Expression of Cholecystokinin, Gastrin, and Their Receptors in the Mouse Cornea

Abstract

Purpose.: Cholecystokinin (CCK) is a neuropeptide that has been identified in trigeminal ganglion neurons. Gastrin (GAST) is a related peptide never explored in the cornea. The presence and role of both gastrointestinal peptides in the cornea and corneal sensory neurons remain to be established. We explored here in mice whether CCK, GAST, and their receptors CCK1R and CCK2R are expressed in the corneal epithelium and trigeminal ganglion neurons innervating the cornea.

Methods.: We used RT-PCR analysis to detect mRNAs of CCK, GAST, CCK1R, and CCK2R in mouse cornea epithelium, trigeminal ganglia, and primary cultured corneal epithelial cells. Immunofluorescence microscopy was used to localize these peptides and their receptors in the cornea, cultured corneal epithelial cells, and corneal nerves, as well as in the cell bodies of corneal trigeminal ganglion neurons identified by retrograde labeling with Fast Blue.

Results.: Mouse corneal epithelial cells in the cornea in situ and in cell cultures expressed CCK and GAST. Only the receptor CCK2R was found in the corneal epithelium. In addition, mouse corneal afferent sensory neurons expressed CCK and GAST, and the CCK1R receptors.

Conclusions.: The presence of CCK, GAST, and their receptors in the mouse corneal epithelium, and in trigeminal ganglion neurons supplying sensory innervation to the cornea, opens the possibility that these neuropeptides are involved in corneal neurogenic inflammation and in the modulation of repairing/remodeling processes following corneal injury.

Introduction

The mammalian cornea is one of the most densely innervated tissues of the body. The main role of their sensory nerves is to detect potentially injurious stimuli, evoking unpleasant sensations and initiating protective reflexes, such as blinking and lacrimation. ^1–3 These corneal sensory nerves, when activated, release neuropeptides that contribute to a local inflammatory response (neurogenic inflammation). ^4,5 Moreover, ocular sensory nerves help to maintain the cornea in an healthy state as well as promote wound healing. ⁶ It is well established that injury of trigeminal nerve fibers innervating the eye causes neuroparalytic keratitis, a corneal pathology characterized by persistent epithelial defect with eventual involvement of the stroma, which can result in corneal ulceration, perforation, and/or stromal melting. ^7,8 In addition, after surgical or chemical sensory denervation of the cornea, healing of corneal wounds is known to be delayed. ^8–10 These observations indicate that chemical factors released by sensory terminals have a role in the normal physiology of the corneal epithelium and in the activation of the proliferation/migration processes triggered by corneal injury. ¹  

A large variety of neuropeptides are expressed in the trigeminal ganglion, and transported centrifugally to peripheral and central terminals from their soma. Sensory neuropeptides include substance P (SP), neurokinin A (NKA), calcitonin gene-related peptide (CGRP), cholecystokinin (CCK), somatostatin (SOM), vasoactive intestinal polypeptide (VIP), galanin (GAL), and neuropeptide Y (NPY). ¹¹ The neuropeptides CGRP, SP, and NKA are the most abundant neuropeptides present in corneal sensory fibers and cell bodies in the trigeminal ganglion. ^12–18  

The SP-specific receptor NK1R is located to corneal epithelial cells ¹⁹ and topical application of SP enhances expression of tight junction proteins that regulate epithelium barrier function, as well as stimulate proliferation and migration rates of corneal epithelial cells after injury, so acting synergically with other growth factors, such as the insulin-like growth factor-1 (IGF-1). ^19–23 Such findings suggest that neuropeptides within sensory nerve terminals may at least have a part in maintaining the functional integrity of the corneal epithelium. 

Other sensory neuropeptides also have been identified in various eye tissues, although their functional roles remain to be established unequivocally. The CCK is present in approximately 10% of the trigeminal ganglion neurons, ²⁴ including corneal ones, where they may colocalize with SP. ²⁵ The CCK-containing nerves also exist in the iris (being involved in miosis), ^26,27 and in the cornea. ¹² Gastrin (GAST), a peptide related to CCK sharing a similar C-terminal region essential for its biological activity, ²⁸ to date has not been reported to exist in the cornea. Significantly, the biological actions of CCK and GAST are mediated by two receptors, CCK1R and CCK2R ²⁹ ; the former binds specifically sulfated CCK peptides, whereas the latter has an equal affinity for GAST and CCK. ^28,30 Therefore, the aim of the present study was to identify and localize CCK and GAST in relation to their receptors in the mouse cornea. 

