Neural serotonin stimulates chloride transport in the rabbit corneal epithelium

Transcription

1 Neural serotonin stimulates chloride transport in the rabbit corneal epithelium S. D. Klyce, K. A. Palkama, M. Harkonen,* W. S. Marshall, S. Huhtaniitty, K. P. Mann, and A. H. Neufeld** Evidence is presented that serotonin acts as a neurotransmitter in the cornea of the adult rabbit. Serotonin was localized to granules in a sparse population of subepithelial corneal nerves by an electron microscopic histochemical procedure. Significant endogenous levels of serotonin and its principal metabolite, 5-hydroxyindoleacetic acid, were detected in the central cornea by a fluorometric assay. Exogenous serotonin stimulated ion transport by the corneal epithelium. This effect was potentiated by monoamine oxidase inhibition and was unaffected by an a-adrenergic receptor antagonist. Serotonin-stimulated ion transport was inhibited by the specific antagonist, methysergide, and by the replacement of Cl~ with an impermeable anion. In tracer experiments, the serotonin-stimulated ion transport was shoion to be caused by increased epithelial Cl~ secretion. The serotonin response was partially inhibited by the ^ adrenergic antagonist, timolol. In a companion article, 16 assay of corneal cyclic AMP showed stimulation of cyclic AMP synthesis by serotonin, inhibition by the specific antagonist, lysergic acid diethylamide, and potentiation by monoamine oxidase inhibition. We postulate that specific serotonergic receptors are present in the corneal epithelium and that activation of these receptors by serotonin released from serotonergic neurons increases the level of cyclic AMP, tohich stimulates active Cl~ secretion by the corneal epithelium. (INVEST OPHTHALMOL VlS SCI 23: , 1982.) Keywords: serotonin, cornea, chloride, transport, corneal nerves, methysergide, epithelium, epinephrine, timolol, rabbit, nialamide From the Lions Eye Research Laboratories, Louisiana State University Eye Center, New Orleans, La., *the Department of Clinical Chemistry, University of Helsinki, Helsinki, Finland, and **the Eye Research Institute of Retina Foundation, Boston, Mass. Supported by U.S.P.H.S. research grants EY03311, EY02360, and EY02377, the Louisiana Lions Eye Research Foundation, S. Jurelius Foundation, and the Medical Research Council of the Academy of Finland. Submitted for publication Aug. 20, Reprint requests: Stephen D. Klyce, Ph.D., LSU Eye Center, 136 South Roman St., New O.'-ans, La Portions of this work were reported at the Annual Meeting of the Association for Research in Vision and Ophthalmology, May 2, 1981, Sarasota, Fla. 1 he modulation of corneal epithelial function by an adrenergic pathway has been studied extensively. Stimulation of the superior cervical ganglion is known to suppress corneal mitotic activity. 2 ' 3 The presence of adrenergic corneal nerves has been detected histochemically in a variety of species. 4 ' 5 Although adrenergic neurons have not been found in adult primate corneas with this histochemical procedure, this negative result could be due to inadequate sensitivity of the method. Recent neuroanatomic evidence has shown that intrastromally injected horserad /82/ $01.20/ Assoc. for Res. in Vis. and Ophthal., Inc. 181

2 Invest. Ophthalmol. Vis. Sci. August Klyce et al. & 1 Fig. 1. Dense core vesicles associated with adrenergic nerves demonstrated with the KMnO4 reaction in the rabbit cornea. Nerves in the central stroma are unmyelinated and are often only partially ensheathed by Schwann cells. Several dense core vesicles appear in the fiber at the top of the figure. (Bar = 100 nm.) ish peroxidase appears a day later in the ipsilateral superior cervical ganglion cell bodies, lending additional support to the view that sympathetic fibers innervate the adult rabbit cornea.6 In addition to the antimitogenic effect of catecholamines, /3-adrenergic agonists have been shown to stimulate Cl~ transport in the frog, rabbit, and human corneal epithelia.7^9 The Cl~ transport response involves activation of adenylate cyclase and produces a specific Cl~ permeability increase at the outer epithelial surface.10 However, addi- tional physiologic roles of the /3-adrenergic pathway in the cornea have not been forthcoming. Although cholera enterotoxin has been shown recently to stimulate epithelial wound healing through an action involving cyclic AMP, there is no evidence that /3-adrenergic stimulation is needed. 11 The sensory fibers may participate in the process of healing, since trigeminal ganglion ablation in the rabbit significantly delays corneal epithelial wound healing and increases epithelial fluorescein permeability.12 The superior cervical ganglion innervates a

