`Cyclosporin A Transport
`
`Verapamil (10, 50, and 100 old) and progesterone (50 nM and
`100 uM) inhibited the conjunctival efflux of 0.3 MM GSA in a
`dose-dependent manner (P < 0.05 by one~way analysis of
`variance; see Fig. 3). No significant inhibition of CSA secretory
`flux was observed with 5 pM verapamil and 5 uM to 10 MM
`progesterone (Fig. 3). Although 100 uM verapamil and proges—
`terone were as effective as 5 rig/ml 4E3 mill) in enhancing CSA
`influx (Table 1), both were less effective than 4E3 mAb in
`reducing (ISA efflux (Table 1). Indeed, 4E3 mAb was the only
`modulator that abolished net GSA secretion.
`
`

`

`1224
`
`Saha et a1.
`
`IOVS, June 1998, Vol. 39, No. 7
`
`cm/s)
`
`CsAbaPapp(E-07
`
`aa
`
`Vg
`
`Verapamil
`
`Progesterone
`
`FIGURE 3. Basolateral-to-apical (ba) transport of cyclosporin A
`(CSA; 0.5 MM) in the absence (I) or the presence of 5 MM (11),
`terone. Error bars represent SEM for n = 4 to 8. Data were
`statistically significant (P < 0.05) by one-way analysis of vari-
`ance (Fisher PLSD) between groups I and III, I and IV, I and V,
`II and III, II and IV, I] and V, 111 and IV, and III and V for
`verapamil, and between groups I and IV, I and V, II and IV, II
`and V, and III and V, and IV and V for progesterone.
`
`ab Papp 1.7 times (Table 2). The net result was a 65% smaller
`pp
`net secretory P“
`and a 2.3 times smaller ratio of efflux to
`influx.
`
`DISCUSSION
`
`The study provides evidence for the existence of a P~gp drug
`efflux pump in cultured RCEs that may restrict the conjunctival
`absorption of CSA and other lipophilic drugs such as verapamil
`and dexamethasone. Western blot analysis using C219 mAb
`
`CsATransported
`
`(pmol)
`
`Time (h)
`
`FIGURE 2. Time course of 0.5 MM cyclosporin A (CSA) bidi-
`rectional transport across cultured conjunctiva] epithelial cell
`layers in the absence (-) or presence (+) of 100 MM verapamil.
`I, apical-to-basal transport (ab; —); El, basal-to-apical transport
`(ba; —); O, ab (+ 100 uM verapamil): 0, ba (+100 p.M vera-
`pamil). Error bars represent SEM for n = 4 to 5. In cases in
`which it is not seen, the error bar is smaller than the size of the
`symbol.
`
`Conjunctival Verapamil and Dexamethasone
`Transport
`
`As in the case of CSA, verapamil and dexamethasone transport
`across the conjunctiva also showed asymmetry in favor of
`efflux (Table 2). Compared with CSA, verapamil and dexameth—
`asone afforded a 30 to 50 times higher Pam, in the ab direction
`and a 10 times higher Papp in the ba direction (Table 2).
`Consequently,
`the ratio of efflux to influx was smaller for
`verapamil and dexamethasone, The presence of 5 ug/ml 4E5
`mAb reduced verapamil ba Papp by 25%, Whereas it increased
`
`TABLE 1. Cyclosporin Transport Across Cultured Rabbit Conjunctival Epithelial Cell Layers Under
`Various Conditions
`
`CsA
`
`(MM)
`
`0.5
`5.0
`
`0.5
`0.5
`
`0.5
`
`0.5
`
`Pa“, (10—7 cm/sec) *
`
`Condition
`
`Control
`10-fold
`Excess
`4°C
`+100 “M
`Verapamil
`+100 p.M
`Progesterone
`+5 ug/ml
`4E3 mAb
`
`'ab
`
`0.9 i 0.04
`1.5 i 03*
`
`2.7 : 0.4%
`2.7 i 0.6:
`
`2.8 : 0.6a
`
`2.5 i 0.2:
`
`[m
`
`8.3 i 0.9
`6.9 : 0.2§
`
`2.7 i 0.1:
`4.1: 0.31:
`
`3.2 i 0.41:
`
`2.6 i 0.71:
`
`Net
`
`' 7.4
`5.4
`
`0
`1.4
`
`0.4
`
`0.1
`
`Ratio1'
`
`9.3
`4.6
`
`1,0
`1.5
`
`1.1
`
`1.0
`
`ba, basal-to—apical transport; ab, apical-to-basal transport; Paw, apparent permeability coefficient; CSA, cyclosporin A; mAb, monoclonal
`antibody.
