Retinaldehyde, a Potent Inhibitor of Gap Junctional Intercellular Communication

Abstract

Retinaldehyde and retinoic acid are derivatives of vitamin A, and retinaldehyde is the precursor for the synthesis of retinoic acid, a well-known inhibitor of gap junctional intercellular communication. In this investigation, we asked the question if retinaldehyde has similar effects on gap junctions. Gap junctional intercellular communication was measured by scrape-loading and preloading dye-transfer methods, and studies were carried out mainly on cultured liver epithelial cells. Retinaldehyde was found to be a more potent inhibitor (dye transfer reduced by 50% at 2.8 μM) than retinoic acid (dye transfer reduced by 50% at 30 μM) and glycyrrhetinic acid (dye transfer reduced by 50% at 65 μM). Both the 11-cis and all-trans forms of retinaldehyde were equally effective. Retinaldehyde inhibited dye transfer of both anionic Lucifer yellow and cationic Neurobiotin. Inhibition by retinaldehyde developed in less than two minutes at 50 μM, but unlike the reported case with retinoic acid, recovery was slower, though full. In addition to liver epithelial cells, retinaldehyde inhibited gap junctional communication in lens epithelial cells, retinal pigment epithelial cells and retinal ganglion cells.

Keywords:

• gap junctions
• glycyrrhetinic acid
• retina
• retinaldehyde
• retinoic acid

INTRODUCTION

Gap junctions are clusters of intercellular channels that permit passage of small molecules of up to one thousand Daltons between neighboring cells (5, 13, 17). A channel is formed by pairing two half channels in the apposed membranes of adjacent cells, each made up of six units of gap junction proteins called connexins. Gap junctions permit rapid passage of ions, metabolites, and signaling molecules between the connected cells and permit the cells to act synchronously.

Gap junctional intercellular communication (GJIC) can be increased or decreased by phosphorylation of connexins by specific protein kinases (18). In addition, a number of lipophilic agents are known, retinoic acid (all-trans form) among them, which inhibit GJIC (13, 28, 32). Retinoic acid is a vitamin A derivative that inhibits GJIC in a nonspecific manner, that is, the inhibition occurs on a wide variety of cells, and apparently does not depend upon the type of connexin protein that makes up the channels (28, 30, 36).

Retinaldehyde is another vitamin A derivative, and a precursor of retinoic acid (22, 25). Its role in vision, as the chromophore of the light receptor protein, rhodopsin, is well known (33). In addition, in the last few years retinaldehyde has been shown to inhibit epithelial cell proliferation (29), to inhibit mammalian replicative DNA polymerase (24), to have antibacterial activity (27), and to inhibit cyclic nucleotide-gated cation channels (7). In view of this broadening role of retinaldehyde as a physiological and pharmacological agent, and given the known role of retinoic acid as an inhibitor of gap junctions, we tested the effect of retinaldehyde on gap junctions in a variety of cultured cells, and show here that retinaldehyde is a potent inhibitor of GJIC, about 10-fold more effective than retinoic acid.

MATERIALS AND METHODS

Cell Culture

WB-F344 rat liver epithelial cells (20), human lens epithelial cells (15), untransformed human retinal pigment epithelial cells (1, 9), and RGC-5 rat retinal ganglion cells (16) were cultured at 37°C in minimum essential medium (GIBCO), pH 7.4, supplemented with 10% fetal bovine serum, in a humidified incubator with 5% CO₂/air mixture. Cells were cultured in 35-mm plastic petri dishes (Corning) and used as confluent monolayers.

Solutions

Stock solutions of all-trans-retinoic acid, all-trans-retinaldehyde, 18α-glycyrrhetinic acid and glycyrrhizic acid (all from Sigma) were made in dimethyl sulfoxide. Stock solution of 11-cis-retinaldehyde (gift from Roche) was made in ethanol. Actual concentrations of retinoic acid and retinaldehyde stocks were determined by measuring optical density of solutions diluted into hexane. The following molar extinction coefficients were used in calculating the concentrations: all-trans-retinoic acid, 45,000 at 351 nm (14); 11-cis-retinaldehyde, 26,360 at 365 nm (3); and all-trans-retinaldehyde, 48,000 at 368 nm (31).