3 * Volume 23 Number 2 Serotonin-stimulated chloride transport 183 Fig. 2, FGD reaction in the rabbit cornea. In this method for localizing 5-hydroxytryptamine, the electron density of the specimen is extremely low except at the dense core vesicles. The periphery of the vesicles (arrows) shows only a weak electron density. Collagen fibers (F) and a stromal keratocyte (K) are detectable. (Bar = 100 nm.) variety of tissues, including the pineal gland. Sympathetic fibers in the pineal gland have been shown to contain both norepinephrine and serotonin13"15; thus serotonin was examined as a potential neurotransmitter in the cornea. In this article we present the evidence that serotonin functions as a neurotransmitter in the cornea and thereby influences corneal epithelial function. The accompanying article describes the evidence that serotonin stimulates the synthesis of cyclic AMP, which mediates the physiologic response. 16 The com- bined evidence supports the novel finding of specific serotonergic receptor-activated, cyclic AMP-mediated stimulation of chloride transport by the corneal epithelium. Methods Adult New Zealand White rabbits (3 to 5 kg) were killed with an overdose of phenobarbital immediately prior to enucleation of the globes. The following pharmacologic agents were used: serotonin oxalate or HC1 (salt) (Sigma Chemical Co., St. Louis, Mo.), nialamide (NM; Sigma), methysergide bimaleate (MSD, Sansert; Sandoz Pharmaceuticals, E. Hanover, N. J.), timolol ma-

4 184 Klyce et al. Invest. Ophthalmol. Vis. Sci. August 1982 Table I. Concentrations of serotonin (5-HT) and 5-HIAA in the cornea compared with levels in other tissues* Compound Cornea Rabbit (n = 8) Iris-ciliary body Cerebral cortex Rat 18 Rabbit 30 uvearetina Hypothalamus 5-HT 5-HIAA 0.41 ± ± ± ± * Values are expressed as means ± S.E.M. in /um/kg wet weight. Table II. Effect of serotonin on corneal electrica parameters* SCC (fia/cm 2 hr) PD (mv) Resistance (kit cm 2 ) Control period Serotonin (10" 4 M) Change p value 2.7 ± ± ± ± ± ± 0.9 > ± ± ± 0.4 <0.01 Values are means ± S.E.M., n = 16 corneas. leate (TIM; Merck, Sharp & Dohme, West Point, Pa.), phentolamine hydrochloride (Ciba Pharmaceutical Co., Summit, N. J.), and /-epinephrine bitartrate (Sigma). Electrophysiology. Corneas were atraumatically dissected and mounted in double-sided Lucite chambers as previously described. 8 Circulation through water jackets maintained the tissue at 35 C. The bathing solution used in these experiments had the following composition: 99.7 mm NaCl, 3.7 mm KC1, 6.92 mm Na 2 SO 4, 20 mm NaHCO :i> 0.6 mm K 2 HPO 4) 25 mm Na/HEPES (ph 7.4), 1.4 mm Ca gluconate, 0.61 mm MgSO 4, and 26 mm glucose (all from Sigma). Cl~-free medium was produced by SO 2 4 ~ substitution, with added sucrose to maintain solution osmolarity (305 mosm). The media were bubbled with 95% O 2 /5% CO 2 and had a ph of 7.4. Corneal short-circuit current (SCC) and resting potential (PD) were measured with dual automatic current/voltage clamps (D. Lee, Inc., Sunnyvale, Calif). The SCC and PD measured across the whole cornea accurately reflect transport processes arising from the epithelial cell layer, since these properties are not altered when the endothelial layer is removed. 