23

BrainResearch, 583 (1992) 23-30 Elsevier Science Publishers B.V.

BRES 17630

Stimulation of inositol phosphate formation in cultured human retinal pigment epithelium Richard B. Crook, Mi-Kyoung Song, Liliana P. Tong, Julie M. Yabu, Jon R. Polansky and Ge Ming Lui Cellular Pharmacology Laboratory, Departmentof Ophthalmology, University of California, San Francisco, San Francisco, CA 94143 (USA) (Accepted 26 November 1991)

Key words: Inositol phosphate; Human retina; Retinal epithelium; Neuropeptide

Several hormones, neurotransmitters, and neuropeptides were screened for the ability to stimulate inositol phosphate formation in cultured human retinal epithelial (RPE) cells. Carbachol, vasopressin and thrombin were found to be effective. Treatment of RPE cells with all three agents produced increases in inositol monophosphate, inositol bisphosphate and inositol trisphosphate in the presence of 10 mM LiCl. Carbachol stimulated a 4-fold increase in the total of inositol phosphates at 1 mM. Studies with cholinergic antagonists showed a rank order of 4 DAMP> QNX > pirenzipene > methoctramine, suggesting the presence of M 3 muscarinic receptors. Vasopressin gave a 2.5-fold stimulation at 10 ~M. Agonists of vasopressin were also tested and gave differential responses. Studies using a VI agonist (PIOVP) and a V2 agonist (DAVP) showed DAVP matching the level of stimulation elicited by vasopressin whereas treatment with PIOVP only reached 50% of the vasopressin response. These data suggested the presence of V2 receptors in the RPE cells. Several proteases were tested for their ability to stimulate RPE inositol phosphates. Thrombin caused a 7-fold increase in inositol phosphate formation at 1 U Iml, whereas trypsin and plasmin elicited smaller responses ( - 2-fold). The thrombin effect was blocked by the thrombin-specific inhibitor, hirudin, but not by other protease inhibitors. Several mediators of inflammation such as bradykinin, histamine and serotonin were also tested, and they were ineffective in stimulating inositol phosphate turnover in the RPE cells.

INTRODUCTION The retinal pigment epithelium (RPE), which is located between the choriocapillaries and the neural retina, plays a vital role in maintaining the structural and functional integrity of the photoreceptors. The RPE is capable of performing a variety of physiological and biochemical functions such as transepithelial movement of nutrients and minerals'Y" and secretion of cytokines and proteolytic enzymes3,19,42 . It also fulfills the phagocytotic function which is essential for the turnover of the rod outer segment 29,49. Since the RPE controls the complex environment on which the photoreceptors rely for survival, it is likely that some signal transduction mechanisms exist by which the RPE senses and monitors the extracellular milieu. Both ,a-adrenergic receptors linked to adenylate cyclase and muscarinic receptors have been demonstrated on human RPE cells". Recently, Friedman et al.22 reported that

muscarinic agonists were able to mobilize intracellular calcium release in human RPE cells. We have begun to study the effects of various neurotransmitters, neuropeptides and other compounds for their ability to stimulate the inositol phosphate / diacylglycerol (Ip /DAG) second messenger pathway in the retinal pigment epithelium. The IP jDAG pathway is triggered by receptormediated activation of phospholipase C 6 . The subsequent formation of inositol trisphosphate (InsP3 ) and diacylglycerol (DAG) from phosphatidylinositol 4,S-bisphosphate forms two separate arms of this signal transduction pathway. InsP3 triggers the release of calcium from endoplasmic reticulum stores, whereas DAG activates protein kinase C. These processes result in elevation of intracellular calcium and protein phosphorylation, leading to the physiological responses of the cells? We report here that carbachol, vasopressin and thrombin all stimulated inositol phosphates in the hu-

Correspondence: Ge Ming Lui, Cellular Pharmacology Laboratory, Department of Ophthalmology, USCF, San Francisco, CA 94143, USA.

