Pharmacological Characterization of a Serotonin Receptor (5-HT₇) Stimulating cAMP Production in Human Corneal Epithelial Cells

Abstract

purpose. To study the mRNA and pharmacology of a serotonin (5-HT) receptor positively coupled to adenylyl cyclase in normal, primary (P-CEPI), and immortalized human corneal epithelial cells (CEPI-17-CL4), by using numerous 5-HT agonists and antagonists. To determine and compare cloned human 5-HT₇ receptor binding affinities of compounds with their functional potency data.

methods. RT-PCR was used to detect the presence of an mRNA for the human 5-HT₇ receptor in CEPI-17-CL4 cells. Receptor-mediated production of cAMP in cultured cells was measured using an enzyme immunoassay. Compound binding affinities were determined using [³H]-lysergic acid diethylamide ([³H]-LSD) binding to cell membranes of human embryonic kidney (HEK-293) cells expressing the cloned human 5-HT₇ receptor.

results. RT-PCR revealed the presence of a 5-HT₇ receptor mRNA in CEPI-17-CL4 cells. Normal P-CEPI cells generated cAMP in response to 5-HT (−log EC₅₀; pEC₅₀ = 7.6), 5-carboxamidotryptamine (5-CT; pEC₅₀ = 7.8), 5-methoxy-tryptamine (pEC₅₀ = 7.0) and 5-methoxy-dimethyl-tryptamine (pEC₅₀ = 5.7). In CEPI-17-CL4 cells, serotonergic agonists also stimulated cAMP production with different potencies (pEC₅₀): 5-CT (7.4) > 5-HT (6.5) ≥ 5-methoxy-tryptamine (6.1) > 5-methoxy-dimethyl-tryptamine (5.4) ≥ 8-OH-DPAT (<5.0) = α-methyl-5-HT (<5.0). Various 5-HT receptor antagonists inhibited cAMP production induced by 5-CT in CEPI-17-CL4 cells with different potencies (pK_i): methiothepin (8.5) > mesulergine (8.1) = metergoline (8.0) > spiperone (7.4) ≥ clozapine (7.2) = SB-258719 (7.2) > mianserin (6.9) > ketanserin (6.3). Antagonist pK_i values in P-CEPI cells were methiothepin (8.7), spiperone (7.4) and SB-258719 (6.6). The rank order of affinity for displacement of [³H]-LSD from the cloned human 5-HT₇ receptor was: methiothepin > ritanserin > mesulergine = clozapine ≥ metergoline = 5-HT > SB-258719 ≥ spiperone > mianserin ≥ ketanserin. The functional agonist and antagonist potency data obtained from CEPI-17-CL4 cells correlated well with cloned human 5-HT₇ receptor binding affinity data (r = 0.69), with P-CEPI cell functional data (r = 0.85), and with functional potency data in the literature for the cloned human 5-HT₇ receptor (r = 0.88).

conclusions. These collective data support the presence of a pharmacologically defined, adenylyl cyclase-coupled 5-HT₇ receptor in the CEPI-17-CL4 cells that may have relevance to physiological and/or pathologic functions of 5-HT₇ receptors in the human cornea.

Serotonin (5-HT) is a major neurotransmitter in the mammalian central and peripheral nervous system. ^{1 2 3} The receptors mediating the many diverse functions associated with 5-HT are currently divided into seven subfamilies (5-HT_1–7) which have all been cloned and pharmacologically characterized to various degrees. ^{1 2 3} These membrane proteins, with the exception of the 5-HT₃ receptor, represent a diverse group of G-protein-coupled receptors. The 5-HT₁ receptor subtypes (5-HT_1A, 5-HT_1B, 5-HT_1D, 5-HT_1E, and 5-HT_1F) are negatively coupled to adenylyl cyclase. The 5-HT₂ (5-HT_2A, 5-HT_2B, and 5-HT_2C) receptor subtypes are coupled to phospholipase C and promote phosphoinositide (PI) turnover and release of intracellular calcium. ⁴ The 5-HT₃ receptor is a ligand-gated ion channel, whereas the 5-HT₄, 5-HT₆, and 5-HT₇ receptors are positively coupled to adenylyl cyclase through the G-protein G_S. ⁵ The 5-HT₇ receptor can be distinguished from the 5-HT₄ and 5-HT₆ receptors by its high-affinity and selectivity for 5-carboximidotryptamine (5-CT). ^{6 7}  

The 5-HT₇ receptor subtype has been cloned from human and several other species ^{8 9} and is the most recent serotonergic receptor subtype to be identified and characterized. The 5-HT₇ receptor exhibits low sequence homology with the other serotonin receptors that are positively coupled to adenylyl cyclase. In the transmembrane domain, for instance, the 5-HT₇ receptor expresses approximately 39% homology with the 5-HT₆ receptor and 46% with the 5-HT₄ receptor. ³ Three human 5-HT₇ splice variants have been identified: the h5-HT_7(a) (long form), h5-HT_7(b) (short form), and 5-HT_7(d), which differ only in their carboxyl terminus regions. ¹⁰ These splice variants are pharmacologically similar in their coupling to adenylyl cyclase. The 5-HT_7(a) and 5-HT_7(b) receptor isoforms predominate in humans. ¹¹ Studies using the h5-HT_7(a) splice variant produced a stimulation of two Ca²⁺/calmodulin-sensitive adenylyl cyclase isoforms by increasing intracellular Ca²⁺. ¹² Human 5-HT₇ receptor splice variants are expressed in numerous peripheral tissues and in the central nervous system, some differentially. ^{3 13} The 5-HT₇ receptor is hypothesized to play a role in several physiological processes including control of circadian rhythms, ^{14 15} the relaxation of smooth muscle, ¹⁶ and the pathophysiology of depression, ¹⁷ migraine headaches, ¹⁸ and schizophrenia. ⁷  