Methods

Animals

We obtained C57BL/6 wild-type mice from the Animal Facilities of the University of Oviedo (Spain). All animals used were adult males (1–6 months) of 25 to 30 g of body weight. They were anesthetized with a mixture of ketamine (80 mg/kg of body mass, Imalgène 500; Merial, Toulouse, France) and xylazine (5 mg/kg of body mass, Rompun; Bayer, Munich, Germany), and euthanized with an overdose of intraperitoneal sodium pentobarbital (Dolethal; Vetoquinol, Paris, France). 

All the procedures were performed according to the guidelines of the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research, and were carried out in accordance with the European Communities Council Directive (86/609/EEC) and the Spanish regulations on the protection of animals used for research. 

Tissue Preparation

The eyes were enucleated immediately after mice deep anesthesia, placed in cold PBS and processed for the immunofluorescence studies. For RT-PCR analysis, epithelium was removed from excised mouse corneas (n = 12) after incubation at 37°C with 2 U Dispase (Invitrogen, Carlsbad, CA) for 1 hour, using fine forceps. For the culture of epithelial cells, eyes were kept in Leibovitz medium (Invitrogen) at 4°C. 

The trigeminal ganglia were excised from euthanized mice and processed for RT-PCR analysis. For immunofluorescence studies, trigeminal ganglia, gut, and stomach were dissected from perfused mice. Perfusion was carried out transcardially with 10% buffered formalin. After removal, organs were post-fixed in the same fixative for one more hour. They then were cryoprotected in 30% phosphate buffered sucrose solution and embedded in OCT medium (Sakura Finetek, Torrance, CA). Serial 10-μm thick sections were cut on a cryostat microtome. 

Culture of Mouse Corneal Epithelial Cells

Corneal buttons from enucleated eyes, including the limbus, were excised from the eye and cut up into small pieces. These explants then were placed on the bottom of a 35-mm cell culture dish, with the epithelial side facing down, and covered gently with a low calcium solution specific for corneal epithelium cell isolation and growth (cnT-50; CellnTec, Bern, Switzerland). ³¹ After 2 to 3 days in culture, we observed at the bottom of the dish single cells leaving the explant and migrating; 2 to 3 days later, discrete colonies of cells were formed. They became confluent in approximately 2 weeks, producing a monolayer. The culture then was fixed and immunostained for CCK, GAST, and their receptors. Immunoreactivity for E-Cadherin was performed to confirm that cultured cells were, indeed, epithelial cells (data not shown). 

RT-PCR Analysis

Total RNA was isolated with Trizol (Invitrogen) from the excised corneal epithelium, trigeminal ganglia, and dissociated cultured cells. Subsequently, RNA was treated with DNase I (Sigma-Aldrich, St. Louis, MO) to avoid DNA genomic contamination, and purified with the RNeasy kit (Qiagen, Valencia, CA); 0.5 to 1 μg of total RNA were reverse transcribed with the first strand cDNA synthesis kit with oligo (dT) High Capacity RNA-to-cDNA Master Mix (Applied Biosystems, Foster City, CA) according to the guidelines of the manufacturer. The corresponding cDNAs were amplified by PCR with appropriate synthetic oligonucleotide primer pairs using the kit ReadyMix Taq PCR Reaction Mix (Sigma-Aldrich). The sequences of the PCR primers (sense and antisense, respectively) were: for CCK, 5′-AGCGGCGTATGTCTGTGCGT-3′ and 5′-CACTGCGCCGGCCAAAATCC-3′; for GAST, 5′-AGCTGACCCAGCGCCACAAC-3′ and 5′-ACTGCGGCGGCCAAAGTCCA-3′; for CCK1R, 5′-AGCAGTCCTGGCAAACATTCCTGC-3′ and 5′-GCATGCGGATCACGCGCTTC-3′; for CCK2R, 5′-GCCGTTTCCTACCTCATGGGGGT-3′ and 5′CCCTTGGTTTCGGACCCGGC-3′; and 5′-CTGTGCCGCCGCCATGTCTCT-3′ and 5′-TGCTGACCGCGGACACGAAG-3′ for ribosomal protein S18 (RPS18, internal control). Molecular biology–grade water was used as negative control. The PCR was performed for 35 cycles with an annealing temperature of 60°C. The products of amplification were visualized by electrophoresis on a 1% agarose gel containing ethidium bromide. 

