Pharmacologic Evidenc foer 5-HT1A Receptors …iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/... · Pharmacologic Evidenc foer 5-HT1A Receptors Associated With Human

Pharmacologic Evidence for 5-HT1A Receptors AssociatedWith Human Retinal Pigment Epithelial Cells in Culture

Mark S. Nash and Neville N. Osborne

Purpose. The authors investigate the possible presence of 5-hydroxytryptamine (5-HT!) typeserotonin receptors negatively coupled to adenylate cyclase activity in cultured human retinalpigment epithelial (RPE) cells.

Methods. Adenylate cyclase activity was assessed by the determination of cellular adenosine3':5' cyclic monophosphate (cAMP) levels and the effects of serotonin on both basal andforskolin-stimulated cAMP levels studied.

Results. Serotonin at 100 /xM had no effect on the basal levels of cAMP in cultured humanRPE cells, but attenuated by 43.6% the stimulation in cAMP production induced by forskolin(5 //M). This effect was dose dependent for serotonin with half-maximal inhibition (EC50)occurring at approximately 1.5 X 10~9 M. The 5-HT, receptor agonists 8-hydroxy [2-di-n-propylamino] tetralin (8-OH DPAT), buspirone, 5-carboxyamidotryptamine, and RU24969mimicked the inhibitory effect of serotonin in a dose-dependent manner. The actions ofserotonin and 8-OH DPAT (10 fjM) were dose-dependently attenuated by the serotonergicantagonists spiroxatrine, propranolol, and spiperone. Pretreatment of RPE cell cultures withpertussis toxin abolished the serotonin-induced reduction of forskolin-elevated cAMP levels.Stimulation of cAMP production by the /3-adrenoceptor agonist isoproterenol at 0.1 fjM, butnot at 10 fjM or 100 fjM, also was attenuated by serotonin (100 /xM), whereas cAMP productioninduced by the adenosine receptor agonist 5'-[N-ethyl]-carboxamidoadenosine (NEGA) at 1//M, 10 fjM, and 100 fjM was unaffected. Serotonin and 8-OH DPAT dose-dependendy inhib-ited isoproterenol-stimulated (0.1 /JM) CAMP production with EC50 values of approximately10 /xM, and pertussis toxin pretreatment partially blocked these effects.

Conclusions. Cultured human RPE cells possess 5-HTjA receptors negatively coupled to cAMPproduction through a pertussis toxin-sensitive G protein. These receptors show differentialeffects on forskolin-, isoproterenol-, and NECA-stimulated cAMP production, which may re-flect a unique spatial distribution of receptor proteins or the phenotypic heterogeneity ofRPE cells diat is the result of or that is preserved in culture. Invest Ophthalmol Vis Sci.1997;38:510-519.

XVetinal pigment epithelial (RPE) cells perform sev-eral functions that are crucial to the functional integ-rity of the photoreceptors, including phagocytosis ofshed outer segment discs,1 maintenance of the ionand fluid homeostasis of the subretinal space,2'3 anduptake and processing of retinoids for use in the visual

From Nujfield Laboratory of Ophthalmology, Oxford University, Oxford, UnitedKingdom.Supported by The Royal National Institute far the Blind and the EuropeanCommunity.Submitted for publication January 19, 1996; revised September 9, 1996; acceptedSeptember 10, 1996.Proprietary interest category: N.Reprint requests: Neville N. Osborne, Nujfield Laboratory of Ophthalmology, OxfordUniversity, Walton Street, Oxford, 0X2 6AW, United Kingdom.

cycle.4 Many of these important functions are modu-lated by mediators of the adenylate cyclase (AC) andphospholipase C (PLC) signal transduction systems.For example, increased adenosine 3':5' cyclic mono-phosphate (cAMP) levels and protein kinase C activityinhibit the ingestion of rod outer segments by cul-tured RPE cells,5"7 subretinal fluid absorption is de-creased by cAMP,89 and ion transport across the RPEcells is modified by factors coupled to AC10 and PLC2'3

activity. The identification of cell surface receptorslinked to these second messenger systems may thusprovide clues as to how RPE cell function is controlledin vivo.

Cultured human RPE possess numerous G-pro-

510Investigative Ophthalmology & Visual Science, February 1997, Vol. 38, No. 2Copyright © Association for Research in Vision and Ophthalmology

Downloaded From: https://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933422/ on 08/08/2018

5-HT,A Receptors Associated With Human RPE 511

tein-coupled receptors that modulate AC and PLC ac-tivity. Beta2 adrenoceptor,11 A2 adenosine,12 and vaso-active intestinal peptide13 receptors are positively cou-pled to AC and stimulate cAMP production, whereasM3 muscarinic acetylcholine,14"16 Hj histamine,16 Vivasopressin,1718 and /32 bradykinin18 receptors are cou-pled to PLC activity and stimulate production of inosi-tol phosphates and intracellular calcium mobilization.Cultured human RPE cells also possess melatonin re-ceptors negatively coupled to AC through an inhibi-tory G protein (Gj) that attenuate the stimulation ofcAMP production by forskolin, a direct activator ofAC.19 Working on cultured rat RPE cells, we recentlyidentified serotonin (5-HT2A) receptors positively cou-pled to inositol phosphate turnover and intracellularCa2+ mobilization.20 However, serotonin (5-hydroxy-tryptamine; 5-HT) fails to influence PLC activity incultured human RPE,1516 suggesting that importantspecies differences may occur in the types of G-pro-tein-coupled receptors associated with RPE cells. Wehave therefore undertaken an investigation into thepossible presence of alternative subtypes of 5-HT re-ceptor and report that cultured human RPE possess5-HT1A receptors negatively coupled to cAMP produc-tion.

