Abstract

Purpose. The objective of this study was to elucidate themechanism of ganciclovir uptake by the rabbit retina and thehuman retinal pigmented epithelium cell line ARPE-19.

Materials and methods. [3H]Adenine, [3H]adenosine,[3H]thymidine, and [3H]ganciclovir were used to elucidatethe mechanism of ganciclovir uptake by the ARPE-19 cellline and the isolated rabbit retinal tissue. Uptake studiesusing ARPE-19 cell line and isolated rabbit retina werecarried out at 37°C and 25°C, respectively, for 5min.

Results. Uptake of [3H]adenine by ARPE-19 cells decreasedby 95% in the presence of unlabeled adenine. Other nucleo-bases such as guanine, thymine, and uracil and the nucleo-sides adenosine, guanosine, thymidine, uridine, and inosinealso reduced uptake of [3H]adenine by the ARPE-19 cells.Although [3H]adenosine and [3H]thymidine uptake wasinhibited by nucleosides, nucleobases did not demonstrateany inhibitory effect, indicating that nucleosides can onlybind to the nucleobase transporter but are not translocated byit. Uptake of the nucleosides and nucleobases by the ARPE-19 cells was sodium and pH independent. [3H]adenosine and [3H]thymidine uptake by the ARPE-19 cells was inhib-ited by nanomolar quantities of nitrobenzylthioinosine.Uptake of [3H]adenine by the isolated rabbit retina was drastically reduced in the presence of unlabeled adenine.Unlabeled thymidine and guanosine, and removal of sodium from the uptake medium, inhibited uptake of[3H]thymidine by the rabbit retina. Nucleosides, nucleobases,and unlabeled ganciclovir did not exhibit any inhibitoryeffect on [3H]ganciclovir uptake by the isolated rabbit retinaor ARPE-19 cells.

Conclusions. Our results indicate that although the rabbitretina and the ARPE-19 cell line express nucleoside andnucleobase transporters, translocation of ganciclovir does not involve any carrier-mediated transport process. Rather,ganciclovir uptake by the rabbit retina and ARPE-19 cells isgoverned primarily by passive diffusion.

Keywords: ARPE-19; ganciclovir; nucleobase; nucleoside;RPE; transporters

Introduction

Human cytomegalovirus (HCMV) retinitis is a very commonopportunistic infection affecting about 15%–42% of patientswith acquired immunodeficiency syndrome (AIDS), oftenleading to blindness if left untreated.1 The infection is knownto appear first in the inner layers of the retina and then spreadoutward to all other layers including retinal pigmentedepithelium (RPE). Viral replication within retinal cells is akey factor in the progression of the disease and occurs in theendothelial cells of retinal blood vessels and in retinal pig-mented and nonpigmented epithelial and retinal glial cells.2,3

Ganciclovir (GCV), a 2¢-deoxyguanosine analog, was thefirst FDA-approved drug available in the United States withsignificant activity against HCMV. Previous studies reportedexcellent in vitro activity against all of the herpesviruses.4,5

GCV is converted to its triphosphate form in HCMV-infectedcells 10–100 times more efficiently than in uninfected cells.GCV triphosphate is known to inhibit HCMV-induced DNA polymerase and also causes DNA chain termination.6

Received: November 3, 2003Accepted: April 23, 2004

Correspondence: Ashim K. Mitra, Division of Pharmaceutical Sciences, School of Pharmacy, University of Missouri-Kansas City, KansasCity, MO, 64110-2499, USA. Fax: (816)-235-5190; E-mail: mitraa@umkc.edu

Mechanism of ganciclovir uptake by rabbit retina and human

retinal pigmented epithelium cell line ARPE-19

Soumyajit Majumdar1, Sreeraj Macha1,2, Dhananjay Pal1 and Ashim K. Mitra1

1Division of Pharmaceutical Sciences, School of Pharmacy, University of Missouri-Kansas City, Kansas City, Missouri, USA;2Drug Metabolism and Pharmacokinetics, Boehringer Ingelheim, Ridgefield, Connecticut, USA

Current Eye Research2004, Vol. 29, Nos. 2–3, pp. 127–136

DOI: 10.1080/02713680490504678 © 2004 Taylor & Francis Ltd.

Cur

r E

ye R

es D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Cal

ifor

nia

Irvi

ne o

n 11

/07/

14Fo

r pe

rson

al u

se o

nly.

128 S. Majumdar et al.

Because the blood–retinal barrier (comprising the RPE and endothelial cells of retinal blood vessels) prevents permeation of therapeutic concentrations of GCV into theposterior chamber following systemic (oral or intravenous)administration,7–9 intravitreal injections, and recently vitrealimplants, of GCV have become the mainstay therapy in CMVretinitis.10–12

Several reports have been published describing vitrealkinetics of GCV after intravitreal administration. In the doseranges studied so far, GCV was found to exhibit dose-independent elimination from the vitreous. Ashton et al.investigated the elimination kinetics of GCV from the pos-terior chamber of the eye and found that the retinal elimina-tion rate constants (vitreal elimination rate constantnormalized to the retinal surface area and volume of distrib-ution) to be very similar in rabbits and diseased humaneyes.13 GCV elimination was found to follow first-order elim-ination kinetics. The authors concluded that the rabbit animalmodel is a good predictor of GCV intravitreal kinetics in thediseased human eye and proposed that elimination of GCVfrom the vitreous chamber probably involves transport acrossthe retina. Hughes et al.14 developed a posterior chambermicrodialysis technique to study the vitreal disposition ofGCV and acyclovir (ACV) following intravitreal administra-tions in New Zealand albino rabbits and observed rapidvitreal clearance of both drugs. Vitreal elimination half-lifeof GCV was 2.62hr. Considering the short half-life observed,the authors proposed that an active process might be involvedin the clearance of GCV across the retina.14 However, inanother study, by Lopez-Cortes et al., passage of GCV fromthe vitreous to the retina was found to demonstrate dose-independent kinetics indicating a passive transretinal diffu-sion mechanism.15 A recent study published by Macha et al.16

reports that the vitreal half-life of GCV in New Zealandalbino rabbits is around 7hr, which is considerably higherthan that observed by Hughes et al. Moreover, GCV demon-strated dose-independent kinetics when administered indoses ranging from 12.5 to 50 mg. These results thus do notsupport the involvement of a carrier-mediated process in theretinal elimination of GCV. However, almost all reports pub-lished on vitreal kinetics of GCV propose transretinal elim-ination as the major clearance pathway.

