Exp. Eye Res. (1997) 65, 645–652

Induction of Human Thioredoxin in Cultured Human Retinal

Pigment Epithelial Cells through Cyclic AMP-dependent Pathway;

Involvement in the Cytoprotective Activity of Prostaglandin E1

MIHO YAMAMOTOa,b, NORIHITO SATOb, HISAO TAJIMAc, KEIZO FURUKEb,

AKIHIRO OHIRAd, YOSHIHITO HONDAa JUNJI YODOIb*

a Department of Ophthalmology, Faculty of Medicine, and b Department of Biological Responses, Institute

for Virus Research, Kyoto University, Kyoto, Japan, c Minase Research Institute, Ono Pharmaceutical Co.

Ltd., Osaka, Japan and d Department of Ophthalmology, School of Medicine, Nagasaki University,

Nagasaki, Japan

(Received Cleveland 11 December 1996 and accepted in revised form 23 June 1997)

Human thioredoxin is one of the oxidative stress-inducible proteins and has a protective function againstoxidant-induced injury. To evaluate the possible involvement of thioredoxin in the cytoprotectivefunction of prostaglandin E

", we analysed the effect of prostaglandin E

"on cellular injury by hydrogen

peroxide and intracellular thioredoxin induction. Cellular survival of human retinal pigment epithelialcell line, established from normal retinal pigment epithelial cells, following exposure to hydrogen peroxidewas markedly improved by pretreatment of 1 µ prostaglandin E

". Thioredoxin expression was

augmented in a dose-dependent manner when retinal pigment epithelial cells were pretreated with10 n–1 µ prostaglandin E

"1 hr before the exposure to hydrogen peroxide. Intracellular cyclic AMP

level was elevated by Prostaglandin E"when the cells were simultaneously exposed to hydrogen peroxide.

Forskolin, an activator of adenylate cyclase, and dibutylyl cAMP, a cyclic AMP analog, could also inducethioredoxin and extend survival of retinal pigment epithelial cells. On the other hand, thioredoxininduction and cellular protection by prostaglandin E

"was blocked by Rp diastereoisomer of cyclic

adenosine 3«, 5«, monophosphorothioate, a competitive inhibitor of cyclic AMP dependent protein kinase.Thioredoxin induction was augmented significantly by pretreatment with prostaglandin I

#, a stimulator

of cyclic AMP dependent signal pathway, while treatment with prostaglandin F#α, a stimulator of inositol

phosphate-dependent signal pathway, failed to enhance thioredoxin. These findings indicate thatprostaglandin E

"has a cytoprotective activity against oxidative injury, partly through thioredoxin

induction via cyclic AMP dependent pathway. # 1997 Academic Press LimitedKey words : prostaglandin E

"; human thioredoxin; oxidative stress ; cyclic AMP; retinal pigment

epithelial cells ; cytoprotection; hydrogen peroxide.

1. Introduction

Adult T cell leukemia derived factor was originally

identified as an interleukin-2 receptor α-chain inducer

abundant in the culture supernatant of a HTLV-1-

transformed T cell line (Yodoi and Tursz, 1991; Yodoi

and Uchiyama, 1992). Subsequent studies revealed

that adult T cell leukemia derived factor is identical to

human thioredoxin (TRX) and that TRX is a multi-

functional redox regulatory protein. The expression of

TRX is augmented in various malignant or activated

cells and in cells under a variety of stresses, such as

ultraviolet or X-ray irradiation and viral infection

(Tagaya et al., 1988; Fujii et al., 1991; Kusama et al.,

1991; Iwai et al., 1992; Nakamura et al., 1992;

Wakita et al., 1992). It has been shown that TRX

scavenges some reactive oxygen intermediates (ROI)

(Mitsui, Hirakawa and Yodoi, 1992) and protects cells

against oxidant-mediated injury (Matsuda et al.,

1991; Gauntt et al., 1994; Hori et al., 1994;

Nakamura et al., 1994; Ohira et al., 1994).

* Address for correspondence and requests for offprints : JunjiYodoi, Department of Biological Responses, Institute for VirusResearch, Kyoto University, Sakyo-ku, 606-01, Kyoto, Japan.

Prostaglandins have a variety of biological activities

(Elkeles et al., 1969; Gorman, 1978). Prostaglandin E"

(PGE") has been shown to be cytoprotective against

gastric mucosal damage by ethanol for which ROI

may be responsible (Robert, 1979; Terano et al.,

1986; Tarnawski et al., 1988; Cherner et al., 1989).

However, the mechanisms of this cytoprotective action

are still unclear. In a separate study, we showed in a

rat model of retinal ischemia}reperfusion that OP-

1206 α-CD, an oral PGE"analogue (Ohno, Morikawa

and Hirata, 1978; Adaikan and Karim, 1981), was

effective to attenuate retinal tissue damage and that

TRX was highly expressed in retinal pigment epithelial

(RPE) cells (Yamamoto et al., 1997). In the current

study, we have examined the possible involvement of

TRX in cytoprotective action of PGE"against RPE cell

injury. RPE cells were treated with hydrogen peroxide

(H#O#) in an in vitro system to cause retinal oxidative

injury and the cytoprotective activity of PGE"

and its

effect on TRX expression were evaluated. Furthermore,

the mechanisms whereby PGE"

induces TRX was

presented by analyses of signaling pathways involving

cyclic AMP (cAMP).

0014–4835}97}110645­08 $25.00}0}ey970370 # 1997 Academic Press Limited

646 M. YAMAMOTO ET AL.

2. Materials and Methods

Materials

H#O#(30%) was obtained from Wako Pure Chemical

Industries, Ltd. (Osaka, Japan). Ham’s F-12 medium

was purchased from ICN K&K Laboratories Inc.