17 Unidirectional fluxes of 36 C1 were measured as described previously, 8 with 200 /A samples taken from the less radioactive side every 15 min. During the flux experiments, the specific activity of the more radioactive side (nominally 30 /xci/meq Cl~) varied by less than 1%. Biochemistry Serotonin and 5-hydroxyindoleacetic acid (5- HIAA) assay in cornea and iris-ciliary body. Tissue concentrations of serotonin and its principal metabolite, 5-HIAA, were determined by the method introduced by Curzon and Green, 18 with modifications. Central mil-thickness corneal tissue samples and pieces of the iris-ciliary body complex (50 to 70 mg) were rinsed, rapidly frozen, weighed at 25 C, and then homogenized in 1 ml of cold acidified n-butanol with a small glass homogenizer. After centrifugation for 5 min at 1100 X g, 0.9 ml of the supernatant was transferred to glass-stoppered tubes containing 1.8 ml ofn-heptane and 145 fi\ of 0.1M HC1 with 0.1% L-cysteine. After shaking for 5 min, the two phases were separated by centrifugation as before. For the determination of serotonin levels, 25 /AI of the lower aqueous phase were pipetted into 0.004% o-phthalaldehyde (OPT) in 10M HC1. This solution was mixed and the tubes were heated in a boiling water bath for 15 min. The tubes were then cooled in water and fluorescence was measured in microcuvettes (150 fj\) with a spectrofluorophotometer (Jasco FP-4). Appropriate standards, reagent blanks, and tissue blanks were used. For the determination of 5-HIAA, 2.3 ml of the organic phase (n-heptane-n-butanol) were acid washed by shaking with 115 /xl of 0.01M HC1 and were centrifuged as before. The organic phase (2 ml) was pipetted into a tube containing 240 fx\

5 Volume 23 Number 2 Serotonin-stimulated chloride transport EPINEPHRINE 10" 7 M 10 i 8 I : 6 5-HT /S.C.C. PDJ OOOOOOOOOOOOoOOOOOo noo o oooooooooooooe Ooo ooooooooooooo! 40 ^ E 20 ~ TIME (min) 0 Fig. 3. Effect of serotonin (5-HT) and epinephrine on epithelial SCC and PD in the isolated rabbit cornea EPINEPHRINE 10" 7 M 8 "NIALAMIDE 10" 4 M 5-HT 10~ 5 M 4 2 -»OOOOOO000O0O<K,OOOOOOOO«OOOO««>C Ooo000«ooooooooo«oooo0oooooo<)<)ooo<)000, ^PD n i i TIME (min) I Fig. 4. Effect of serotonin (5-HT) and epinephrine on the epithelial SCC and PD in a cornea pretreated with NM. 0.5M phosphate buffer, ph 7.0, and the tube was shaken for 10 min and centrifuged. Two 60 fi\ aliquots of the aqueous phase were pipetted into two tubes, labeled A and B. To tube A were added 6 /u,l of 1% cysteine solution, and 6 /xl of 0.02% sodium periodate solution were added to tube B; 120 fx\ of concentrated HC1 were then added to both tubes. Subsequently, 6 fjd of 0.1% OPT (in methanol) and 6 ^tl of 0.02% periodate solution were added to tube A. After 30 min, 6 fi\ of 1% cysteine and 6 fx\ of 0.1% OPT solutions were added to tube B. Both tubes were then placed in a boiling water bath for 10 min and cooled, and the fluorescence at 470 nm was measured with an excitation wavelength of 360 nm. The blank reading (tube B) was subtracted from the test reading (tube A). Appropriate standards and blanks were prepared similarly. Histochemistry The formaldehyde - glutaraldehyde - dichromate (FGD) reaction for serotonin. Small wedges were cut from central rabbit corneas and processed for

6 186 Klyce et al Invest. Ophthahnol. Vis. Sci. August 1982 Table III. Effect of serotonin on NM-treated corneas* SCC (/xa/cm 2 hr) PD (mv) Resistance (kfl cm 2 ) Control period Serotonin (10-4 M) Change p value 3.5 ± ± ± ± ± ± 0.7 < ± ± ±0.5 * Values are means ± S.E.M., n = 30 corneas. The incubation medium contained NM (10~ 4 M) throughout the experiments. Table IV. Effect of MSD on the NM-potentiated serotonin response SCC (MA/cm 2 hr) PD (mv) Resistance (kfl cm 2 ) (1) Control period (2) Serotonin (10-4 M) (3) MSD (10" 4 M) p values (1 vs. 2) (2 vs. 3) (1 vs. 3) 3.6 ± ± ± 0.2 > ± ± ± 2.0 <0.01 < ± ± ± 0.6 <0.01 <0.01 *Values are means ± S.E.M., n = 21 corneas. The incubation medium contained NM (10 ''M) throughout the experiments. the electron microscopic demonstration of serotonin according to the procedures described by Jaim-Etcheverry and Zieher. 