Transretinal elimination of GCV could be either acrossthe retinal blood vessels or across the outer blood–retinalbarrier (i.e., RPE) into the choroidal circulation. A major dif-ference between the rabbit and human retina is that humanretina is richly supplied with retinal blood vessels, whereasrabbit retina is predominantly avascular.17 When GCV elimination from the vitreous chamber was normalized toretinal surface area, elimination kinetics in the two models(rabbits and humans) demonstrated good correlation.13 Thus,it appears that passage across the RPE, and not the retinalblood vessels, is the major elimination route for GCV afterintravitreal administration.

Although RPE appears to be the major pathway of GCVelimination, it is not clear, as yet, whether any carrier-

mediated process is involved. The objective of this study wasto delineate whether any transporter is involved in GCVuptake by the retinal cells, including RPE. HCMV is knownto reside and actively replicate in the cytoplasm of retinalcells. Thus, an understanding of the mechanism of GCVuptake from the vitreous humor by the retinal cells would aidin the design of future antiviral agents and prodrugs targetedagainst HCMV.

To elucidate the retinal permeation mechanism of GCV,uptake studies were carried out with isolated intact rabbitretina and a human RPE-derived cell line, ARPE-19. A 5-min ex vivo retinal uptake technique using excised intactrabbit retina was employed to elucidate the mechanism ofGCV uptake by the retinal cells. The rabbit model is wellaccepted and has widely been used to study ocular metabo-lism, distribution, and pharmacokinetics of drug moleculesincluding GCV.13,14,16 Isolated rabbit retinas have beendemonstrated to be robust and to remain viable for 2–3hrunder ex vivo conditions. Several reports have described theuse of a similar model to study the physiological action ofdrugs on the neural retina.18–23 Recently, we reported that the ex vivo model correlates well with the in vivo rabbitmodel for studying retinal permeation mechanism of drugmolecules.24

The ARPE-19 cell line was employed as the in vitromodel in this study to delineate the mechanism of GCVuptake by RPE cells (which form the outer blood–retinal barrier). This cell line has extensively been investi-gated and found to mimic the human RPE with respect tomorphology, enzyme expression, and polarization of carrierproteins.25–31

Nucleoside and nucleobase transporters have beenreported to be involved in the transport of GCV across cel-lular membranes. However, the permeation mechanism ofGCV in most tissues investigated so far involved passive dif-fusion and not any carrier-mediated transport process.32–34

Our broad objective was first to establish whether the isolatedrabbit retina and the ARPE-19 cell line expresses any nucleoside/nucleobase transporter and then investigatewhether GCV is a substrate for these transporters or anyother carrier-mediated transport process.

Materials and methods

Materials

ARPE-19 cells were purchased from American Type CultureCollection (Manassas, VA, USA) at passage number 21.[3H]Thymidine (26Ci/mmol) was procured from AmershamPharmacia Biotech (Piscataway, NJ, USA). [3H]GCV (9.5Ci/mmol) and [3H]adenine (14Ci/mmol) were procuredfrom Moravek Biochemicals (Brea, CA, USA). D-MEM/F-12 and 10% fetal bovine serum (FBS) were purchased fromGibco-BRL-Invitrogen (Grand Island, NY, USA). Cultureplates were obtained from Costar (Corning, NY, USA). Allother chemicals were purchased from Sigma Chemical

Cur

r E

ye R

es D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Cal

ifor

nia

Irvi

ne o

n 11

/07/

14Fo

r pe

rson

al u

se o

nly.

Ganciclovir uptake by rabbit retina and ARPE-19 cells 129

Company (St Louis, MO, USA) and used without furtherpurification.

Animals

Adult male New Zealand albino rabbits weighing 2–2.5kgwere supplied by Myrtle’s Rabbitry (Thompson Station, TN,USA). Experiments with rabbits conformed to the tenets ofthe Association for Research in Vision and Ophthalmology(ARVO) statement on the Use of Animals in Ophthalmic andVision Research.

Cell culture

ARPE-19 cells were used between passages 22 and 30.Culture media consisted of D-MEM/F-12 containing 10%fetal bovine serum, 15mM hydroxyl ethyl piperazine ethanesulfonicacid (HEPES), 29mM sodium bicarbonate, peni-cillin (100mg/ml), and streptomycin (100 mg/ml). Themedium was replaced every other day. Cells were maintainedat 37°C in a humidified atmosphere of 5% CO2 and 90% relative humidity. For uptake studies, cells were plated at a density of 500000 cells/well on 12-well culture plates and incubated at 37°C. This culture media was used for thefirst 5 days, and for subsequent days culture media contain-ing 1% FBS was employed.

Uptake experiments

Uptake studies

Uptake studies were conducted based on our earlier pub-lished method32 with slight modifications. Briefly, 28–30days postseeding, medium was removed and cells werewashed twice with Dulbecco’s phosphate-buffered saline(DPBS), pH 7.4, containing 130mM NaCl, 2.5mM KCl, 7.5mM Na2HPO4, 1.5mM KH2PO4, 1mM CaCl2, 0.5mMMgSO4 and 5mM glucose. Solutions were spiked with 0.5mCi/ml of radioactive compounds in the presence andabsence of various competing substrates (nucleosides,nucelobases, nucleoside analogs, probenecid, and pyruvicacid) to establish the expression of specific transporters.Uptake was conducted for 5min at 37°C and was initiated byincubating the cells with 2ml of the drug solution.