(Plainview, NY, U.S.A.). Fetal calf serum (FCS) was

obtained from Cell Culture Laboratory (Cleveland, OH,

U.S.A.), while skim milk was purchased from Difco

(Detroit, MI, U.S.A.). Bovine serum albumin, trichloro-

acetic acid and Commasie brilliant blue were pur-

chased from Nakalai Tesque (Kyoto, Japan). Protein

assay dye reagent concentrate was from Bio-Rad

Laboratories (Richmond, CA, U.S.A.). Mouse anti-

human ADF mAb was obtained from Fuji Lebio

(Tokyo, Japan). Biotinylated anti-mouse IgG, avidin-

alkalinephosphatase complex were purchased from

Vector Laboratory (Buringame, CA, U.S.A.). PGE",

prostaglandin I#

(PGI#) and prostaglandin F

#α (PGF

#α)

was kindly provided from Ono Pharmaceutical Co. Ltd.

(Osaka, Japan). Rp diastereoisomer of cyclic adenosine

3«, 5« monophosphorothioate (Rp-cAMPS) was pur-

chased from BIO-LOG (La Jolla, CA, U.S.A.). The

following reagents were purchased from Sigma

Chemical Co. (St Louis, MO, U.S.A.) : 3-(4,5-Dimethyl-

2-thiazolyl)-2,5-diphenyl-2H tetrazolium bromide

(MTT), Nonidet P-40 (NP-40), phenylmethylsulfonyl

fluoride (PMSF), aprotinin, 2-mercaptoethanol (2-ME),

forskolin (FK), and dibutylyl cAMP (dBcAMP), 1-

methyl-3-iso-butylxanthine (IBMX).

Cells and Cell Culture

The RPE cell line (K-1034) was established from a

26-year old human eye which was enucleated because

of an orbital angioma. This cell line retained many

original morphological characteristics, despite some

changes in chromosomal count and pigmentation

(Kigasawa et al., 1994). The cells were used at the

70–80th passage level in this study. The cells at this

passage level were observed to actively phagocytize

latex beads (1 µm in diameter), which revealed that

the cells preserved the characteristic function of RPE.

The cells were cultured to subconfluency in Ham’s F-

12 medium supplemented with 10% FCS at 37°C and

5% CO#

in humidified air (Gauntt et al., 1994).

MTT Assay

Cellular viability under the indicated conditions was

evaluated by MTT assay (Mosmann, 1983). This

method is useful to measure cytotoxicity, proliferation

or activation, because tetrazolium ring of MTT is

cleaved in active mitochondria and this reaction occurs

only in viable cells. Briefly, 1¬10& cells}well were

plated on a 96-microwell plate (Corning). After over-

night incubation, cells were pretreated with various

concentration of PGE"

or 1 µ FK. After the 1 hr

pretreatment, various doses of H#O#

(0, 10, 100, or

200 µ in the final dose) were applied into each well

(total volume was 100 µl a well). After 48 hr incuba-

tion, 10 µl of MTT solution (5 mg}ml MTT in PBS) was

added into each well and incubated at 37°C and 5%

CO#

in air for 4 hr. Finally 100 µl of 0±04 N HCl-iso-

propanol was added and mixed thoroughly to dissolve

MTT-formazan salt. Coloration was measured with a

scanning multiwell spectrophotometer (MICROPLATE

READER model 3550 Bio-Rad, Richmond, CA) with a

test wavelength of 595 nm and a reference wave-

length of 650 nm. Each assay was carried out in

triplicate.

Evaluation of Intracellular TRX Induction by PGE"

and}or H#O#

Intracellular TRX expression was determined by

western blot analysis. Cells (1¬10&}dish) were cul-

tured in FCS-free Ham’s F-12 medium overnight. After

pretreatment with PGE"for 1 hr, cells were incubated

in the presence of 10 µ of H#O#. After 24 hr in

culture, cells were harvested from dishes with 0±25%

trypsin and immersed in a solubilizing buffer (0±5%

NP-40, 10 m Tris–HCl, 150 m NaCl, 1 m PMSF,

0±11 IU ml−" Aprotinin, 0±02% NaN$, in nanopure

water), and centrifuged. The whole amount of protein

in each lysate was quantified by the DC protein assay

(Bio Rad, Richmond, CA, U.S.A.). Cell lysates were

electrophoresed on 15% SDS polyacrylamide gel under

reducing conditions. After electrotransfer to an

Immobilon polyvinyliden difluoride membrane

(Millipore, Bedford, MA, U.S.A.) under a constant

current of 200 mA, the membrane was blocked with

F. 1. Effect of PGE"on the toxicity of H

#O#. The toxicity

of H#O#was determined by %OD reduction calculated as fol-

lows: %OD reduction¯ (T®S)¬100}(T®B), S, absorbancein the presence of H

#O#; T, absorbance in the absence of

H#O#; B, absorbance of medium alone. Mean of triplicate

culture is adopted for each factor. Each bar represents themean³.. of three separate experiments. Values indicated(*P!0±01, **P!0±005) are significantly different from thegroup of (PGE

"; 0 ) in each concentration of H

#O#. PGE

"conc () : +, 0 ; *, 1¬10−'.