14 Initial fixation was carried out at 4 C in 8% paraformaldehyde buffered to ph 7.4 with 0.1M Na cacodylate. Subsequently, the tissues were fixed in cacodylatebuffered 3% glutaraldehyde for 3 hr at 4 C and, after a buffered rinse in 0.32M sucrose solution (2 x 20 min), were incubated overnight at 4 C in 2.5% potassium dichromate (K 2 Cr 2 O 7 ) buffered to ph 5 with 0.1M phosphate/citrate. Potassium permanganate (KMnO J reaction for adrenergic vesicles. Wedges from the same rabbit corneas as above were treated in ice-cold 3% KMnO 4 solution at ph 7 with acetate buffer for 45 min as detailed by Richardson 19 and rinsed in icecold buffer (3 x 10 min). After the fixation procedures described above, the specimens were dehydrated in a graded ethanol series and embedded in epon/araldite. Sections were analyzed with a Model 109 Zeiss electron microscope. Results With the KMnO 4 method, nerve endings with granulated vesicles were visible. The fibers were located in the anterior central stroma, and the diameter of the vesicles was about 40 nm (Fig. 1). The FGD technique revealed nerve fibers containing granulated vesicles (about 60 nm in diameter) containing the serotonin-specific reaction product; these were observed in both the central anterior corneal stroma and near the basal epithelial cells (Fig. 2) but were relatively sparse. The levels of serotonin and 5-HIAA measured in the rabbit cornea and iris-ciliary body were compared with levels in rabbit uvea-retina 20 and in the rat brain 18 (Table I). The serotonin/5-hiaa ratio measured in cornea (3.7) was similar to that in both uvearetina and in brain but was substantially less than the ratio reported for the iris-ciliary body complex (7.3). The concentration of serotonin in the cornea was about one third that of the iris ciliary body. In preliminary experiments, serotonin added to the tear side of the isolated cornea had little or variable effect. When serotonin was added to both sides of the isolated cornea, the response was relatively consistent. However, in a few of the experiments with serotonin or serotonin plus NM (vide infra), no response to serotonin was observed, although the corneal electrical parameters appeared normal. These data were included in the statistical analyses. The only criterion used for the rejection of data was an initial PD of less than 10 mv, since a low PD indicates substantial damage to the epithelium during dissection.

7 Volume 23 Number 2 Serotonin-stimulated chloride transport < 6 3. NIALAMIDE 10" 4 M 5-HT 10" 4 M METHYSERGIDE 10" 4 M EPINEPHRINE TIME (min) Fig. 5. Inhibition of the serotonin response with MSD. A response to epinephrine can be elicited in the presence of MSD. 10 PHENTOLAMINE METHYSERGIDE 10" 5 M 10 ~ 4 M 8 NIALAMIDE 10 " 4 M TIME (min) 90 0 O: Fig. 6. Alpha-receptor blockade with phentolamine has no effect on the SCC stimulated by serotonin. Serotonin (10 /xm) added to both sides of the isolated rabbit cornea produced a significant (26%) increase in the SCC (Fig. 3, Table II). After the serotonin response, corneas were still sensitive to the /3-adrenergic agonist, epinephrine, which is known to stimulate epithelial Cl secretion in this tissue. 7 ' 8 Serotonin decreased epithelial resistance by 13% but had little effect on corneal PD. The corneal response to serotonin was markedly enhanced by pretreatment with the monoamine oxidase inhibitor, NM (Fig. 4, Table III). Again, the SCC could be stimulated by posttreatment with epinephrine. In the presence of NM, serotonin increased the SCC by 57%. This significant potentiation of the serotonin response by monoamine oxidase inhibition suggests that the serotoninsensitive receptors are spatially close to the enzyme monoamine oxidase, a situation apparently different from that of the epinephrine-sensitive receptors. With NM pretreatment, serotonin produced a 41% reduction in epithelial resistance and a marginally significant depression of corneal potential. The fraction of the SCC that was stimulated by serotonin could be completely blocked (p > 0.7) with MSD, a known competitive inhibitor of serotonin receptors in other tissues (Fig. 5, Table IV). Under these conditions, SCC could still be stimulated by epinephrine (Fig. 5), indicating that the /3-adrenergic receptors have a low affinity for MSD. However, the reduction in corneal resistance produced by serotonin was not completely reversed by MSD posttreatment, suggesting nonspecific action of serotonin and/or MSD on general epithelial ion permeability. Corneal PD declined during the serial addition of serotonin, MSD, and epinephrine; this was probably due to an effect similar to that postulated for the decline in resistance.