At the end of the incubation period, solution was aspiratedoff and cells were washed twice with 2ml of ice-cold stopsolution (0.52g/L HEPES and 15.64g/L KCl) to arrest cel-lular uptake. One milliliter of 0.3N NaOH containing 0.1%Triton-X solution was then added to each well and leftovernight to solubilize the cells. Five hundred microliters ofthat solution was then transferred to scintillation vials con-taining 5ml scintillation cocktail. Cellular radioactivity wasquantified using a scintillation counter (Model LS-6500;Beckman Coulter, Fullerton, CA, USA). Twenty microlitersof solution from each well was then sampled, and proteincontent was measured by the method of Bradford35 withbovine serum albumin as the standard (Bio-Rad protein

estimation kit, Hercules, CA, USA). Nonspecific bindingwas corrected by carrying out the uptake at 4°C.

pH dependence studies

Solution pH was varied from 5.0 to 7.4 for pH dependencestudies (pH adjusted with 2-(N-morpholino) ethanesulfonicacid; MES). In all cases, uptake studies were conductedaccording to the procedure described earlier.

Inhibition and sodium dependency studies

Studies investigating the effect of metabolic inhibitors(ouabain and sodium azide) and nitrobenzylthioinosine(NBT), a specific nucleoside transporter inhibitor, werecarried out by preincubating the cells for 30min with bufferalone (control) or in the presence of the inhibitors.

To determine the sodium dependence of the uptakeprocess, NaCl and Na2HPO4 in DPBS were replaced withcholine chloride and K2HPO4, respectively, in equimolarquantities.

Uptake studies using isolated intact rabbit retina

New Zealand male albino rabbits weighing 2–2.5kg(Myrtle’s Rabbitry) were anesthetized with ketamine HCl (35mg/kg) and xylazine (5mg/kg) administered intramuscu-larly, and then euthanized by an overdose of sodium pento-barbital administered through the marginal ear vein. Eyeswere enucleated immediately, and the posterior segments(retina/choroid/sclera) were carefully removed after dissect-ing along the corneal-scleral limbus and promptly washedwith DPBS, pH 7.4, to remove any traces of blood, vitreoushumor, and extraneous matter. It was then shaped in the formof a cup, using a glass well, with the retina facing inside. Exvivo uptake experiments were carried out at 25°C (highertemperatures occasionally resulted in retinal detachment) for5min with [3H]adenine and [3H]thymidine as model sub-strates for nucleoside and nucleobase transporters, respec-tively (earlier reports indicate the involvement of adenine andthymidine transporters in the translocation of GCV acrosscell membranes34), and [3H]GCV in the presence and absenceof nucleosides and nucleobases.

Solutions were prepared immediately prior to the initia-tion of an uptake study. Retinal uptake was initiated byadding 300ml radiolabeled drug solution (0.5 mCi/ml) to thetissue (vitreal side) in the absence and presence of compet-ing substrates. Uptake was terminated after 5min by aspirat-ing off the drug solution and washing the tissue with ice-coldstop solution (0.52g/L HEPES and 15.64g/L KCl). After twowashings, retina/choroid was carefully separated from thesclera and lysed overnight with 1ml 0.1% Triton-X solutionin 1N NaOH. Aliquots (500 ml) were then transferred to scintillation vials containing 5ml scintillation cocktail(Fisher Scientific, Fair Lawn, NJ, USA). Samples were ana-lyzed for radioactivity using a scintillation counter (Beckman

Cur

r E

ye R

es D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Cal

ifor

nia

Irvi

ne o

n 11

/07/

14Fo

r pe

rson

al u

se o

nly.

130 S. Majumdar et al.

Instruments Inc., Model LS-6500). Uptake data was normal-ized to the protein content of each well. Cell lysate protein content was measured by the method of Bradford usingbovine serum albumin as the standard (Bio-Rad protein estimation kit).

Statistical analysis

All experiments were conducted at least in quadruplicate,and results are expressed as mean ± standard deviation. Statistical analysis between two groups was carried out with Student’s t test. A difference between mean values was considered statistically significant if the p value was£0.05.

Results

Effect of nucleosides, nucleobases, and nucleosideanalogs on uptake of [3H]adenine by ARPE-19 cells

Uptake studies were carried out with 0.5 mCi/ml of[3H]adenine in the presence and absence of various nucleo-sides (adenosine, guanosine, uridine, inosine, and thymidine)and nucleobases (unlabeled adenine, guanine, hypoxanthine,thymine, and uracil) to examine the substrate specificity ofthe transporter (Table 1). All competing substrates wereadded at 1mM concentrations. Unlabeled adenine drasticallyinhibited uptake of [3H]adenine, reducing rate of uptake from 555.6 ± 14.5 fmols·min-1·mg-1 of protein to 63.9 ±1.6 fmols·min-1·mg-1 protein. Substantial inhibition was alsoachieved with the other nucleobases (guanine, uracil,

thymine, and hypoxanthine). Both purine and pyrimidinenucleosides exhibited significant inhibition of [3H]adenineuptake (Table 1).

Sodium, pH, and energy dependence of [3H]adenineuptake by ARPE-19 cells

Uptake of 0.5 mCi/ml of [3H]adenine was carried out in thepresence and absence of sodium to determine whether the process was sodium dependent. Absence of sodium in theuptake medium did not have any significant effect on[3H]adenine uptake, indicating that the process was sodiumindependent (Fig. 1). Change in medium pH (in the rangestested) did not demonstrate any effect on the uptake process.Inhibitors for Na+/K+-ATPase (ouabain 1mM) and metabolicinhibitors (sodium azide 1mM) were used to characterizefurther the transport system. Significant differences were notobserved in the uptake of [3H]adenine between control cellsand cells preincubated with sodium azide or ouabain, indi-cating that uptake was pH, sodium, and energy independent(Fig. 1).