Additionally, a control PCR with genomic DNA was performed to check the ability of our primers to amplify specifically cDNA. No amplification was ever observed (data not shown). 

Retrograde Labeling of Corneal Sensory Neurons

In anesthetized mice, 2 mm diameter discs of Spongostan (Johnson & Johnson, New Brunswick, NJ) embedded on the retrograde tracer Fast Blue (Polysciences, Inc., Eppelheim, Germany) 5% to 10% in distilled water were applied onto both corneas during 1 hour. Eyes were washed extensively afterwards with physiologic saline. We confirmed that Fast Blue did not spread outside the cornea by the absence of fluorescence outside the cornea when the eye was exposed to UV light. Five days later, to allow transportation of the tracer to the neuron's somas in the trigeminal ganglion, the mouse was again anesthetized and perfused. Trigeminal ganglia were dissected and processed as described above. Fast Blue-labeled corneal neurons were identified by their fluorescence emission when excited with UV light. 

Immunofluorescent Techniques

Immunofluorescence staining was carried out in paraffin-embedded corneas, in 4% paraformaldehyde-fixed mouse primary corneal epithelial cell cultures, and in frozen trigeminal ganglion, gut, and stomach, using conventional procedures. Before immunofluorescence staining, antigen retrieval was performed in paraffin-embedded corneas to be used for anti-CCK labeling (Santa Cruz Biotechnologies, Dallas, TX). For this purpose, sections were incubated in boiling citrate buffer (0.1 M, pH 6.0) for 20 minutes, cooled at room temperature, and finally washed 3 times with PBS. 

For immunofluorescent microscopy, tissue sections and fixed cultures were immersed in a solution containing 10% goat/donkey serum in PBS, to block unspecific secondary antibody binding, and incubated overnight with anti-CCK (1:100; Santa Cruz Biotechnologies) affinity purified goat polyclonal IgG primary antibody, anti-CCK (1:800; Abcam, Cambridge, UK), rabbit polyclonal primary IgG serum, anti-GAST (1:100; Abcam), rabbit polyclonal primary IgG serum, anti-CCK1R (1:200–1:1000; Abcam), anti-CCK2R (1:50; Santa Cruz Biotechnologies) affinity purified goat polyclonal primary IgG antibody, and neuronal class III beta-tubulin (TUBB3, 1:100; Covance, Princeton, NJ) mouse monoclonal IgG primary antibody. Receptors and peptides were visualized by indirect immunofluorescence, in which secondary antibodies were donkey anti-rabbit Alexa Fluor 594 (Invitrogen) to label TUBB3 in mouse corneas, and donkey anti-goat or goat anti-rabbit Alexa Fluor 488 (Invitrogen) for cultured epithelial cells, corneas, trigeminal ganglia, gut, and stomach. Nuclei were counterstained using 2 μg/mL 4′6-diamidino-2-phenylindole (DAPI; Invitrogen) in all cases except for trigeminal ganglion, due to the overlap between Fast Blue and DAPI emission. All micrographs were taken using a conventional Leica DM 6000B microscope and an inverted Leica DMI 6000B fluorescence microscope, both equipped with a Leica CTR 6000 filter set and a Leica DFC310 FX camera (Leica Microsystems, Inc., Buffalo Grove, IL). 

Preadsorption Control

To verify the specificity of primary antibodies, a preadsorption test was used. ³² Briefly, the antibodies were incubated with their specific blocking peptide at concentrations five times higher than those of the corresponding antibody. They were: CCK, CCK1R, CCK2R (Santa Cruz Biotechnologies), and gastrin-17 (Sigma-Aldrich), 4 hours at room temperature. After that, immunofluorescent staining was performed following the protocol previously described, using the preadsorbed antibody, which can no longer bind to specific antigens in the section. 

