TFF Peptides in the Human Efferent Tear Ducts

Abstract

purpose. To determine whether the lining epithelium of the human lacrimal sac and nasolacrimal duct synthesizes TFF peptides (formerly P-domain peptides, trefoil factors), a family of mucin-associated secretory peptides.

methods. Expression of TFF peptides in human lacrimal sac and nasolacrimal ducts was monitored by reverse transcription–polymerase chain reaction and Western blot analysis. Antisera specific for TFF peptides were used in immunohistochemical analysis to determine the presence and distribution of all three TFF peptides in epithelia of the lacrimal passage. The samples investigated originated from tissue obtained during surgery (18 patients) and postmortem tissue (10 specimens).

results. mRNA expression of TFF1 and TFF3, but not TFF2, was detected in human lacrimal sac and nasolacrimal duct. TFF1 was detected in only approximately 50% of the investigated probes, whereas TFF3 was present in all samples. Immunohistochemistry revealed TFF1 (if present) to be associated with goblet cells forming intraepithelial mucous glands. TFF3 occurred in epithelial cells of the lacrimal sac and the nasolacrimal duct as well as in the acinar cells of subepithelial serous glands, but appeared to be absent in goblet cells.

conclusions. The epithelium of the nasolacrimal ducts synthesizes TFF3 and in some cases also TFF1. In contrast to the human conjunctiva, in which TFF3 is detectable only in goblet cells, TFF3 of the lacrimal sac and nasolacrimal duct is produced in large amounts by epithelial cells as well as by serous glands, but not—or in small amounts only—by goblet cells. This is comparable with localization of TFF3 in the major salivary glands. Thus, TFF3 may have a special function in tear transport through the lacrimal passage comparable to its function on the ocular surface, because the peptide, together with TFF1, may contribute to the rheologic properties of the tear film. Moreover, the TFF peptides may also influence epithelial healing with their motogenic properties.

Tear fluid is drained by the efferent tear ducts into the inferior meatus of the nose. The function of the drainage structures regarding tear fluid composition is still unknown. Many factors involved in tear outflow have been hypothesized, dependent on the unique anatomic configuration of the efferent tear ducts. Besides an active lacrimal pump mechanism functioning by contraction of the orbicularis eye muscle, ¹ suggestions have included a “wrung out” mechanism governed by a system of helically arranged fibrillar structures, ² the action of a cavernous body surrounding the lacrimal sac and the nasolacrimal duct, ³ and physical factors, such as capillarity, ⁴ gravity, ^{5 6} respiration, ⁷ evaporation, ⁸ and reabsorption of tear fluid through the lining epithelium of the efferent tear ducts. ⁹  

The lining epithelium of the lacrimal sac and the nasolacrimal duct is a pseudostratified, columnar epithelium rich in goblet cells. The goblet cells are integrated in the epithelium as solitary cells or, particularly in the lacrimal sac, in a characteristic arrangement of several cell groups. ^{10 11} Small seromucous glands are detected in the lamina propria in addition to epithelial and goblet cells. These glands are situated in the region of the fundus of the lacrimal sac. Their excretory ducts penetrate the lining epithelium and exit into the lumen of the sac. ¹⁰  

The tear film overlying the ocular surface is composed of three layers: an outer lipid layer secreted by the meibomian glands; an intermediate aqueous layer secreted by the lacrimal gland and the accessory lacrimal glands; and an inner mucus layer containing mucins as its major structural component. ¹² These mucins are thought to influence the rheologic properties of the ocular mucus. ^{13 14} The rheologic properties are defined by tear break-up time, which changes in various pathologic conditions (e.g., in patients with dry-eye symptoms). Alterations of mucin have been reported in the conjunctival epithelia of such patients. ¹⁵  

