Localization of Thrombomodulin in the Anterior Segment of the Human Eye

Abstract

purpose. To localize thrombomodulin (TM) in the anterior segment of the human eye. TM is a vascular endothelial cell surface glycoprotein that acts as a cofactor for the thrombin-catalyzed activation of the anticoagulant protease zymogen, protein C.

methods. Immunohistochemical methods were used to detect TM expression in corneal epithelial cells, the lens epithelial cells, and other cells in the anterior segment of the eye. The expression of TM was also examined in cultured human corneal epithelial cells.

results. TM was expressed in corneal epithelial cells, corneal endothelial cells, and nonpigmented ciliary epithelial cells, which are in direct contact with the aqueous humor. TM was also expressed in cultured corneal epithelial cells and showed cofactor activity. The amount of the antigen in the cultured corneal cells was approximately one tenth of that in human umbilical vein endothelial cells, but its specific cofactor activity (75%) was comparable to that of TM in human umbilical vein endothelial cells. The trabecular meshwork and endothelial cells lining Schlemm’s canal also showed positive staining for TM.

conclusions. The TM in the cells that are in contact with the aqueous humor appears to be involved in maintaining the fluidity of the aqueous humor. In contrast, TM in cells that are not in contact with the aqueous humor may function in regulating cell proliferation and/or differentiation.

Thrombomodulin (TM) acts as a cofactor for the thrombin-catalyzed activation of protein C and is present on the surface of vascular endothelial cells. ^{1 2} Activated protein C functions as an anticoagulant and operates by inactivating factors Va and VIIIa that are indispensable for the coagulation system. In addition, the binding of TM to thrombin inhibits fibrinogen clotting, ³ blocks the activation of factors V and VIII, ³ inhibits platelet activation of thrombin, ⁴ and accelerates the inhibition of thrombin by antithrombin III. ⁵  

TM is widely distributed on the endothelial cells of capillaries, arteries, veins, and lymphatic vessels ^{2 6} and is also present in syncytiotrophoblast cells, ⁶ synovial lining cells, ^{7 8} monocytes, ⁸ platelets, ⁹ and neutrophils. ¹⁰ Soluble TM fragments are also present in blood and urine. ¹¹ These sites are in keeping with the physiological role played by TM in maintaining the fluidity of the blood in the blood vessels. TM acts not only in the blood vessels but also in the lymphatic vessels, subarachnoidal cavities, and the synovial cavity to maintain the fluidity of the lymph, cerebrospinal fluid, and synovial fluid, respectively. ^{7 8}  

Subsequent to the localization of TM in vascular endothelial cells, TM was detected in other cell types that are not in contact with the blood. ^{7 8 9 10 11 12 13 14} In particular, it is of interest that TM is also present in epidermal keratinocytes ¹² and at cell-to-cell contacts including the lung bud epithelium, the neural epithelium in mouse embryos ¹³ and desmoglein I, a chief desmosomal adhesion molecule in human skin. ¹⁴ What physiological role TM might play in tissues other than vascular endothelial cells is still undetermined. However, these nonvascular sites suggest an association of TM to cell adhesion, differentiation, and/or proliferation. 

Human TM consists of 557 amino acid residues arranged in five distinct domains: from the NH₂-terminal, a C-type lectin-like (Ala¹-Asp²²⁶) domain, an epidermal growth factor (EGF)–like (Cys²²⁷-Cys⁴⁶²) domain, an o-glycosylation–rich (Asp⁴⁶³-Ser⁴⁹⁷) domain, a transmembrane (Gly⁴⁹⁸-Leu⁵²¹) domain, and a cytoplasmic (Arg⁵²²-Leu⁵⁵⁷) domain. ^{15 16} The EGF-like domain is essential for cofactor activity and is composed of a six repeat constitution, and thrombin and protein C bind on the fifth and fourth EGF-like constitutions, respectively. ^{17 18} The attachment of chondroitin sulfate glycosaminoglycan (CSGAG) to the o-glycosylation–rich domain of rabbit TM contributes to the affinity of TM to thrombin. ¹⁹  

TM extracted from cultured human endothelial cells is separated as heterogeneous molecules of 105- to 130-kDa bound CSGAG or as a single molecule of 105 kDa without CSGAG by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) under reducing conditions. ^{20 21} It is also known that the attachment of acidic phospholipids, such as phosphatidylserine or phosphatidylethanolamine, to the plasma membrane accelerates the affinity of protein C to the fourth EGF-like structure. ^{22 23 24} Therefore, activation of protein C by thrombin is efficiently accelerated by the TM molecule with CSGAG on the surface of the endothelium. ²⁴  

The localization of TM has been evaluated in various organs and tissues, but detailed studies on the human eye have not been reported. We have determined the loci of TM in the anterior segment of the human eye and will discuss the roles TM may play in these structures. The relationship between TM expression and the migration of the epithelial cells during wound healing in the cornea and limbus was also studied. 