To block the CCK2R binding site of the anti-CCK2R antibody and also the CCK binding domain of the blocking peptide (in case that this still was preserved), we preadsorbed simultaneously the anti-CCK2R antibody with its blocking peptide and with Tyr[SO3H] cholecystokinin octapeptide (Sigma-Aldrich). 

Characterization of Corneal Neurons in the Trigeminal Ganglion

Counts of retrogradely-labeled corneal trigeminal ganglion neurons and of corneal neurons immunolabeled with specific antibodies were performed on the fluorescence microscope using a ×20 objective magnification. For plotting the retrogradely-labeled corneal neurons distribution within the trigeminal ganglion the ImageJ software (National Institutes of Health [NIH], Bethesda, MD) was used. Only neurons with a visible nucleus were counted. Multiple series (one in every six sections) were examined, and a minimum of three trigeminal ganglia were used. 

Results

Expression of CCK and GAST by Corneal Epithelial Cells

The RT-PCR analysis of excised mouse corneal epithelium evidenced the existence of CCK and GAST mRNAs (316 and 338 base pairs [bp], respectively; Fig. 1). Moreover, CCK and GAST immunoreactivities were observed in the epithelium of intact mouse corneas. Superficial corneal epithelium cells exhibited an irregular, weak staining, whereas staining of basal cells was more continuous and intense, with the maximal intensity at the junction between basal epithelial cells and Bowman's membrane (Figs. 2A–C). This immunostaining did not overlap with TUBB3 immunostaining (Neuronspecific class III beta-tubulin, a marker for neurons), excluding the possibility that peptides were located within corneal nerves (Fig. 3). The GAST immunolabeling was eliminated completely by preadsorption with gastrin-17 (Fig. 4), thereby confirming the specificity of the labeling. 

Figure 1

 

The RT-PCR amplification of CCK, GAST, CCK1R, CCK2R transcripts in mouse corneal epithelium. The corresponding sizes of fragments are: for CCK 316 bp, for GAST 338 bp, for CCK2R 417 bp, and for RPS18 (positive control) 417 bp. The CCK1R was not amplified.

Figure 2

 

Expression of CCK, GAST, and CCK2R in mouse corneal epithelium. Fluorescent microscopy images of the intact corneal epithelium in sections of the whole mouse eye immunolabeled with anti-CCK antibodies from Santa Cruz Biotechnologies (A) and from Abcam ([B], see Methods), anti-GAST antibody (C) and anti-CCK2R antibody (D). Epithelial cells nuclei are stained in blue and peptide labeling appears in green.

Figure 3

 

Double immunolabeling of CCK/TUBB3 in mouse corneal epithelium. The CCK immunostaining is marked in green and TUBB3 in red. No colocalization of CCK with TUBB3 (a specific neuronal marker) was observed, suggesting that the CCK found in the corneal epithelium is not located within peptidergic sensory nerve fibers.

Figure 4

 

Preadsorption control of GAST in mouse corneal epithelium and trigeminal ganglion. Immunostaining at the corneal epithelium (A, B) and the trigeminal ganglion (C, D) was abolished completely when the sections were incubated with GAST antiserum after preadsorption with the blocking peptide (BP) gastrin-17.

Contrarily, CCK immunostaining was not abolished using a similar preadsorption strategy. Still, CCK labeling was observed using two different antibodies that recognize different epitopes (one mapping near the C-terminus and the other mapping the N-terminus of CCK). Moreover, these two antibodies labeled the same areas in the epithelium, near the Bowman's membrane (Figs. 2A, 2B). These observations strongly suggested that the CCK labeling was specific. We also found that cultured mouse primary corneal epithelial cells contained CCK and GAST mRNAs (Fig. 5A), and that the expressed peptides could be detected immunohistochemically in these cells (Figs. 5B, C). 

Figure 5

 

Expression of CCK and GAST in corneal epithelial cells in culture. (A) RT-PCR showing the predicted products of 316 bp for CCK, 338 bp for GAST, and 418 bp for the RPS18 housekeeping gene used to validate results. (B, C) Fluorescent microscopy images of corneal epithelial cells in culture, showing immunofluorescence for CCK (B) and GAST (C). Arrows mark the nonimmunoreactive cells.