INDUCTION OF TRX BY PGE1 IN RPE CELLS 647

(A)

(B)

(C)

F. 2. (A) Effect of H#O#on the expression of TRX in RPE

cells. Amount of TRX was compared by the relative intensityof each immunostaining band in Western blot (control¯100). Each cell lysate was applied to contain 3 µg protein}lane. Results shown are mean³.. of three differentsamples. Values indicated (*P!0±05, **P!0±01) are sig-nificantly different from PGE

"(®) group in each concentra-

tion. D, PGE"(­) ; *, PGE

"(®). (B) Western blot analysis of

TRX using monoclonal anti-TRX antibody. Each lane repre-sents the lysate of RPE cells incubated with PGE

"(n) and

H#O#(µ) for 24 hr as follows: cells with no treatment (Lane

1), PGE"; 100 n, (Lane 2), H

#O#; 10 µ (Lane 3), and

PGE"; 100 n, H

#O#; 10 µ (Lane 4). Each cell lysate was

applied to contain 5 µg protein}lane. (C) Effect of PGE"

onthe expression of TRX in RPE cells. Amount of TRX wascompared by the relative intensity of each immunostainingband western blot (control¯100). Each cell lysate wasapplied to contain 3 µg protein}lane. Results shown are themean³.. of three different samples. Values indicated(*P!0±05, **P!0±01) are significantly different from thegroup of [PGE

"; 0 (), H

#O#

(®)]. *, H#O#

(­) ; V, H#O#

(®).

7% skim milk and 3% BSA and then was incubated

with 0±35 µg}ml of mouse anti-human TRX mono-

clonal antibody. As a control, the same concentration

of non-immune mouse IgG was used instead of the

primary antibody. Subsequently, the membranes were

incubated with a horseradish peroxidase-conjugated

anti-mouse IgG antibody and developed by HRP-ECL

methodology (Amersham, Buckinghamshire, U.K.).

The optical density of each band was measured using

MICRO COMPUTER IMAGING DEVICE (Imaging Res.

Inc., Ontario, Canada) and expressed as the relative

intensity compared with that of the control.

Measurement of Intracellular cAMP Concentration in

RPE Cells

Cells (1¬10&}dish) were treated with or without

PGE"for 1 hr in FCS-free Ham’s F-12 medium contain-

ing 0±1 m IBMX as described above, and then were

incubated with H#O#at 37°C for 10 min. The incuba-

tion time with 1 µ FK or 10 µ dBcAMP was also

10 min. The medium was aspirated and the reaction

was stopped with 2 ml of 5% (w}v) trichloroacetic

acid. The supernatants were assessed for the concen-

tration of cAMP with BIOTRAK enzyme-immunoassay

kit (Amersham, Buckinghamshire, U.K.).

Effect of FK and dBcAMP on TRX Induction

Cells (1¬10&}dish) were cultured in FCS-free Ham’s

F-12 medium overnight and were treated with each

concentration of either FK or dBcAMP. After 24 hr

incubation, cells were harvested and western blot

analysis was performed.

Effect of Rp-cAMPS on Inhibition of TRX Induction by

PGE"

Cells (1¬10&}dish) were cultured with FCS-free

Ham’s F-12 medium overnight and were treated with

each concentration of Rp-cAMPS for 1 hr. Next, the

cells were treated with PGE"

(1 µ) and after 1 hr

H#O#

was added. Cells were harvested after 24 hr

incubation and TRX induction was detected by

western blot analysis.

Effect of Prostaglandin I#

(PGI#) and Prostaglandin F

#α

(PGF#α) on TRX Induction

Cells (1¬10&}dish) were cultured with FCS-free

Ham’s F-12 medium overnight and were treated with

each concentration of either PGI#or PGF

#α for 1 hr as

in the previous experiment. After 24 hr incubation

with H#O#

at 37°C, cells were harvested for western

blot analysis.

Statistical Analysis

Values are given as the mean³... The Student’s

t-test was used for statistical evaluation. The

significance of differences of P!0±05 was considered

statistically significant for each analysis.

648 M. YAMAMOTO ET AL.

3. Results

Protective Effect of PGE"

on RPE Cells Against Injury

Induced by H#O#

We analysed the cytoprotective effect of PGE"

to

reduce cellular damage induced by oxidative insult. A

preliminary study showed that 1-hr pretreatment with

1 µ PGE"

was optimal for the induction of cyto-

protective activity. Thereafter, we evaluated the effect

of 1 µ PGE"

on cellular viability following exposure

to different concentrations of H#O#. The sensitivity of

cultured RPE cells against H#O#was effectively reduced

by preincubation with 1 µ PGE"

(Fig. 1). Cytotoxic

effects were dose-dependent in the range of

10–200 µ H#O#

in the control cells without PGE"

treatment. The protective effect of 1 µ PGE"was most

evident when RPE cells were exposed to 200 µ H#O#.

More than 80% of the RPE cells pretreated with PGE"

remained alive after 48 hr incubation with 200 µ

H#O#, whereas less than 10% of RPE cells survived

without PGE"

pretreatment at the same condition.

TRX Induction of PGE"

and}or H#O#

To evaluate the expression of intracellular TRX after

treatment with PGE"

alone, H#O#

alone or both PGE"

and H#O#, western blot analysis was performed. TRX

induction was significantly augmented in the group

treated with PGE"at each concentration of H

#O#when

compared to the group without PGE"

treatment,

although TRX was induced dose-dependently in both

groups with increasing concentrations of H#O#

[Fig.

2(A), (B)]. Enhancement of TRX expression was

dependent on the concentration of PGE"

when RPE

F. 3. Effect of PGE", FK and dBcAMP on the induction of

intracellular cAMP. The concentration of intracellular cAMPafter the stimulation by each agent indicated conditions inthe text determined by ELISA. Bars represent the meanconcentration with standard error in each group. Mean and..s of triplicate cultures are shown. This is representative ofthree separate experiments that gave similar results.