24 man RPE cells. Studies with specific agonists and antagonists suggest that the RPE muscarinic receptor is the M 3 subtype. Thrombin stimulation of the IP IDAG pathway could be effectively blocked by the specific thrombin inhibitor, hirudin. Inactivation of the proteolytic site of thrombin resulted in diminished stimulation of the inositol phosphate pathway, suggesting that thrombin action may be dependent on its proteolytic activity.

with a rubber policeman. The inositol phosphates were separated from lipids as describedv", Ion exchange chromatography by the method of Berridge" was performed as previously described!', Total inositol phosphates {inositol monophosphate [InsPtl, inositol bisphosphate [InsPzl and inositol trisphosphate [lnsP31were eluted from 1.2 ml AG-1 x 8 columns with 8 ml of 1 M HCOONH 4/0.1 M HCOOH. InsP\, InsPz, and InsP3 were individually eluted as described 11. Elution characteristics of InsP\, InsPz and InsP3 were determined using tritiated standards (Amersharn). One ml of sample was added to 7.8 ml of Scintiverse II (Fisher Scientific) and counted in a Packard Scintillation counter with 35% counting efficiency. Data presentation

MATERIALS AND METHODS Chemicals

Carbamylcholine chloride (carbachol), arginine-vasopressin, substance P, serotonin, histamine, prostaglandin I z, angiotensin and bradykinin were from Sigma Chemical Co. (St. Louis, MO). 4-Diphenylacetoxy-N-methylpiperidine methiodide (4 DAMP), methoctramine, pirenzipine and 3-quinuclidinylxanthene-9-carboxylate (QNX) were from Research Biochemicals (Natick, MA). d(CHz)s [Tyr Et Valz]AVP (SKF 101926)was a generous gift from Ms. Julia Christie of Smith Kline and French Laboratories (King of Prussia, PA). Bombesin was from Bachem Bioscience (Philadelphia, PA). Deamino [ArgBI vasopressin (dAVP) and [Phe z,I1e3,OrnB)vasopressin (PIOVP) were from Peninsula Laboratories (Belmont, CA). [3HIMyoinositol (10-20 CijmmoO was from Amersham Corp. (Arlington Heights, IL). All other chemicals were reagent grade from Fisher Scientific (Pittsburgh, PA). Tissue culture supplies

Dulbecco's modified Eagle's medium (DME-HI6), phosphate buffered saline (PBS), STV (saline, 0.05% trypsin, 0.02% EDTA, pH 7.4) solution, and penicillin-streptomycin (Pen-Strep) were obtained from Gibco (Grand Island, NY). Fetal calf serum (FCS) was obtained from Sterile Systems, Inc. (Logan, UT). Tissue culture plates were from Falcon Plastics (Lincoln Park, NJ). Fungizone was from Squibb (Princeton, NJ). Glutamine, dextran, gentamicin and gelatin were from Sigma Chemical Co. Basic FGF (bFGF) was purified from bovine pituitary as described/".

All experiments were carried out in duplicate, and were done at least twice with comparable results. Data shown are representative experiments whose values are shown by the error bars bracketing the means. Where duplicate values fall inside the figure symbols, error bars are not shown.

RESULTS

Several hormones, neuroactive agents and other compounds known to stimulate the IP IDAG pathway in other cell types were tested for the ability to stimulate inositol phosphate formation in confluent RPE cultures. Carbachol (1 mM), vasopressin (1 p.M) and thrombin (I U Imn all were effective (Fig. 1). Several other compounds known to stimulate the IP IDAG pathway in other ocular cell types, such as histamine!' and bombesin 13 failed to elicit a significant response in RPE cells. Angiotensin, serotonin, substance P, norepinephrine and prostaglandin 12 were also ineffective at the concentrations tested. Stimulation of InsP 1, InsP z and InsP 3 by carbachol, vasopressin and thrombin is shown in Fig. 2. All three compounds evoked early stimulations (~2 min) of

Cell culture

Human RPE cells were obtained and grown as described'" on extracellular matrix (ECM) in 24-well multiwells. ECM was prepared as described by Gospodarowiz et al zs. The RPE cells used in these experiments were stocks derived from 19- to 21-week aborted human fetuses. All experiments were performed within the guidelines of the Human Experimentation Committee at UCSF. Briefly, cultures of RPE cells were maintained in DME-HI6 supplemented with 15% FCS, 300 ngyrnl glutamine, 50 ngyrnl gentamicin, 100 UIml PenStrep, and 2.5 ngyrnl Fungizone. bFGF at concentrations of 1 ngz'ml was added to the cultures every other day. Confluent plates of RPE cells were passaged at a split ratio of 1: 64 for experimental studies. Usually we used the culture at 3-4 weeks post-confluence at which time they started to regain pigmentation. bFGF was withdrawn as soon as the cells reached confluency and the FCS concentration was lowered to 10%, with medium being changed every other day. Inositol phosphate measurement