Materials and Methods

Materials

Reagents were purchased from Wako Pure Chemicals (Osaka, Japan), unless otherwise indicated. Isolation of human placental TM and preparation of monoclonal mouse anti-human TM IgGs (TMmAb-2, -11, and -20) were performed by previously described methods. ^{25 26} Q-Sepharose Fast Flow, Sephacryl S-300, and Protein A-Sepharose were purchased from Pharmacia Biotec (Uppsala, Sweden). Bovine serum albumin (BSA), gentamicin, human antithrombin III, and human thrombin (4000 NIH units/mg) were obtained from Gibco (Grand Island, NY). Protein C was purchased from Calbiochem (La Jolla, CA); horseradish peroxidase from Toyobo (Osaka, Japan); 3-aminopropyltriethoxysilane from Tokyo Kasei (Tokyo, Japan); Cellulofine from Seikagaku Kogyo, (Tokyo, Japan); Bacto complete Freund’s adjuvant (CFA) from Difco (Detroit, MI); Dulbecco’s modified Eagle’s medium/F12 (DMEM/F12) and fetal calf serum (FCS) from Flow Laboratories (Irvine, CA); and Boc-Leu-Ser-Thr-Arg-MCA and 7-amino-4-methylcoumarin from the Peptide Institute (Osaka, Japan). 

Recombinant Human TM and Rabbit Anti-human Recombinant TM IgG

The cell line, CHO-K1-RS7TM-neo-No.29, was donated by the Research Institute, Daiichi Pharmaceutical (Tokyo, Japan). Recombinant human TM (rTM), which consists of 491 amino acids from the N-terminal Ala¹ to C-terminal Ala⁴⁹¹ but does not have the transmembrane and cytoplasmic domains of native human TM, was isolated as described by Nawa et al. ²⁷ Briefly, this cell line was cultured in 36 ml of GIT medium (Wako) in 150 cm² culture bottles at 37°C in 5% CO₂-air. After attaining confluence, the culture medium was replaced with fresh medium daily, and the cultures were continued for 1 week. The culture medium was centrifuged to remove cells, and the pH was adjusted to 7.5 with 10 N NaOH. The rTM in the supernatant was purified by Q-Sepharose Fast Flow column chromatography, anti-human TM monoclonal IgG (TMmAb-20)-conjugated Cellulofine chromatography, and Sephacryl S-300 column chromatography. The eluate was monitored by absorbance at 280 nm. The concentration of rTM in each fraction was measured by enzyme immunoassay (EIA), as previously described. ²⁵ The isolated rTM migrated as a sharp single band at 67 kDa under nonreducing condition on SDS-PAGE. The specific cofactor activity for thrombin-dependent protein C activation was 0.48 picomoles of protein C formed per minute per nanogram rTM. 

Four hundred micrograms rTM emulsified with equal volume of CFA was injected subcutaneously into a male rabbit, and 1 month later, 300 μg rTM in booster emulsion was injected in the same way. The antiserum was collected 14 days after the boosting, and polyclonal rabbit anti-human rTM IgG was purified by a Protein A-Sepharose column chromatography according to an established method. ²⁸ Eight hundred forty nanograms of the IgG inhibited by 50% the cofactor activity of 30 ng TM in thrombin-dependent protein C activation. All experiments were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Immunohistochemical Examination of TM in the Anterior Segment of the Human Eye

Three human eyes obtained at the Osaka City General Hospital were used. One eye was enucleated at the time of an extensive resection of a maxillary cancer with orbital invasion. Two eyes were enucleated due to a rupture of the globes that could not be repaired. The eye from the donor with cancer was obtained from a 55-year-old man, and the injured eyes from a 33-year-old man and a 42-year-old woman. Informed consent was obtained from all patients. All research procedures involving humans were in accordance with institutional guidelines and the Declaration of Helsinki. 