(A)

(B)

F. 4. (A) Effect of FK on the expression of TRX in RPEcells. Each lane represents lysate of human RPE cellsincubated with FK for 24 hr as follows: without FK (Lane 1),100 n FK (Lane 2) and 1 µ FK (Lane 3). Each cell lysatewas applied to contain 5 µg protein}lane. (B) Effect ofdBcAMP on the expression of TRX in RPE cells. Each lanerepresents lysate of human RPE cells incubated withdBcAMP for 24 hr; without dBcAMP (Lane 1), 1 µdBcAMP (Lane 2) and 10 µ dBcAMP (Lane 3). Each celllysate was applied to contain 5 µg protein}lane.

cells were exposed to 10 µ H#O#for 24 hr [Fig. 2(C)].

It should be noted that treatment with PGE"

alone

failed to augment TRX induction in RPE cells [Fig. 2(B)

lane 2, Fig. 2(C)]. In the presence of 10 µ H#O#, the

concentration of PGE"required for one half of maximal

effect (&!

) on TRX induction was 10 n.

To exclude the possibility that TRX may have been

induced by a reaction between PGE"and H

#O#during

incubation, we evaluated TRX induction when H#O#

was added after removing PGE"

by washing. After

pretreatment with 1 µ PGE"for 1 hr, RPE cells were

washed with fresh medium 3 times and then exposed

to 10 µ H#O#for 24 hr. Again, in this experiment the

induction of TRX was augmented significantly as in

the previous experiment (data not shown), indicating

that the interaction of PGE"

and H#O#

does not affect

TRX induction even though chemical reactants might

be generated.

Elevation of Intracellular cAMP Concentration by PGE"

and H#O#

PGE"

has been known to activate multiple in-

tracellular signaling pathways, which include the

elevation of cAMP and the activation of cAMP

dependent protein kinases (PKA) (Lombroso et al.,

1984; Best et al., 1987; Dutta-Roy and Sinha, 1987;

Hashimoto, Negishi and Ichikawa, 1990). As

expected, intracellular cAMP levels in the cells treated

with 1 µ PGE"alone was higher than the untreated

control (P!0±01), while the treatment with 10 µ

H#O#

alone did not increase the intracellular cAMP

level. Of interest was the finding that intracellular

cAMP levels in cells treated by 1 µ PGE"

was

significantly augmented by addition of 10 µ H#O#in

comparison with the cells treated by PGE"alone (P!

INDUCTION OF TRX BY PGE1 IN RPE CELLS 649

T I

Inhibitory effects of Rp-cAMPS on the cytoprotective action of PGE"

and FK against H#O#-mediated injury

H#O#

(µ)Rp-cAMPS (®)

PGE"

(®)Rp-cAMPS (®)

PGE"

(­)Rp-cAMPS (­)

PGE"

(®)Rp-cAMPS (­)

PGE"

(­)Rp-cAMPS (®)

FK (­)Rp-cAMPS (­)

FK (­)

100 42±8³7±7 22±6³11±2* 42±3³5±2 38±5³7±9 25±2³4±3* 50±1³4±2200 92±1³2±3 16±7³6±4** 94±1³2±3 90±1³4±5 31±5³3±4** 92±3³1±3

Cytotoxicity was evaluated by MTT assay. Values represent %OD reduction calculated as in the legend for Fig. 1. *P!0±01, **P!0±005compared with [Rp-cAMPS (®), PGE

"(®)] cells.

F. 5. Inhibitory effect of Rp-cAMPS on TRX induction byPGE

". Lane 1 and 2 represent cell lysates of non-treated cells

and 1 µ PGE"and 10 µ H

#O#-treated cells as negative and

positive controls, respectively. In Lane 3–6, cells werestimulated with 1 µ PGE

"and 10 µ H

#O#after treatment

of various doses of Rp-cAMPS as follows: 100 n Rp-cAMPS(Lane 3), 1 µ Rp-cAMPS (Lane 4), 10 µ Rp-cAMPS (Lane5), 100 µ Rp-cAMPS (Lane 6). Each cell lysate was appliedto contain 10 µg protein}lane.

0±01). Enhanced cAMP levels induced by PGE"

and

H#O#were comparable to those induced by 1 µ FK or

10 µ dBcAMP.

Effect of FK and dBcAMP on the Expression of TRX

FK has been known to directly activate the catalytic

subunit of adenylate cyclase in a reversible manner

(Parker, Sullivan and Wedner, 1974) and dBcAMP is

a cAMP analogue which can penetrate the cell mem-

brane (Henion, Sutherland and Posternak, 1967). We

assessed if FK or dBcAMP could induce TRX in RPE

cells. Both 1 µ FK and 10 µ dBcAMP elicited

significant stimulatory effects on the induction of TRX

expression in RPE cells [Fig. 4(A), (B)]. One micromolar

FK has also been shown to protect RPE cells against

cytotoxic effect of 100 or 200 µ H#O#

(Table I). No

cytotoxic effect of FK or dBcAMP was observed in RPE

cells at the concentrations employed here.

Inhibition by Rp-cAMPS on Both the Induction of TRX

by PGE"

and the Cytoprotective Action of PGE"

or FK

against H#O#-mediated Injury

Since the effects of PGE"are not solely mediated by

cAMP (Dutta-Roy and Shinha, 1987; Coleman, Smith

and Narumiya, 1994; Namba et al., 1994), we next

tested if the induction of TRX by PGE"

could be

inhibited by blocking the cAMP-dependent pathway.