Dye-Transfer Measurements

Scrape-loading of the dye was done according to a standard procedure (8). Briefly, cells were washed twice in PBS (phosphate-buffered saline containing calcium chloride and magnesium chloride, both at 0.1 g/liter) and incubated for 30 minutes (unless indicated otherwise) at 37°C with the desired concentration of the test substance. For each treatment, three 35-mm culture dishes were used. The cells were washed and overlaid with Lucifer yellow (0.5 mg/ml) (Sigma) or Neurobiotin (25 mg/ml) (Vector) and three to four scrapes were made in each dish and dye transfer was allowed to occur for four minutes. In some experiments, Texas Red-dextran was also included in the dye solution to serve as an impermeant fluorophore. The dye solution was then aspirated out and the dishes were rinsed twice with PBS. The cells were fixed in 4% paraformaldehyde for 15–30 minutes. After fixing, the cells treated with Lucifer yellow were washed with PBS, cover slipped and photographed with a digital camera in phase and UV modes using the 20x or 40x objective of a Nikon epifluorescence microscope. Cells scrape-loaded with Neurobiotin were visualized according to a reported method (2); briefly, cells were permeabilized with 2% BSA/0.25% Triton x-100 in PBS for 15 minutes at 37°C, washed twice with PBS and incubated with 100-fold diluted FITC-conjugated streptavidin (Pierce) for 30 minutes at 37°C, finally washed with PBS, cover slipped and photographed. The cells remained confluent after the treatment and the processing.

Preloading was done according to a standard procedure (10). Briefly, a confluent monolayer of liver epithelial cells in a 35-mm dish was labeled for 30 minutes at 37°C in a solution of 5 μM calcein AM (calcein acetoxymethyl ester; Molecular Probes) and 10 μM DiI (1,1′-dioctadecyl-3,3,3′-3′-tetramethylindocarbocyanine perchlorate; Molecular Probes) in 0.3 M glucose. The cells were then washed twice with 0.3 M glucose, trypsinized for 3 min, and suspended in growth medium (DMEM +10% fetal bovine serum). Labeled cells were diluted 1:1000 with unlabeled cells in growth medium without or with a desired concentration of all-trans-retinaldehyde and about two million cells were plated in each of three 35-mm dishes so that they formed a confluent monolayer when settled. After 90 min, the cells were fixed for 10 min in 4% paraformaldehyde, cover slipped and photographed.

Data Analysis

Data analysis was done on digital images of labeled cells using UTHSCSA Image Tool (University of Texas Health Sciences Center, San Antonio, Texas, USA). For dishes in which scrape-loading was done with both Texas Red-dextran and Lucifer yellow, cells labeled with either dye were counted in 16–24 randomly picked 7-cm × 2-cm fields of the digital images and the extent of dye coupling was determined by the ratio of labeling with permeant dye (Lucifer yellow) to impermeant dye (Texas Red-dextran). Coupling was considered inhibited when the ratio was reduced, and 100% inhibited when it was reduced to one. In some experiments, an impermeant dye was not used and the data shown represent mean and standard deviation of the number of labeled cells. For preloading experiments, dye coupling was defined as the average number of recipient cells (labeled with calcein only) per donor cell (labeled with DiI and calcein).