MATERIALS AND METHODS

Materials

Postmortem human eyes were obtained from BristolEye Bank (Bristol, UK) after removal of the corneafor transplant surgery. [2, 8-3H]-adenosine 3':5' cyclicmonophosphate (33 Ci/mmol) was purchased fromAmersham International (Amersham, United King-dom). Fetal bovine serum (European Community ap-proved), Hams-FlO, fungizone (amphotericin B), gluta-mine, 0.25% trypsin solution, and 24-multiwell plates(NUNC) were from Gibco (Paisley, United Kingdom),and 25 cm2 and 75 cm2 tissue culture flasks were fromFalcon (Oxford, United Kingdom). The 8-hydroxy [2-di-n-propylamino] tetralin (±8-OH DPAT) and mians-erin were from Research Biochemicals International(St. Albans, United Kingdom); 5-carboxyamidotryp-tamine (5-CT), and sumatriptan from Glaxo(Greenford, United Kingdom); RU24969 from Roussel-UCLAF (Paris, France); buspirone from Bristol-MeyersSquibb (Wallingford, CT); methysergide, metergoline,and SDZ 21009 from Sandoz (Basel, Switzerland); MDL72222 from Marrion Merrel Dow (Cincinnati, OH);and ketanserin, spiperone, and spiroxatrine from Jans-sen Pharmaceuticals (Geel, Belgium). All other stan-dard chemicals and biochemicals were obtained fromSigma (Poole, United Kingdom) or Merck (Lutterw-orth, United Kingdom).

Methods

Primary cultures of human RPE cells were establishedas described previously19 and grown to confluence in25-cm2 flasks containing Hams-FlO culture medium(Hams-FlO [with L-glutamine], 10% fetal bovine se-rum, 0.4% glucose, 2-mM glutamine, 2.5 /Ltg/ml am-photericin B, and 100 //g/ml gentamycin). Cultureswere passaged with a ratio of 1:3 in 75-cm2 flasks, andcells between passages 2 through 4 used for experi-mentation. Cells were immunostained routinely withthe monoclonal antibody K 8.13 (Sigma) to test forthe presence of cytokeratins and to confirm the purityand nature of the cells present because potential con-taminating cells do not express cytokeratins.21

The effects of drug treatment on cellular cAMPlevels were determined as described by Nash and Os-borne.19 Briefly, culture medium was removed fromcells in 24-multiwell plates, replaced with serum-freeHams-FlO, and cells were incubated at 37°C/5% car-bon dioxide for 2 to 4 hours. Cells then were incu-bated in 200 ji\ of experimental buffer (Hams-FlOwith 20-mM Hepes, pH 7.4) for 5 minutes at 37°Cafter which 10 /A of drug was added. An experimentallayout of 5 minutes preincubation with antagonist, afurther 5-minute treatment with serotonergic agonist,and then incubation with forskolin, isoproterenol, or5'-[A^ethyl]-carboxamidoadenosine (NECA) for 5minutes was used throughout. The reaction then wasterminated by boiling for 3 minutes. Fifty microliteraliquots were removed and cAMP levels determinedusing the method of Brown et al,22'23 which uses aspecific cAMP-binding protein isolated from bovineadrenal cortex to detect cellular cAMP by competitionwith a standard amount of tritiated cAMP. Details ofthe methodology can be found in the study by Nashand Osborne.19

Statistical Analysis

Statistical significance was determined using Student'sHest for paired data, and a P < 0.05 was consideredsignificant. Half-maximal (EC50) values were calcu-lated by determination of the concentration of agonistrequired to achieve half the maximum response ob-served at saturating drug concentrations.

RESULTS

The biochemical studies were performed on culturedRPE cells derived from 13 different human donorsaged between 14 and 65 years old. Some variationswere observed in the responsiveness to test drugs be-tween particular cell cultures. To minimize any com-plications in interpretation of results arising from suchvariation, results were compiled, as far as possible,


512 Investigative Ophthalmology & Visual Science, February 1997, Vol. 38, No. 2

TABLE l. Effect of Serotonin on Basal andForskolin (5 /xM)-stimulated Levels ofcAMP in Cultured Human RPE Cells

Drugs Added cAMP (pmol/well/5 min)

BasalSerotonin (100Forskolin (5

+ Serotonin (100

3.95 ± 0.973.65 ± 0.84

61.91 ± 3.4434.90 ± 3.82*

cAMP = cyclic adenosine monophosphate; RPE = retinalpigment epithelial.Results are expressed as mean ± SEM values from seven separateexperiments carried out in triplicate.* P < 0.001 versus forskolin addition alone using Student'spaired Mest.

from experimental data obtained from two or morecell lines to give an average response for the observedeffects.

The effect of 5-HT on cAMP metabolism in cul-tured human RPE cells is listed in Table 1. Exposureof human RPE cells to 100-//M 5-HT for 10 minuteshad no effect on the basal levels of cAMP. Forskolinat 5 fiM, however, elevated cellular cAMP levels by1,467.3% after 5 minutes. In contrast to its effects onbasal levels, a 5-minute preincubation with 100-//M 5-HT attenuated the ability of forskolin to increasecAMP production by 43.6%. This effect of 5-HT wasmimicked by 5-CT, RU24969, and sumatriptan (gen-eral 5-HT! receptor agonists) and 8-OH DPAT andbuspirone (specific 5-HTiA receptor agonists) (Fig. 1)with an order of efficacy of 8-OH DPAT (52.0%) >5-CT (47.8%) > 5-HT (44.3%) > buspirone (41.3%)> RU24969 (38.4%) > sumatriptan (34.4%). The re-duction in forskolin-stimulated cAMP production wasdose dependent for 5-HT, 8-OH DPAT, and buspirone(Fig. 2A), and 5-CT and RU24969 (Fig. 2B) with EC50

values of 1.5 X 10"9M, 1.9 X 10"9M, 1.0 X 10~8M,1.7X 10"n M, and 7.3 X 10"9 M, respectively. Significantinhibition of forskolin-induced cAMP levels at nano-molar concentrations was observed for both 8-OHDPAT (20.8%) and 5-CT (25.9%), and for 5-CT, sig-nificant inhibition still was observed at 0.01 nM (Fig.2B). The order of potency for the agonists was 5-CT> 5-HT « 8-OH DPAT > RU24969 > buspirone.