These three eyes were fixed immediately after enucleation in phosphate-buffered 4% paraformaldehyde (pH 7.4) for 2 days, dehydrated with a graded ethanol series, and embedded in paraffin wax. Serial sections (approximately 3 μm) of the eyes were cut in the plane parallel to the ocular axis and mounted on silane-coated (3-aminopropyltriethoxysilane; Tokyo Kasei) glass slides. Deparaffinized and hydrated sections were treated with 0.1% trypsin in Tris-HCl buffer (pH 7.6) containing 0.1% calcium chloride for 10 minutes at room temperature (activation of immunoreactivity), 0.3% H₂O₂ in aqueous solution for 10 minutes (internal peroxidase blocking), and 5% normal goat serum in PBS (0.01 M phosphate buffer at pH 7.2 containing 0.9% NaCl) for 60 minutes at room temperature (blocking of second antibody adsorption). The slides were then incubated in mouse monoclonal anti-human TM IgG (TMmAb-20, 1:250–1:500) overnight at 4°C. Further incubations were performed with biotinylated goat anti-mouse IgG (Immu-Mark Universal Kit, ICN Pharmaceuticals, Costa Mesa, CA) and streptavidin-peroxidase–conjugated biotin complex in Tris-HCl buffer (pH 7.6; Dako, Glostrup, Denmark) for 30 minutes at room temperature. Detection of peroxidase was accomplished by incubation in 0.02% diaminobenzidine and 0.002% H₂O₂ in 0.05 M Tris-HCl buffer (pH 7.6) for 5 minutes. After each step, sections were washed thoroughly with PBS. Then, all sections were counterstained with hematoxylin, dehydrated, cleared, and mounted with coverslips in a routine way. For control, tissue sections were incubated either with primary antibody preabsorbed by excess antigen (rTM) (rTM-neutralized mouse monoclonal anti-human TM IgG; concentrations used were 4.36 × 10⁻³ mg protein/ml for anti-TM antibody and 4.36 × 10⁻² mg protein/ml for rTM) or with PBS instead of primary antibody under the same conditions as that used for the experimental slides. 

Cell Culture

Simian virus (SV)40-immortalized human corneal epithelial cells (HCECs) were supplied by Santen Pharmaceutical (Nara, Japan). HCECs were grown in hormone-supplemented epithelial medium (5 μg/ml insulin, 0.1 μg/ml cholera toxin, 10 ng/ml EGF, 40 ng/ml gentamicin, and 15% FCS in DMEM/F12) at 37°C in 5% CO₂ humidified atmosphere. ²⁹ The cells were subcultured in 100-mm dishes or 24-well plates and grown to confluence in hormone-supplemented epithelial medium. 

Human umbilical vein endothelial cells (HUVECs) were harvested from human umbilical cord veins by the method of Jaffe et al. ³⁰ and were cultured for three passages in HuMedia-EG2 containing 10 ng/ml human EGF, 1 μg/ml hydrocortisone, 50 μg/ml gentamicin, 50 ng/ml amphotericin B, 5 ng/ml human β-FGF and 10 μg/ml heparin (Kurabo, Osaka, Japan). HUVECs were subcultured in Type-1 collagen-coated 100-mm diameter dishes or gelatin-coated 24-well plates (Iwaki Glass, Tokyo, Japan) in HuMedia-EG2. 

Analysis of TM mRNA by RT-PCR

Fresh human corneal epithelium specimens were obtained from a 62-year-old woman and a 66-year-old man. During vitreous surgery for age-related macular degeneration, removal of neovascular membranes resulted in choroidal bleeding. To stop this bleeding, the intraocular pressure was elevated, and the bleeding was stopped. The high intraocular pressure led to edema of the corneal epithelium and made fundus observation difficult. The edematous corneal epithelium was removed from the cornea and used for the TM mRNA analysis with consent of the two patients. 

Total RNA was prepared by the guanidinium thiocyanate method ³¹ from the two extirpated human corneal epithelia, HCECs (35-mm diameter dish), and HUVECs (35-mm diameter dish). Analysis of TM mRNA was performed by reverse transcription–polymerase chain reaction (RT-PCR) as previously described, ³² with slight modification. Primers (primers 1 and 2 for the forward and reverse strand sequences, respectively) corresponding to nucleotide numbers from A of the ATG codon of TM and β-actin genes were as follow: 970-989 (primer 1 in EGF-like domain 3: 5′-GAGGACGTGGATGACTGCAT-3′) and 1423-1442 (primer 2 in EGF-like domain 6: 5′-TCACAGTCGGTGCCAATGTG-3′) of TM; and 969-988 (primer 1 in exon 3: 5′-GTACGTTGCTATCCAGGCTG-3′) and 1239-1258 (primer 2 in exon 3: 5′-TGGCCATCTCTTGCTCGAAG-3′) ofβ -actin. All primers for PCR were purchased from Amersham Pharmacia Biotech (Tokyo, Japan). Total RNA (1.0 μg) was subjected to cDNA synthesis using a preamplification system (SuperScript; GIBCO Life Technology, Gaithersburg, MD) by priming with oligo(dT). ^{12 13 14 15 16 17 18}  