`‘ Apparent permeability coefficient; mean : SEM for n = 4 to 6.
`’r ba/ab Pupp
`tip < 0.05 from control.
`§p > 0.05 from control.
`
`

`

`IOVS, June 1998, Vol. 39, No. 7
`
`Conjunctival Drug Efflux Pump
`
`1225
`
`TABLE 2. Apparent Permeability Coefficient Values for the Transport of p—Glycoprotein Substrates Across Cultured
`Rabbit Conjunctival Cell Layers
`
`—6
`
`Pa”, (10
`
`cm/sec)1'
`
`Solute
`
`log PC‘
`
`ab
`
`CsA
`Verapamil
`Verapamil + 4135 mb
`Dexamethasone‘
`
`5.0
`3.8
`
`2.9
`
`0.09 i 0.0
`2.8 i 0.1
`4.8 i‘ 0.1
`4.1 i 0.1
`
`lm
`
`0.83 : 0.09§
`9.3 : 0.6§
`7.1 i 0.3“
`6.6 i- O.5§
`
`Net
`
`7.4
`6.5
`2.3
`2.5
`
`Ratio:
`
`9.3
`3.4
`1.5
`1.6
`
`ba, basal-to-apical transport; ab, apical-to-basal transport; P
`antibody.
`‘ Logarithm of the n-octanol/pH 7.4 buffer partition coefficient.
`TMean : SEMforn = 4t08.
`fl: ba/ab Pup”.
`§ P < 0.05 from ab flux.
`‘P > 0.05 from ab flux.
`
`app!
`
`apparent permeability coefficient; CSA, cyclosporin A; mAb, monoclonal
`
`against P—gp revealed a 170-kDa protein in the lysates of freshly
`isolated and cultured RCEs (Fig. 1), as was the case in Caco~2
`cells.30 Despite equal amounts of protein, the density of the
`170-kDa band in freshly isolated RCEs was less than that af-
`forded by the cultured cells. A possible explanation may be the
`dilution of epithelial cell protein, hence P-gp, by nonepithelial
`cell protein in the case of the tissue. Another possible expla-
`nation may be the overexpression of P-gp in cultured cells that
`was induced by hormones alone or in conjunction with xenO»
`biotics in the medium. This has been shown to be the case for
`
`rat hepatocyte primary culturesgz’fi’i and for Caco-2 cells.35 In
`any event, the location of the P—gp pump in the cell is probably
`apical, which is characteristic of other cell types. Nevertheless,
`the basolateral location of this or a related P-gp drug efflux
`pump cannot be ruled out. Indeed, the multidrug resistance—
`related protein, another drug efflux pump, has been found
`recently in the basolateral membrane of the pig kidney cell line
`LLc-PKi.36
`‘
`Probably because of the P-gp drug efflux pump, the trans-
`port of CSA, verapamil, and dexamethasone across the con-
`junctiva was asymmetric (9.5, 3.4, and 1.6 times higher, re-
`spectively; see Table 2) for efflux than influx. Further evidence
`for the involvement of P-gp in restricting the conjunctiva]
`absorption of CSA includes: saturation of P—gp with increasing
`CSA concentration, as indicated by a 67% increase in CSA
`influx when increasing the CSA concentration from 0.5 to 5
`MM (Table 1); a threefold increase in the influx of CSA in the
`presence of 100 uM verapamil, 100 uM progesterone, and 5
`jig/ml 4E3 mAb (Table 1); and the abolition of net CSA secre-
`tion at 4°C (Table 1). Partly because of the P-gp drug efflux
`pump, the ab Pam, of CSA, verapamil, and dexanlethasone was
`1 to 2 orders of magnitude smaller than that of betaxolol of
`comparable lipophilicity (log PC = 5.4).“ The 4°C and the 4155
`mAb results suggest that approximately 30% of CSA transport
`in either direction is P-gp independent. Given that the PH“, of
`CSA under these conditions (2.5—2.7 X 10—7 cm/sec) is ap—
`proximately three times higher than that of mannitol (1 X 10~7
`cm/sec), such a P—gp-independent pathway is probably trans
`cellular.