Effect of nucleosides, nucleobases, and nucleosideanalogs on uptake of [3H]adenosine and [3H]thymidineby ARPE-19 cells

Uptake studies were carried out with 0.5 mCi/ml of[3H]adenosine and [3H]thymidine (Table 1) in the presenceand absence of various nucleosides (adenosine, guanosine,and thymidine) and nucleobases (adenine, guanine, thymine,and uracil). All competing substrates were added at 1mM

Table 1. Effect of nucleosides and nucleobases on the uptake of [3H]adenine (0.5 mCi/ml),[3H]adenosine (0.5 mCi/ml), and [3H]thymidine (0.5 mCi/ml) by ARPE-19 cells.

Inhibitors Uptake of Uptake of Uptake of[3H]adenine as a [3H]adenosine [3H]thymidine as

percentage of control as a percentage a percentage of control(%) of control (%)

(%)

Effect of nucleosidesThymidine 1 mM 54.4 (2.6)* 36.9 (5.8)* 6.5 (2.6)*Uridine 1 mM 49.9 (4.6)* –Adenosine 1 mM 51.5 (3.6)* 3.7 (0.2)* 29.6 (1.8)*Guanosine 1 mM 47.9 (1.2)* 49.7 (3.3)* 59.3 (12.5)*Inosine 1 mM 37.4 (4.3)* – –

Effect of nucleobasesGuanine 1 mM 63.6 (6.7)* 108.2 (11.5) 107.9 (7.6)Adenine 1 mM 5.5 (1.2)* 89.5 (10.8) 102.6 (4.8)Uracil 1 mM 70.1 (1.7)* 93.4 (10.4) 89.2 (10.4)Thymine 1 mM 75.2 (2.4)* 98.7 (7.7) 95.5 (6.9)Hypoxanthine 1 mM 51.3 (1.2)* – –

Data reported as mean ± SD (n = 4).* p < 0.05.

Cur

r E

ye R

es D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Cal

ifor

nia

Irvi

ne o

n 11

/07/

14Fo

r pe

rson

al u

se o

nly.

Ganciclovir uptake by rabbit retina and ARPE-19 cells 131

concentrations. Unlabeled adenosine and thymidine drasti-cally inhibited uptake of [3H]adenosine and [3H]thymidine,respectively. [3H]adenosine uptake was reduced to 44.1 ±2.2 fmols·min-1·mg-1 of protein from 1180.2 ± 16.5fmols·min-1·mg-1 of protein by unlabeled adenosine, whereasunlabeled thymidine reduced uptake of [3H]thymidine from175.7 ± 20.1 fmols·min-1·mg-1 of protein to 20.2 ± 3.5 fmols·min-1·mg-1 of protein. Both purine and pyrimidine nucleo-sides demonstrated substantial inhibition in the uptake of[3H]adenosine and [3H]thymidine (Table 1). However, inhi-bition of [3H]adenosine and [3H]thymidine uptake was notobserved in presence of any of the nucleobases.

Sodium pH and energy dependence of [3H]thymidineand [3H]adenosine uptake by ARPE-19 cells

Uptake of 0.5 mCi/ml of [3H]adenosine and [3H]thymidinewere carried out in the presence and absence of sodium todetermine whether the process was sodium dependent.Absence of sodium in the medium did not have any signifi-cant effect on the uptake of [3H]thymidine or [3H]adenosine,indicating that the uptake process was sodium independent(Fig. 1). Medium pH (in the range tested) also did not demon-strate any effect on the uptake process. Inhibitors for Na+/K+-ATPase (ouabain 1mM) and metabolic inhibitors (sodiumazide 1mM) were added to characterize further the transportsystem. Significant differences were not observed in uptakeof [3H]thymidine or [3H]adenosine between control cells andcells preincubated with sodium azide or ouabain, indicatingthat uptake was pH, sodium, and energy independent (Fig. 1).

Effect of NBT on uptake of [3H]adenine, [3H]thymidine,and [3H]adenosine by ARPE-19 cells

To confirm the expression of nucleoside transporters a spe-cific nucleoside transporter inhibitor, nitrobenzylthioinosine(NBT), was selected to examine its inhibitory effect on theuptake of 0.5 mCi/ml of [3H]adenine, [3H]thymidine, and[3H]adenosine by ARPE-19 cells. Though 100nM NBT dras-tically reduced [3H]thymidine and [3H]adenosine uptake by94% and 96%, respectively, it did not significantly alter[3H]adenine uptake by the ARPE-19 cells (Fig. 2).

[3H]Ganciclovir uptake by ARPE-19 cells

Uptake studies were carried out with 0.5mCi/ml of [3H]GCVin the presence and absence of various nucleosides (adeno-sine, guanosine, and thymidine), nucleobases (adenine,guanine and thymine), probenecid and pyruvic acid (organicanion transporter inhibitors), and unlabeled GCV. None ofthe compounds examined produced any inhibitory effect on[3H]GCV uptake (Table 2). Rate of uptake of [3H]GCV bythe ARPE-19 cells was observed to be 29 ± 3 fmols·min-1·mg-1 of protein.

Uptake of nucleosides, nucleobases, and GCV byisolated intact rabbit retina

Uptake studies were carried out by adding 300 ml of solutioncontaining 0.5 mCi/ml [3H]adenine or [3H]thymidine in thepresence and absence of 1mM unlabeled adenine, thymidine,

0

20

40

60

80

100

120

140Adenine

Adenosine

Thymidine

Upt

ake

as a

% o

f co

ntro

l

Control pH6.0

pH5.0

Ouabain 1mM

SodiumAzide 1mM

SodiumFree

Figure 1. Uptake of [3H]adenine, [3H]adenosine, and [3H]thymidine (0.5 mCi/ml) by ARPE-19 cells as a percentage of control. Uptake studieswere carried out at pH 7.4, pH 6.0, pH 5.0, and in the absence of sodium. Effect of preincubation with 1 mM ouabain and sodium azide priorto uptake was also studied (pH 7.4). Uptake at pH 7.4 served as the control for all the studies. Data reported as mean ± SD (n = 4).