PCR was performed in a reaction mixture containing cDNA, primers 1 and 2, deoxynucleotides, and 1.25 U Taq polymerase (Takara Shuzo, Kyoto, Japan) in a total volume of 50 μl of 10 mM Tris-HCl (pH 8.3) containing 50 mM KCl and 1.5 mM MgCl₂. The PCR conditions (denaturization, annealing, and extension) were as follows: 93°C for 30 seconds, 61°C for 1 minute, and 72°C for 1.5 minutes for TM (473 bp, expected product size), and 93°C for 30 seconds, 57°C for 1 minute, and 72°C for 1.5 minutes for β-actin (290 bp). The repeat cycles were 32 and 28 for TM and β-actin, respectively. After amplification, an aliquot of each reaction mixture was subjected to electrophoresis on a 2% agarose gel, the gels were stained with ethidium bromide (0.1 μg/ml) and photographed on a light box. The sequences of PCR products were identified by the direct sequencing method for DNA by Takara Custom Services (Takara Shuzo). 

Measurement of Cofactor Activity and Antigen Level of TM

Adherent cultures of HCECs or HUVECs (100-mm dish) were washed three times with ice-cold Ca²⁺- and Mg²⁺-free phosphate-buffered saline (PBS[−]), and cells were collected by scraping in PBS(−). Cell suspensions were centrifuged at 500g for 5 minutes, and the pellets were resuspended in 100 μl of 20 mM Tris-HCl (pH 8.0) containing 0.5% Triton X-100 and 0.15 M NaCl. After incubation for 5 minutes at room temperature (mixing by vortex mixer every minute), Triton X-100 insoluble materials were removed by centrifugation (12,000g, 5 minutes). The cofactor activity of TM in cell lysate was determined by its ability to accelerate thrombin-dependent protein C activation as described previously. ^{33 34} Briefly, cell lysate was incubated with 50 μg/ml human protein C, 1 NIH unit/ml thrombin, 1 mM CaCl₂ and 0.1% BSA in 50 mM Tris-HCl (pH 8.0) containing 0.15 M NaCl for 30 minutes at 37°C. The activation of protein C was terminated by addition of a mixture of antithrombin III (final concentration 2 units/ml) and heparin (final concentration 8 units/ml). A mixture of 0.2 ml of the reaction solution and 0.2 ml of 400 μM Boc-Leu-Ser-Thr-Arg-MCA, synthetic substrate of activated protein C, was incubated for 30 minutes at 37°C, and the reaction was terminated by adding acetic acid to a final concentration of 12% (vol/vol). The liberated 7-amino-4-methylcoumarin was then measured by using a fluorescence spectrophotometer with excitation at 380 nm and emission at 460 nm. One picomole of activated protein C releases 19.6 picomoles of aminomethylcoumarin per milliliter per minute under these conditions. 

The protein concentrations of cell lysates were determined by Bradford’s method using BSA as the standard. ³⁵ The concentrations of TM antigen in HCEC and HUVEC lysates were measured by EIA using monoclonal antibodies (TMmAb-2, TMmAb-11, and TMmAb-20), as described previously. ^{25 34} Isolated human placental TM was used as a standard. 

Immunoblot Analysis

Cells were washed three times with PBS(−), fixed in ice-cold PBS(−) containing 3% formaldehyde on ice, and collected in PBS(−) by scraping. The cell suspensions were then centrifuged at 500g for 5 minutes, and the pellets were lysed in 300 μl of 1% SDS solution at room temperature. The lysates were centrifuged at 12,000g for 5 minutes at 4°C, and the supernatants were saved. Samples were heated with 0.625 M Tris-HCl (pH 6.8), containing 1% SDS and 10% glycerol in the presence or absence of 2.5% 2-mercaptoethanol in boiling water bath for 3 minutes. Aliquots of samples were subjected to SDS-PAGE on 7.5% acrylamide gel. After SDS-PAGE, proteins were electrotransferred to polyvinylidene difluoride (PVDF) membranes (Millipore, Bedford, MA). After blocking nonspecific binding by incubation with 1% skim milk in Tris-buffered saline (TBS; 20 mM Tris-HCl [pH 7.5], containing 0.15 M NaCl), the membranes were then incubated with a rabbit anti-rTM IgG (8.5 μg/ml), rTM-neutralized rabbit anti-rTM IgG (8.5 μg/ml) or preimmune rabbit IgG (8.5 μg/ml) in TBS containing 0.05% Tween-20 and 0.1% skim milk for 1 hour at room temperature. The rTM-neutralized anti-rTM IgG was prepared by incubation with the antibody (42.5 μg per 187 μg rTM at a molar ratio of approximately 1:10 in TBS containing 0.05% Tween-20 at 37°C for 60 minutes. After the membranes were washed with TBS containing 0.05% Tween-20, they were incubated with goat anti-rabbit IgG conjugated with horseradish peroxidase (Wako) in TBS containing 0.05% Tween-20 and 0.1% skim milk for 1 hour at room temperature. After the membranes were again washed with TBS containing 0.05% Tween-20, the peroxidase activity was developed using 0.02% diaminobenzidine in TBS supplemented with 0.2% hydrogen peroxide to yield a brown band. The reaction was stopped by washing with distilled water. 