8 188 Klyce et al. Invest. Ophthalmol. Vis. Sci. August 1982 E < 4 a. NIALAMIDE 10" 4 M 5-HT TIMOLOL METHYSERGIDE 10" 4 M EPINEPHRINE MgUUU'BSo 10"? M / TIME (min) Fig. 7. Effect of TIM on the SCC stimulated with serotonin. The biphasic response to TIM shown here was typical, suggesting an action on two receptors or processes. stroma tears C\'~~-~ c^ TIM corneal nerve fic- EPI cr Fig. 8. Proposed scheme for serotonin action in the cornea. SER, serotonin; MAO, monoamine oxidase; S, serotonin receptor; AC, adenylate cyclase; j8, /3-adrenergic receptor; EPI, epinephrine; LSD, lysergic acid diethylamide; camp, cyclic AMP. Because the epithelial cells are electrically coupled, the epithelium may be regarded as a functional syncytium (cf. ref. 31) and may be represented as a single cell. See text for additional details. The a-adrenergic receptor antagonist, phentolamine, was without effect on corneal electrical parameters either before or after treatment with NM and serotonin (Fig. 6). However, MSD still inhibited the serotoninstimulated SCC in the presence of phentolamine. Unexpectedly, the potent, nonselective, /3i, /3 2 -adrenergic receptor antagonist, TIM, strongly inhibited the serotonin-stimulated SCC in a biphasic manner. After TIM treatment, the tissue was unresponsive to posttreatment with either MSD or epinephrine (Fig. 7). Pretreatment of corneas with 10~ 5 M

9 Volume 23 Number 2 Serotonin-stimulated chloride transport 189 Table V. Effect of TIM on the serotonin corneal response potentiated with NM* SCC (fialcm 2 hr) PD (mv) Resistance (kfl cm 2 ) TIM (10-3 M) Serotonin (10" 4 M) Change p value 3.5 ± ± ± 0.1 < ± ± ± 1.1 > ± ± ± 0.5 >0.7 Values are means ± S.E.M., n = 7 corneas. The incubation medium contained NM (10" 4 M) throughout the experiments. Table VI. Effect of Cl~ Control Serotonin (10" 4 M) Change p value removal on the NM-potentiated serotonin SCC (fialcm 2 hr) 2.3 ± ± ± 0.1 >0.7 PD (mv) 25.1 ± ± ± 0.9 >0.8 response* Resistance (k l cm 2 ) 10.9 ± ± ± 0.4 >0.8 *VaIues are means ± S.E.M., n = 5 corneas. The incubation medium was chloride-free Ringer solution and contained NM (10~ 4 M) throughout the experiments. TIM reduced the subsequent serotonin response. This behavior of TIM is unusual because TIM effects on other than /3-adrenergic receptors have not been previously reported. In tissues that were pretreated with NM and TIM, serotonin stimulated SCC by only 9% (Table V), whereas no significant changes in either corneal resistance or PD could be demonstrated. It is important to note that the serotonin response was not completely blocked by TIM. In a companion series of experiments, TIM (10~ 9 M to 10~ 7 M) partially inhibited both the serotonin and epinephrine stimulation of SCC. No dose of TIM was found that would selectively block the epinephrine response without affecting the serotonin response. In three of the series reported above (Tables III to V) NM alone significantly increased the SCC in the control periods over the average SCC in NM-untreated corneas (Table II). This suggests a sparing effect of NM on endogenous corneal monoamines. In the absence of Cl~ in the bathing medium, serotonin had no statistically significant effect in any of the measured electrical parameters (Table VI). To substantiate the inference that serotonin stimulated Cl~ transport, net 36 C1~ fluxes were compared with SCC (Table VII). The net changes in SCC produced by serotonin and MSD were not significantly different (p > 0.8) from the changes in net Cl transport. During the experiment, the 36 C1~ flux in the passive (tears to aqueous) direction increased marginally, which may correspond to nonspecific effects of these agents on general epithelial permeability (vide supra). Also suggestive of passive permeability effects was the incomplete reversal by MSD of the reduction in corneal resistance produced by serotonin. These effects may be associated with nonspecific effects on epithelial permeability, although conductance