Cur

r E

ye R

es D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Cal

ifor

nia

Irvi

ne o

n 11

/07/

14Fo

r pe

rson

al u

se o

nly.

132 S. Majumdar et al.

or guanosine (Fig. 3A). Unlabeled adenine drasticallyreduced uptake of [3H]adenine by the rabbit retina to 68.9 ±8 fmols·5min-1·mg-1 of protein from 147 ± 44.6 fmols·5min-1 ·mg-1 of protein. However no significant inhibition of

[3H]adenine uptake was observed in the presence of thymi guanosine, NBT, or in the absence of sodium (Fig. 3A).

Thymidine and guanosine reduced rabbit retinal uptake of[3H]thymidine from 12.8 ± 3.1 fmols·5min-1·mg-1 of proteinto 4.7 ± 0.1 fmols·5min-1·mg-1 of protein and 6.54 ±0.6 fmols·5min-1·mg-1 of protein, respectively. On the otherhand, adenine did not demonstrate any significant effect on[3H]thymidine uptake. Though 1 mM NBT did not signifi-cantly inhibit [3H]thymidine uptake, absence of sodium in themedium produced a 65% drop in the uptake (Fig. 3A).

Unlabeled GCV, adenine, or thymidine did not exhibit anyinhibitory effect on retinal uptake of 0.5 mCi/ml [3H]GCV(Fig. 3, B). [3H]GCV uptake remained fairly constant at 38.8 ± 5.3 fmols·5min-1·mg-1 of protein.

Discussion

Natural nucleosides and nucleoside analogs, such as GCV,have been demonstrated to be substrates of nucleoside/nucleobase transporters. Naturally occurring nucleosides,although hydrophilic, are efficiently transported across cellmembranes by nucleoside transport systems. Several nucle-oside analogs have been shown to be substrates of nucleo-side nucleobase transporters. ACV and GCV have been

0

20

40

60

80

100

120 Control

NBT 100nM

Thymidine Adenosine Adenine

Upt

ake

as a

% o

f co

ntro

l

* *

*P < .05.

Figure 2. Uptake of [3H]thymidine, [3H]adenosine, and [3H]adenine (0.5 mCi/ml) in the presence or absence (control) of 100 nM nitrobenzylthioinosine (NBT) by ARPE-19 cells. Data reported as mean ± SD (n = 4).

Table 2. Uptake of [3H]GCV (0.5 mCi/ml) by ARPE-19 cells in the presence or absence (control) of nucleosides, nucleobases,probenecid, and unlabeled GCV.*

Inhibitors Uptake of [3H]GCV SDas a percentage of control (%)

Adenosine 1 mM 88.7 10.4Guanosine 1 mM 89.1 7.9Thymidine 1 mM 86.5 8.2Guanine 1 mM 103.3 10.6Adenine 1 mM 96.5 11.4Thymine 1 mM 95.3 8.5Probenecid 1 mM 104.3 8.3Pyruvic acid 1 mM 106.8 7.2Unlabeled GCV 1 mM 102.1 7.5

GCV, ganciclovir.* Concentration of all competing substrates was 1 mM. Datareported as mean ± SD (n = 4).

Cur

r E

ye R

es D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Cal

ifor

nia

Irvi

ne o

n 11

/07/

14Fo

r pe

rson

al u

se o

nly.

Ganciclovir uptake by rabbit retina and ARPE-19 cells 133

known to be transported by the nucleobase and nucleosidetransporters in human erythrocytes and placenta.33,34,36,37

Further, a thymidine analog, 5-iodo-2¢-deoxyuridine (IDU),was shown to be a substrate of the Na+-dependent nucleosidetransporter expressed on the brush border membrane vesiclesfrom human kidney.38.39 However, ACV was not translocatedby the nucleoside transporter present on the rat jejunum andrabbit cornea.40,41 Recent reports indicate that besides thenucleoside and nucleobase transport systems, nucleosideanalogs are also translocated by organic anion transporters.42,43

Nucleoside transporters are classified into two categoriesbased on their sodium dependency39,44–49: sodium-

independent equilibrative nucleoside transporters (facilitateddiffusion) and concentrative (sodium-dependent) nucleosidetransporters. Equilibrative nucleoside transport systems arefurther classified into two subtypes based on their sensitivityto nitrobenzylthioinosine (NBT). Equilibrative sensitivetype, also referred to as “es,” is inhibited by nanomolar con-centrations of NBT (ki 0.1–1nM), whereas the equilibrativeinsensitive type, commonly referred to as “ei,” is inhibitedby micromolar concentrations of NBT. Both the “es” and “ei”types of equilibrative transporters exhibit broad substratespecificities for the purine and pyrimidine nucleosides.

Sodium-dependent concentrative systems comprise 5 sub-divisions, N1–N5. This class of nucleoside transporters is

Upt

ake

as a

% o

f co

ntro

l

Control Adenine Thymidine Guanosine Sodium Free

NBT 0

20

40

60

80

100

120

140 Adenine

ThymidineA

*

*

**

*P < .05.

Figure 3. Rabbit retinal uptake of (A) [3H]adenine and [3H]thymidine and (B) [3H]GCV in the presence or absence (control) of sodium,adenine (1 mM), thymidine (1 mM), guanosine (1 mM), 1 mM nitrobenzylthioinosine (NBT), or unlabeled GCV (1 mM). Data reported as mean± SD (n = 4).

Thymidine Adenine Control GCV 0

20

40

60

80

100

120

140

160

Upt

ake

of [

3 H]G

CV

as

a %

of

co

ntro

l

B

Cur

r E

ye R

es D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Cal

ifor

nia

Irvi

ne o

n 11

/07/

14Fo

r pe

rson

al u

se o

nly.