To further characterize the effect of serotonin onforskolin-stimulated cAMP levels, a range of serotoner-gic antagonists were assessed for their ability to attenu-ate the response (Fig. 3). A 5-minute preincubationwith the specific 5-HT1A receptor antagonist spiroxa-trine or the mixed 5-HT1A/2A antagonist spiperone in-hibited 5-HT action by 57.2% and 41.2%, respectively.The /3-adrenergic and serotonergiciA/1B antagonistpropranolol also significantly blocked the 5-HT-in-duced reduction of forskolin-elevated cAMP levels by

53.9%. At 0.1 fjM, neither spiroxatrine, spiperone, norpropranolol alone influenced forskolin stimulation(95.2% ± 3.1%, 96.1% ± 3.8%, and 96.7% ± 1.0%,respectively). The more general 5-HT! receptor antag-onists methysergide, SDZ 21009, and metergoline(Sandoz) all attenuated the 5-HT response to varyingdegrees, although the effect of SDZ 21009 was statisti-cally insignificant. In contrast, the 5-HT2A/2c receptorantagonist ketanserin, the general 5-HT2 receptor an-tagonist mianserin, and the 5-HT3 receptor antagonistMDL 72222 (Marrion Merrel Dow) were without ef-fect. Dose dependency was observed for the effect ofspiroxatrine (Fig. 4A), spiperone (Fig. 4B), and pro-pranolol (Fig. 4C) on both the 5-HT and 8-OH DPAT(10 /iM)-induced reductions of forskolin-stimulatedcAMP levels. Taking the maximal inhibition as 100%half-maximal inhibition (IC50) of 5-HT action for spir-oxatrine, spiperone, and propranolol was observed at1.4 X 10"8 M, 2.2 X 10"7 M, and 1.6 X 10"8 M, respec-tively, and 2.0 X 10~8 M, 1.6 X 10~7 M, and 4.0 X 10"8

M for 8-OH DPAT action, respectively.

The effect of 5-HT on forskolin action was foundto be pertussis toxin-sensitive (Fig. 5). An 18-hour in-cubation with pertussis toxin (100 ng/ml) failed toinfluence basal levels of cAMP and those induced byforskolin. However, the treatment potentiated the

Agonist (10p.M)

FIGURE l. Effect of various serotonin (5-HT[) receptor ago-nists on the stimulation of adenosine 3':5'-cyclic monophos-phate production by forskolin (5 fxM) in cultured humanretinal pigment epithelial cells. The results are mean ± stan-dard error of the mean value from three separate experi-ments performed in triplicate. The effect of drug comparedwith forskolin stimulation alone was statistically significant(P < 0.05) in all cases by Student's t-test for paired data.


5-HT1A Receptors Associated With Human RPE

70

.3

oco

DU "

5 0 -

40 -

30-

20 -

10-

0 -

-10 -

B.

Li

i VrX^l.h i ±

i i i i i i-16 -14 -12 -10 -8 -6 -2

Log [Agonist] M

FIGURE 2. (A) Dose response curves are shown for the effectof 5-HT (O) and the 5-HT1A receptor agonists, 8-OH DPAT(•) and buspirone (•) on the stimulation of cAMP produc-tion by forskolin (5 /J,M) in cultured human retinal pigmentepithelial cells. (B) Dose response curves are shown for theeffect of the general 5-HT, receptor agonists, 5-CT (O), andRU24969 (•). The data are presented as mean ± standarderror of the mean value from between 4 and 10 separateexperiments performed in triplicate. cAMP = adenosine3':5' cyclic monophosphate.

stimulation of cAMP production by the /?-adrenocep-tor agonist isoproterenol by approximately 32.4% andcompletely attenuated the 5-HT (100 /JM)-induced re-duction of forskolin-stimulated cAMP production.

Cultured human RPE cells have been shown pre-viously to possess /52-adrenoceptorn and A2 adeno-sine12 receptors. Thus, the influence of 5-HT on the

513

ability of isoproterenol and NEGA to stimulate cAMPproduction was determined (Fig. 6). Isoproterenolstimulates cAMP production in cultured human RPEcells with an EC50 value of approximately 0.2 /xM anda saturating concentration of 10 (JM, whereas NECAhas an EC50 value of approximately 1 //M and satura-tion at approximately 10 //M (data not shown). Inter-estingly, serotonin had little or no effect on maximallystimulating concentrations of isoproterenol (100 and10 //M) but inhibited by 64.8% the stimulation ofcAMP production with 0.1-//M isoproterenol. In con-trast, 5-HT had no effect on the stimulation of cAMPsynthesis through the adenosine receptor, whether ac-tivated by maximal or submaximal concentrations ofNECA. Elevation of cAMP levels by 0.1-//M isoprotere-nol was dose-dependently inhibited by both serotoninand 8-OH DPAT, although, unlike their effect on for-skolin, significant inhibition was witnessed only at con-centrations greater than micromolar. However, at 1mM, both drugs achieved near-complete attenuationof isoproterenol-stimulated cAMP levels (Fig. 7). Thestimulation of cAMP production by isoproterenol at100 fjM was unaffected by preincubation with a rangeof concentrations of 5-HT and, in addition, no effectof 5-HT or 8-OH DPAT was observed on NECA (1//M)-stimulated cAMP production across a range ofagonist concentrations (data not shown).