RESULTS

Retinaldehyde Inhibits Dye Transfer Between Liver Epithelial Cells

Figure 1A shows dye transfer in cells scrape-loaded with Lucifer yellow and Texas Red-dextran. The cells were either untreated or treated with 10 μM all-trans-retinaldehyde. As shown, retinaldehyde inhibited dye transfer almost completely, reducing dye coupling from 7.1 ± 1.0 in the control to 1.1 ± 0.1 in the treated (n = 3). Dimethyl sulfoxide, in which retinaldehyde was dissolved, did not have any effect on dye transfer at 0.5%. The concentration of the solvent carried into the assays from stock solutions was below this concentration in all experiments. Neither the solvent nor the inhibitor changed the pH of the assay medium.

1 Inhibition of dye transfer by retinaldehyde as measured by the scrape-loading method. (A) Dye transfer was measured between liver epithelial cells treated for 30 min with vehicle (control) or with 10 μM all-trans-retinaldehyde and scrape-loaded with Lucifer yellow and Texas Red-dextran. (B) Dye transfer of scrape-loaded Neurobiotin was measured between liver epithelial cells treated for 30 min with vehicle (control) or 10 μM retinaldehyde. (See Color Plate V).

V CELL COMMUNICATION & ADHESION VOLUME 11, NUMBER 1. COLOR PLATE V. See S. Pulukuri and A. Sitaramayya, Figure 1.

Retinaldehyde Inhibits Transfer of Neurobiotin

In order to test if retinaldehyde's effect is limited to the transfer of anionic molecules like Lucifer yellow, we measured its influence on the transfer of cationic Neurobiotin. As shown in Figure 1B, 10 μM retinaldehyde strongly inhibited the transfer of Neurobiotin, as it did with Lucifer yellow (Figure 1A).

Verification of Inhibition by Preloading Method

The influence of all-trans-retinaldehyde on dye transfer was investigated also by the preloading method. As shown in Figure 2, transfer of calcein from labeled cells to unlabeled cells was completely inhibited by 10 μM retinaldehyde. Dose-response analysis (Table 1) showed that partial inhibition was observed at concentrations below 2 μM and complete inhibition at 5 μM. Retinaldehyde appeared to cause stronger inhibition in the preloading method than in the scrape-loading method (Figure 1) probably because of the longer incubation period (90 min in preloading and 30 min in scrape-loading).

2 Inhibition of dye transfer by retinaldehyde as measured by the preloading method. Liver epithelial cells preloaded with DiI and calcein AM were mixed with unlabeled cells and plated in the absence or presence of 10 μM retinaldehyde. Dye transfer was measured 90 min later in the monolayers formed. (See Color Plate VI).

VI CELL COMMUNICATION & ADHESION VOLUME 11, NUMBER 1. COLOR PLATE VI. See S. Pulukuri and A. Sitaramayya, Figure 2.

Inhibition of dye coupling by retinaldehyde

Retinaldehyde	Dye coupling
(μM)	(Number of recipient cells / donor cell)
None (control)	6.5 ± 2.3 (41)
0.5	5.1 ± 1.7 (36)
2.0	1.2 ± 1.2 (33)
5.0	No recipient cells observed

Dye coupling was measured by the preloading method. The numbers in parentheses indicates the number of DiI–labeled cells counted.

Dose Response of Inhibition by Retinaldehyde and Other Inhibitors of GJIC

These experiments were done using the scrape-loading method with Lucifer yellow as the tracer. Figure 3 shows reduction of dye transfer in response to treatment with different concentrations of retinaldehyde as well as retinoic acid and glycyrrhetinic acid, two well-known inhibitors of GJIC. Retinaldehyde reduced the dye transfer by 50% at 2.8 μM, retinoic acid, at 30 μM, and glycyrrhetinic acid, at 65 μM. Retinaldehyde inhibited dye transfer fully at concentrations below 20 μM, similar to the observation made in Figure 1A; however, consistent with earlier reports, dye transfer was not comparably reduced by glycyrrhetinic acid or retinoic acid even at higher concentrations (4, 28, 32). Glycyrrhizic acid, an analog of glycyrrhetinic acid, was used as a control for glycyrrhetinic acid (32); it had no effect on dye transfer at concentrations up to 100 μM.