Recently, it has been shown that goblet cells of porcine and human conjunctiva secrete TFF peptides ^{16 17 18 19} (formerly P-domain peptides, trefoil factors), ²⁰ which are, together with mucins, typical constituents of mucus gels (e.g., from the gastrointestinal and the respiratory tracts and the uterus; for reviews, see Refs. ^{21 22 23} ). Three TFF peptides have been characterized in mammals, including human beings: TFF1 (formerly pS2), TFF2 (formerly hSP), and TFF3 (formerly hP1.B/hITF). They are characterized by the TFF motif, a three-looped structure held tightly together by disulfide bonds based on six cysteine residues. One such motif is found in TFF1 and TFF3, whereas TFF2 possesses two TFF domains. Besides their occurrence as major secretory products of many mucin-producing cells, ^{22 23} they have been detected in the brain. ²⁴  

The physiological functions of TFF peptides are multiple. ²³ They promote migration of intestinal, corneal, or bronchial epithelial cells in vitro, ^{25 26 27} they have antiapoptotic properties, ^{28 29 30} and they induce cell scattering. ^{31 32} TFF3 has also been detected as a new neuropeptide of the hypothalamopituitary axis. ^{24 33} All TFF peptides are known for their protective or healing effects in vivo, in particular for the gastrointestinal mucosa, and they are aberrantly secreted during various chronic inflammatory diseases (for reviews, see Refs. ²² , ²³ , ³⁴ ). The presumable impact of TFF peptides on the rheologic properties of tear film ^{16 17 18 19} and their physiological functions relevant to the surface integrity of mucous epithelia led us to a detailed analysis of all three TFF peptides in human epithelial cells of the lacrimal sac and the nasolacrimal duct. 

Materials and Methods

Ocular Sample Preparation

Eighteen biopsy specimens of lacrimal systems (from 6 male and 12 female patients, aged 5–78 years) obtained during surgical procedures and 10 lacrimal systems (5 male, 5 female, aged 53–88 years) obtained from cadavers donated to the Department of Anatomy, Christian Albrecht University of Kiel, Germany, were prepared. Surgical material was obtained with the permission of a medical ethics committee and used in accordance with the Declaration of Helsinki. All 18 live donors had tear duct stenosis of different causes without dacryocystitis. Limited information was available on the cadaveric donors, but they were individuals known to be free of recent trauma, eye or nasal infections, and diseases potentially involving or affecting lacrimal function. Except for the size of the removed lacrimal systems, there were no individual differences between the freshly obtained specimens and cadaveric specimens. After dissection, one half of each tissue biopsy specimen obtained surgically was immediately frozen in liquid nitrogen, and the other half was fixed in 4% formalin, dehydrated in graded concentrations of ethanol, and embedded in paraffin. Two samples (one from the lacrimal sac and one from the nasolacrimal duct) of both right and left efferent tear duct systems were obtained from 10 cadavers. Four samples were obtained from each cadaver. The samples from seven right efferent tear duct systems were frozen in liquid nitrogen. All other samples from 3 right and 10 left efferent tear duct systems were fixed in 4% formalin, dehydrated in graded concentrations of ethanol, and embedded in paraffin. 

RNA Extraction and Reverse Transcription–Polymerase Chain Reaction Analysis

For RT-PCR, frozen samples (20 mg) from all 18 patients and from the right efferent tear ducts (one sample each from the lacrimal sac and the nasolacrimal duct) of 4 cadavers were crushed in an agate mortar under liquid nitrogen, and the RNA was isolated using a guanidium thiocyanate protocol. RNA purification via CsCl ultracentrifugation and RT-PCR analysis monitoring expression of TFF1, TFF2, and TFF3 were performed essentially as described previously, ³⁵ with 30 amplification cycles (Taq DNA polymerase; Qiagen GmbH, Hilden, Germany). As a control, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) transcripts were amplified in a parallel reaction with a specific primer pair. ³⁵  

Antisera and Western Blot Analysis

The following antisera-monitoring TFF peptides were used. Anti-TFF1: A polyclonal rabbit antiserum against the carboxyl-terminal region of human TFF1 was purchased from Novocastra (Newcastle, UK) and used for Western blot analysis. Alternatively, a monoclonal antiserum against the 30 C-terminal amino acids of human TFF1 was purchased from Zymed Laboratories (South San Francisco, CA) and used for immunohistochemistry. Anti-TFF2: The production of polyclonal rabbit antiserum anti-hTFF2-1 against the C-terminal region of TFF2 has been described. ²⁴ Anti-TFF3: The affinity-purified polyclonal rabbit antiserum anti-hTFF3-2 ³⁵ against the C terminus of human TFF3 was used for Western blot analysis, whereas a polyclonal rabbit antiserum against the C terminus of rat TFF3 (anti-rTFF3-1) ³⁶ was used for immunohistochemistry. 