Results

Immunohistochemical Localization of TM

The cornea, limbus, bulbar conjunctiva, anterior chamber angle, ciliary body, and lens were examined (Fig. 1A ). 

Cornea.

In the central area of the cornea, the surface layer, the superficial cells, and the basal cells of the epithelium showed positive immunostaining for TM (Fig. 1B) , whereas the wing cells were negative. In the peripheral cornea, positive staining for TM was present on the surface layer and in the cytoplasm of the superficial cells, wing cells, and basal cells (Fig. 1C) . All the corneal endothelial cells exhibited positive staining (Fig. 1D) . The corneal keratocytes were also TM positive (Fig. 1D) . 

Limbus.

When subjected to immunostaining, the cytoplasm of the basal cells reacted moderately to strongly to the anti-TM antibody (Fig. 1E) . The cytoplasm of the wing cells and superficial cells also showed positive reactions but to different degrees (Fig. 1E) . 

Bulbar Conjunctiva.

Immunohistochemical staining with anti-TM antibody showed that the superficial cells and basal cells of the conjunctival epithelium were stained positively but to different degrees (Fig. 1F) . The capillary endothelial cells in the lamina propria mucosa were also TM positive (Fig. 1F) . 

Anterior Chamber Angle.

All the layers of the trabecular meshwork and the internal and external walls of the endothelial cells in Schlemm’s canal reacted with different degrees of positivity to anti-TM antibody (Fig. 1G) . 

Lens.

The epithelium on the anterior surface of the lens and the lens fiber in contact with the epithelium showed TM positivity (Fig. 1H) . 

Ciliary Body.

The nonpigmented ciliary epithelial cells and the capillary endothelium in the stroma showed a TM-positive reaction (Fig. 1I) . The stromal cells of the ciliary processes showed a weak or negative reaction for TM antibody (Fig. 1I) . Ciliary muscle cells showed weak to moderate TM-positive reactions (Fig. 1A) . Because the pigmented epithelial cells are rich in melanin, it was not possible to determine whether any anti-TM staining was present. 

Iris.

The iris is also rich in melanocytes (Fig. 1A) , and therefore, the results of immunologic reactions could not be determined. 

Controls.

Negative controls showed no positive reactions in any of the structures that were TM-positive in the experimental preparations (Figs. 1J 1K) . 

Staining with Polyclonal Rabbit Anti-human TM

The same distribution of TM antigen in the anterior segment of the human eye was observed when polyclonal rabbit anti-human TM IgG instead of mouse monoclonal anti-human TM IgG (TMmAb-20) was used for the immunohistochemical staining (data not shown). 

Expression of TM in HCECs and Cultured HCECs

To confirm the expression of TM in HCECs, expression of TM mRNA in the extirpated human corneal epithelium and cultured HCECs was determined by RT-PCR with gene-specific primer pairs (Fig. 2) . A single PCR product was detected in both the human corneal epithelium (lane 1) and HCECs (lane 3). The size of PCR product was identified with that of HUVECs as a positive control (473 bp; lane 5). The sequence of the PCR product was confirmed to be the same as the sequence of human TM gene from 970 to 1442 bases, as published in the GenBank database. No PCR product was observed in RT-PCR without reverse transcriptase (lanes 2, 4, and 6), indicating that the present RT-PCR condition detected the targeted RNA (TM mRNA), and the PCR product was not due to genomic DNA. These results suggest that TM mRNA is expressed in human corneal epithelium and cultured HCECs as well as HUVECs. 