As for the 5-HT-induced inhibition of forskolin-

5-HT(lOpM)

+ Antagonist(0.1 pM)

FIGURE 3. Effect of various antagonists (0.1 fjM) on the5-HT-(10 (JM) induced reduction of forskolin- (5 (JM) stimulatedcAMP levels in cultured human retinal pigment epithelialcells. The results are mean ± standard error of the meanvalue from between three and eight separate experimentsconducted in triplicate. *P < 0.05 when compared with 5-HT effect alone using Student's Hest for paired data. cAMP= adenosine 3':5' cyclic monophosphate.



Oo

•§p

T3

•3

.3

oco

O

T3

O

oo

Log [Spiroxatrine] M

-14 -12 -10

Log [Spiperone] M

-14 -12 -10

Log [Propranolol] M

FIGURE 4. Dose-dependent antagonism by spiroxatrine (A),spiperone (B), and propranolol (C) of the 10-̂ M inducedeffect of 5-HT (•) and 8-OH DPAT (O) on forskolin (5//M) stimulation of cAMP production in cultured humanretinal pigment epithelial cells. Data are expressed as mean± standard error of the mean values from between fourand seven experiments performed in triplicate. cAMP =adenosine 3':5' cyclic monophosphate.

elevated cAMP levels, the effect of 5-HT (100 //M)on isoproterenol (0.1 fjM) action was pertussis toxin-sensitive, as was the effect of 8-OH DPAT (100 //M)(Table 2). However, in these experiments, the pertus-sis toxin pretreatment was only partially effective in

attenuating the agonist-induced inhibition of isopro-terenol-stimulated cAMP production.

DISCUSSION

Cultured rat RPE cells have been shown recently to pos-sess 5-HT2A receptors that, when activated, increase inosi-tol phosphate turnover and mobilize intracellular Ca2+

stores.20 However, no serotonin effect on the PLC signaltransduction system is observed in human RPE cell cul-tures.1516 The data presented here show that, in contrastto their effect on rat RPE cells, serotonin receptors nega-tively couple to cAMP production in cultured humanRPE cells. Fourteen different serotonin receptors cur-rendy have been characterized, and these have beenclassified into 7 different families.24"26 The 5-HT! recep-tor family is negatively coupled to AC activity, whereas5-HT4, 5-HT6, and 5-HT7 receptors are positively cou-pled to the enzyme. The 5-HT2 receptors are positivelycoupled to PLC activity, and the 5-HT3 receptor is aligand-gated cation channel. No second messenger sys-tem coupled to the 5-HT5 receptors has yet been identi-fied. Cultured human RPE cells thus possess a receptor

6.50 - -

6.00 - -

5.50 - -

• | 5.00 - -

«o 4.50 - -

73 4 . 0 0 - -

^ 0.80 - -o

ja 0.75 - -

| 0 .70 - -

0.65 - -

0.60 - -

0.55 - -

I I Control• • +PTX100ng/ml

I1

Basal Forskolin Forskolin Isoproterenol(5pM) c + . (lOOpM)

Serotonin p

(lOOpM)FIGURE 5. Effect of pertussis toxin on the serotonin-inducedreduction of forskolin-stimulated cAMP levels and the stimu-lation of cAMP production by isoproterenol in cultured hu-man retinal pigment epithelial cells. Data are mean ± stan-dard error of the mean value from three separate experi-ments carried out in triplicate. *P < 0.01 when comparedwith isoproterenol stimulation in the absence of pertussistoxin by Student's Kest for paired data. **P < 0.05 whencompared with forskolin stimulation of cAMP productionalone by Student's Hest for paired data. cAMP = adenosine3':5' cyclic monophosphate.


5-HT,A Receptors Associated With Human RPE 515

2•aen

1600 -

1400 --

*- 1200

" 1000

sa.cu

800 - -

600 - -

4 0 0 - -o

JS 200

i3 0

I I Isoproterenol^M NECAH I Agonist

+ 5-HT(100MM)

-7 -5 -4 -6 -5

Log [Agonist] M

-4

FIGURE 6. Effect of 5-HT on the stimulation of cAMP produc-tion by various concentrations of isoproterenol and NECA.Data are expressed as mean ± standard error of the meanvalue from three separate experiments conducted in tripli-cate. *P < 0.01 when compared with stimulation of cAMPproduction by isoproterenol alone. cAMP = adenosine 3':5'cyclic monophosphate.

of the 5-HT! family, which decreases AC activity. At pres-ent, there are five members of the 5-HTi family: the 5-HT|A, 5-HT)Da, 5-HT1D/3 (and its rodent homologue the5-HTIB receptor), 5-HTiE, and 5-HT]F subtypes. The re-sults on cultured human RPE cells described here arein good agreement with the pharmacology predicted of5-HT1A receptors.