Despite the wealth of information about the many different roles of 5-HT in the central and peripheral nervous systems (see above), the role of 5-HT in the physiology and pathophysiology of the eye is less well explored and understood. The demonstration of serotonergic innervation of the eye ¹⁹ and the presence of 5-HT in the aqueous humor ²⁰ has prompted some ocular serotonergic studies. For instance, 5-HT_1A receptors have been found in the iris-ciliary body (ICB) of the rabbit, ²¹ and prejunctional 5-HT₃ receptors modulating [³H]-norepinephrine release have been discovered in bovine and human ICBs. ²² However, although early studies conducted in rabbit corneal tissues revealed the presence of a 5-HT receptor coupled positively to adenylyl cyclase, ²³ the pharmacological characterization of this rabbit receptor was not completed because of the unavailability of suitable, potent, and selective 5-HT receptor agonists and antagonists. We hypothesized that human corneal epithelium may also express such a 5-HT receptor and initiated functional studies on normal human corneal epithelial cells. However, because of the difficulty of obtaining and propagating these cells, we also explored the use of a previously immortalized human corneal epithelial cell line (CEPI-17-CL4) that exhibits many of the key genetic, phenotypic, morphologic, pharmacological, and physiological features of the normal primary corneal epithelial cells and tissue. ^{24 25} In the current communication, we describe the detailed pharmacological properties of the 5-HT receptor that promoted the generation of cAMP in the human CEPI-17-CL4 cells, with confirmation of some data in normal human corneal epithelial cells. In addition, we compared the functional potencies of the various serotonergic compounds determined in the CEPI-17-CL4 cells with the receptor binding affinities of these compounds at the cloned human 5-HT₇ receptor. These studies have helped us identify this adenylyl cyclase–coupled human corneal epithelial 5-HT receptor as the 5-HT₇ subtype. 

Materials and Methods

Tissue Culture

Normal, primary human corneal epithelial cells (P-CEPI; passages 2–4; Cascade Biologicals, Portland, OR) isolated from three human donors (males; ages 37–55, who had no ocular diseases) were cultured in medium containing 1% human corneal growth supplement, 100 U/mL penicillin G, 100 μg/mL streptomycin sulfate, and 0.25 μg/mL amphotericin B (EpiLife; Cascade Biologicals). Culture plates were coated with 0.1% gelatin to help the cells adhere to the bottom of the plates. These P-CEPI cells were difficult to grow and propagate. 

The morphologic, pharmacological, and genetic characterization of the simian virus (SV)-40 immortalized human corneal epithelial cells (CEPI-17-CL4 cell) has been previously reported. ^{24 25} These cells (passages 58–158) were cultured in keratinocyte growth medium (KGM) with 0.15 mM CaCl₂. Amphotericin B and gentamicin were replaced by penicillin (100 U/mL) and streptomycin (100 μg/mL). ^{24 26} Media and other supplements were purchased from BioWhittaker (Walkersville, MD). 

cAMP Production in Cultured Cells

Agonist-dependent cAMP formation was measured by a previously described method using a sensitive enzyme immunoassay (EIA). ^{27 28} In brief, compounds of interest were diluted in ethanol so that the final ethanol concentration was 1%, a concentration well tolerated by the cells. Cells were seeded at a concentration of 6000 cells per well. On reaching confluence, the cells were rinsed twice with 0.5 mL Dulbecco’s modified Eagle’s medium (DMEM)/F-12. The cells were preincubated for 20 minutes in the presence or absence of 5-HT receptor antagonists in DMEM/F-12 containing 0.8 mM ascorbate and 1.0 mM of the phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine (IBMX; Sigma-Aldrich, St. Louis, MO) at 23°C. Serotonin receptor agonists were added at the end of this period, and the reaction was allowed to proceed for another 15 minutes at 23°C. When the effects of 5-HT antagonists were investigated, the agonist, 5-carboxyamidotryptamine (5-CT; 100 nM), was used to stimulate cAMP production. After aspiration of the reaction medium, ice cold 0.1 M acetic acid (150 μL, pH 3.5) was added for the termination of cAMP synthesis and cell lysis. Finally, ice cold 0.1 M sodium acetate (225 μL, pH 11.5–12.0) was added to neutralize the samples before analysis by the EIA. ^{27 28}  

cAMP Measurements

cAMP production was measured using an EIA kit purchased from Amersham Pharmacia Biotech (Piscataway, NJ). This assay was conducted according to the package insert in an automated manner using a robotic workstation (Biomek 2000; Beckman Instruments, Fullerton, CA). ^{27 28}  