Since there is a partial homology between CCK and GAST, ^28,30 the absence of cross reactivity between the used antibodies was tested in other mouse tissues. As can be seen in Figures 6A and 6C, in the gut the characteristic labeling for CCK-positive cells (I-cells, arrows) was detected, but not for GAST (Fig. 6F). Contrarily, in the stomach labeling for GAST (G-cells, arrows, Fig. 6E), but not for CCK was observed (Figs. 6B, 6D). 

Figure 6

 

Immunolabeling of CCK and GAST in the mouse gastrointestinal tract. Microtome sections of mouse gut (A, C, F) and stomach (B, D, E) evidencing the selectivity of the antibodies for CCK or GAST in marking their respective peptides. Discrete CCK-positive cells (i.e., I-cells, arrows) sparsely distributed in the mouse gut were labeled in (A) with the Abcam antibody (ab) and in (C) with the Santa Cruz Biotechnology antibody (sc, [C]). The CCK-positive cells for both antibodies were absent in the stomach (B, D). The GAST-positive cells, stained in green, were abundant in the stomach (E), but absent in the gut (F).

Expression of CCK/GAST Receptors by Corneal Epithelial Cells

The expression of CCK and GAST by corneal epithelial cells suggested that both peptides are secreted by these cells. Hence, we explored whether the cornea contains their receptors CCK1R and CCK2R. While RT-PCR showed the presence of the transcripts of CCK and GAST in the corneal epithelium, a clear expression could be found only for CCK2R, but not for CCK1R (Fig. 1). This was confirmed with immunohistochemistry, which evidenced CCK2R immunoreactivity in the corneal epithelium, while CCK1R staining was absent. The CCK2R immunoreactivity was scattered in the superficial corneal epithelium cell layers and more continuous in basal cells, close to the Bowman's membrane, as occurred with CCK immunoreactivity (Fig. 2D). 

In cultured corneal epithelial cells, no CCK1R or CCK2R receptor immunoreactivities were detected and no evidence was found for expression of their mRNAs (Fig. 5A). 

Specificity of the anti-CCK2R antibody in the corneal epithelium was explored using the preadsorption strategy (see Methods). Preincubation of the antibody with the blocking peptide did not prevent immunostaining; in fact, combination of the blocking peptide and the antibody resulted in a more intense staining than when the antibody was added alone, thus casting doubts about its specificity. Nevertheless, preincubation of the antibody with the blocking peptide plus Tyr[SO3H] cholecystokinin octapeptide before staining of the corneas totally prevented CCK2R immunolabeling (Fig. 7); thus, suggesting that the blocking peptide still conserved its CCK-binding domain and that the antibody was specific. 

Figure 7

 

Preadsorption control of CCK2R in mouse corneal epithelium. Schematic diagram of the preadsorption experiment. Corneal sections were incubated with the anti-CCK2R antibody (A), anti-CCK2R antibody preadsorbed with the blocking peptide (B), and the anti-CCK2R antibody preincubated with the blocking peptide, and with CCK octapeptide, simultaneously (D). Only in this last case was the signal completely eliminated. A control of anti-CCK2R antibody plus CCK octapeptide (C) was performed to show that CCK alone did not abolish the capacity of the antibody to bind the CCK2R present in the corneal epithelium. The left panel illustrates schematically the possible interaction between the anti-CCK2R antibody and the blocking peptide during preadsorption alone and in the presence of the CCK analog Tyr[SO3H] CCK octapeptide. We hypothesized that being the CCK2R-antibody blocking peptide a fragment of the CCK2 receptor, it may contain the site where CCK binds its receptor. If this were the case, when used alone the blocking peptide can bind the CCK present in the cornea, resulting in an intense immunostaining, In the presence of the CCK octapeptide, immunolabeling would be totally prevented.

Expression of CCK and GAST by Corneal Trigeminal Ganglion Neurons

The CCK has been reported previously to be present in trigeminal ganglion neurons. ^24,25 Indeed, RT-PCR analysis of trigeminal ganglia of mice confirmed a low, but clearly evident expression of GAST mRNA (Fig. 8). The CCK and GAST also were expressed in trigeminal ganglion neurons innervating the cornea. Fast Blue retrogradely transported from the cornea to the trigeminal ganglion stained a discrete number of putative corneal neurons (Fig. 9) round in shape, and of various sizes that were homogeneously distributed within the medial side (ophthalmic region) of the ganglion, ^13,33,34 and represented 0.95% of the total number of trigeminal ganglion neurons counted in 5 to 7 sections obtained from ganglia of 3 different mice. 