134 S. Majumdar et al.

capable of translocating nucleosides against a concentrationgradient. The subclasses are defined based on their substratespecificity. The N1 type is selective for the purine nucleo-sides and uridine, N2 is selective for the pyrimidine nucleo-sides and adenosine, N3 has broad specificity for both purineand pyrimidine nucleosides, whereas N4 is pyrimidine selec-tive but also transports adenosine and guanosine but notinosine. The N5 system is NBT sensitive and prefers guano-sine as the substrate.39

On the other hand, various types of nucleobase trans-porters have not been well characterized with respect to their substrate specificity or factors affecting transporteractivity.48,50

Isolated rabbit retinal uptake studies revealed that the per-meation of [3H]adenine was significantly reduced in the pres-ence of unlabeled adenine. However, thymidine or adenosinedid not demonstrate any inhibitory effect on the uptake of[3H]adenine by the rabbit retina indicating that the nucleo-base transporter was not involved in the translocation ofnucleosides. Moreover, uptake was not inhibited by NBT orthe absence of sodium.

[3H]thymidine uptake by the rabbit retina was drasticallyreduced (64% decrease) in the presence of unlabeled thymi-dine. Both purine and pyrimidine nucleosides and deletion ofsodium from the uptake buffer markedly decreased uptake of[3H]thymidine. However, neither adenine nor NBT exhibitedany inhibitory effect on retinal uptake of [3H]thymidine. Thedata thus suggest the expression of a sodium-dependentnucleoside transporter (expression of equilibrative nucleo-side transporters cannot be ruled out from this study and wasbeyond the scope of this research) and an adenine nucleobasetransporter on the rabbit retina. However, uptake of [3H]GCVwas not affected in the presence of unlabeled GCV, adenine,or thymidine, although GCV was previously observed to bea substrate of adenine and thymidine transporters expressedby placental cells.34 Our results thus demonstrate that rabbitretinal uptake of GCV does not involve any carrier-mediatedprocess.

ARPE-19 cell line was also found to express nucleosideand nucleobase transporters. [3H]Adenine uptake was inhibited in the presence of unlabeled adenine. Both nucleo-sides and nucleobases were observed to diminish uptake of[3H]adenine. Uptake was sodium, pH, and energy indepen-dent and was not inhibited in the presence of NBT. Uptakeof [3H]thymidine and [3H]adenosine was decreased in thepresence of both purine and pyrimidine nucleosides. Pres-ence or absence of sodium and metabolic inhibitors did notexhibit any effect on [3H]thymidine or [3H]adenosine uptakeby the ARPE-19 cells. However, presence of nanomolar con-centrations of NBT drastically reduced the translocation ofboth nucleosides. The nucleoside transporter expressed bythe ARPE-19 cells is thus probably of the equilibrative sen-sitive type.

Interestingly, nucleobases did not inhibit uptake of[3H]adenosine or [3H]thymidine, although nucleosides wereobserved to inhibit uptake of [3H]adenine. Moreover,

although NBT inhibited uptake of [3H]thymidine and[3H]adenosine, NBT did not exhibit any inhibitory effect on[3H]adenine uptake. Our results thus suggest that uptake ofnucleobases and nucleosides by the ARPE-19 cells involveseparate nucleobase and nucleoside transporter systems.Inhibition of [3H]adenine by the nucleosides indicates thatnucleosides can probably bind to the nucleobase transportersbut are not translocated by it.

Uptake of [3H]GCV by the ARPE-19 cells was notaffected by the presence of any of the nucleobases or nucleosides examined. Further, probenecid and pyruvic acid did not inhibit [3H]GCV uptake by ARPE-19, thus rulingout the involvement of any organic anion transporter. Unlabeled GCV also did not demonstrate any inhibitoryeffect on [3H]GCV uptake, thus indicating that carrier-mediated processes were probably not involved in GCVuptake by ARPE-19 cells.

Considering GCV’s mechanism of action, it is evident that for GCV to be effective it has to be taken up into theintracellular compartment of the CMV-infected retinal cells.Although intravitreal injections of GCV may provide suffi-cient vitreal concentrations, efficacy of GCV will be gov-erned by its concentration in the CMV-infected retinal cellcytoplasm. Because GCV, like other nucleobases, nucleo-sides, and their analogs, is a relatively hydrophilic molecule,transcellular permeation is limited, and its penetration into cells depends on specific transport proteins expressed onthe cellular membrane. Our results indicate that permeationof GCV into the retinal cells is mainly by passive diffusionalmechanism. Thus, intracellular concentrations of GCV inside the retinal cell cytoplasm will be limited, althoughnucleoside and nucleobase transporters are expressed by theretinal cells. It was observed that uptake of GCV in bothmodels was significantly lower than that of [3H]adenine,whereas it was lower than that of [3H]adenosine and[3H]thymidine in the ARPE-19 cells. In the isolated retina ex vivo studies, the total uptake of GCV was observed to begreater than that of [3H]thymidine. A possible explanationcould be the involvement of sodium-dependent thymidinetransporters, which generally display high affinity and lowcapacity.

Transporter targeted prodrug derivatization approach maybe a viable strategy to increase intracellular GCV concen-trations in retinal cells. Several transporters such as aminoacid, peptide, and monocarboxylate, besides nucleoside andnucleobase transporters, are known to be expressed on retinalcells, including the RPE. Rational targeted prodrug design,taking into consideration capacities of various transporters,may significantly improve the therapeutic efficacy of GCVand other nucleoside analogs.

In conclusion, this study reports for the first time themechanism of GCV uptake by the rabbit retina and thehuman retinal pigmented epithelium cell line ARPE-19.Although distinct nucleoside and nucleobase transporters areexpressed both on the rabbit retina as well as on the ARPE-19 cells, GCV uptake does not seem to involve any carrier-

Cur

r E

ye R

es D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Cal

ifor

nia

Irvi

ne o

n 11

/07/

14Fo

r pe

rson

al u

se o

nly.

Ganciclovir uptake by rabbit retina and ARPE-19 cells 135

mediated process but rather seems to take place by passivediffusion in both models.