Receptor Binding Studies

Membranes from human embryonic kidney (HEK-293) cells expressing the cloned human 5-HT₇ receptor (PerkinElmer Life Sciences, Boston, MA) were diluted to 16 μg protein/mL in 4°C 50 mM Tris buffer (pH 7.4). The membranes were resuspended using a tissue disrupter (Polytron; Brinkman Instruments, Westbury, NY) equipped with a generator (<20 seconds; PTA10TS). Drug dilutions were made in 10:10 dimethyl sulfoxide-ethanol (vol/vol), using the robotic workstation (model 2000; Biomek). The diluted compounds (50 μL) were then added to a 96-well deep block. A volume of 400 μL of receptor preparation was manually added to the 96-well block. The workstation was then used to add 50 μL (2 nM final concentration) of [³H]-lysergic acid diethylamide ([³H]-LSD; Perkin Elmer Life Sciences). Nonspecific binding was defined using 10 μM methiothepin. The 96-well blocks were sealed and incubated in a shaking water bath for 60 minutes at 37°C. The 96-well blocks were then transferred to a harvester (MachIV; TomTech, Hamden, CT), and the incubations were terminated by rapid vacuum filtration using a glass fiber filter mats (Whatman GF/B; Whatman, Clifton, NJ) previously soaked in 0.3% polyethylenimine. The samples were counted on a beta-counter (Wallac Bigspot; Perkin Elmer) for 3 minutes at approximately 50% efficiency. 

RT-PCR Analysis of 5-HT₇ Receptor mRNA

Oligonucleotide PCR Primers.

Sense and anti-sense primers ⁶ for the cloned human 5-HT₇ receptor were synthesized at the Central Facility of the Institute of Molecular Biology and Biotechnology at McMaster University (Hamilton, Ontario, Canada). All primers were designed to span intron–exon boundaries to distinguish between amplification of mRNA and genomic DNA and were based on published cDNA sequences. ⁶ All primers were designed after BLAST retrieval of human cDNA sequence of the human 5-HT₇ receptor from a GenBank search (http://www.ncbi.nlm.nih.gov/Genbank; provided in the public domain by the National Center for Biotechnology Information, Bethesda, MD), as for other receptors we have worked on in the past. ²⁹ The sense primer sequence was 5′-GGA ACA GAT CAA CTA CGG CAG AGT-3′; the antisense primer sequence was 5′-TCT ATT GCT TTA CTG AGC ACT GTC-3′. 

cDNA Synthesis.

Total RNA was isolated from CEPI-17-CL4 cells using the standard guanidine thiocyanate procedure. ²⁹ This RNA was then converted into cDNA in a 10-μL reverse transcription reaction containing 0.5 μg of total RNA; 1× first-strand buffer (75 mM KCl; 50 mM Tris-HCl, [pH 8.3]; 3.0 mM MgCl₂); 1.7 mM MgCl₂; 1 mM each dNTP; 10 mM dithiothreitol [dTT]; 2.5 μM oligo (dT)₁₈, and 5 U/μL reverse transcriptase [SuperScript II Reverse Transcriptase; Invitrogen-Gibco, Gaithersburg, MD]). Reactions were incubated at 42°C for 60 minutes, heated at 95°C for 5 minutes and then cooled at 4°C for a minimum of 5 minutes and a maximum of 30 minutes. 

Polymerase Chain Reaction.

PCR was performed on 5 μL of cDNA preparation, to which was added 44 μL of a PCR master mix containing 1× PCR buffer (55 mM KCl; 13 mM Tris-HCl [pH 8.3]); 1 mM MgCl₂; 10% dimethylsulfoxide [DMSO]; 1.25 U/50 μL DNA polymerase (AmpliTaq Gold; Applied Biosciences Foster City, CA) with a thermocycler (1400 GeneAmp; Applied Biosciences) and 0.2 μM each sense and antisense primer in a total volume of 50 μL. A hot-start PCR method was performed in the thermocycler, using the following parameters: an initial denaturing step of 10 minutes at 95°C, denaturing at 94°C for 30 seconds, annealing at the optimal temperature (55°C) for 30 seconds, and extending at 72°C for 1 minute. The final polymerization step was extended an additional 7 minutes. Forty PCR cycles were performed, and suitable precautions were taken to avoid product contamination. PCR setup, amplification, and product processing were performed using dedicated equipment in separate rooms. In addition, several control reactions were routinely run in parallel during RT-PCR analysis, including RT reactions run in the absence of the reverse transcriptase enzyme, to confirm the absence of genomic DNA and/or cDNA contamination, and RT reactions without RNA to check for reagent contamination. PCR amplification of 1.5 ng of human genomic DNA served as a negative control. Positive control RT-PCR reactions were performed using purchased total human lung or brain RNA (BD Biosciences-Clontech, Palo Alto, CA). PCR amplification reactions were evaluated through electrophoresis of 12 μL of PCR product on a 1.5% agarose gels containing 1 μg/mL ethidium bromide and visualized by UV transillumination on an imager (GeneGenius Imager; Synoptics Ltd., Cambridge, UK). Initial product identification was made by comparison to the positive control and the molecular weight ladder. Endonuclease digestion was used to confirm product identity. Briefly, digestion of the 5-HT₇ receptor mRNA amplification product was performed using the BclI restriction endonuclease enzyme in a final reaction volume of 25 μL. After digestion, products were resolved by 2.5 hours of electrophoresis at 90 V on a 2.0% agarose-TBE gel stained with 1 μg/mL ethidium bromide. Gels were visualized and photographed GeneGenius and GeneSnap software. Confirmation of appropriate splice products was made by comparison to the molecular weight ladder and to the positive control. All RT-PCR experiments were performed at least three times. 