Figure 8

 

RT-PCR amplification of CCK, GAST, CCK1R, and CCK2R transcripts in mouse trigeminal ganglion. RT-PCR analysis of CCK, GAST, CCK1R, and CCK2R of trigeminal ganglion lysate. Products of expected size were generated: CCK (316 bp), GAST (338 bp), CCK1R (351 bp), CCK2R (417 bp), and RPS18 (417 bp).

Figure 9

 

Presence of CCK, GAST, CCK1R, and CCK2R immunolabeling in the soma of corneal trigeminal ganglion neurons. (A) Schematic diagram showing the procedure used to label trigeminal ganglion neurons innervating the cornea with Fast Blue applied onto the corneal surface (see Methods). (B) Sections of the whole trigeminal ganglion, showing somas of retrogradely-labeled corneal neurons marked in blue. (C) Example at larger magnification of corneal neurons labeled with Fast Blue in a section of the trigeminal ganglion. (D) Neurons showing immunoreactivity to CCK (CCK-IR) in the same section. (E) Merged image of (B, C), to evidence the presence of CCK-IR in the soma of one of the Fast Blue-labeled corneal neuron (marked with a red asterisk) and its absence in the neuron (marked with a white asterisk). (F–H) Immunolabeling for GAST (GAST-IR, [J]) of a corneal trigeminal ganglion neuron. (I–K) Immunolabeling for CCK1R (CCK1R-IR) of a corneal neuron marked with a red asterisk. A second corneal neuron, lacking CCK1R-IR is marked with a white asterisk. Nuclei of CCK-IR neurons that were not labeled with Fast Blue are marked with hash key. (L–N) A trigeminal ganglion neuron exhibiting immunolabeling for CCK2R (marked with a hash key) surrounded by corneal trigeminal neurons, which are non-CCK2R-IR (marked with an asterisk).

Of the total number of trigeminal ganglion neurons, 15% were immunoreactive for CCK; CCK-positive corneal neurons represented 24.5% of the population of corneal neurons. In the case of GAST, all trigeminal ganglion neurons, including those with a corneal origin, were immunoreactive for the peptide. The neuronal labeling for both peptides was observed primarily in the cytoplasm. 

Trigeminal GAST immunolabeling was abolished completely by preadsorption with gastrin-17 (Figs. 4C, 4D), as occurred in the epithelium. In the case of CCK, the signal in the trigeminal ganglion was not eliminated by preadsorption with its blocking peptide as also was the case in the epithelium. Unfortunately, only one of the CCK antibodies used in the cornea is suitable for staining CCKergic neurons according to the manufacturer (Abcam). Therefore, the approach used in the cornea for specificity detection could not be applied to trigeminal ganglion neurons. 

Expression of CCK1R and CCK2R in Corneal Trigeminal Ganglion Neurons

The RT-PCR analysis confirmed the presence of mRNAs of CCK1R and CCK2R in mice trigeminal ganglion (Fig. 8) in agreement with previous observations. ³⁵ Immunoreactivity for CCK1R was present in 21.9% of the whole population of trigeminal ganglion neurons. The proportion of CCK1R corneal immunoreactive neurons was rather similar (23%). Only 8% of the trigeminal neurons showed immunoreactivity for CCK2R, while this was observed in only one corneal neuron. Labeling for CCK1R was seen in the nucleus (Figs. 9J, 9K) whereas for CCK2R-IR staining appeared in the neuron's cytoplasm (Figs. 9M, 9N). 

Trigeminal CCK1R nuclear immunoreactivity was eliminated completely by preadsorption with its blocking peptide (Fig. 10); thus, supporting the specificity of the labeling. This was confirmed further by the positive labeling within the nucleus with an antibody that binds a different domain of receptor (data not shown). We were unable to block CCK2R labeling in the trigeminal ganglion. 