Human tissue from the lacrimal sac and the nasolacrimal duct of three right efferent tear ducts of cadavers (standardized ratio: 100 mg wet weight/400 μL buffer containing 1% SDS and 4% 2-mercaptoethanol) was extracted as described in detail elsewhere, ³⁷ and the protein content was measured with a protein assay based on the Bradford dye-binding procedure (Bio-Rad, Hercules, CA). Total protein (20 μg) was analyzed by Western blot as reported previously. ³⁷ In brief, proteins were resolved by reducing 15% SDS-polyacrylamide gel electrophoresis, electrophoretically transferred at room temperature for 1 hour at 0.8 mA/cm² onto 0.1-μm pore size nitrocellulose membranes (BA79 from Schleicher & Schüll, Hamburg, Germany) and fixed with 0.2% glutaraldehyde in phosphate-buffered saline for 30 minutes. The filter was washed, blocked in 1% bovine serum albumin and 1% dry milk for 1 hour and incubated with the respective polyclonal TFF antiserum at a 1:1000 dilution for 1 hour. Staining was with diaminobenzidine/Ni²⁺ as a substrate for horseradish peroxidase after incubation with a peroxidase-conjugated secondary antibody (dilution 1:2000; Vector Laboratories, Burlingame, CA) for 45 minutes. 

General Histology and Immunohistochemistry

Mucins were stained using Alcian blue 8GX at pH 2.5. Nuclei were counterstained with hematoxylin. Immunohistochemical staining was performed with antibodies against TFF1 (anti-TFF1, 1:5000 dilution) and TFF3 (anti-rTFF3-1, 1:1000 dilution). These were applied with a standard peroxidase-labeled streptavidin-biotin technique, using conventional methods with trypsinization. After counterstaining with hemalum, the sections were finally mounted in aqueous medium (Aquatex; Roche, Mannheim, Germany). For colocalization studies of TFF1 and TFF3 a double-staining system (En Vision; Dako, Glostrup, Denmark) was used. Sections were first stained with anti-TFF1 revealing a brown reaction product and afterward with anti-rTFF3-1 revealing a red reaction product. Two negative control sections were used in each case. One was incubated with the second antibody only, the other with the primary antibody only. Sections of human submandibular gland and human jejunal mucosa were used for positive control. Furthermore, specificity of the TFF1 and TFF3 staining was tested by competition with the corresponding synthetic peptide—that is, 1 mL TFF1 blocking peptide (1:500 dilution, Santa Cruz Biotechnology, Santa Cruz, CA) was preabsorbed with 10 μg human TFF1 for 16 hours at 4°C and 1 mL anti-rTFF3-1 (1:5000 dilution) was preabsorbed with 10 μg synthetic FKPLQEAECTF (representing the C terminus of human TFF3) for 16 hours at 4°C and then used for immunohistochemistry. All slides were examined by microscope (Axiophot; Carl Zeiss, Oberkochen, Germany). Photomicrographs were then obtained (Ektachrome 64; Eastman Kodak, Rochester, NY). 

Results

No differences in results were obtained between cadaveric and fresh biopsy tissues. This was determined by comparison of both types of tissues in RT-PCR analysis and immunohistochemistry. Western blot analysis was performed only on cadaveric tissue, because there was not enough surgically obtained tissue for this method. 