Compounds

Serotonin receptor agonists and antagonists were either obtained commercially from Sigma-Aldrich/ RBI or Tocris (Ballwin, MO) or were synthesized by the Medicinal Chemistry Group at Alcon Research, Ltd. (Ft. Worth, TX). 

Data Analysis

Sample optical density (OD) readings were compared with the standard curve and the cAMP content of each sample was evaluated by linear regression analysis using spreadsheet program (Excel; Microsoft, Redmond, WA). Functional potency (EC₅₀, IC_50,; and K _i) was computed with a sigmoidal-fit function of the Origin software package (Microcal Software, Inc., Northampton, MA), ²⁹ which was also used to perform linear regression analysis. Ligand binding and functional antagonist IC₅₀ values were calculated using ActivityBase (IDBS, Surrey, UK) and converted to equilibrium dissociation constants (K _i) as previously described. ^{24 26 29} All data were subsequently converted to their –log values (e.g., pK_i; pEC₅₀) to permit comparisons with the literature information and to allow construction of suitable correlation plots. All functional and binding data are expressed as the mean ± SEM. 

Results

Initial studies of agonist-stimulated cAMP production were performed in normal, primary human corneal epithelial (P-CEPI) cells. 5-CT (pEC₅₀ = 7.8 ± 0.2), 5-HT (pEC₅₀ = 7.6 ± 0.2), 5-methoxy-tryptamine (pEC₅₀ = 7.0 ± 0.01), and 5-methoxy-dimethyl-tryptamine (pEC₅₀ = 5.7 ± 0.4) concentration-dependently induced cAMP accumulation in these P-CEPI cells. Basal levels of cAMP in confluent monolayers of P-CEPI cells were 1.6 ± 0.4 pmol/well (n = 10) and were elevated to 9.0 ± 1.4 pmol/well (n = 10) under maximum stimulation by these agonists. Thus, these compounds stimulated cAMP by 7.6 ± 1.1-fold above basal levels (n = 10). However, due to the difficulty of obtaining a steady large supply of normal human corneal epithelial cells and the problems associated with their slow growth rate and thus the difficulty of propagating these cells, subsequent studies were conducted with the previously immortalized human corneal epithelial cell line, CEPI-17-CL4. ^{24 25 26}  

To define the pharmacological characteristics of the 5-HT receptor stimulating cAMP in the CEPI-17-CL4 cells, we used a broad panel of well-known agonist and antagonists that exhibit some level of selectivity for the different 5-HT receptor subtypes. Thus, numerous 5-HT receptor agonists stimulated cAMP production in CEPI-17-CL4 cells in a concentration-dependent manner (Fig. 1) . Basal levels of cAMP were 1.0 ± 0.1 pmol/well (n = 44), and these were increased to 9.1 ± 1.2 pmol/well by maximum stimulation by the different agonists used in the CEPI-17-CL4 cells. Thus, the various agonists stimulated cAMP by 10.8 ± 1.1-fold (n = 44) above basal levels, comparable to that observed in the P-CEPI cells. Table 1 shows a summary of the data for the six agonists tested in the CEPI-17-CL4 cells. The rank order of potency (pEC₅₀) for these compounds was: 5-CT (pEC₅₀ = 7.4) > 5-HT (pEC₅₀ = 6.5) > 5-methoxy-tryptamine (pEC₅₀ = 6.1) > 5-methoxy-dimethyl-tryptamine (pEC₅₀ = 5.4) ≥ 8-OH-DPAT (pEC₅₀ < 5.0; Table 1 ). Whereas 5-CT, 5-HT, and 5-methoxy-tryptamine were full agonists, 5-methoxy-dimethyl-tryptamine, 8-OH-DPAT, and α-methyl-5-HT were partial agonists (Table 1) . Similarly, 5-HT₄ receptor partial agonists, RS-67333, and RS-67506 were very weak or inactive agonists in these cells (Table 1) . 