Rp-cAMPS is a potent competitive inhibitor of cAMP

dependent protein kinase (PKA) I and II (Rothermel

and Parker Botelho, 1988) and thus blocks cAMP-

dependent signals at the downstream of elevation of

cAMP. When cells were pretreated with Rp-cAMPS

before addition of PGE", the stimulatory effect of PGE

"

on TRX induction was completely inhibited. In the

cells pretreated with 100 µ Rp-cAMPS for 1 hr, 1 µ

PGE"did not show protective effect against 100 µ or

200 µ H#O#

exposure (Fig. 5). Similarly, Rp-cAMPS

abolished the protective action of FK against cellular

injury caused by 100 or 200 µ H#O#

(Table I).

PGI#

but not PGF#α Enhances TRX Induction

To prove cAMP-dependent signaling process was

involved in the TRX induction by prostaglandins, it

was important to assess TRX expression in RPE cells

treated with other prostaglandins. In this regard, we

examined the effect of two additional prostaglandins,

i.e. PGI#

and PGF#α, on TRX induction in RPE cells

injured by oxidative stress. PGI#acts through a cAMP-

dependent intracellular signaling pathway which is

mediated by the receptor of PGI#, IP (Best et al., 1977;

Hashimoto et al., 1990; Lombroso et al., 1984; Dutta-

Roy and Sinha, 1987). On the other hand, PGF#α

interacts with an FP receptor to stimulate phos-

phoinositide hydrolysis, which results in elevated Ca#+

transient signals (Woodward and Lawrence, 1994).

TRX induction was augmented in a dose-dependent

manner (1 n–1 µ) when cells were treated with

PGI#, as seen with PGE

". However, no significant en-

hancement of TRX induction was observed when cells

were treated with any tested dose of PGF#α, which does

not involve a cAMP dependent pathway (Fig. 6).

4. Discussion

In recent studies PGE"

has been shown to protect

cells or tissues from a variety of harmful insults. Our

previous findings showed that PGE"

could attenuate

retinal damage following ischemia and reperfusion in

vivo (Yamamoto et al., 1997). However, it has been

difficult to confirm in animal models whether the

cytoprotective effect of PGE"is a result of direct effect

on cells or indirect effect mediated by improved blood

supply, because PGE"

has many biological activities

such as vasodilatory and antiplatelet actions (Elkeles

et al., 1969; Gorman, 1978). In this study, we

demonstrated that PGE"reduced H

#O#-induced cellular

injury in cultured RPE cells in association with up-

regulated expression of TRX. TRX has been shown to

protect cells against oxidative injury. For example, in

our previous studies we have shown that recombinant

TRX protected cells against injury caused by H#O#

or

650 M. YAMAMOTO ET AL.

F. 6. Effect of PGI#

and PGF#α on TRX induction in RPE cells. Amount of TRX was compared by the relative intensity of

each immunostaining band in western blot (control¯100). Control : [PG(®), H#O#(10 µ)]. Each cell lysate applied to contain

3 µg protein}lane. Results shown are mean³.. of three different samples. Values indicated (*P!0±05, **P!0±01) aresignificantly different from the group of control. E, PGE

"; *, PGI

#; +, PGF

#α.

TNF-α (Matsuda et al., 1991; Nakamura et al., 1994).

More recently, it was demonstrated that intracellular

levels of TRX in some carcinomas were shown to

correlate with the sensitivity against several ROI-

generating anticancer drugs (Yokomizo et al., 1995;

Sasada et al., 1996; Yamada et al., 1996). Fur-

thermore, a finding by others that a thiol alkylating

agent suppresses the protective activity of prosta-

glandins in gastric mucosal cells (Robert, Eberle and

Kaplowitz, 1984) indicates that cellular thiols in-

cluding TRX are critically involved in the cyto-

protective action of prostaglandins. These findings

correlatively indicate that PGE"

has direct cyto-

protective activity against oxidative insults.

We further analysed the exact mechanism of the

direct cytoprotective action of PGE"

concerning the

induction of TRX. Measurement of intracellular cAMP

levels showed that intracellular cAMP accumulation

was significantly enhanced by treatment with both

PGE"and H

#O#when compared to the treatment with

PGE"alone. In this regard, up-regulation of TRX was

observed only when cells were stimulated with both

PGE"

and H#O#

while the treatment with PGE"

alone

failed to induce TRX. Furthermore, FK and dBcAMP,

both of which could induce as same levels of cAMP as

treatment with both PGE"

and H#O#, also enhanced

TRX expression. These observations correlatively

suggest that TRX induction in RPE cells is dependent

upon the levels of intracellular cAMP and that the

concentration of cAMP induced by PGE"alone may be

insufficient for intracellular TRX induction. Addition-

ally, both PGE"

and FK failed to elicit cytoprotective

activity as well as TRX augmentation when cells were

pretreated with Rp-cAMPS, a competitive inhibitor of

cAMP dependent protein kinases. Treatment of RPE

cells with PGI#, a stimulator of cAMP-dependent

pathway like PGE", could augment intracellular TRX

induction while treatment with PGF#α, which

functions through cAMP-indpendent pathway, failed

to show this property. We also found that PGI#

had

cytoprotective effects against oxidative stress in vivo as

well as in vitro in our separate study (unpublished

data). These observations suggest strongly that in-

tracellular TRX induction via a cAMP-dependent

signal pathway may play an important role on the

cytoprotective effect of PGE".