RT-PCR Analysis

RNA was isolated from surgically obtained efferent tear duct tissue from 18 patients and from the right efferent tear ducts of four cadavers (four samples from the lacrimal sac and four samples from the nasolacrimal duct) and cDNA was amplified by the use of specific primer pairs, ³⁵ testing for TFF1, TFF2, or TFF3 transcripts (Fig. 1) . As a control, GAPDH transcripts were amplified. TFF1-specific amplification products were visible at a low level after separation on an agarose gel in nearly one half of the investigated specimens, whereas TFF3-specific amplification products were clearly detected in every sample. By contrast, expression of TFF2 was not detectable in human nasolacrimal ducts. As controls, TFF1, TFF2, and TFF3 transcripts were monitored in parallel reactions with cDNA samples from human stomach and colon (Fig. 1) . 

Western Blot Analysis

Nasolacrimal duct samples and lacrimal sac samples were dissected from the right efferent tear ducts of three cadavers, and extracts were tested for TFF peptides by Western blot analysis (Fig. 2) . Relatively small amounts of TFF1 were detected in all samples. In contrast, the TFF content differed greatly between these samples. There was a tendency toward higher TFF3 concentrations in the nasolacrimal ducts compared with the lacrimal sacs from the same individuals. TFF2 was not detectable at all with this assay. 

General Histology and Immunohistochemistry

Sections (7 μm) from 18 surgically obtained efferent tear duct samples and from 13 lacrimal sac samples and 13 nasolacrimal ducts samples obtained from cadavers were analyzed. Reactivity of TFF1 was observed in nearly one half of the specimens investigated, and TFF3 was detectable in each specimen. The cellular localization of TFF1 and TFF3 revealed a different pattern. If present, staining of TFF1 was found mostly only perinuclearly in goblet cells organized as intraepithelial glands and sometimes also weakly in their cytoplasm (Fig. 3A) . Although TFF1 staining looked like nuclear staining in several cases (compare Fig. 3D ) scrolling with the micrometer screw of the microscope through the sections showed clearly that the staining was perinuclear and not nuclear. Moreover, the staining was highly specific and was inhibited with the corresponding synthetic peptide (Fig. 3B) . The TFF1-positive cells were also Alcian blue–positive due to their characteristic mucin contents (Fig. 3C) . TFF3 occurred in epithelial cells of the lacrimal sac and the nasolacrimal duct, revealing strong staining (Figs. 3E 3F) , as well as in acinar cells of subepithelial serous glands (Fig. 3G) . Reactivity for TFF3 was also visible perinuclearly in goblet cells but was absent in their stored secretory product (Fig. 3F) . Moreover, TFF3 was visible in some cells inside the lamina propria which had a fibroblast-like appearance. There were variations in intensity between different epithelial cells, goblet cells, and especially serous cells of subepithelial serous glands. The latter revealed strong nuclear staining of some serous cells in single sections (Fig. 3G) . Competition experiments with the corresponding synthetic peptide revealed that the nuclear staining could not be inhibited, whereas the cytoplasmic staining of TFF3 was inhibitable (Fig. 3H) . Double staining of TFF1 and TFF3 revealed that both peptides were related to different cell types (Fig. 3D) : TFF1 to goblet cells that form intraepithelial mucous glands and TFF3 mostly to columnar epithelial cells. TFF2 was not detected in all the samples of the nasolacrimal ducts tested with the polyclonal antiserum anti-hTFF2-1 that successfully detected TFF2 in human stomach biopsy specimens (data not shown). 

Discussion

In this study, RT-PCR, Western blot analysis, and immunohistochemistry were used to investigate the biosynthesis of TFF peptides on the mucosal side of the lacrimal sac and nasolacrimal duct. Biopsy material as well as tissue from cadavers was studied and compared. The data suggest that the human efferent tear ducts synthesize and store the secretory peptides TFF1 and TFF3, but not TFF2. 

TFF3 is actually produced by the columnar epithelial cells of the efferent tear ducts as well as the serous cells of seromucous glands, but not—or in small amounts only—by goblet cells. This result is in agreement with a similar observation concerning TFF3 in the major salivary glands, ³⁸ especially in submandibular and sublingual glands, but is surprising, in that the distribution pattern is in clear contrast to that in the human and porcine conjunctiva, ^{16 17} in which the peptide is typically localized in goblet cells. Furthermore, in most other epithelia—for example, that of the gastrointestinal and respiratory tract ^{22 23} —TFF3 is normally found in goblet cells. 