Figure 2 shows a representative plot of antagonism of 5-CT (100 nM)–stimulated cAMP production by various 5-HT receptor antagonists in CEPI-17-CL4 cells. Table 2 shows a comparison of functional antagonist data obtained for CEPI-17-CL4 cells with those obtained for the recombinant human 5-HT₇ receptor and native 5-HT₇ receptor in various cells and tissues from the literature. The rank order of potency of the 5-HT receptor antagonists for the CEPI-17-CL4 cell cAMP assay was methiothepin (pK_i = 8.5) > mesulergine (pK_i = 8.1) = metergoline (pK_i = 8.0) > spiperone (pK_i = 7.4) ≥ ritanserin (pK_i = 7.3) ≥ clozapine (pK_i = 7.2) = SB-258719 (pK_i = 7.2) > mianserin (pK_i = 6.9) > MDL-100907 (pK_i = 6.6) > ketanserin (pK_i = 6.3), and so forth (Table 2) . Limited antagonist studies in the P-CEPI cells yielded the following antagonist potencies: methiothepin (pK_i = 8.7), spiperone (pK_i = 7.4), and SB-258719 (pK_i = 6.6). All these functional agonist–antagonist data from P-CEPI and CEPI-17-CL4 cells compared well with the agonist and antagonist potency data reported for the 5-HT₇ receptor in the literature (described later). 

Because of the difficulty of obtaining large preparations of CEPI-17 cell membranes, all binding studies were conducted with the cloned human 5-HT₇ receptor expressed in HEK-293 cell membranes. Thus, to compare the pharmacological characteristics of the 5-HT receptor stimulating cAMP in the CEPI-17-CL4 cells with that of the cloned human 5-HT₇ receptor, we also determined the binding affinities of a broad panel of serotonergic agents, mainly antagonists, that exhibit some level of selectivity for the different 5-HT receptor subtypes at this human 5-HT₇ receptor. Figure 3 shows a representative plot of [³H]-LSD binding data obtained from cloned human 5-HT₇ receptor preparations (expressed in HEK-293 cells) in the current studies. The pK_i values for 5-HT, 5-methoxy-dimethyl-tryptamine and 8-OH-DPAT were 8.5, 7.9, and 8.5, respectively (Table 2) . Table 2 also shows a summary of the 5-HT antagonist binding data compared with functional antagonist data. The rank order of potency for displacement of [³H]-LSD from the cloned human 5-HT₇ receptor in the current studies was methiothepin (pK_i = 11.1) > ritanserin (pK_i = 10.3) > mesulergine (pK_i = 8.8) = clozapine (pK_i = 8.8) ≥ metergoline (pK_i = 8.6) > SB-258719 (pK_i = 8.0) ≥ spiperone (pK_i = 7.9) > mianserin (pK_i = 7.2) ≥ ketanserin (pK_i = 7.1; Table 2 ). 

A correlation plot of pEC₅₀;pK_i data for the CEPI-17-CL4 cell functional data and pK_B or pK_i data from the average cAMP data from 5-HT_7(a), 5-HT_7(b), and 5-HT_7(d) receptors expressed in HEK-293 cells ¹⁰ indicated a high level of correlation (r = 0.88, slope = 1.05, P < 0.0001) of these two sets of data and thus the pharmacological characteristics of the receptors involved in these systems (Figs. 4A 4B) . The functional cAMP data for numerous agonists and antagonists obtained from CEPI-17-CL4 cells were also well correlated with the [³H]-LSD binding affinity data obtained from the cloned human 5-HT₇ receptor (expressed in HEK-293 cells) from our studies coupled with those reported in the literature (r = 0.69; P < 0.001; Fig. 4C ). Likewise, the CEPI-17-CL4 functional data correlated well with the 5-HT₇ receptor-mediated response data in the literature involving cAMP production through cloned human 5-HT₇ receptors and rat astrocyte 5-HT₇ receptors, and canine cerebral artery relaxation data, coupled with [³H]-5-HT release from bovine ICB data (r = 0.71, P < 0.0001; Fig. 4B ). Finally, as expected, the CEPI-17-CL4 functional data correlated highly with agonist and antagonist data obtained from P-CEPI cells (r = 0.85, slope = 0.81, P < 0.02; Fig. 4D ). 

Molecular biological studies using RT-PCR techniques revealed the presence of the mRNA for the human 5-HT₇ receptor in CEPI-17-CL4 cells (Fig. 5) , thus confirming and supplementing the pharmacological data presented herein for these cells. In addition, the mRNA for the human 5-HT₇ receptor was also detected in biopsy specimens of human conjunctival and corneal tissues (Senchyna M, et al., unpublished observations, 2000), using RT-PCR assays. Limited RT-PCR studies to study mRNAs for other 5-HT receptors revealed the absence of 5-HT_1A, 5-HT_2A, 5-HT_2B, and 5-HT_2C mRNAs in CEPI-17-CL4 cells (data not shown). 