3 Dose dependence of inhibition of GJIC. Liver epithelial cells were treated for 30 min with different concentrations of retinaldehyde (▴), retinoic acid (•) or glycyrrhetinic acid (▪), followed by measurement of GJIC by the scrape-loading dye-transfer method. Lucifer yellow was used as the dye.

Both All-Trans and 11-Cis Isomers of Retinaldehyde were Effective Inhibitors

Retinaldehyde is present in the body mostly in two isoforms, the 11-cis and the all-trans. All of the above studies were done with all-trans-retinaldehyde. In order to test if the inhibition of GJIC is isoform specific, we also tested the influence of 11-cis-retinaldehyde and observed that it was as effective as all-trans-retinaldehyde. At 10 μM, all-trans-retinaldehyde reduced dye coupling from 7.1 ± 1.0 cells to 1.1 ± 0.1 cells and 11-cis-retinaldehyde reduced it to 1.2 ± 0.2 cells (n = 3).

Retinaldehyde Did Not Influence Cell Viability

The effect of retinaldehyde on GJIC could have been due to cytotoxicity. In order to test this, a monolayer of cultured cells was treated with 10 μM all-trans-retinaldehyde or vehicle for 30 min, washed with PBS and cell viability was determined by the Trypan blue exclusion method. No difference was found between treated and untreated cells: with more than two million cells counted per sample, 94.7 ± 2.4% of cells were found viable in the control and 95.3 ± 2.3% in the treated (n = 3). Also, there appeared to be no long-term toxic effect of the treatment as both control and treated cells appeared healthy 24 hr after the treatment and their dye transfer ability was similar.

The effect of retinaldehyde could possibly be simply due to the aldehyde group. In order to test this possibility, dye transfer was measured following treatment for 30 min with 100 μM formaldehyde. It was observed that formaldehyde had no influence on dye transfer: dye coupling was 7.1 ± 1.0 in the control and 7.0 ± 0.2 in the treated (n = 3).

Time Course of Inhibition and Recovery

In order to gain insight into the mechanism of action of retinaldehyde, the kinetics of inhibition and recovery from inhibition were investigated. As shown in Figure 4, inhibition was very rapid; dye transfer was reduced by about 50% in about 2 min and 85% in 6 min. However, upon removal of retinaldehyde and incubation in retinaldehyde-free growth medium, the ability to transfer Lucifer yellow recovered more slowly; half-maximal activity recovered in about 2 hours, and near complete recovery in about 6 hours.

4 Kinetics of inhibition and recovery. Dye transfer was measured between liver epithelial cells following treatment with 50 μM retinaldehyde for different periods of time up to 30 min. Cells treated for 30 min with retinaldehyde were washed (washout), incubated in retinaldehyde-free growth medium for various periods of time and dye transfer was measured. Lucifer yellow was used as the dye.

Effect of Retinaldehyde on GJIC in Other Cultured Cells

In order to determine if the effect of retinaldehyde is limited to liver epithelial cells, its effect on dye transfer was investigated also on cultured human lens and pigment epithelial cells and on rat retinal ganglion cells. These experiments were done by scrape-loading Lucifer yellow in the case of lens and pigment epithelial cells. Neurobiotin was used for ganglion cells which did not transfer Lucifer yellow. Table 2 shows that retinaldehyde reduced dye transfer between all three cell types.

Effect of retinaldehyde on dye transfer in different cell types

Retinaldehyde (μM)	HLE	HRPE	RGC
None, control	6.2 ± 1.8	4.8 ± 2.0	14.5 ± 1.6
10	2.9 ± 1.4	1.9 ± 1.1	5.9 ± 1.2
25	1.7 ± 1.0	1.9 ± 1.0	3.0 ± 0.9
50	1.5 ± 1.4	0.9 ± 0.9	1.4 ± 0.5

HLE, Human lens epithelial cells; HRPE, human retinal pigment epithelial cells; RGC, rat retinal ganglion cells. Dye transfer was measured by the scrape-loading method using Lucifer yellow for HLE and HRPE and Neurobiotin for RGC. Data shown are mean and standard deviation of the number of labeled cells from 24 fields for HLE and HRPE and 12 fields for RGC.