However, it is also possible that any other pathways

also regulate TRX induction as well as cAMP-

dependent process. Recently, expression of the TRX

gene by oxidative stress has been demonstrated to be

controlled by a specific cis-acting regulatory element

(Taniguchi et al., 1996). It is an intriguing possibility

that cAMP-dependent pathway interacts oxidative

stress-dependent pathway of TRX induction. Thus, the

effect of PGE"

seen in this study may be more

adequately described as ‘priming’ in which RPE cells

are rendered to express more TRX upon subsequent

exposure to H#O#. Alternatively it is required to

examine whether TRX can be regulated by other

pathways separate from cAMP and further inves-

INDUCTION OF TRX BY PGE1 IN RPE CELLS 651

tigations are being undertaken to clarify this issue in

our laboratory.

It has been shown that oxidative stress, such as UV

irradiation or exposure to H#O#

(Lennon, Martin and

Cotter, 1991; Martin and Cotter, 1991; Uckun et al.,

1992), resulted in apoptosis of various cells and that

apoptotic cell death induced by oxidative stress was

successfully protected by anti-oxidants (Hockenbery et

al., 1993; Sandstrom and Buttke, 1993). In this study

we found that the moderate doses (100–200 µ) of

H#O#did not induce cell death until 48 hr of incubation

while the higher dose of over 300 µ H#O#

induced

necrosis within 12 hr of incubation (data not shown).

Our similar observation using a human T cell line, i.e.,

Jurkat cells, showed that apoptotic cell death was

induced by treatment with moderate doses of

100–200 µ diamide, a sulfhydryl-specific oxidant,

later than 48 hr of culture, while as high as 400 µ

diamide induced necrosis in the early phase (Sato et

al., 1995). Thus, prolonged survival of RPE cells with

PGE"-treatment may be caused by protective effect of

PGE"

from induction of apoptosis. Intracellular

changes caused by oxidative stress may be induced in

the very early phase, which would result in cell death

in the late phase, because only first 30 min of exposure

to diamide could induce apoptotic cell death which is

observed in the late phase (unpublished observation).

Various intracellular events including cAMP induc-

tion, which occurred immediately after exposure to

H#O#, may dictate whether RPE cells live or die and

TRX induced consequently may work as a protective

factor against apoptotic cell death, as seen in our

previous studies which demonstrated that exogenous

TRX inhibited cell death caused by anti-Fas Ab

(Matsuda et al., 1991) and promoted the prolonged

survival of murine neuronal cells in primary culture

(Hori et al., 1994).

In general, eukaryotic cells are equipped with

several defensive enzymes against oxidative stress,

such as superoxide dismutase, catalase, and

glutathione-dependent enzymes. The involvement of

these enzymes in the cytoprotective action of PGE"will

be also subjected to further investigation.

We described here the induction of TRX via a

cAMP-dependent pathway as a possible novel mech-

anism by which PGE"

protects cells against oxidative

injury. As shown in the present study, TRX may act as

an anti-oxidative or anti-apoptotic factor in injured

tissues where oxidative stress such as ischemia}reperfusion and UV-exposure, plays pathological roles.

Therefore, PGE"

could be a therapeutic agent which

can elicit the endogenous induction of TRX against

oxidative injuries or apoptotic cell damages. Our

current findings provide justification to extend the use

of PGE"as a therapeutic agent against these diseases.

Acknowledgements

The authors thank Dr S. Narumiya for helpful discussionsand suggestions, Dr H. Masutani and Dr S. Yamamoto forassistance and advice and Dr J. L. VanCott for critical

reading. We are also grateful to Drs H. Ohno, D. Fukushimaand their colleagues for supplying the drugs, their experttechnical assistance and helpful discussions and MsY. Kanekiyo for her editorial assistance.

This work was supported by the Grant-in Aid for ScientificResearch from the Ministry of Education, Science andCulture of Japan and the grant and drugs from OnoPharmaceutical Co. Ltd.

References

Adaikan, P. G. and Karim, S. M. (1981). Some pharma-cological effects of 17(S)methyl-omega-homo-trans-delta2-PGE

"(ONO 1206). Prostaglandins Med. 6,

449–58.Best, L. C., Martin, T. J., Russel, R. G. and Preston, F. E.

(1977). Prostacyclin increases cyclic AMP levels andadenylate cyclase activity in platelets. Nature 267,850–2.

Cherner, J. A., Naik, L., Tarnawski, A., Brzozowski, T.,Stachura, J. and Singh, G. (1989). Ability of prosta-glandin to reduce ethanol injury to dispersed chief cellsfrom guinea pig stomach. Am. J. Physiol. 256, G704–14.

Coleman, R. A., Smith, W. L. and Narumiya, S. (1994).International union of pharmacology classification ofprostanoid receptors : properties, distribution and struc-ture of the receptors and their subtypes. Pharmacol. Rev.46, 205–29.

Dutta-Roy, A. K. and Sinha, A. K. (1987). Purification andproperties of prostaglandin E

"}prostacyclin receptor of

human platelets. J. Biol. Chem. 262, 12685–91.Elkeles, R. S., Hampton, J. R., Harrison, M. J. and Mitchell,

J. R. (1969). Prostaglandin E"

and human platelet.Lancet 2, 111.

Fujii, S., Nanbu, Y., Nonogaki, H., Konishi, I., Mori, T.,Masutani, H. and Yodoi, J. (1991). Coexpression ofadult T-cell leukemia-derived factor, a humanthioredoxin homologue, and human papillomavirusDNA in neoplastic cervical squamous epithelium. Cancer68, 1583–91.