The maximal reduction of forskolin-stimulatedcAMP production induced by 5-HT is similar to valuesobtained from human iris ciliary processes,27 guinea

-10 -8 -6

Log [Agonist] M

FIGURE 7. Dose response curves for die effect of 5-HT (•)and 8-OH DPAT (O) on die stimulation of cAMP produc-tion by isoproterenol (0.1 //M) in cultured human retinalpigment epithelial cells. The data are expressed as standarderror of the mean values from three independent experi-ments carried out in triplicate. cAMP = adenosine 3':5'cyclic monophosphate.

pig hippocampus,28 and selected cell lines expressinghuman 5-HT1A receptors,29'30 although significantlylower than the maximal-induced effect in transfectedNational Institutes of Health (NIH)-3T331 and certainHeLa cell lines.29 The 5-HT-induced effect was mim-icked by 8-OH DPAT, buspirone, 5-CT, RU24969, andsumatriptan (Fig. 1), and these drugs act similarly onthe stimulated AC in hippocampus from a number ofspecies32"35 and NIH-3T3 cells transfected with human

TABLE 2. Effect of Pertussis Toxin (100 ng/ml) on the 5-HT- and 8-OH DPAT-induced Reductions of Isoproterenol-stimulated cAMPLevels in Cultured Human RPE Cells

% Stimulation of cAMP ProductionRelative to Basal Levels

Drugs Added - PTX + PTX

ControlIsoproterenol (0.1

+ 5-HT (100+ 8-OH DPAT (100 /M)

1046.0 ± 60.3150.9 ± 8.1*649.0 ± 65.6*

17.6 ± 3.51053.7 ± 39.1194.8 ± 12.7f (+29.1%)788.7 ± 80.8f (+21.5%)

cAMP = cyclic adenosine monophosphate; RPE = retinal pigment epithelial.Cells were pretreated with pertussis toxin for 18 hours prior to experimentation. Results are mean ±SEM values from three separate experiments conducted in triplicate.* P < 0.02 versus isoproterenol stimulation of cAMP production by Student's paired Rest.f P < 0.05 versus drug effect on isoproterenol-induced stimulation in the absence of PTX.


516 Investigative Ophthalmology &: Visual Science, February 1997, Vol. 38, No. 2

5-HT1A receptors.31'36 The ability of 8-OH DPAT tomimic 5-HT action has been considered diagnostic of5-HTIA receptors,37'38 although the recent discoverythat the newly cloned 5-HT7 receptor shows high af-finity for 8-OH DPAT39 has led to a reappraisal of itsdiagnostic role. However, 5-HT7 receptors are posi-tively coupled to AC39 and thus unlikely to mediatethe effect of 8-OH DPAT on cAMP metabolism inhuman RPE cells. The observed inhibition of for-skolin-stimulated cAMP production by 8-OH DPAT incultured human RPE is thus strong evidence for thepresence of 5-HTiA receptors in these cells. Buspironebinds with high affinity and selectivity for 5-HT]A re-ceptors40 and reduces forskolin-stimulated cAMP pro-duction in hippocampus32'33 and NIH-3T3 cells ex-pressing human 5-HT]A receptors.31 In contrast,buspirone has no effect on stimulated cAMP levels inChinese hamster ovary cells transfected with human5-HTiA receptors30 and is an antagonist at 5-HT1A re-ceptors in primary cultures of cortical neurons.34 Inhuman RPE cells, buspirone reduced forskolin-stimu-lated cAMP levels by 41.3% (Fig. 1) and thus the hu-man RPE 5-HT1A receptors show greater similarity tothose in mouse hippocampus than those in corticalneurons.34 Sumatriptan has been reported to havehigh affinity and selectivity for 5-HTiB/iD receptors41

but has also been shown to interact with calf hippo-campus 5-HT1A receptors to reduce forskolin-stimu-lated AC.35 A similar effect is evident in human RPEcells (Fig. 1).

The potency order obtained for the action of theagonists on forskolin-stimulated cAMP production(Fig. 2) is in fair agreement with that obtained frommouse and calf hippocampus33'34 and from NIH-3T3cells expressing human 5-HT]A receptors.31 However,important differences are evident in the response ofthese 5-HTiA receptors to the agonists. Most notableis the high potency (low EC50 value) of 5-HT for itsreceptor, which is comparable to that of 8-OH DPAT,in cultured human RPE. In NIH-3T3 cells transfectedwith human 5-HTJA receptors, the potency of 5-HTwas 10 times less than that for 8-OH DPAT31 than it wasin hippocampus.33'34 Furthermore, the half-maximaleffect of 5-CT was observed to be 100-fold lower incultured human RPE cells than that witnessed in othersystems, and the potencies of buspirone and RU24969were at least 10-fold that observed previously (Fig.o \ 31,33,34

The action of 5-HT on forskolin-stimulated cAMPproduction was antagonized by spiroxatrine, spiper-one, propranolol, methysergide, and metergoline(Fig. 3). Spiperone classically has been used to antago-nize 5-HT1A receptor-mediated responses and potentlyattenuates inhibition of forskolin-stimulated AC activ-ity in hippocampus,32"34 and HeLa29 and CHO30 cells

transfected with the human 5-HT]A receptor. Beta ad-renoceptor receptor antagonists also are known tohave high affinity for 5-HTiA receptors, reflecting thesimilarity of these two receptor types, and propranololinhibits 5-HT and 8-OH DPAT action on 5-HTjA recep-tors in rat hippocampus.42 Spiroxatrine has high affin-ity and selectivity for 5-HT1A receptors but has beendescribed variously as an antagonist43'44 and a partialagonist.33'35 In human RPE cells, spiroxatrine appearsto act as a full antagonist and is more potent than arepropranolol and spiperone at inhibiting both 5-HTand 8-OH DPAT action (Fig. 4). The partial agonistsmethysergide and metergoline have been describedas agonists at 5-HTiA receptors in hippocampal neu-rons,33'34 but in human RPE cell cultures act as antago-nists as found in mouse cortical and striatal neu-rons.33'45

The 5-HT-induced inhibition of forskolin-stimu-lated cAMP production was observed to be pertussistoxin-sensitive (Fig. 5) as expected for receptors nega-tively coupled to through Gs proteins. Basal levels ofcAMP were unaffected by pertussis toxin, whereas iso-proterenol-stimulated levels were found to be potenti-ated. Pertussis toxin preincubation attenuated 5-HTaction by near 100%. To prove conclusively that per-tussis toxin is inhibiting G; protein action, however,adenosine diphosphate-ribosylation of the a\ subunitmust be shown.