Figure 10

 

Preadsorption control of CCK1R in mouse. Nuclei of trigeminal ganglion neurons immunostained for CCK1R (A). Labeling was totally abolished when the anti-CCK1R antibody was preincubated with its blocking peptide (B). Nuclei of nonlabeled neurons are marked with a red asterisk.

Discussion

To our knowledge, this is the first study to show the existence of mRNA for CCK and GAST in the intact mouse corneal epithelium and in cultured corneal epithelial cells, and to provide immunohistochemical evidence of their protein expression. In the corneal epithelium, immunolabeling for CCK and GAST was localized primarily in the basal cell layer, near Bowman's membrane. 

Antibodies have been used widely to identify proteins in cell cultures and in intact tissues. However, this experimental strategy requires controls that guarantee the reliability of the labeling. ³² We did preliminary experiments to select the most appropriate fixative and performed preadsorption controls to assess the specificity of the antibodies used. This approach allows elimination of the ability of the primary antibody to bind specifically the antigen in the sample through its previous binding to the corresponding blocking peptide. A limitation of adsorption controls is that the blocking peptide, even after blocking the primary antibody, can retain an ability to bind to other proteins in the cells, thus giving false-positive labeling. ³² In our experiments, GAST labeling was eliminated completely by preadsorption with its blocking peptide, thereby confirming the specificity of the antibody employed. In contrast, adsorption controls did not block CCK labeling. Nevertheless, in the corneal epithelium two different antibodies that recognized different epitopes of CCK were identified in the same areas of the epithelium close to the Bowman's membrane, thus supporting the interpretation that labeling of CCK was specific. 

Our study also showed immunoexpression of CCK2R, but not of CCK1R in the cornea. The CCK2R corneal immunoreactivity was seen close to the Bowman's membrane. Staining of this receptor was prevented only when preadsorption was carried out simultaneously with the blocking peptide and Tyr[SO3H] CCK octapeptide, suggesting that the blocking peptide maintains the CCK binding site of the receptor and that this domain is different from the anti-CCK2R antibody binding site, which allows the complex anti-CCK2R antibody-blocking peptide to bind to the CCK present in the Bowman's membrane. Altogether, these data reinforced our previous observation that CCK is located close to the Bowman's membrane, and that the antibody used for the identification of this peptide was specific. 

Notably, no mRNA or protein of CCK2R was found in cultured corneal epithelial cells, suggesting that this receptor's gene is finely regulated. Corneal cells “in vivo” are exposed to autocrine, paracrine, and exocrine factors from corneal tissue and tears that are absent in cultured cells, and this may determine the differential expression of the receptor's gene. The CCK2Rs are known to have similar affinities for CCK and GAST, which complicates the experimental exploration of their functional interaction with each peptide. Moreover, the presence of CCK and GAST together with their receptor in epithelial cells and corneal neurons also makes it difficult to define the individual functional roles in the corneal epithelium. This also is the case for SP and CGRP, two corneal neuropeptides that are implicated in neurogenic inflammatory responses, trophic maintenance of epithelial cells, and/or wound healing. ^19–23  

The CCK neuropeptide serves as a hormone in the gastrointestinal tract, but also functions as an endogenous neuropeptide. ^29,36 In the spinal cord, CCK exerts a pronociceptive influence, ³⁷ antagonizing opiate analgesia produced by foot shock and morphine. ³⁸ The anti-opioid properties of CCK in the central nervous system (CNS) ^39,40 prompted the suggestion that endogenously released CCK could act physiologically as a specific opiate antagonist in the CNS regions involved in pain modulation, thus behaving as an endogenous “anti-analgesic” peptide in neuropathic pain conditions. ⁴¹ Peripherally, it is well established that CCK released from duodenal epithelium upon the arrival of nutrients activates afferent terminals from vagal neurons of the nodose ganglion via CCK1R. ⁴² However, the functional significance of CCK and its receptors in the soma, peripheral axons, and target tissues of peripheral sensory ganglion neurons innervating somatic surface tissues, like the skin and mucosae or the cornea, is enigmatic. 