DISCUSSION

Vitamin A and its derivatives, including retinaldehyde, are usually present in the body bound to carrier proteins or receptors (11, 21, 25, 33, 35). Retinaldehyde, in its 11-cis form, is present in the outer segments of retinal photoreceptor cells at 2–3 mM concentration (19), most of it bound to rhodopsin and converted to the all-trans form upon light activation. When dark adapted retina is activated by bright light, there is a potential for rapid release of a large amount of this all-trans-retinaldehyde (26, 34). Our goal in the present experiments was to test if free retinaldehyde could influence GJIC.

The most significant observation from these experiments is that retinaldehyde is a potent inhibitor of GJIC. With dye transfer reduced by half at about 3 μM concentration, it is 10 and 20 times more effective than retinoic acid (30 μM) and glycyrrhetinic acid (65 μM), respectively. Our results on retinoic acid and glycyrrhetinic acid are in agreement with earlier reports of their effects on GJIC, though they were shown to be more or less effective in some reports (4, 6, 28, 32, 36–38). In general, the magnitude of inhibition depends upon concentration of the inhibitor as well as the length of incubation. At and below 10 μM concentration and after longer periods of incubation, retinoic acid is also known to stimulate GJIC (4, 23). In the present study, all three inhibitors were tested under identical conditions and retinaldehyde appeared to be a much more effective inhibitor than retinoic acid and glycyrrhetinic acid. Unlike retinoic acid, retinaldehyde did not stimulate GJIC at low concentrations (between 2 and 10 μM).

The molecular mechanism of inhibition of GJIC by retinaldehyde is not clear. It is, however, not due to cytotoxicity, and also not simply due to the aldehyde group. Retinaldehyde may function like other lipid-soluble substances that inhibit gap junctions, by partitioning into the membrane and affecting its physical properties (13, 32). The rapid kinetics of inhibition and relatively slower recovery resembles the effects of oleamide, another inhibitor of GJIC, thought to function by perturbing the lipid environment of connexins (12). It is, however, also possible that retinaldehyde binds with high affinity to some connexins, explaining the slow recovery after washout.

Retinaldehyde inhibits dye transfer in different cell types tested. However, whether gap junctions made of different connexins exhibit distinct sensitivity to retinaldehyde remains to be investigated.

Both the 11-cis and all-trans forms of retinaldehyde inhibited GJIC just as they both were shown to inhibit cyclic GMP-gated channels in photoreceptor cells (7). This lack of stereospecificity permits both to be used as pharmacological tools in the study of gap junctions, but the physiological significance may be restricted to the effects of the all-trans form. The 11-cis form is present at 2–3 mM concentration in the photoreceptor cells of retina, bound to rhodopsin (19). When released after exposure of retina to bright light, it is in the all-trans form. Given the potential for release in large amounts, and its accumulation for a period of time before conversion back to the 11-cis form (26, 34), all-trans-retinaldehyde might influence GJIC between neighboring photoreceptor and pigment epithelial cells. In the present study, we have already shown that all-trans-retinaldehyde inhibits GJIC between pigment epithelial cells; future studies will determine its potential effects on gap junctions between photoreceptor cells.

ACKNOWLEDGEMENTS

This study was supported by grants from the National Eye Institute (EY 07158 and EY 014803). We thank Drs. James E. Trosko and Brad L. Upham of Michigan State University, East Lansing, MI, USA, for demonstrating to us the technique of scrape-loading dye-transfer and its use as a quantitative tool. We also thank Dr. Diane F. Matesic of Mercer University, Atlanta, GA, USA, for many helpful discussions.