Three to 4 weeks after the RPE cells reached confluency, the cells were labeled with 15 JLCijrnl [3Hlrnyoinositol for 20-26 h at 37°C. The cells were then changed to medium (without FCS) containing 10 mM LiCI as described by Berridge? and incubated for 5 min at 37°C. Bioactive compounds were added and the cells were further incubated at 37°C for varying times. In antagonist studies, antagonists were added 2-5 min prior to the addition of agonists, To terminate an experiment, medium was aspirated and the cell monolayers were lysed with 100% methanol and scraped off the surface

4000

3000 ~ ~

:J ::1:

..J-

« ::1:

o tJ

2000

>--

1000

CONT THR CAR VAS HIS

ANG SER

ERA SEP NCR

BOM PG1

2

Fig. 1. Effect of various agents on inositol phosphate formation in RPE cells. Confluent cells were given 15 p,Cijml [3H]myoinositolfor 20-26 h, the medium was changed to M199 containing 10 mM LiCI and cells were given the following compounds for 30 min at 37°. The incubations were terminated and total inositol phosphates determined as described in Materials and Methods. CONT, inositol; THR, 1 UIml thrombin; CAR, 1 mM carbachol; VAS, 1 JLM vasopressin; HIS, 100 p,m histamine; ANG, 10 p,m angiotensin; SER, 100 p,M serotonin; BRA, 10 p,M bradykinin; SBP, 10 p,M substance P; NOR, 10 p,M norepinephrine; BOM, 10 p,M bombesin; PGI 2 , 10 p,M prostaglandin I z. Data are means of duplicate cultures whose spread is indicated by the error bars.

30000T

~

A

.>:

200001 10000 01

160 ("tl

0.. til

z

I

I

t

:::M~ I

I

I

4

8

12

I

I

I~

1001 0

'=oj

I

.

I

-----l

2500'-

12000T

/

12500 I

I

I

B

10000 8000

/'

I

I

I

C

6000

I

I

I

I

I

20

10

90

7011" I I

I

16 20 24

I

I

2832

50 0

4

8

12

TIME

16 20 24

28 32

701 I I 048

I

I

I

12 16 20

I

24

I

I

28 32

eMI N)

Fig. 2. Effect of carbachol, vasopressin and thrombin on stimulation of InsP 1 , InsP z and InsP 3 formation. Confluent cells treated as described in the legend to Fig. 1 were given 10 mM LiCI for 30 min and either (A) 1 mM carbachol, (B) llA-M vasopressin or (C) 1 U Iml thrombin for various lengths of time. The incubations were stopped and InsP1 , InsPz and InsP3 analyzed as described (ref. 11).

~

26 InsP3 followed by somewhat slower increases in InsP2 (3-5 min). InsP 1 in all cases rose continuously for at least 30 min after a lag of one to several minutes. These kinetics are consistent with the general concept that InsP3 is formed prior to InsP 2 and Insp 16 • The dependence of inositol phosphate formation on vasopressin and carbachol concentration is shown in Fig. 3A Vasopressin achieved a maximal increase of 2.2-fold at 1 iLM, with an EC so (half-maximal concentration) of about 100 nM. Carbachol achieved about a 4-fold increase at 1 mM, with an EC so of roughly 30 iLM. These ECsos and maximal concentrations are similar to those of these agents on other ocular cells':', The effect of thrombin concentration on inositol phosphate formation is shown in Fig. 3B. Thrombin caused the largest stimulation of any compound tested: a 7-fold increase at 1 U /ml (10 nM), with an EC so of 0.3 U 1m!. Thrombin is a serine protease which shares common active site characteristics with other serine proteases", To assess the specificity of the inositol phosphate response to thrombin , we surveyed a number of proteases for the ability to stimulate the IP IDAG pathway in RPE cells (Fig. 4A). Plasmin and trypsin were found to stimulate RPE inositol phosphates, whereas collagenase , papain , chymotrypsin, thermolysin and Ca2+-proteinase had no effect at the concentrations' tested. Further study of the responses to trypsin and plasmin (Fig. 4B) showed that at concentrations which had no deleterious effects on cell attachment or viability (as measured by trypan blue exclusion, data not shown) both proteases elicited about a two-fold increase in inositol phosphate formation in RPE cells. Agonist, antagonist and inhibitor studies Cholinergic antagonists were employed to establish the rank order of effectiveness in inhibiting ca'rbacholspecific stimulation of inositol phosphates in RPE cells. 3500