Discussion

There is a relative paucity of information on the effects of serotonergic agonists on ocular cells such as the corneal epithelium. The initial work of Neufeld et al. ²³ in the rabbit cornea suggested that this tissue contained a serotonergic receptor but the investigators were unable to identify the subtype of this receptor pharmacologically because of the unavailability of potent and 5-HT receptor-selective compounds. 5-HT was shown to increase cAMP levels in incubated rabbit corneas, which subsequently resulted in Cl⁻ secretion. ³⁶ A physiological or pathologic role for 5-HT on the ocular surface was further suggested by the detection of 5-HT in human tears. ³⁷  

We hypothesized that of the family of 5-HT receptors, the receptor type mediating the aforementioned effects in the cornea involved either the 5-HT₄, 5-HT₆, or 5-HT₇ receptor because these are the only 5-HT receptors known to stimulate cAMP production by activating adenylyl cyclase. ^{1 2 3 4} Accordingly, the current studies have provided functional pharmacological evidence for the presence of a 5-HT₇ receptor that is positively coupled to adenylyl cyclase in the CEPI-17-CL4 cells and also in normal primary human corneal epithelial cells. We observed a high potency for 5-CT (pEC₅₀ = 7.4–7.8) in P-CEPI and CEPI-17-CL4 cells, a serotonergic-agonist known to exhibit a relatively high potency and selectivity for the 5-HT₇ receptor. ^{1 2 3 4 5 6 7 8 9 10 11} The rank order of potency in the current CEPI-17-CL4 cells studies of 5-CT > 5-HT > 8-OH-DPAT for the stimulation of adenylyl cyclase was highly corroborated by similar cAMP data from both the human 5-HT_7(a) ³⁰ and 5-HT_7(b) splice variants ³³ expressed in HEK-293 cells (Table 1) . Similarly, Adham et al. ⁵ reported an identical rank order of potency (5-CT > 5-HT > 5-methoxy-tryptamine > 5-methoxy-dimethyl-tryptamine > 8-OH-DPAT > α-methyl-5-HT) for stimulation of cAMP production by the recombinant human 5-HT_7(a) receptor expressed in murine fibroblasts. The profile of receptor antagonist potency data for the inhibition of 5-CT–stimulated cAMP production also supported the presence of functional 5-HT₇ receptors in the CEPI-17-CL4 cells. We observed a rank order of antagonist potency of methiothepin > spiperone = metergoline = clozapine = SB-258719 > ketanserin in the CEPI-7-CL4 cells, and similar findings were noted for the inhibition of 5-HT–stimulated cAMP production in murine fibroblasts expressing the recombinant human 5-HT_(7a) receptor isoform. ⁵ In addition, our agonist and antagonist potency data from the CEPI-17-CL4 cells compared well with data obtained for the 5-HT₇ receptor mediating relaxation of the canine cerebral arteries (Tables 1 2) . 

Further evidence supporting the notion that a functionally active 5-HT₇ receptor mediates production of cAMP in the CEPI-17-CL4 cells included the following: The 5-HT₄ receptor-selective antagonists, ² RS-23597, RS-39604, and SB-203186, were weak or inactive at inhibiting 5-CT–stimulated cAMP production in CEPI-17-CL4 cells (pK_i < 5), thus ruling out the presence of 5-HT₄ receptors in these cells; likewise, the relatively weak antagonist potency of the 5-HT₆ receptor antagonist, Ro-04-6790, ^{22 38} in the CEPI-17-CL4 cells ruled out the presence of a 5-HT₆ receptor in these cells. However, SB-258719, ^{22 30 39} a 5-HT₇ receptor antagonist exhibiting a 100-fold selectivity for the 5-HT₇ receptor, ^{22 30 39} was a potent inhibitor of 5-CT–stimulated cAMP production (pK_i = 7.2; Table 2 ). This finding, along with other high antagonist potency data (e.g., for methiothepin, metergoline, and spiperone) typical of 5-HT₇ receptors in other systems (Table 2) , supports the identification of a 5-HT₇ receptor in the CEPI-17-CL4 (and P-CEPI) cells. This conclusion was further supported by the detection of an mRNA of the cloned human 5-HT₇ receptor in CEPI-17-CL4 cells (Fig. 5) . In addition, mRNA corresponding to the cloned human 5-HT₇ receptor was observed in human conjunctival/corneal biopsy tissues (Senchyna M, et al., unpublished observations, 2000). 

Results of the current [³H]-LSD–binding studies performed using the cloned human 5-HT₇ receptor were similar to the binding profiles reported in the literature. Displacement of [³H]-HT for the three human 5-HT_7(a), 5-HT_7(b) and 5-HT_7(d) splice variants produced pK_i values of between 8.8 and 8.9 for serotonin and 7.4 and 7.6 for 8-OH-DPAT, ¹⁰ compared with pK_i values of 8.3 and 7.0 for the same compounds, respectively, in our studies. Krobert et al. ¹⁰ observed a rank order of affinity for 5-HT receptor antagonists of methiothepin > mesulergine > clozapine > spiperone > ritanserin > ketanserin for all three human 5-HT_7(a), 5-HT_7(b), and 5-HT_7(d) splice variants. These results compared favorably with our current studies, with the exception of a higher potency in our studies for ritanserin. However, this disparity may be attributed to differences in assay conditions between the two studies. 

The expression of 5-HT₇ receptors in human corneal epithelial cells suggests a role for this receptor is maintaining the normal physiology of the ocular surface. Serotonin has been implicated in the regulation of aqueous humor dynamics in the eye ^{19 20 21 22} and prejunctional 5-HT₇ receptors have been hypothesized to play a role in intraocular pressure regulation. ²² Serotonin, which is present in human tears, ³⁷ increases cAMP levels in corneas, which results in increased Cl⁻ transport. ³⁶ All these findings suggest that the presence of functional 5-HT₇ receptors in corneal epithelial cells may play a role in fluid/mucin secretion on the ocular surface. However, further work is necessary to expand on these findings. 