`
`Like CSA, the transport of verapamil across cultured RCEs
`also showed directional asymmetry (Table 2). The net secre—
`tory verapamil PuPp was reduced by 65% in the presence of S
`jig/ml 4E3 mAb, which suggests that verapamil itself was a
`
`substrate for P-gp. Such a possibility is consistent with the
`finding of Saitoh and Aungst.2| It is impossible to estimate the
`relative affinity of CSA and verapamil for P-gp on the basis of
`the ratio of Palms in the ba and ab directions at a single
`concentration (0.5 jiM). In inhibition studies, verapamil inhib-
`ited CSA ba Pm, at concentrations of 10 11M to 100 MM, which
`were 20 to 200 times higher than that of CSA (0.5 MM).
`Therefore, verapamil may block CSA efflux by acting as a
`competitive P—gp substrate. Relative to CSA and verapamil of
`comparable lipophilicity,
`the ratio of efflux and influx for
`dexamethasone was smaller (1.6; see Table 2). It is possible
`that dexamethasone efflux was already saturated at 0.5 MM,
`which was 20 times higher than the 25 nM used by Ueda22 to
`establish dexamethasone as a physiological substrate of P—gp.
`The unexpectedly lower Papp for CSA in comparison with
`that for verapamil and dexamethasone of comparable lipophi-
`licity needs to be explained. For a lipophilic peptide similar to
`CSA, the principal determinant of passive transport across the
`cell membrane seems to be desolvation of the solute, rather
`than lipophilicity, for it to enter the hydrophobic, non-hydro-
`gen— bonding interior of the membrane.”38 Therefore, as the
`number of potential solute—solvent hydrogen bonds increase,
`peptide permeability will decrease. The presence of potential
`hydrogen-bonding functional groups similar to amide (—NH-
`CO—) and alcohol (-OH) in the CSA structure suggests a pro-
`nounced impeding influence of the associated hydrogen-bond-
`ing potential on the entry of CSA into the membrane interior of
`the conjunctival cells. This factor together with the P-gp drug
`Ipp
`efflux pump probably contribute to the unusally small CSA PK
`observed in the ab direction in the conjunctiva.
`
`SUMMARY
`
`Western blot analysis and the transport behavior of model P-gp
`substrates under typical conditions for detecting the existence
`of a P-gp drug efflux pump support such a possibility in RCEs.
`The present study underscores a possible role for P—gp in
`restricting conjunctival CSA absorption en route to the poste-
`rior segment after topical application. Further work is required
`to determine whether the conjunctival absorption of other
`lipophilic ophthalmic drugs would be affected similarly as for
`CSA and to determine whether P—gp also exists in corneal
`epithelial cells.
`
`

`

`1226
`
`Saha et al.
`
`IOVS, June 1998, V01. 39, N0. 7
`
`Acknowledgments
`The authors thank John Xu for technical assistance with the western
`blot analysis.
`
`References
`l.
`
`Ling V. P-glycoprotein and resistance to anticancer drugs. Cancer.
`1992;69:2603-2609.
`Holash JA, Stewart PA. The relationship of astrocyte-like cells to
`the vessels that contribute to the bloodocular barriers. Brain Res.
`1993;629:218-224.