Gauntt, C. D., Ohira, A., Honda, O., Kigasawa, K., Fujimoto,T., Masutani, H., Yodoi, J. and Honda, Y. (1994).Mitochondrial induction of adult T cell leukemia derivedfactor (ADF}hTX) after oxidative stress in retinalpigment epithelial cells. Invest. Ophthalmol. Vis. Sci. 35,2916–23.

Gorman, R. R. (1978). Prostaglandins, thromboxanes andprostacyclin. In: Biochemistry and Mode of Action ofHormones. (Ed. H. V. Rickenberg). Pp. 81–107.Baltimore, MD: University Park.

Hashimoto, H., Negishi, M. and Ichikawa, A. (1990).Identification of prostacyclin receptor coupled to theadenylate cyclase system via a stimulatory GTP-bindingprotein in mouse mastocytoma P-815 cells. Prosta-glandins 40, 491–505.

Henion, W. F., Sutherland, E. W. and Posternak, T. (1967).Effects of derivatives of adenosine 3«, 5«-phosphate ofliver slices and intact animals. Biochem. Biophys. Acta.148, 106–13.

Hockenbery, D. M., Oltvai, Z. N., Yin, X.-M., Milliman, C. L.and Korsmeyer, S. J. (1993). Bcl-2 functions in an anti-oxidant pathway to prevent apoptosis. Cell 75, 241–51.

Hori, K., Katayama, M., Sato, N., Ishii, K., Waga, S. andYodoi, J. (1994). Neuroprotection by glial cells throughadult T cell leukemia-derived factor}human thioredoxin(ADF}TRX). Brain. Res. 652, 304–10.

Iwai, T., Fujii, S., Nanbu, Y., Nonogaki, H., Konishi, I., Mori,T., Masutani, H. and Yodoi, J. (1992). Expression ofadult T-cell leukemia-derived factor, a human thio-redoxin homologue, in the human ovary throughoutthe menstrual cycle. Virchows. Arch. A. Pathol. Anat.Histopathol. 420, 213–17.

652 M. YAMAMOTO ET AL.

Kigasawa, K., Soushi, S., Tanaka, Y. and Obazawa, H.(1994). Morphologic and chromosomal study of ahuman retinal pigment epithelial cell line. Jpn. J.Ophthalmol. 38, 10–5.

Kusama, K., Saitoh, T., Masutani, H., Nakamura, H., Yodoi,J. and Moro, I. (1991). Adult T cell leukemia derivedfactor (ADF) in oral epithelial lesions. J. Oral. Pathol.Med. 20, 421–4.

Lennon, S. V., Martin, S. J. and Cotter, T. G. (1991). Dose-dependent induction of apoptosis in human tumor celllines by widely diverging stimuli. Cell Prolif. 24,203–14.

Lombroso, M., Nicosia, S., Paoletti, R., Whittle, B. J.,Moncada, S. and Vane, J. R. (1984). The use of stableprostaglandins to investigate prostacyclin (PGI2)-bind-ing sites and PGI2-sensitive adenylate cyclase in humanplatelet membranes. Prostaglandins 27, 321–33.

Martin, S. J. and Cotter, T. G. (1991). Ultraviolet B irradi-ation of human leukemia HL-60 cells in vitro inducesapoptosis. Int. J. Radiat. Biol. 59, 1001–16.

Matsuda, M., Masutani, H., Nakamura, H., Miyajima, S.,Yamauchi, A., Yonehara, S., Uchida, A., Irimajiri, K.,Horiuchi, A. and Yodoi, J. (1991). Protective activity ofadult T cell leukemia-derived factor (ADF) againsttumor necrosis factor-dependent cytotoxity on U937cells. J. Immunol. 147, 3837–41.

Mitsui, A., Hirakawa, T. and Yodo, J. (1992). Reactiveoxygen-reducing and protein-refolding activities ofadult T cell leukemia-derived factor}human thio-redoxin. Biochem. Biophys. Res. Commun. 186,1220–26.

Mosmann, T. (1983). Rapid colorimetric assay for cellulargrowth and survival : application to proliferation andcytotoxicity assays. J. Immunol. Methods 65, 55–63.

Nakamura, H., Masutani, H., Tagaya, Y., Yamauchi, A.,Inamoto, T., Nanbu, Y., Fujii, S., Ozawa, K. and Yodoi,J. (1992). Expression and growth-promoting effect ofadult T-cell leukemia-derived factor. A human thio-redoxin homologue in hepatocellular carcinoma. Cancer69, 2091–7.

Nakamura, H., Matsuda, M., Furuke, K., Kitaoka, Y., Iwata,S., Toda, K., Inamoto, T., Yamaoka, Y., Ozawa, K. andYodoi, J. (1994). Adult T cell leukemia-derived factor}human thioredoxin protects endothelial F-2 cell injurycaused by activated neutrophils or hydrogen peroxide.Immunol. Let. 42, 75–80.

Namba, T., Oida, H., Sugimoto, Y., Kakizuka, A., Negishi,M., Ichikawa, A. and Narumiya, S. (1994). cDNAcloning of a mouse prostacyclin receptor. Multiplesignaling pathways and expression in thymic medulla.J. Biol. Chem. 269, 9986–92.

Ohira, A., Honda, O., Gauntt, C. D., Yamamoto, M., Hori, K.,Masutani, H., Yodoi, J. and Honda, Y. (1994). Oxidativestress induces adult T cell leukemia derived factor}thioredoxin in the rat retina. Lab. Invest. 70, 279–85.