Acknowledgment

This work was supported by National Eye Institute grants2R01 EY 09171-09 and 2R01 EY 10659-08.

References

1. Freeman WR, Lerner CW, Mines JA, Lash RS, Nadel AJ,Starr MB, Tapper ML. A prospective study of the ophthal-mologic findings in the acquired immune deficiency syn-drome. Am J Ophthalmol. 1984;97:133–142.

2. Burd EM, Pulido JS, Puro DG, O’Brien WJ. Replication ofhuman cytomegalovirus in human retinal glial cells. InvestOphthalmol Vis Sci. 1996;37:1957–1966.

3. Bodaghi B, Michelson S. Cytomegalovirus: Virologicalfacts for clinicians. Ocul Immunol Inflamm. 1999;7:133–137.

4. Andrei G, Snoeck R, Schols D, Goubau P, Desmyter J, DeClercq E. Comparative activity of selected antiviral compounds against clinical isolates of humancytomegalovirus. Eur J Clin Microbiol Infect Dis. 1991;10:1026–1033.

5. Shigeta S, Konno K, Baba M, Yokota T, De Clercq E. Com-parative inhibitory effects of nucleoside analogues on dif-ferent clinical isolates of human cytomegalovirus in vitro.J Infect Dis. 1991;163:270–275.

6. Mar EC, Chiou JF, Cheng YC, Huang ES. Inhibition of cellular DNA polymerase alpha and humancytomegalovirus-induced DNA polymerase by the triphosphates of 9-(2-hydroxyethoxymethyl)guanine and 9-(1,3-dihydroxy-2-propoxymethyl)guanine. J Virol.1985;53:776–780.

7. Arevalo JF, Gonzalez C, Capparelli EV, Kirsch LS, GarciaRF, Quiceno JI, Connor JD, Gambertoglio J, Bergeron-LynnG, Freeman WR. Intravitreous and plasma concentrationsof ganciclovir and foscarnet after intravenous therapy inpatients with AIDS and cytomegalovirus retinitis. J InfectDis. 1995;172:951–956.

8. Kuppermann BD, Quiceno JI, Flores-Aguilar M, ConnorJD, Capparelli EV, Sherwood CH, Freeman WR. Intravit-real ganciclovir concentration after intravenous administra-tion in AIDS patients with cytomegalovirus retinitis:Implications for therapy. J Infect Dis. 1993;168:1506–1509.

9. Cunha-Vaz JG. The blood-retinal barriers. Doc Ophthal-mol. 1976;41:287–327.

10. Cadman J. Improved ganciclovir regimens. GMHC TreatIssues. 1997;12:28–29.

11. Cadman J. Ganciclovir implants: One year later. GMHCTreat Issues. 1997;11:3–6.

12. Vazquez E. Saving sight. Posit Aware. 1997;8:22–23.13. Ashton P, Brown JD, Pearson PA, Blandford DL, Smith TJ,

Anand R, Nightingale SD, Sanborn GE. Intravitreal

ganciclovir pharmacokinetics in rabbits and man. J OculPharmacol. 1992;8:343–347.

14. Hughes PM, Krishnamoorthy R, Mitra AK. Vitreous dis-position of two acycloguanosine antivirals in the albino andpigmented rabbit models: A novel ocular microdialysistechnique. J Ocul Pharmacol Ther. 1996;12:209–224.

15. Lopez-Cortes LF, Pastor-Ramos MT, Ruiz-Valderas R,Cordero E, Uceda-Montanes A, Claro-Cala CM, Lucero-Munoz MJ. Intravitreal pharmacokinetics and retinal con-centrations of ganciclovir and foscarnet after intravitrealadministration in rabbits. Invest Ophthalmol Vis Sci.2001;42:1024–1028.

16. Macha S, Mitra AK. Ocular disposition of ganciclovir and its monoester prodrugs following intravitreal admin-istration using microdialysis. Drug Metab Dispos.2002;30:670–675.

17. Yu DY, Cringle SJ. Oxygen distribution and consumptionwithin the retina in vascularised and avascular retinas andin animal models of retinal disease. Prog Retin Eye Res.2001;20:175–208.

18. Frambach DA, Valentine JL, Weiter JJ. Topical phenyle-phrine for mydriasis affects rabbit retinal pigment epithe-lial transport. Invest Ophthalmol Vis Sci. 1989;30:343–344.

19. Frambach DA, Valentine JL, Weiter JJ. Initial observationsof rabbit retinal pigment epithelium-choroid-sclera prepa-rations. Invest Ophthalmol Vis Sci. 1988;29:814–817.

20. Koyano S, Araie M, Eguchi S. Movement of fluorescein andits glucuronide across retinal pigment epithelium-choroid.Invest Ophthalmol Vis Sci. 1993;34:531–538.

21. Jensen RJ, Rizzo JF, 3rd, Ziv OR, Grumet A, Wyatt J.Thresholds for activation of rabbit retinal ganglion cellswith an ultrafine, extracellular microelectrode. Invest Ophthalmol Vis Sci. 2003;44:3533–3543.

22. Neal M, Cunningham J, Lever I, Pezet S, Malcangio M.Mechanism by which brain-derived neurotrophic factorincreases dopamine release from the rabbit retina. InvestOphthalmol Vis Sci. 2003;44:791–798.

23. Baptiste DC, Hartwick AT, Jollimore CA, Baldridge WH,Chauhan BC, Tremblay F, Kelly ME. Comparison of theneuroprotective effects of adrenoceptor drugs in retinal cellculture and intact retina. Invest Ophthalmol Vis Sci.2002;43:2666–2676.

24. Majumdar S, Macha S, Nashed Y, Mitra A. Expression ofPeptide Transporters on the Rabbit Retina: A Strategy toImprove Retinal Delivery of Ganciclovir. Letters in DrugDesign & Discovery. 2004;1:73–77.

25. Han YH, Sweet DH, Hu DN, Pritchard JB. Characterizationof a novel cationic drug transporter in human retinalpigment epithelial cells. J Pharmacol Exp Ther.2001;296:450–457.