`'l‘hiebaut F, Tsuruo T, Hamada H, Gottesman MM, Pastan I, Will-
`ingham MC. lmmunohistochemical localization in normal tissues
`of different epitopes in the multidrug transport protein P170:
`evidence for localization in brain capillaries and crossreactivity of
`one antibody with a muscle protein. JHistocbem Cytochem. 1989;
`37:1 59 —164.
`. Thiebaut F, Tsuruo T, Hamada H, Gottesman MM, Pastan I, Will-
`ingham MC. Cellular localization of the multidrug-resistance gene
`product p-glycoprotein in normal human tissues. Proc Natl Acad
`Sci USA. 1989;813:7735—7738.
`. Cordon-Cardo C, O‘Brien JP, Casalas D, Rittman-Grauer L, Biedler
`JL, Melamed M. Multidrug—resistance gene (p—glycoprotein) is ex-
`pressed by endothelial cells at blood-brain barrier sites. Proc Natl
`Acad Sci USA. 1989;86:695-698.
`Augustijns PF, Bradshaw TP, Gan LL, Hendren RW, ’l'hakker DR.
`Evidence for a polarized efflttx system in Caco~2 cells capable of
`modulating cyclosporin A transport. Biochem Biophys Res Co-
`mtm. 1993;197:360 —365.
`. Tstlji A, Tamai I, Sakata A, Tenda Y, Terasaki T. Restricted transport
`of cyclosporin A across the blood-brain barrier by a multidrug
`transporter, p-glycoprotein. Biocbem Pbm‘macol. 1993;46:1096—
`1099.
`. Sparreboom A, Van AsperenJ, Mayer U, et al. Limited oral bioavail-
`ability and active epithelial excretion of paclitaxel (Taxol) caused
`by p-glycoprotein in the intestine. Proc Natl Acad Sci USA. 1997;
`94:2031-2035.
`. Hunter J, Hirst BH, Simmons NL. Drug absorption limited by
`p—glycoprotein—mediated secretory drug transport in human intes-
`tinal epithelial Caco-2 cell layers. Pharmacol Res. 1993;10:743-
`749.
`Burton PS, Conradi RA, Hilgers AR, Ho NFH. Evidence for a polar-
`ized efflux system for peptides in the apical membrane of CacoZ
`cells. Biochem Biophys Res Cormm. 1993;190:760 -766.
`Tsuji A, Terasaki T, Takabatake Y, et al. P—glycoprotein as the drug
`efflux pump in primary cultured bovine brain capillary endothelial
`cells. Life Sci. 1992;51:1427—1437.
`HunterJ, Hirst BH, Simmons N1. Epithelial secretion of vinblastine
`by human intestinal adenocarcinoma cell (HCT-8 and T84) layers
`expressing p-glycoprotein. Br] Cancer. 1991;64:437—444.
`Sharma RC, lnoue S, RoitelmanJ, Schimke RT, Simoni R1). Peptide
`transport by the multidrug resistance pump] Biol Chem. 1992;
`267:5731-5734.
`Ahmed l, Patton TF. Disposition of timolol and inulin in the rabbit
`eye following corneal versus non-corneal absorption. Int] Pbarm.
`1987;38:9-21.
`Hayakawa E, Chien DS, lnagaki K, Yamamoto A, Wang W, Lee VHL.
`Conjunctival penetration of insulin and peptide drugs in the albino
`rabbit. Pbarmacol Res. 1992;92769—775.
`Horibe Y, Hosoya K, Kim K, Ogiso T, Lee VHL. Polar solute
`transport across the pigmented rabitt conjunctiva: size depen-
`dence and the influence of 8-bromo cyclic adenosine monophos-
`phate. Pharmacol Res. 1997;14:1246—1251.
`Sun L, Basu SK, Kim K, Lee VHL. Arginjne vasopressin transport
`and metabolism in the pigmented rabbit conjunctiva. EurJPbafin
`Sci. l998;6:47~52.