Taken together, our pharmacological and molecular biological data have demonstrated that P-CEPI and CEPI-17-CL4 cells express a functional 5-HT₇ receptor positively coupled to adenylyl cyclase. Furthermore, these observations suggest that CEPI-17-CL4 cells represent a suitable and predictive in vitro cellular system for further study of the biochemical and pharmacological aspects of the human ocular 5-HT₇ receptor. 

 Supported by Alcon Research, Ltd.

 Submitted for publication December 17, 2002; revised December 17, 2002; accepted May 5, 2003.

 Disclosure: J.Y. Crider, Alcon Research, Ltd. (E, F); G.W. Williams, Alcon Research, Ltd. (E, F); C.D. Drace, Alcon Research, Ltd. (E, F); P. Katoli, Alcon Research, Ltd. (E, F); M. Senchyna, None; N.A. Sharif, Alcon Research, Ltd. (E, F)

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Agonist-induced cAMP production in CEPI-17-CL4 cells. The cells were incubated with the agonists (six concentrations) for 15 minutes, the cells lysed and the cAMP generated determined with an EIA procedure. Representative functional concentration–response curves for several serotonergic receptor agonists in the CEPI-17-CL4 cell adenylyl cyclase assay are shown. Data points represent means of two determinations. Summary data from several such experiments are shown in Table 1 .

Table 1.

 

Agonist-Stimulated Production of cAMP in CEPI-17-CL4 Cells Compared with Data in the Literature

Agonist	Reported Receptor Selectivity	Agonist Potency (pEC50) in CEPI-17-CL4 Cells (% Max. Response)	Literature Agonist Potencies (pEC50) for 5-HT7 Receptor-Mediated cAMP Production or Tissue Relaxation
5-CT	5-HT7	7.4 ± 0.1 (100%)	7.7 30 ; 7.8 31 ; 8.0 5 ; 8.39 22 ; 7.24 32 ; 8.7 33
5-HT	Nonselective	6.5 ± 0.1 (100%)	6.8 30 ; 6.68 31 ; 7.1 5 ; 6.69 22 ; 6.0 32 ; 7.5 33
5-methoxy-tryptamine	Nonselective	6.1 ± 0.1 (123 ± 20%)	6.86 31 ; 6.88 5 ; 6.5 22 ; 5.03 32 ; 8.12 33
5-methoxy-dimethyl-tryptamine	Nonselective	5.4 ± 0.1 (53 ± 7%)	5.5 5
8-OH-DPAT	5-HT1A	<5.0 (17 ± 6%)	6.1 30 ; 4.2 5 ; <5 22 ; 5.25 33
α-methyl-5-HT	5-HT2	<5.0 (13 ± 4%)	4.0 5 ; 5.4 22 ; <5 32
RS-67333	5-HT4	<5.0	—
RS-67506	5-HT4	<5.0	—

Figure 2.

 

Inhibition of agonist-induced cAMP production in CEPI-17-CL4 cells. Various serotonergic antagonists were preincubated with the cells for 20 minutes before the addition of 5-CT (100 nM). The assays were continued for another 15 minutes, the cells lysed, and the cAMP generated determined with an EIA procedure. Representative functional inhibition curves for a range of serotonergic receptor antagonists in the CEPI-17-CL4 cells are shown. Data points represent means of two determinations. Summary data from several such experiments are shown in Table 2 .

Table 2.

 

Antagonism of 5-CT–Mediated cAMP Production in CEPI-17-CL4 Cells and Competition for [³H]-LSD Binding to Cloned Human 5-HT₇ Receptors: Comparison with 5-HT₇ Receptor Functional Data in the Literature

Antagonist	Reported 5-HT Receptor Selectivity	Inhibition of cAMP Production in CEPI-17-CL4 Cells (pKi)	Inhibition of cAMP Production or Tissue Relaxation Via 5-HT7 Receptors (pKi; pKb) from the Literature	[3H]-LSD Binding to Cloned Human 5-HT7 Receptor (pKi)
Methiothepin	Nonselective	8.5 ± 0.1	8.1 30 ; 9.29 10 ; 8.45 33 ; 7.98 31	11.1 ± 0.4
Mesulergine	5-HT2C	8.1 ± 0.2	7.5 30 ; 8.18 10 ; 7.58 31 ; 7.3 32	8.8 ± 0.4
Metergoline	5-HT1, 5-HT2	8.0 ± 0.3	8.60 10 ; 7.1 33	8.6 ± 0.1
Spiperone	5-HT1A, 5-HT2A	7.4 ± 0.1	7.95 10 ; 6.03 33 ; 6.6 32 ; 6.9 32	7.9 ± 0.2
Ritanserin	5-HT1, 5-HT2	7.3 ± 0.1	7.74 10 ; 7.2 31	10.3 ± 0.2
Clozapine	5-HT7, 5-HT6	7.2 ± 0.03	7.2 30 ; 6.65 33 ; 7.03 31 ; 7.1 32	8.8 ± 0.1
SB-258719	5-HT7	7.2 ± 0.3	7.0 30 ; 7.72 10	8.0 ± 0.1
Mianserin	Nonselective	6.9 ± 0.1	6.05 33 ; 6.4 31	7.2 ± 0.0
MDL-100907	5-HT2A	6.6 ± 0.2	—	—
Ketanserin	5-HT2	6.3 ± 0.1	6.3 30 ; 6.95 10 ; 5.3 33	7.1 ± 0.1
BW501C67	5-HT2A	6.1 ± 0.2	—	—
WB-4101	5-HT1A	5.5 ± 0.2	—	—
Way 100635	5-HT1A	5.5 ± 0.2	5 31	—
Xylamidine	5-HT2A	5.4 ± 0.03	—	—
RS-127445	5-HT2B	5.4 ± 0.2	—	—
BMY-7378	5-HT1A; α1d	5.3 ± 0.1	—	—
RS-102221	5-HT2C	<5.00	—	—
SB-203186	5-HT4	<5.00	—	—
SB-242084	5-HT2C	<5.00	—	—
Y-25130	5-HT3	<5.00	—	—
RS-23597-190	5-HT4	<5.00	—	—
RS-39604	5-HT4	<5.00	—	—
Ro-04-6790	5-HT6	<5.00	—	—