Studies on the action of 5-HT on isoproterenol-and NECA-stimulated cAMP production providedsome unexpected results. No effect of 5-HT on NECAaction could be observed with either maximal or sub-maximal concentrations of NECA (Fig. 6) or differentconcentrations of 5-HT (data not shown). A similardifferential effect on second-messenger productionwas observed by us for the effect of epidermal growthfactor on the stimulation of cAMP production in hu-man RPE cells where epidermal growth factor wasfound to potentiate the effects of forskolin and isopro-terenol but not of NECA.46 Two explanations for theseeffects are possible. The failure of 5-HT and 8-OHDPAT to modulate NECA-stimulated cAMP produc-tion might result from their respective receptors beingsequestered from each other in the human RPE cellmembrane. The RPE are highly polarized and areknown to show polarized and localized expression ofmembrane proteins (e.g., the Na+/K+ ATPase).4

Thus, the restriction of movement of membrane re-ceptors and effector systems to specific domains is notimprobable. If the A2 adenosine receptor signal trans-duction system is sequestered from the 5-HTJA recep-tors, then this also may explain why the inhibitoryeffect 5-HT on forskolin action fails to reach 100%because a certain level of AC will sequester with theadenosine receptors.


5-HT1A Receptors Associated With Human RPE 517

The recent description of different phenotypesof human RPE cells in culture might offer an alter-native explanation.47'48 Two distinct subpopula-tions of cultured RPE cells have been identified,one showing an epithelial phenotype and the otherfusiform,47 and these subsequently were found toshow heterogeneity in the presence of phosphopro-teins in the cell periphery.48 Using bovine RPE ex-plants, Burke et al48 also identified similar pheno-typic variations in RPE cells in situ. It is possible,therefore, that subpopulations of human RPE cellsin culture might express a different complement ofcell surface receptors. Thus, if one subpopulationexpresses adenosine receptors but not 5-HT1A re-ceptors, no effect of 5-HT on NECA-stimulatedcAMP production would be observed. The pres-ence of two subpopulations also would explain theinability of 5-HT to inhibit fully the forskolin-in-duced response because if only one phenotype ex-presses 5-HT1A receptors, then 100% inhibition ofcAMP production could not be achieved. The abil-ity to separate these two phenotypes of human RPEcell based on differential adhesion47 offers an at-tractive method to assess possible heterogeneity inthe possession of cell surface receptors in the fu-ture.

The action of 5-HT on isoproterenol-stimu-lated cAMP levels is harder to understand becauseat high concentrations, both 5-HT and 8-OH DPATinhibit the action of low concentrations of isopro-terenol (Fig. 7). One possible explanation for theseeffects is that like adenosine receptors, the /3-adre-noceptors are sequestered from the 5-HT1A recep-tors due to either of the possibilities discussedabove and that high concentrations of 5-HT and 8-OH DPAT are able to directly inhibit isoproterenolbinding to the /?-adrenoceptor, thus reducingcAMP production. This, however, would not ex-plain the partial attenuation of the 5-HT-inducedresponse by pertussis toxin. Another explanationmight include the coupling of a low density of 5-HTtA receptors to the /?2-adrenoceptors or the in-teraction of the receptors with different subtypesof G-protein subunit or ACs. However, currently,there is no evidence for either of these possibilities.The actual explanation of this interesting phenom-enon must therefore await future research.

In conclusion, the results presented show thatcultured human RPE cells possess functional 5-HT1A receptors that mediate inhibition of for-skolin-stimulated cAMP production. This is in con-trast to cultured rat RPE cells, which express 5-HT2A receptors coupled to inositol phosphate pro-duction and [Ca2+]i mobilization.20 The role of the5-HT1A receptors associated with human RPE cells

is unknown, although the wide range of effects me-diated by cAMP in vitro suggests they may be in-volved in the control of RPE cell function. Al-though it is unknown how serotonin may reach theRPE cells to trigger responses, several possibilitiesexist. One source may be directly from the seroto-nin-accumulating amacrine cells situated in the in-ner retina,49 whereas another could be from thechoroidal blood supply.50 An alternative source isfrom the centrifugal serotonergic nerve fibers,which originate in the raphe nuclei or the suprachi-asmatic nucleus.51'52 These fibers terminate in theouter plexiform layer in the rat retina,52 fromwhere serotonin could reach the RPE cells moreeasily than from the inner plexiform layer or thechoroid. The presence of this retinopetal pathwaymight suggest some degree of central control ofthe circadian rhythms of the outer retina.

Key Words

adenylate cyclase, cAMP (adenosine 3':5' cyclic monophos-phate), pertussis toxin, retinal pigment epithelial cells, 5-HT]A serotonin receptors

References

1. Bok D. Retinal photoreceptor disc shedding and pig-ment epithelial phagocytosis. In: Ryan SJ, ed. Retina.St. Louis: Mosby; 1994:81-94.

2. Joseph DP, Miller SS. Apical and basal membrane iontransport mechanisms in bovine retinal pigment epi-thelium. JPhysiol (Lond). 1991;435:439-463.