The antigen levels and cofactor activity of TM were measured in cultured HCECs instead of fresh human corneal epithelium. In Japan, the experimental use of eyes retained by the eye bank is legally prohibited, so that adequate amounts of fresh human corneal epithelium could not be used for experiments. HCECs contained 1.1 ± 0.1 ng TM/10⁵ cells or 54.8 ± 1.2 ng TM per milligram cell protein, and TM cofactor activity of 48.4 ± 1.2 picomoles of protein C activated per minute per milligram cell protein. These values are approximately one tenth of the antigen and cofactor levels of TM in cultured HUVECs (Table 1) . However, the ability of the TM expressed by HCECs to accelerate protein C activation by thrombin was comparable to that of HUVECs, because the specific activities of the lysates of HCECs and HUVECs were 0.88 ± 0.02 and 1.20 ± 0.07 picomoles of protein C activated per minute per nanogram TM, respectively. The cofactor activity of TM in HCECs lysates was completely inhibited by the presence of a rabbit polyclonal anti-rTM IgG (1 μM) as was the activity of TM in HUVEC lysates (data not shown). These results indicate that the HCEC-expressed TM maintains its cofactor activity for protein C activation by thrombin. 

To determine the molecular weight of the TM expressed by HCEC, Western blot analysis was performed by using a rabbit polyclonal anti-rTM IgG (Fig. 3) . HUVEC lysates were used as control. HCEC and HUVEC lysates were prepared after fixation with 3% formaldehyde to avoid degradation of intact TM during extraction from the cells. Under nonreducing conditions, a single band of 78 kDa appeared in the blots of HCEC lysate (Fig. 3 , lane 1) and a diffuse, high-molecular-mass band of 78 to 100 kDa was seen in the HUVEC lysate (Fig. 3 , lane 2). After reduction, the blots of HCEC (Fig. 3 , lane 3) and HUVEC (Fig. 3 , lane 4) lysates revealed a single band at 105 kDa and a diffuse molecular mass band of 105 to 120 kDa, respectively. No staining was observed in the Western blot analysis under both reducing and nonreducing conditions by a preimmune rabbit IgG or anti-rTM IgG neutralized with rTM (data not shown). These results indicated that the apparent molecular weight of the TM expressed in HCECs was 78 kDa before reduction and 105 kDa after reduction. 

Discussion

The results of this study have shown that TM is localized in cells in direct contact with the vascular system and also in cells not associated with the vascular system in the anterior segment of the human eye. Thus, TM was found in the corneal epithelial cells, corneal endothelial cells, limbus, bulbar conjunctiva, lens epithelial cells, nonpigmented ciliary body epithelial cells, trabecular meshwork, and the endothelial cells lining Schlemm’s canal. The same results were observed even when rabbit polyclonal anti-human TM IgG instead of mouse monoclonal anti-human TM IgG (TMmAb-20) was used for the immunohistochemical staining. This is contrary to the report of the absence of TM in the rabbit cornea ³⁶ and the report of the presence of TM in only the corneal endothelial cells in rats. ³⁷ This discrepancy in the localization of TM between human and animal corneas may be due to species difference or to the specificity of the antibodies. 

The expression of TM in corneal epithelial cells was confirmed by both the RT-PCR method on human corneal epithelium and cultured HCECs and the EIA of cultured HCEC lysates. The quantity of TM antigen (in nanograms per cell) in cultured HCECs was approximately one tenth that in HUVECs; however, the specific cofactor activity per TM molecule was only slightly lower in HCECs than that in cultured HUVECs. The lower expression of TM in cultured HCECs than in HUVECs may be due to the effect of the immortalization of the corneal epithelial cells by SV40. 

The HUVEC-expressed TM molecules are of heterogeneous size ranging from 78 to 100 kDa under nonreducing conditions and 105 to 120 kDa under reducing conditions (Fig. 2 , lane 2 and 4), as previously reported. ^{20 21} It is known that the heterogeneity is due to a difference in the number of CSGAG molecules that are bound to the o-glycosylation-rich domain of TM. ^{20 21} The TM expressed in HCEC, however, was detected as a sharp single band of 78 kDa under nonreducing conditions and 105 kDa under reduced conditions. These findings suggest that the TM in HCECs probably has a different CSGAG content than that of HUVECs. 