Ohno, H., Morikawa, Y. and Hirata, F. (1978). Studies on15-hydroxyprostaglandin dehydrogenase with variousprostaglandin analogues. J. Biol. Chem. 84, 1485–94.

Parker, C. W., Sullivan, T. J. and Wedner, H. J. (1974).Cyclic AMP and immune response. Adv. Cyclic NucleotideRes. 4, 1–79.

Robert, A. (1979). Cytoprotection by prostaglandins. Gastro-enterology 77, 761–7.

Robert, A., Eberle, D. and Kaplowitz, N. (1984). Role ofglutathione in gastric mucosal cytoprotection. Am. J.Physiol. 247, G296–304.

Rothermel, J. D. and Parker Botelho, L. H. (1988). Amechanistic and kinetic analysis of the interactions ofthe diastereoisomers of adenosine 3«, 5«-(cyclic)phosphorothioate with purified cyclic AMP-dependentprotein kinase. Biochem. J. 251, 757–62.

Sandstrom, P. A. and Buttke, T. M. (1993). Autocrine pro-duction of extracellular catalase prevents apoptosis of

the human CEM T cell line in serum-free medium. Pro.Natl. Acad. Sci. U.S.A. 90, 4708–12.

Sasada, T., Iwata, S., Sato, N., Kitaoka, Y., Hirota, K.,Nakamura, K., Nishiyama, A., Taniguchi, Y.,Takabayashi, A. and Yodoi, J. (1996). Redox control ofresistance to cis-diamminedichloroplatinum (II)(CDDP): protective effect of human thioredoxin againstCDDP-induced cytotoxicity. J. Clin. Invest. 97, 2268–76.

Sato, N., Iwata, S., Nakamura, K., Hori, T., Mori, K. andYodoi, J. (1995). Thiol-mediated redox regulation ofapoptosis : Possible roles of cellular thiols other thanglutathione in T cell apoptosis. J. Immunol. 154,3194–203.

Tagaya, Y., Okada, M., Sugie, K., Kasahara, T., Kondo, N.,Hamuro, J., Matsushima, K., Dinarello, C. A. and Yodoi,J. (1988). IL-2 receptor(p55)}Tac-inducing factor.Purification and characterization of adult T cellleukemia-derived factor. J. Immunol. 140, 2614–20.

Taniguchi, Y., Taniguchi-Ueda, Y., Mori, K. and Yodoi, J.(1996). A novel promoter sequence is involved in theoxidative stress-induced expression of the adult T-cellleukemia-derived factor (ADF)}human thioredoxin(Trx) gene. Nucl. Acid Res. 24, 2746–52.

Tarnawski, A., Brzozowski, T., Sarfeh, I. J., Krause, W. J.,Ulich, T. R., Gergely, H. and Hollander, D. (1988).Prostaglandin protection of human isolated gastricglands against indomethacin and ethanol injury.Evidence for direct cellular actions of prostaglandin. J.Clin. Invest. 81, 1081–9.

Terano, A., Ota, S., Mach, T., Hiraishi, H., Stachura, J.,Tarnawski, A. and Ivey, K. J. (1986). Prostaglandinprotects against taurocholate-induced damage to ratgastric mucosal cell culture. Gastroenterology 92,669–77.

Uckun, F. M., Tuel-Ahlgren, L., Song, C. W., Waddick, K.,Myers, D. E., Kirihara, J., Ledbetter, J. A. and Schieven,G. L. (1992). Ionizing radiation stimulates unidentifiedtyrosine-specific protein kinases in human B-lympho-cyte precursors, triggering apoptosis and clonogenic celldeath. Proc. Natl. Acad. Sci, U.S.A. 89, 9005–9.

Wakita, H., Yodoi, J., Masutani, H., Toda, K. and Takigawa,M. (1992). Immunohistochemical distribution of adultT-cell leukemia-derived factor}thioredoxin in epithelialcomponents of normal and pathologic human skinconditions. J. Invest. Dermatol. 99, 101–7.

Woodward, D. F. and Lawrence, R. A. (1994). Identificationof a single (FP) receptor associated with prostanoid-induced signals in Swiss 3T3 cells. Biochem. Pharmacol.47, 1567–74.

Yamada, M., Tomida, A., Yoshikawa, H., Taketani, Y. andTsuruo, T. (1996). Increased expression of thioredoxin}adult T-cell leukemia-derived factor in cisplatin-re-sistant human cancer cell lines. Clin. Cancer Res. 2,427–432.

Yamamoto, M., Ohira, A., Honda, O., Sato, N., Furuke, K.,Yodoi, J. and Honda, Y. (1997). The analysis oflocalization of Adult T-cell Leukemia Derived Factor inthe transient ischemic rat retina after the treatmentwith OP-1206 α-CD, a Prostaglandin E

"analogue. J.

Histochem. Cytochem. 45, 63–70.Yodoi, J. and Tursz, T. (1991). ADF, a growth-promoting

factor derived from adult T cell leukemia and hom-ologous to thioredoxin: involvement in lymphocyteimmortalization by HTLV-1 and EBV. Adv. Cancer Res.57, 381–411.

Yodoi, J. and Uchiyama, T. (1992). Diseases associated withHTLV-1: virus, IL-2 receptor dysregulation and redoxregulation. Immunol. Today 13, 405–11.

Yokomizo, A., Ono, M., Nanri, H., Makino, Y., Ohga, T.,Wada, M., Okamoto, T., Yodoi, J., Kuwano, M. andKohno, K. (1995). Cellular levels of thioredoxinassociated with drug sensitivity to cisplatin, mitomycinC, doxorubicin, and etoposide. Cancer Res. 55, 4293–6.