3. Quinn RH, Miller SS. Ion transport mechanisms innative human retinal pigment epithelium. Invest Oph-thalmol VisSci. 1992;33:3513-3527.

4. Bok D. The retinal pigment epithelium: A versatilepartner in vision. J Cell Sci Suppl. 1993; 17:189-195.

5. Edwards RB, Flaherty PM. Association of changes inintracellular cyclic AMP with changes in phagocytosisin cultured rat pigment epithelium. Curr Eye Res.1986;5:19-26.

6. Hall MO, Abrams TA, Mittag TW. The phagocytosisof rod outer segments is inhibited by drugs linked tocyclic adenosine monophosphate production. InvestOphthalmol Vis Sci. 1993;34:2392-2401.

7. Hall MO, Abrams TA, Mittag TW. ROS ingestion byRPE cells is turned off by increased protein kinase Cactivity and by increased calcium. Exp Eye Res. 1991;52:591-598.

8. Miller SS, Hughes BA, Machen TE. Fluid transportacross retinal pigment epithelium is inhibited by cyclicAMP. Proc Natl Acad Sci USA. 1982; 79:2111-2115.

9. Marmor MF, Negi A. Pharmacologic modification ofsubretinal fluid absorption in the rabbit eye. Arch Oph-thalmol. 1986; 104:1674-1677.

10. Frambach DA, Fain GL, Farber DB, Bok D. Beta ad-renergic receptors on cultured human retinal pig-ment epithelium. Invest Ophthalmol Vis Sci. 1990;31:1767-1772.



11. Friedman Z, Hackett SF, Campochiaro PA. Character-ization of adenylate cyclase in human retinal pigmentepithelial cells in vitro. Exp Eye Res. 1987; 44:471-479.

12. Friedman Z, Hackett SF, Linden J, Campochiaro PA.Human retinal pigment epithelial cells in culture pos-sess A2-adenosine receptors. Brain Res. 1989; 492:29-35.

13. Koh S-WM, Yue BYJT, Edwards RB, Newkirk C, ResauJH. Evidence of a functional VIP receptor in culturedhuman retinal pigment epithelium. Curr Eye Res.1995; 14:1009-1014.

14. Friedman Z, Hackett SF, Campochiaro PA. Humanretinal pigment epithelial cells possess muscarinic re-ceptors coupled to calcium mobilization. Brain Res.1988;446:11-16.

15. Osborne NN, FitzGibbon F, Schwartz G. Muscarinicacetylcholine receptor-mediated phosphoinositideturnover in cultured human retinal pigment epithe-lium cells. Vision Res. 1991;31:1119-1127.

16. Feldman EL, Randolph AE, Johnston GC, DelMonteMA, Greene DA. Receptor-coupled phosphoinositidehydrolysis in human retinal pigment epithelium. /Neurochem. 1991; 56:2094-2100.

17. Friedman Z, Delahunty TM, Linden J, CampochiaroPA. Human retinal pigment epithelial cells possess VIvasopressin receptors. Curr Eye Res. 1991; 10:811-816.

18. Feldman EL, Randolph AE. Peptides stimulate phos-phoinositide hydrolysis in human retinal pigment epi-thelium. Invest Ophthalmol Vis Sci. 1993; 34:431-437.

19. Nash MS, Osborne NN. Pertussis toxin-sensitive mela-tonin receptors negatively coupled to adenylate cy-clase associated with cultured human and rat retinalpigment epithelial cells. Invest Ophthalmol Vis Sci.1995;36:95-102.

20. Osborne NN, Fitzgibbon F, Nash M, Liu NP, Leslie R,Cholewinski A. Serotonergic, 5-HT2, receptor-medi-ated phosphoinositide turnover and mobilization ofcalcium in cultured rat retinal pigment epitheliumcells. Vision Res. 1993;33:2171-2179.

21. McKechnie NM, Boulton M, Robey HL, Savage FJ,Grierson I. The cytoskeletal elements of human reti-nal pigment epithelium: In vitro and in vivo. / Cell Sci.1988;91:303-312.

22. Brown BL, Ekins RP, Albano JD. Saturation assay forcyclic AMP using endogenous binding protein. AdvCyclic Nucleotide Res. 1972;2:25-40.

23. Brown BL, Albano JD, Ekins RP, Sgherzi AM. A simpleand sensitive saturation assay method for die measure-ment of adenosine 3':5'-cyclic monophosphate. Bio-chemj. 1971; 121:561-562.

24. Boess FG, Martin IL. Molecular biology of 5-HT recep-tors. Neuropharmacology. 1994; 33:275-317.

25. Hoyer D, Clarke DE, Fozard JR, et al. InternationalUnion of Pharmacology classification of receptors for5-hydroxytryptamine (Serotonin). Pharmacol Rev.1994; 46:157-203.

26. Saudou F, Hen R. 5-hydroxytryptamine receptor sub-types in vertebrates and invertebrates. Neurochem Int.1994;25:503-532.

27. Barnett NL, Osborne NN. The presence of serotonin

(5-HT1) receptors negatively coupled to adenylate cy-clase in rabbit and human iris-ciliary processes. ExpEye Res. 1993;57:209-216.

28. DeVivo M, Maayani S. Stimulation and inhibition ofadenylyl cyclase by distinct 5-hydroxytryptamine re-ceptors. Biochem Pharmacol. 1990; 40:1551-1558.

29. Fargin A, Raymond JR, Regan JW, Cotecchia S, Lef-kowitz RJ, Caron MG. Effector coupling mechanismsof the cloned 5-HT1A receptor. J Biol Chem.1989; 264:14848-14852.