Several publications describe occasional supranuclear TFF1 or TFF3 immunoreactivity. ^{39 40 41} Also, the present immunohistochemical results revealed nuclear TFF3 staining in single sections (Fig. 3G) that was not inhibited with a corresponding synthetic peptide (Fig. 3H) , whereas all cytoplasmic staining was inhibited. This points to the presumption that the nuclear staining of TFF3 is due to an unspecific reaction and therefore seems to be artificial. By contrast, the present immunohistochemical results also reveal TFF1 staining in direct proximity to the nucleus in several sections where the peptide was detected (Fig. 3A) . However, this staining was inhibited with a corresponding synthetic peptide (Fig. 3B) and therefore appeared to be specific. 

The distribution pattern of TFF3 in the efferent tear ducts does not overlap with that of TFF1. This finding is in agreement with most other mucous epithelia. TFF1 of the efferent tear ducts is localized in goblet cells. This situation can be compared with that in the conjunctiva. ^{16 17} By contrast, TFF1 in the stomach is typically present in gastric surface cells. ^{37 42 43} The finding that TFF1 was present only in trace amounts in human efferent tear ducts is reminiscent of the situation in salivary glands ³⁸ and of descriptions in previous reports. ^{42 44}  

Each mucin-producing cell type has been shown to secrete a characteristic TFF peptide-mucin combination, ^{22 23} probably reflecting the complex physiological needs of the environment of these cells. The localization of TFF1 and TFF3 in the human conjunctival goblet cells matches precisely that of the secretory mucin MUC5AC. ^{13 45 46} The localization of TFF3 in serous cells of the submandibular glands is coincidental with that of mucin MUC7. ^{47 48} Thus, the presence of the mucins MUC5AC and MUC7 in the efferent tear ducts is within the limits of expectation. However, a mucin classification of the efferent tear ducts is still missing and will be of great interest. 

Based on the cosecretion of TFF peptides with mucins, it is postulated that TFF peptides interact with mucins as link peptides, influencing the rheologic properties of these complex viscous biopolymers. ⁴⁹ Preliminary studies with TFF2 and TFF3 underline this hypothesis, in that both increase the viscosities of mucin preparations. ⁵⁰ The dimeric structure of TFF1 ⁵¹ and TFF3 ⁵² would be ideally suited to form an entangled network ^{53 54} with the mucins of the efferent tear ducts. The precise nature of the interaction between TFF peptides and mucins is currently not known; however, recent studies show that TFF1 interacts with the von Willebrand factor (vWBF) C-terminal domains of MUC2 and MUC5AC, indicating that the protective effect of TFF peptides may operate by organizing the complex mucus layer. ⁵⁵  

The differences in TFF-peptide distribution between conjunctiva and efferent tear ducts are interesting with regard to tear transport through the efferent tear duct passage. The change in the composition of the mucus layer inside the efferent tear ducts may have a significant impact on the rheologic properties effecting the drainage of tears through the efferent tear duct passage. 

TFF1 and TFF3 in the efferent tear ducts are generally expected to have protective effects as in the gastrointestinal tract (for review, see Refs. ²² , ²³ ). It is interesting in this regard that the columnar epithelial cells of the efferent tear ducts also produce a broad spectrum of antimicrobial peptides including lysozyme, lactoferrin, secretory phospholipase A₂, human β-defensin 1, and human-inducible β-defensin 2. ^{56 57} An interaction of TFF3 with the antimicrobial peptides is possible, because both are products of the same cells. Remarkably, a combined secretion of TFF peptides and lysozyme has also been observed in a specific glandlike structure termed the ulcer-associated cell line (UACL). ³⁴ The antiapoptotic effect of TFF3 ^{28 30} should also be considered in this context. 