`Hosoya K, Kompella UB, Kim K, Lee VHL. Contribution of Na"—
`glucose cotransport to the shortcircuit current in the pigmented
`rabbit conjunctiva. Curr Eye Res. 1996;16:1050—1055.
`
`2.
`
`.
`
`VI
`
`6.
`
`10.
`
`11.
`
`12.
`
`‘13.
`
`14.
`
`15.
`
`16.
`
`17.
`
`18.
`
`19.
`
`20.
`
`21.
`
`22.
`
`23.
`
`24.
`
`25.
`
`26.
`
`27.
`
`28.
`
`29.
`
`30.
`
`31.
`
`32.
`
`33.
`
`34.
`
`35.
`
`36.
`
`37.
`
`38.
`
`Oh JO, Minasi P, Grabner G, Ohashi Y. Suppression of secondary
`herpes simplex uveitis by cyclosporine. Invest Ophthalmol Vis
`Sci. 1985;26:494 -500.
`Whitcup SM, Chan CC, Luyo DA, Bo P, Li Q. Topical cyclosporine
`inhibits mast cell-mediated conjunctivitis. Invest Ophthalmol Vis
`Sci. 1996;357:2686-2695.
`Saitoh H, Aungst BJ. Possible involvement of multiple p-glycopro—
`tein-mediated efflux systems in the transport of verapamil and
`other organic cations across rat intestine. Pbarmacol Res. 1995;
`12:1304-1310.
`Ueda K. Human p—glycoprotein transports cortisol, aldosterone,
`and dexamethasone, but not progesterone.J Biol Chem. 1992;267:
`24248—24252.
`Saha P, Kim K, Lee VHL. A primary culture model of rabbit
`conjunctival epithelial cells exhibiting tight barrier properties.
`Curr Eye Res. 1996;15:1163—1169.
`Saha P, Uchiyama T, Kim K, Lee VHL. Permeability characteristics
`of primary cultured rabbit conjunctival epithelial cells to low
`molecular weight drugs. Curr Eye Res. 1996;15:1170—1174.
`Arceci RJ, Steiglitz K, Croop JM, Bras J, Schinkel A, Baas F. A
`monoclonal antibody to an external epitope of the p-glycoprotein.
`Cancer Res. 1992;53:310 -317.
`Georges E, Zhang JT, Ling V. Modulation of ATP and drug binding
`by monoclonal antibodies against p-glycoprotein. ] Cell Physiol.
`1991;148:479—484.
`Toth K, Vaughan MM, Slocum HK, et al. Comparison of an immu-
`noperoxidase “sandwich” staining method and western blot de-
`tection of p-glycoprotein in human cell lines and sarcomas. Am]
`Pathol. 1992;140:1009-1016.
`Muller C, Bailly JD, Goubin F, et al. Verapamll decreases p~glyco-
`protein expression in multidrug-resistant human leukemic cell
`lines. Int] Cancer. 1994;56:749 —754.
`Ambasch K, Cabantchik I, Slotki
`IN. Effects of hypotonic and
`hypoionic media on drug pumping by p-glycoprotein expressed in
`epithelial and nonepithelial cell lines. ] Cell Physiol. 1995;164:
`1 17- 122.
`
`Hosoya K, Kim K, Lee VHL. Age-dependent expression of p-glyco—
`protein gp17O in Caco-Z cell monolayers. Pbarmacol Res. 1996;
`13:885- 890.
`'
`Artursson P, KarlssonJ. Correlation between oral drug absorption
`in humans and apparent permeability coefficients in human intes-
`tinal epithelial (Caco-2) cells. Biochem Biophys Res Comun. 1991;
`175:880—885.
`Adson A, Raub TJ, Burton PS, et al. Quantitative approaches to
`delineate paracellular diffusion in cultured epithelial cell monolay-
`ers] Pharm Sci. 19943315294536.
`Bailly JD, Pourquier P, Jaffrezou JP, et al. Effect of 5736—c0ndi-
`tioned medium and recombinant cytokines on P—glycoprotein ex-
`pression in a human GM-CSF—dependent leukemic my