30. Raymond JR, Albers FJ, Middleton JP. Functional ex-pression of human 5-HT1A receptors and differentialcoupling to second messengers in CHO cells. NaunynSchmiedebergs Arch Pharmacol. 1992;346:127-137.

31. Varrault A, BockaertJ, Waeber C. Activation of 5-HT1A

receptors expressed in NIH-3T3 cells induces focusformation and potentiates EGF effect on DNA synthe-sis. MolBiol Cell. 1992; 3:961-969.

32. DeVivo M, Maayani S. Characterization of the 5-HTIA

receptor-mediated inhibition of forskolin-stimulatedadenylate cyclase activity in guinea pig and rat hippo-campal membranes. / Pharmacol Exp Ther. 1986;238:248-253.

33. Schoeffter P, Hoyer D. Centrally acting hypotensiveagents with affinity for 5-HTiA binding sites inhibitforskolin-stimulated adenylate cyclase activity in calfhippocampus. Br J Pharmacol. 1988;95:975-985.

34. Dumuis A, Sebben M, Bockaert J. Pharmacology of 5-hydroxytryptamine-lA receptors which inhibit cAMPproduction in hippocampal and cortical neurons inprimary culture. MolPharmacol. 1988;33:178-186.

35. Boddeke HW, Fargin A, Raymond JR, Schoeffter P,Hoyer D. Agonist/antagonist interactions with clonedhuman 5-HTiA receptors: Variations in intrinsic activ-ity studied in transfected HeLa cells. Naunyn Schmiede-bergs Arch Pharmacol. 1992;345:257-263.

36. Varrault A, Journot L, Audigier Y, Bockaert J. Trans-fection of human 5-hydroxytryptaminel A receptors inNIH-3T3 fibroblasts: Effects of increasing receptordensity on the coupling of 5-hydroxytryptaminel A re-ceptors to adenylyl cyclase. Mol Pharmacol. 1992;41:999-1007.

37. Gozlan H, El-Mestikawy S, Pichat L, Glowinski J, Ha-mon M. Identification of presynaptic serotonin autor-eceptors using a new ligand: 3H-PAT. Nature.1983;305:140-142.

38. Middlemiss DN, Fozard JR. 8-Hydroxy-2-(di-n-propy-lamino)-tetralin discriminates between subtypes of die5-HT[ recognition site. Eur J Pharmacol. 1983;90:151-153.

39. Meyerhof W, Obermuller F, Fehr S, Richter D. A novelrat serotonin receptor: Primary structure, pharmacol-ogy, and expression pattern in distinct brain regions.DNA Cell Biol. 1993; 12:401-409.

40. Zifa E, Fillion G. 5-Hydroxytryptamine receptors. Phar-macol Rev. 1992; 44:401-458.

41. Peroutka SJ, McCarthy BG. Sumatriptan (GR 43175)interacts selectively with 5-HT1B and 5-HT1D bindingsites. Eur J Pharmacol. 1989; 163:133-136.

42. Oksenberg D, Peroutka SJ. Antagonism of 5-hydroxy-


5-HTu Receptors Associated With Human RPE 519

tryptaminelA (5-HT1A) receptor-mediated modula-tion of adenylate cyclase activity by pindolol and pro-pranolol isomers. Biochem Pharmacol. 1988; 37:3429-3433.

43. Nelson DL, Taylor EW. Spiroxatrine: A selective sero-toninlA receptor antagonist. Eur J Pharmacol. 1986;124:207-208.

44. Albert PR, Zhou QY, Van-Tol HH, Bunzow JR, CivelliO. Cloning, functional expression, and mRNA tissuedistribution of the rat 5-hydroxytryptaminelA recep-tor gene. JBiol Chem. 1990;265:5825-5832.

45. Weiss S, Sebben M, Kemp DE, Bockaert J. Serotonin 5-HT| receptors mediate inhibition of cyclic AMP produc-tion in neurons. Eur J Pharmacol. 1986; 120:227-230.

46. Liu NP, Fitzgibbon F, Nash M, Osborne NN. Epider-mal growth factor potentiates the transmitter-inducedstimulation of C-AMP and inositol phosphates in hu-man pigment epithelial cells in culture. Exp Eye Res.1992;55:489-497.

47. McKay BS, Burke JM. Separation of phenotypically

distinct subpopulations of cultured human retinal pig-ment epithelial cells. Exp Cell Res. 1994;213:85-92.

48. Burke JM, Skumatz CMB, Irving PE, McKay BS.Phenotypic heterogeneity of retinal pigment epi-thelial cells in vitro and in situ. Exp Eye Res.1996;62:63-73.

49. Brunken WJ, Jin XT, Pis-Lopez AM. The propertiesof the serotonergic system in the retina. In: OsborneNN, Chader GJ, eds. Progress in Retinal Research. Perga-mon Press: Oxford; 1993; 12:75-99.

50. Rapport MM, Green AA, Page IH. Serum vasoconstric-tor (serotonin): Isolation and characterization. J BiolChem. 1948; 176:1243-1251.

51. Villar MJ, Vitale ML, Parisi MN. Dorsal raphe seroto-nergic projection to the retina. A combined peroxi-dase tracing-neurochemical/high-performance liquidchromatography study in the rat. Neuroscience 1987;22:681-686.

52. Schutte M. Centrifugal innervation of the rat retina.VisNeurosci. 1995; 12:1083-1092.


Pharmacologic Evidenc foer 5-HT1A Receptors …iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/... · Pharmacologic Evidenc foer 5-HT1A Receptors Associated With Human

Documents