26. Dunn KC, Aotaki-Keen AE, Putkey FR, Hjelmeland LM.ARPE-19, a human retinal pigment epithelial cell line withdifferentiated properties. Exp Eye Res. 1996;62:155–169.

27. Aukunuru JV, Sunkara G, Bandi N, Thoreson WB, Kompella UB. Expression of multidrug resistance-associ-ated protein (MRP) in human retinal pigment epithelial

Cur

r E

ye R

es D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Cal

ifor

nia

Irvi

ne o

n 11

/07/

14Fo

r pe

rson

al u

se o

nly.

136 S. Majumdar et al.

cells and its interaction with BAPSG, a novel aldose reduc-tase inhibitor. Pharm Res. 2001;18:565–572.

28. Bridges CC, Kekuda R, Wang H, Prasad PD, Mehta P,Huang W, Smith SB, Ganapathy V. Structure, function, andregulation of human cystine/glutamate transporter in retinalpigment epithelial cells. Invest Ophthalmol Vis Sci.2001;42:47–54.

29. Bridges CC, El-Sherbeny A, Ola MS, Ganapathy V, SmithSB. Transcellular transfer of folate across the retinalpigment epithelium. Curr Eye Res. 2002;24:129–138.

30. Rajan PD, Kekuda R, Chancy CD, Huang W, Ganapathy V,Smith SB. Expression of the extraneuronal monoaminetransporter in RPE and neural retina. Curr Eye Res.2000;20:195–204.

31. Philp NJ, Wang D, Yoon H, Hjelmeland LM. Polarizedexpression of monocarboxylate transporters in humanretinal pigment epithelium and ARPE-19 cells. Invest Ophthalmol Vis Sci. 2003;44:1716–1721.

32. Majumdar S, Tirucherai GS, Pal D, Mitra AK. Functionaldifferences in nucleoside and nucleobase transportersexpressed on the rabbit corneal epithelial cell line (SIRC) and isolated rabbit cornea. AAPS Pharm Sci.2003;5:E15.

33. Mahony WB, Domin BA, Zimmerman TP. Ganciclovir per-meation of the human erythrocyte membrane. BiochemPharmacol. 1991;41:263–271.

34. Henderson GI, Hu ZQ, Yang Y, Perez TB, Devi BG, FrostoTA, Schenker S. Ganciclovir transfer by human placentaand its effects on rat fetal cells. Am J Med Sci.1993;306:151–156.

35. Bradford MM. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem.1976;72:248–254.

36. Mahony WB, Domin BA, McConnell RT, Zimmerman TP.Acyclovir transport into human erythrocytes. J Biol Chem.1988;263:9285–9291.

37. Henderson GI, Hu ZQ, Johnson RF, Perez AB, Yang Y,Schenker S. Acyclovir transport by the human placenta. J Lab Clin Med. 1992;120:885–892.

38. Gutierrez MM, Giacomini KM. Substrate selectivity,potential sensitivity and stoichiometry of Na(+)-nucleosidetransport in brush border membrane vesicles from humankidney. Biochim Biophys Acta. 1993;1149:202–208.

39. Balimane PV, Sinko PJ. Involvement of multiple trans-porters in the oral absorption of nucleoside analogues. AdvDrug Deliv Rev. 1999;39:183–209.

40. Meadows KC, Dressman JB. Mechanism of acycloviruptake in rat jejunum. Pharm Res. 1990;7:299–303.

41. Majumdar S, Gunda S, Mitra A. Functional expression ofa sodium dependent nucleoside transporter on rabbitcornea: Role in corneal permeation of acyclovir and idoxuridine. Curr Eye Res. 2003;26:175–183.

42. Takeda M, Khamdang S, Narikawa S, Kimura H,Kobayashi Y, Yamamoto T, Cha SH, Sekine T, Endou H.Human organic anion transporters and human organiccation transporters mediate renal antiviral transport. JPharmacol Exp Ther. 2002;300:918–924.

43. Hasegawa M, Kusuhara H, Endou H, Sugiyama Y. Con-tribution of organic anion transporters to the renal uptakeof anionic compounds and nucleoside derivatives in rat. J Pharmacol Exp Ther. 2003;305:1087–1097.

44. Hosoya K, Horibe Y, Kim KJ, Lee VH. Nucleoside trans-port mechanisms in the pigmented rabbit conjunctiva.Invest Ophthalmol Vis Sci. 1998;39:372–377.

45. Griffith DA, Jarvis SM. Nucleoside and nucleobase trans-port systems of mammalian cells. Biochim Biophys Acta.1996;1286:153–181.

46. Baldwin SA, Mackey JR, Cass CE, Young JD. Nucleosidetransporters: molecular biology and implications for thera-peutic development. Mol Med Today. 1999;5:216–224.

47. Kiss A, Farah K, Kim J, Garriock RJ, Drysdale TA,Hammond JR. Molecular cloning and functional character-ization of inhibitor-sensitive (mENT1) and inhibitor-resistant (mENT2) equilibrative nucleoside transportersfrom mouse brain. Biochem J. 2000;352 Pt 2:363–372.

48. Kraupp M, Marz R. Nucleobase and nucleoside transportin mammalian cells. Wien Klin Wochenschr. 1995;107:677–680.

49. Ward JL, Sherali A, Mo ZP, Tse CM. Kinetic and pharma-cological properties of cloned human equilibrative nucleo-side transporters, ENT1 and ENT2, stably expressed innucleoside transporter-deficient PK15 cells. Ent2 exhibitsa low affinity for guanosine and cytidine but a high affin-ity for inosine. J Biol Chem. 2000;275:8375–8381.

50. de Koning H, Diallinas G. Nucleobase transporters(review). Mol Membr Biol. 2000;17:75–94.

Cur

r E

ye R

es D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Cal

ifor

nia

Irvi

ne o

n 11

/07/

14Fo

r pe

rson

al u

se o

nly.