As shown, TM was expressed in the corneal endothelial cells, nonpigmented ciliary epithelial cells, trabecular meshwork, and the endothelial cells lining Schlemm’s canal. All these cells are in direct contact with the aqueous humor. In pathologic states with hemorrhage through rhexis into the anterior chamber as by blunt eye injury, the blood does not clot but flows out of the anterior chamber. However, inflammatory changes in the anterior chamber, such as anterior uveitis, ³⁸ leads to fibrin formation. Inflammatory cytokines such as tumor necrosis factor (TNF)-α and interleukin (IL)-1β reduce the expression of TM and induce expression of tissue factors, a trigger for blood coagulation in various cells including endothelial cells and monocytes. ^{32 39 40 41} In vivo experiments have shown that murine cellular TM exerts a protective effect on thrombin-induced thromboembolism in mice ⁴² and infusion of isolated TM prevents hematologic abnormalities caused by the injection of tissue factors. ⁴³ This suggests that the TM expressed in endothelial cells of the cornea, trabecular meshwork and Schlemm’s canal and the nonpigmented ciliary epithelial cells may participate as an anticoagulant of blood in the aqueous humor. A decrease in TM expression in cells that contact the aqueous humor may be induced by the presence of cytokines, such as TNF-α and IL-1β, in the anterior chamber during inflammation. It is therefore important to investigate the expression of TM at several sites of eyes under pathologic conditions, such as inflammation and corneal injury. 

The expression of TM was observed in the basal cells, wing cells, and superficial cells of the peripheral corneal epithelium and limbus. These epithelial cells are formed by the differentiation and migration of multipotent stem cells in the limbal basal layer to produce the stratified squamous epithelium. ^{44 45} In the epidermis of the skin, TM has been reported to be absent in the basal layer and the surface cornified layer but strongly expressed in the keratinocytes of the suprabasal spinous layer. ¹² This difference may be associated with the tendency of the corneal epithelium to differentiate and proliferate. These findings suggest that the basal cells may be homologous to the keratinocytes in the epidermis. 

TM has been shown to be expressed at cell-to-cell contacts including the lung bud epithelium and the neural epithelium in mouse embryos, ¹³ and in desmoglein I, a chief desmosomal adhesion molecule, in human skin. ¹⁴ An immunohistochemical study of TM distribution in normal skin and the skin obtained from patients with acantholytic dermatoses revealed a high correlation between desmoglein I and the TM immunostaining pattern. This suggests a role for TM in intercellular adhesion of keratinocytes. ¹⁴ TM has an extracellular amino terminal domain resembling many of the lectinlike domains of adhesive molecules. ⁴⁶ This domain is not required for TM cofactor activity, ⁴⁷ which suggests that TM plays a role in cell adhesion and cell-to-cell interaction during epidermal differentiation. ¹² The EGF-like domain of TM also stimulates proliferation of fibroblasts. ⁴⁸ The strong TM expression in basal and wing cells in the limbus supports an association of TM with cell adhesion, differentiation, and/or proliferation. It is thus important to investigate the relationship between TM expression and the migration of the epithelial cells during wound healing in the cornea and limbal conjunctiva. 

There was a strong, unexpected expression of TM in the lens epithelial cells. TM localization in lens has never been shown in previous reports in rabbit and rat eyes. ^{36 37} The embryonic lens is formed at approximately 7 weeks and grows continuously throughout life, and only the epithelial cells have the ability to divide. ⁴⁹ Lens epithelial cells elongate at the lens equator, lose their nucleus, and become lens fibrocytes. ⁴⁹ It is possible that the TM in the lens epithelial cells also plays a role in adhesion, differentiation, and/or proliferation of these cells, as discussed earlier. 

To summarize our observations, TM in the cells, including corneal endothelial cells, nonpigmented ciliary epithelial cells, trabecular meshwork, and the endothelial cells lining Schlemm’s canal, which are in contact with the aqueous humor, appears to be involved in maintaining the fluidity of the aqueous humor. In contrast, TM in cells, including corneal epithelial cells, and limbus and lens epithelial cells, which are not in contact with the aqueous humor, may function in regulating cell proliferation and/or differentiation. 

 Submitted for publication January 31, 2000; revised June 13, 2000; accepted June 23, 2000.

 Commercial relationships policy: N.

 

Figure 1.

 