In the present experiments, we found no colocalization of neuronal anti-tubulin III and peptide immunofluorescence in the cornea. The number of sensory fibers found in transversal sections of the corneal epithelium was low. Moreover, our data indicated that CCK-positive corneal neurons represent 24.5% of the population of corneal trigeminal neurons. Hence, low content and a scarce number of CCK-positive fibers may be a reason to explain our failure to localize the peptide in nerve fibers. Finally, we used a polyclonal antibody that recognizes an epitope mapping near the C-terminus of CCK of human origin (a region within amino acids 50 to 100 of the human CCK protein according to the manufacturer). Hence, the possibility that this antibody recognizes larger forms of CCK, such as the pro-peptide, which is located in the neuronal soma, and misses shorter forms, like CCK-8, which is the expected form in the sensory fibers of the cornea, also must be considered. Additionally, expression of CCK at detectable levels may require stimulation as occurs with the expression of other peptides. 

There is evidence that CCK and SP coexpress in DRG neurons. ⁴³ In the skin, SP is present in sensory nerves, but also can be found in epidermal keratinocytes, mast cells, fibroblasts, and other cutaneous immunocompetent cells, many of which also express the SP receptor NK1R. ^44,45 Moreover, not only corneal sensory nerves, but also corneal epithelial cells and keratocytes contain SP. ^18,46 Likewise, we found CCK/GAST in the soma of corneal neurons as well as in basal layer cells of the intact corneal epithelium and in cultured corneal epithelial cells. Hence, it is tempting to speculate that, as occurs with SP, ^1,20 CCK and GAST could be released during corneal injury and inflammation. ^47,48  

The SP participates in tissue repair and remodeling, ⁴⁹ and enhances corneal epithelial cell proliferation and migration. ^{17,19–21,50} Moreover, SP behaves as an injury-inducible messenger to attract to the site of injury CD29(+) stromal-like cells, thereby accelerating wound healing. ⁵¹ Perhaps CCK/GAST also participate in the modulation of the complex cellular processes triggered by injury, especially the epithelium-stroma interactions, ^52–54 as would be suggested by the accumulation of these peptides and their receptor nearby the Bowman's membrane. Noteworthy, in the gastric mucosa, CCK given subcutaneously reduced dose-dependently the ulcer area. ⁵⁵  

There is indirect evidence that during inflammation, SP and CCK interact. Acutely released SP in the skin sensitizes and activates nociceptive and pruritic nerve endings through an action on keratinocyte and immunocompetent cells. Exposure to SP markedly enhanced expression of CCK2R, but not CCK1R in human and murine keratinocyte. On the other hand, itch-related scratching behavior evoked by SP injection in the skin of mice was suppressed by topical application of CCK, suggesting that CCK down-modulates SP-induced release of itch-inducing agents by dermal cells, epidermal keratinocyte, and mast cells. ⁵⁶  

Finally, in searching a functional role for peripheral CCK/GAST, it is worth noting that inflammatory mediators released at the injured tissue during inflammation induce secretion of opioid peptides from immune cells. Opioid peptides bind to, and activate peripheral opioid receptors present in sensory nerve fibers, decreasing their excitability and/or inhibiting release of pro-inflammatory neuropeptides, thus contributing to analgesia. ⁵⁷ On another hand, peripheral nerve injury causes a marked increase of CCK levels in DRG neurons. ²⁴ Considering the “anti-opioid” role described for CCK, ^37,41,58,59 the possibility exists that in injured tissues, including the cornea, increased CCK levels antagonize peripheral opioid receptors, thus counteracting the inhibitory effects of peripheral opioids on nociceptor sensitization during inflammation. 

Acknowledgments

The authors thank Natalia Vazquez Moreno, Manuel Chacón Rodríguez, Almudena Íñigo, Enol Artime, and Paola Braga for their skillful technical assistance, and Miguel Coca-Prados and Neville Osborne for their critical review and comments. 

Supported by Fundación Mª Cristina Masaveu Peterson, Fundación Ramón Areces, and Instituto de Salud Carlos III PI FIS 110288 FEDER, SAF2011-22500, and BFU2008-04425 (Spain). 

Disclosure: A.F. Gonzalez-Coto, None; C. Alonso-Ron, None; I. Alcalde, None; J. Gallar, None; Á. Meana, None; J. Merayo-Lloves, None; C. Belmonte, None 

Footnotes

 AFG-C and CA-R contributed equally to the work presented here and should therefore be regarded as equivalent authors.