An eventual full understanding of the molecular function of TFF peptides at the mucosal surface of the efferent tear duct passage probably will provide further insight into the occurrence of dacryocystitis, which often leads to residual functional impairment with epiphora. The factors controlling the production of efferent tear-duct–associated TFF peptides are unknown, and it is likely that some infection risk factors such as old age, changes in hormonal status (postmenopausal women), or immunodeficiency are associated with downregulation of TFF-peptide production. This may explain the wide variations in the TFF3 content of the efferent tear ducts in particular. Another hypothesis suggests that a preexisting stenosis—a narrowing of the efferent tear duct passage—downregulates the production of TFF peptides. The normally constant flow of tears could be a positive feedback signal for production, which comes to a halt if tears are not drained into the nose. However, this does not explain why dacryocystitis never develops in some patients with epiphora due to postsaccal stenosis. 

Cases of functional dacryostenosis—presence of epiphora in spite of efferent tear duct passages shown to be patent by syringing—may result not only from the occurrence of a nonfunctional segment in the efferent tear duct passage ⁵⁸ but also from the downregulation of TFF peptides and mucins in this segment. Moreover, changes in TFF peptide production may contribute to the occurrence of dacryoliths in the efferent tear duct passage. TFF peptide secretion may also be influenced by alterations in glycosylation of goblet cell and epithelial cell mucins, because they occur in patients with dry eye symptoms. ¹⁵ However, no data are available as yet concerning the synthesis and secretion of ocular TFF peptides during pathologic conditions. 

 Supported by German Research Foundation Grant Pa 738/1-3 (FPP) and by Chemical Industry Fund Grants 0163615 and 0500058 (WH).

 Submitted for publication February 4, 2002; revised April 30, 2002; accepted June 11, 2002.

 Commercial relationships policy: N.

 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.

 

RT-PCR analysis. TFF1, TFF2, or TFF3 expression was monitored in the epithelium of human efferent tear ducts (two cadavers—lane 1: lacrimal sac; lane 2: nasolacrimal duct) and four specimens from surgical material (lanes 3–6). Total RNA from stomach or colon was analyzed as a positive control for TFF1, TFF2, or TFF3 transcripts (lane c). The integrity of the cDNAs was tested by amplification of the GAPDH transcript. The molecular size standard is shown at left.

Figure 2.

 

Detection of TFF peptides in postmortem tissue of three nasolacrimal ducts (ND_1–3) and three lacrimal sacs (LS_1–3). Extracts (20 μg total protein) from three different individuals were analyzed by SDS-polyacrylamide gel electrophoresis (15%) and subsequent Western blot, using the following polyclonal antisera against TFF peptides: anti-human TFF1 (TFF1), affinity-purified anti-hTFF2-1 (TFF2) or affinity-purified anti-hTFF3-2 (TFF3). Human stomach or duodenum extracts were used as the positive control (c). The molecular size standard is shown at left.

Figure 3.

 

TFF1 and TFF3 in human nasolacrimal ducts. (A) Localization of TFF1 (red) to goblet cells of the lacrimal sac by using the monoclonal TFF1 antiserum. (B) Competition with the corresponding synthetic peptide of TFF1 revealed that the staining could be inhibited (compare with A). (C) Comparable section to (A) stained with Alcian blue (pH 2.5) and counterstained with hemalum. (D) Double staining of TFF1 and TFF3 in human nasolacrimal ducts. TFF1 (brown) and TFF3 (red) did not co-localize. TFF1 was localized in goblet cells, whereas TFF3 stained columnar epithelial cells. (E, F) Localization of TFF3 (red) to columnar epithelial cells of the lacrimal sac (E) and the nasolacrimal duct (F) using antiserum anti-rTFF3-1. TFF3 was also visible perinuclearly in goblet cells (arrows). (G) Localization of TFF3 (red) to serous cells of serous glands using antiserum anti-rTFF3-1. (H) Competition with the corresponding synthetic peptide of TFF3 revealed that the nuclear staining was not inhibited, whereas the cytoplasmic staining of TFF3 was inhibitable (compare with Fig. G). (A, C–G) Counterstaining with hemalum. Scale bar: (A–D, F–H) 27.5 μm; (E) 43.3 μm.

The authors thank Regine Worm for excellent technical assistance as well as Monica Berry and Antony P. Corfield, University of Bristol, Mucin Research Group, Bristol, United Kingdom, for helpful discussions and comments on the manuscript.