and ionic permeability of untreated corneas often increase gradually during prolonged in vitro incubation. Discussion Serotonin has been shown to be a functional neurotransmitter in central nervous tissue. In addition, this autacoid affects a variety of physiologic processes in the cardiovascular system, alimentary tract, and in both exocrine and endocrine glands. Serotonergic receptors have been pharmacologically classified as serotonin-1 and serotonin-2 receptors, 21 but the present data on cornea are not extensive enough to match the putative corneal serotonin receptor with these classifications. Studies at the membrane level in tissues

10 190 Klyce et al. Invest. Ophthalmol. Vis. Sci. August 1982 Table VII. Effect of serotonin (5-HT) and MSD on 3(i Cl~ fluxes and electrical parameters of the rabbit corneal epithelium A Mi Cl~ fluxes Flux period Aqueous to tears (neq/cm 2 hr) Tears to aqueous (neq/cm 2 hr) Net B (neq/cm 2 hr) Control 5-HT (10-4 M) change p value c MSD (10-4 M) change p value ± 9(9) 170 ± ± ± 6(7) 57 ± < ± > ± ± ± 8 60 ± ± 19 <0.05 A Values expressed as the mean ± S.E. (number of corneas). The incubation medium contained NM (10-4 M) throughout the experiments. B Calculated from paired corneas. C 5-HT vs. control periods, t test. "MSD vs. 5-HT periods, t test. other than cornea indicate that a major functional response to serotonin is an increase in ion permeability. 22 Such is the functional response produced by epinephrine 10 and apparently by serotonin (present study) in the stimulation of Cl~ secretion by the rabbit corneal epithelium. In many of the physiologic responses elicited by serotonin, cyclic AMP acts as a second messenger in biomembrane receptormediated events. Whereas serotonin may be responsible for the diarrhea symptomatic of the carcinoid syndrome via an action on cyclic AMP-mediated intestinal Cl~ secretion, the evidence in support of this hypothesis is equivocal. 23 " 25 In the eye, serotonin has been implicated as a potential activator of corneal neovascularization. 26 Serotonin was without effect on lens epithelial ion transport but was found to inhibit lens ATPase activity. 27 Evidence for a unique serotonin receptor in the cornea. The serotonin-stimulated increase in Cl~ transport is completely blocked by MSD while sparing the responsiveness of the /3-adrenergic receptors to epinephrine activation, and the serotonin-stimulated increase in cyclic AMP synthesis is completely blocked by lysergic acid diethylamide. 16 These findings support the concept of the existence of specific corneal serotonergic receptors. Furthermore, the difference in temporal response of the corneal SCC to epinephrine and serotonin suggests a spatial separation of the /3-adrenergic and serotonergic receptors. The response to epinephrine is fast (calculations show that the response may be limited only by the diffusion time through the tearside unstirred layer), whereas the time of the response to serotonin is considerably longer, suggesting that this receptor population is deeper in the epithelium, possibly closer to corneal nerve terminals. The observations that TIM acts as an incomplete but potent antagonist of serotonin stimulation of epithelial Cl~ transport but not of cyclic AMP synthesis are difficult to reconcile with what is known about TIM. Specifically, there is no evidence that TIM acts as an antagonist at any other than the /3-adrenergic receptors. Hence, the ability of TIM to block serotonin-stimulated chloride transport (but not at the serotonergic receptor) suggests that it acts on a nonreceptor locus in the corneal Cl~ transport system. Source of endogenous serotonin in the cornea. The concentration of serotonin measured in the central rabbit cornea is relatively high for a sparsely innervated tissue. It has been estimated from morphometric studies in the adult rabbit that the corneal nerves (both sensory and sympathetic fibers combined) occupy a tissue volume fraction of about 0.01% (R. W. Beuerman, personal communication), which corresponds to about 6 nl. If the 0.41 /umol/kg wet weight (Table I)