Light micrographs of the anterior segment of the eye immunohistochemically stained with mouse monoclonal anti-human TM IgG. (A) Low-power micrograph of the anterior segment of the eye, immunostained with anti-TM IgG. Central (1) and peripheral (2) areas of cornea, corneal endothelium, (3) limbus, (4) conjunctiva, (5) anterior chamber angle, (6) and ciliary body (7). (B) Corneal epithelium in the central area, stained with anti-TM IgG (magnified image of A-1). The surface coat and superficial cells showed moderate positive reactions to TM, and the basal cells showed weak positive reactions, but the wing cell layer was negative. No positive reaction was observed in the subepithelial Bowman’s membrane. (C) Corneal epithelium in the peripheral area stained with anti-TM IgG (magnified image of A-2). The corneal surface and superficial cells showed moderate to strong TM-positive reactions. The cytoplasm of the wing and basal cells exhibited moderate positive reactions. (D) Corneal endothelium stained with anti-TM IgG (magnified image of A-3). Corneal endothelial cells showed strong TM-positive reactions. Corneal keratocytes showed moderate TM-positive reactions. (E) Limbal area stained with anti-TM IgG (magnified image of A-4). The cytoplasm of epithelial basal cells exhibited moderate to strong positive reactions for TM. The cytoplasm of some superficial cells showed a moderate to strong reaction (arrow). Weak to moderate positive reactions were observed in the wing cells. (F) Conjunctiva stained with anti-TM IgG (magnified image of A-5). In the mucosal epithelium, basal cells showed weak positive reaction but superficial cells (arrow) reacted moderately. The lamina propria mucosa and capillary endothelial cells (arrowhead) showed moderate to strong positive reactions. Polygonal cells in the intermediate layer of the mucosal epithelium and free cells and fixed cells scattered in the lamina propria mucosa were negative for immunostaining. (G) Anterior chamber angle stained with anti-TM IgG (magnified image of A-6). Endothelial cells in the inner wall (arrow) and outer wall (arrowhead) of Schlemm’s canal react moderately. The trabecular meshwork (∗) also exhibited weak to moderate positive reactions. (H) Lens stained with anti-TM IgG. The lens epithelium showed moderate to strong positive reactions. The lens fiber in contact with the epithelium showed weak positive reactions. The lens capsule (arrow) was not stained. (I) Ciliary body stained with anti-TM IgG (magnified image of 1A-7). Nonpigmented ciliary epithelial cells (arrow) showed moderate to strong positive reactions. Stromal cells of ciliary processes showed weak or negative reactions. Pigmented epithelial cells (∗) were brown to black because of the presence of melanin in the cytoplasm. (J) Conjunctiva immunohistochemically stained with anti-TM IgG preabsorbed by excessive specific antigen (rTM). All histologic structures that showed TM-positive reactions in (F) were not stained. (K) Epithelium in the limbal area immunohistochemically stained using PBS instead of primary antibody. All histologic structures showing TM-positive reactions in (E) exhibited negative reactions. Part of the basal cell layer was brown because of the presence of melanin (arrow). Bar, (A) 500 μm; (B through K) 50 μm.

Figure 2.

 

RT-PCR analysis of TM mRNA in the human corneal epithelium and cultured HCECs. The expression of TM mRNA detected by RT-PCR on RNA samples from the human corneal epithelium and cultured HCECs was analyzed on a 2% agarose gel stained with ethidium bromide. TM in HUVECs and β-actin were used as a positive control in each sample. The expected PCR product sizes are as follows: TM, 473 bp; β-actin, 290 bp. M, DNA marker. Lanes 1 and 2: Human corneal epithelium; lanes 3 and 4: HCECs; lanes 5 and 6: HUVECs. Lanes 1, 3 and 5: RT performed with reverse transcriptase; lanes 2, 4, and 6: RT performed without reverse transcriptase.

Table 1.

 

Thrombomodulin Antigen Levels and Cofactor Activity in Cell Lysates

Lysate	Antigen			Cofactor Activity*	Specific Activity, †
		Ng/mg Protein	Ng/105 Cells
HCEC	54.8 ± 1.2	1.1 ± 0.1	48.4 ± 1.2	0.88 ± 0.02
HUVEC	317.7 ± 50.2	12.1 ± 0.3	381.4 ± 23.1	1.20 ± 0.07

 Values represent the mean ± SD of four determinations. HCEC, human corneal epithelial cells; HUVEC, human umbilical vein endothelial cells.

*  Expressed as picomoles of protein C activated per minute per milligram protein.

†  Expressed as picomoles of protein C activated per minute per nanogram TM.

Figure 3.

 

Western blot analysis of TM. Cell lysates were subjected to 7.5% SDS-PAGE under nonreducing (A) or reducing (B) conditions. Immunoblots were performed using a rabbit anti-human rTM polyclonal IgG. Lane 1: Protein (9.3 μg) from HCEC lysates; lane 2: protein (1.9 μg) from HUVEC lysates; lane 3: protein (37.2 μg) from HCEC lysates; lane 4: protein (15.2 μg) from HUVEC lysates. Left: Molecular markers.

The authors thank Kaoru Araki–Sasaki, Division of Ophthalmology, Toyonaka Municipal Hospital and Eiichi Shirasawa, Santen Pharmaceutical Co., Ltd., Osaka, Japan, for supplying cultured corneal epithelial cells and Mitsuyo Maeda, First Department of Anatomy, Osaka City University Medical School for her cooperation in preparation of sections.