Figure 3.

 

[³H]-LSD binding to cloned human 5-HT₇ receptors expressed in HEK-293 cell membranes. Competition assays were conducted with a range of serotonergic agents (eight concentrations) using [³H]-LSD (2 nM final). Data points represent means of two determinations in duplicate. Summary affinity data from several such experiments are shown in Table 2 .

Figure 4.

 

(A) Correlation of serotonergic agonist and antagonist potency data from CEPI-17-CL4 cells and functional literature data for the cloned human 5-HT₇ receptors. The correlation plot depicts functional potency data (pEC₅₀; pK_i) (cAMP production) from the CEPI-17-CL4 cells and similar pEC₅₀; pK_B; pK_i data from the adenylyl cyclase (cAMP production) assays conducted on the cloned human 5-HT_7(a), 5-HT_7(b), and 5-HT_7(d) receptors expressed in HEK-293 cells. ^{5 10 30 33} (B) Correlation of serotonergic agonist and antagonist potency data from CEPI-17-CL4 cells and functional data in the literature for the cloned human 5-HT₇ receptor and the 5-HT₇ receptor in various cells and tissues of different species. The correlation plot depicts functional potency data (pEC₅₀; pK_i) from the CEPI-17-CL4 cells and similar pEC₅₀; pK_B; pK_i data from the adenylyl cyclase assays conducted on the cloned human 5-HT_7(a), 5-HT_7(b), and 5-HT_7(c) receptors expressed in HEK-293 cells, ^{5 10 30 33} and functional assays performed on native 5-HT₇ receptor in brain astrocytes, ³¹ bovine ICB, ²² and canine cerebral arteries. ³² (C) Correlation of the functional agonist and antagonist data from CEPI-17-CL4 cells and that of binding data from the current studies ([³H]-LSD) and from the literature ([³H]-LSD and [³H]-5-HT) using the cloned human 5-HT₇ receptor. ^{6 9} Assays to measure cAMP accumulation in CEPI-17-CL4 cells and [³H]-LSD binding to the cloned human 5-HT₇ receptor were conducted. The correlation details were changed to the following when the two outlying data points (for ritanserin and methiothepin) were omitted from the linear regression analysis: r = 0.82, slope = 0.96, P < 0.001. (D) Correlation of the functional agonist–antagonist data from P-CEPI and CEPI-17-CL4 cells from the current studies. Assays to measure cAMP accumulation in CEPI-17-CL4 cells and P-CEPI cells were conducted. r = 0.85, slope = 0.81, P < 0.02.

Figure 5.

 

Visualization of the mRNA for human 5-HT₇ receptor in CEPI-17-CL4 cells and other cells and tissues. Total RNA isolated from CEPI-17-CL4 cells was reverse transcribed into cDNA. Similar cDNAs were generated for other control tissues (e.g., rat brain, hamster uterus). PCR was performed on the cDNAs, using the sense and antisense primers for the human 5-HT₇ receptor and the PCR products size fractionated on agarose gels, visualized under UV light, and then photographed. The PCR product size for the human 5-HT₇ receptor was 772 bp. Sense and antisense primers for the human 5-HT₇ receptor were constructed as previously described ^{6 35} (see the Methods section for sequences) and the PCR procedure conducted as for other receptor systems. ^{29 35} Lanes 1, 2: nonpigmented ciliary epithelial cells; lanes 3, 4: CEPI-17-CL4 cells; lanes 5, 6: hamster uterus smooth muscle; lane 7: molecular weight ladder; lane 8: no RNA added (negative control 1); lane 9: no reverse transcriptase added (negative control 2); lane 10: human brain; lane 11, 12: G3PDH house-keeping enzyme. Long arrows: position of the 5-HT₇ mRNA; short arrows: position of the house-keeping enzyme (G3PDH).

The authors thank Jesse May for a critical review of the manuscript and Tom Dean for his generous support which enabled us to conduct these studies.