11 Volume 23 Number 2 Serotonin-stimulated chloride transport 191 sec neqlcm' l lhr fialcm 2 PD (mv) Resistance (kil cm 2 ) 132 : t 9(16) 203 : t dt 11 < : t 7-72 dt ± 0.24(16) 5.4 ± ± ± ± < ± > ± ± ± measured in the cornea were evenly distributed throughout the corneal nerve volume, the neural serotonin concentration would be 4.1 mm. Mast cell-like elements have been reported in the corneal epithelium 28 " 30 and have been variously designated dendritic cells, polygonal cells, Langerhans' cells, and epithelial secretory cells. Because mast cells normally contain serotonin and other biogenic amines, it is possible that a fraction of the corneal serotonin derives from this source. Whereas a few of these cells have been demonstrated peripherally in healthy corneas of rabbit, guinea pig, and man, they are apparently absent from the central zone of the cornea. Therefore, corneal nerves could provide a major source of endogenous serotonin in the normal adult rabbit. Proposed scheme for serotonin action in the cornea. A model consistent with the results presented in this study and the accompanying article 16 is illustrated in Fig. 8. The histochemical and biochemical evidence indicates that serotonin is contained within a population of dense core vesicles in corneal nerve fibers. As a neurotransmitter in the cornea, serotonin would be released from these fibers and would activate postjunctional serotonergic receptors. Excess serotonin would be taken up by the nerves and restored in vesicles or taken up by the epithelial cells and metabolized by monoamine oxidase, as suggested by the fact that the serotonin response was markedly potentiated by NM inhibition of monoamine oxidase. As demonstrated in other tissues, serotonin may act in an autoregulatory role directly on the nerve to modulate the further release of neurotransmitter. Binding of serotonin to its receptor activates adenylate cyclase to synthesize the intracellular second messenger, cyclic AMP, which mediates the physiologic response. Previous studies have shown that methylxanthine inhibition of cyclic nucleotide phosphodiesterase potentiates the effect of cyclic AMP on Cl~ transport in the cornea 7 and that stimulation of Cl~ transport via cyclic AMP is caused by a specific increase in the Cl~ permeability of the apical or surface epithelial cell membranes. 9 Whereas adrenergic activation of Cl~ transport appears to be mediated by /3-adrenergic receptors located on the corneal epithelial surface, serotonergic activation of Cl~ transport appears to be a distinct process mediated by a spatially separate group of serotonergic receptors. Although there are two apparent control mechanisms for Cl~ transport in this epithelium, at present there is not sufficient information to predict which of them is of greater functional significance in vivo. The generous gifts of timolol by Merck, Sharp & Dohme and of methysergide by Sandoz Pharmaceuticals, and the helpful discussion of this work with Dr. Roger Beuerman are gratefully acknowledged. Statistical analyses were performed using the LSU/Lions Eye Research Computer Facility. REFERENCES 1. Klyce SD, Palkama KA, Marshall WS, Huhtaniitty S, and Mann KP: Evidence for serotonergic recep-

12 192 Klyce et al. Invest. Ophthalmol. Vis. Sci. August 1982 tors in comeal epithelium. INVEST OPHTHALMOL VIS SCI 20(ARVO Suppl.):194, Friedenwald JS and Buschke W: The effects of excitement, of epinephrine and of sympathectomy on the mitotic activity of the comeal epithelium in rats. Am J Physiol 141:689, Mishima S: The effects of denervation and the stimulation of the sympathetic and the trigeminal nerve on the mitotic rate of the comeal epithelium in the rabbit. Jpn J Ophthalmol 1:65, Laties A and Jacobowitz D: Histochemical study of the adrenergic and the cholinergic innervation of the anterior segment of the rabbit eye. INVEST OPHTHALMOL 3:592, Tervo T and Palkama A: Adrenergic innervation of the rat comeal epithelium. INVEST OPHTHALMOL 15:147, Klyce SD: Cl transport in rabbit cornea. In Cl Transport in Biological Membranes. Zadunaisky JA, et al., editors. New York, Academic Press, Inc. (in press, 1982). 