4 0CXl

A

B

3 000 CA ReAOtOL

2500

3 0 00

2000 2000

1000

10 00

500

°--I--+---_r_--+--+--t----t 10 -9 - B - 7 - 6 -5 -4 - 3 LOG [ AGON IST] ("1 )

o+-_

-

--+--

.0 '

.10

1,0

-+-----i 10

100

[ TH ROMBIN) ( ~ ITSJML)

Fig. 3. Effect of carbachol, vasopressin and thrombin concentration on total inositol phosphate formation. Cells prepared as in the legend to Fig. 1 were given 10 IJlM LiCI and various concentrations of carbachol, vasopressin or thrombin for 30 min. The incubations were stopped and total inositol phosphates measured as in Fig. I. A: vasopressin (4) and carbachol (0). B: thrombin.

A

4000

3 0 00 --l

~

:2 a, :.>

20 0 0

1000

CONT THR Pl.A TRY COL PAP CHY THM CA- PR

B 900

:::; 700

:2

-... ~

11.

~ 50 0

300 -I-----+----='-----"'-t----r---+_ .01 0.1 1.0 10 .001 [ PROT EASE] (UNITS/ ML)

Fig. 4. Effect of various proteases on total inositol phosphates formation. Cells prepared as in Fig. 1 were exposed to 10 mM LiCI and various proteases for 30 min. Incubations were terminated and inositol phosphates assayed as in Fig. I. Concentrations were in unitsyrnl as defined by the vendor. A: CONT, control; THR , 1.0 V I ml thrombin, PLA, 1.0 U I ml plasmin; TRY, 1.0 V I ml trypsin; COL, 1.0 Uyrnl collagenase; PAP, 1.0 V lml papain; CRY, 0.9 V lml a-chymotrypsin; THM, 1.0 Uyrnl thermolysin; CA-PR, 1.0 U I ml Ca2 +-protease. B: plasmin and trypsin dose-responses. Various concentrations of plasmin ( 0 ) or trypsin (e) were given to cells and inositol phosphates measured.

Pirenzipine, an M 1 specific antagonist.", methoctramine, an M, antagonist 35.37, QNX, an M 2-M 3 antagonist" and 4-DAMP, an M 3 specific antagonist'? were tested. Fig. 5 shows that 4-DAMP was the most potent antagonist, with an IC so (half-maximal inhibitory concentration) of 3 nM and complete inhibition of carbachol stimulation at 100 nM. QNX showed an IC so of 30 nM and a complete inhibition at 10 J.LM Pirenzipine was much less potent, with an IC so of 3 J.LM and complete inhibition at 100 iLM, while methoctramine showed the least antagonism, with an IC so at 25 J.LM and complete inhibition at 1 mM. These data suggest the presence of M 3 muscarinic receptors in human RPE cells. To characterize the vasopressin receptor on RPE cells, two agonists of vasopressin were tested. Deamino [Arg8]vasopressin (dAvP), a potent Vz-specific agonist in anirnals'", and [Phe Z ,Ile 3,Orn8]vasopressin (PIOVP), a V1 agonist in animals'" , were compared to vasopressin for their ability to stimulate the inositol phos-

27 4000

20 00

LEUPEP

3000

15 00 ~

..J

}:

--.

100 0

}:

.(l.

~

1000

50 0

o 4--

--+-

- 10

-9

-

+--

-+--

-

I-----l--6

-7

-8

--+-4

-5

0 +----4-

-----i

-9

-3

--+--

-6

-+---