7. Chalfie M, Neufeld AH, and Zadunaisky JA: Action of epinephrine and other cyclic AMP-mediated agents on the chloride transport of the frog cornea. INVEST OPHTHALMOL 11:644, Klyce SD, Neufeld AH, and Zadunaisky JA: The activation of the chloride transport by epinephrine and DB cyclic-amp in the cornea of the rabbit. IN- VEST OPHTHALMOL 12:127, Fischer FH, Schmitz L, Hoff W, Schartl S, Liegl O, and Wiederholt M: Sodium and chloride transport in the isolated human cornea. Pfluegers Arch 373:179, Klyce SD and Wong RKS: Site and mode of adrenaline action on chloride transport across the rabbit comeal epithelium. J Physiol 266:777, Jumblatt MM, Fogle JA, and Neufeld AH: Cholera toxin stimulates adenosine 3', 5'-monophosphate synthesis and epithelial wound closure in the rabbit cornea. INVEST OPHTHALMOL Vis Sci 19:1321, Beuerman RW and Schimmelpfennig B: Sensory denervation of the rabbit cornea affects epithelial properties. Exp Neurol 69:196, Pellegrino De Iraldi A and Gueudet R: Catecholamine and serotonin in granulated vesicles of the nerve endings in the pineal gland of the rat. Int J Neuropharmacol 8:9, Jaim-Etcheverry J and Zieher LM: Ultrastructural cytochemistry and pharmacology of 5-hydroxytryptamine in adrenergic nerve endings. III. Selective increase of norepinephrine in the rat pineal gland consecutive to depletion of neuronal 5-hydroxytryptamine. J Pharmacol Exp Ther 178:42, Jaim-Etcheverry G and Zieher LM: Stimulationdepletion of serotonin and noradrenaline from vesicles of sympathetic nerves in the pineal gland of the rat. Cell Tissue Res 207:13, Neufeld AH, Ledgard SE, Jumblatt MM, and Klyce SD: Serotonin-stimulated cyclic AMP synthesis in the rabbit corneal epithelium. INVEST OPHTHALMOL VIS SCI 23:193, Klyce SD: Transport of Na, Cl, and water by the rabbit corneal epithelium at resting potential. Am J Physiol 228:1446, Curzon G and Green R: Rapid method for the determination of 5-hydroxytryptamine and 5-hydroxyindoleacetic acid in small regions of rat brain. Br J Pharmacol 39:653, Richardson KC: Electron microscopic identification of autonomic nerve ending. Nature 210:756, Levene RZ: Mast cells and amines in normal ocular tissues. INVEST OPHTHALMOL 1:531, Snyder SH, Bruns RF, Daly JW, and Innis RB: Multiple neurotransmitter receptors in the brain: amines, adenosine, and cholecystokinin. Fed Proc 40:142, Douglas WW: Autacoids. In The Pharmacological Basis of Therapeutics. Gilman AG, Goodman L, and Gilman A, editors, ed. 6. New York, 1980, Macmillan Publishing Co., 1980, p Donowitz M, Charney AN, and HefFernan JM: Effect of serotonin treatment on intestinal transport in the rabbit. Am J Physiol 232:E85, Kisloff B and Moore EW: Effect of serotonin on water and electrolyte transport in the in vivo rabbit small intestine. Gastroenterology 71:1033, Sheerin HE: Serotonin action on short-circuit current and ion transport across the rabbit ileal mucosa. Life Sci 24:1609, Zauberman H, Michaelson IC, Bergman F, and Maurice DM: Stimulation of neovascularization of the cornea by biogenic amines. Exp Eye Res 8:77, Candia OA, Lanzetta PA, Alvarez LJ, and Gaines W: Inhibition of ionic transport and ATPase activities by serotonin analogues in the isolated toad lens. Biochim Biophys Acta 602:389, Teng CC: The fine structure of the corneal epithelium and basement membrane of the rabbit. Am J Ophthalmol 51:278, Segawa K: Electron microscopic studies on the human corneal epithelium: dendritic cells. Arch Ophthalmol 72:650, Brown J, Soderstrom CW, and Winklemann RK: Langerhans' cells in guinea pig cornea: response to chemical injury. INVEST OPHTHALMOL 7:668, Klyce SD: Electrical profiles in the corneal epithelium. J Physiol 226:407, 1972.