August 1999
Volume 40, Issue 9
Free
Cornea  |   August 1999
Corneal Epithelial–Specific Cytokeratin 3 is an Autoantigen in Wegener’s Granulomatosis–Associated Peripheral Ulcerative Keratitis
Author Affiliations
  • Irena Reynolds
    From the Musculoskeletal Research Group,
  • Andrew B. Tullo
    University Department of Ophthalmology, and
  • Sally L. John
    ARC ERU, University of Manchester, and Central Manchester Health Care NHS Trust, Manchester, United Kingdom.
  • P. J. Lennox Holt
    From the Musculoskeletal Research Group,
  • M. Chantal Hillarby
    From the Musculoskeletal Research Group,
Investigative Ophthalmology & Visual Science August 1999, Vol.40, 2147-2151. doi:
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Irena Reynolds, Andrew B. Tullo, Sally L. John, P. J. Lennox Holt, M. Chantal Hillarby; Corneal Epithelial–Specific Cytokeratin 3 is an Autoantigen in Wegener’s Granulomatosis–Associated Peripheral Ulcerative Keratitis. Invest. Ophthalmol. Vis. Sci. 1999;40(9):2147-2151.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

purpose. In a previous investigation it was demonstrated that circulating antibodies to a 66-kDa corneal epithelial antigen (BCEA-A) are associated with peripheral ulcerative keratitis (PUK) in patients with Wegener’s granulomatosis (WG). The aim of this study was to identify BCEA-A.

methods. The 66-kDa antigen was purified from a bovine corneal epithelial protein extract, using DE52 ion exchange chromatography. Purified protein was used to raise rabbit polyclonal antibodies. These antibodies were used to screen a bovine corneal epithelial cDNA expression library. Positive clones were purified and sequenced. Clones were identified by DNA sequence homology searches of the GenBank DNA database.

results. A cDNA clone that demonstrated strong binding to both the rabbit polyclonal antibody and patient sera, showed 85% homology to rabbit cytokeratin 3 (K3). K3 is a basic cytokeratin specific to corneal epithelium. No bovine DNA sequence for K3 is available. However, bovine K3 is larger than rabbit K3, with a molecular weight of 66 kDa. Immunofluorescence using both patient sera and the rabbit antibody demonstrated a cytoplasmic binding pattern on human corneal epithelium.

conclusions. This evidence suggests that the 66-kDa autoantigen (BCEA-A) associated with PUK in WG is cytokeratin 3, and this may form the basis of a diagnostic/prognostic test.

Wegener’s granulomatosis (WG) is a rare inflammatory disease of unknown etiology that is characterized by vasculitis of the upper and lower respiratory tract, often in combination with glomerulonephritis. WG also can affect any other organ system, including the skin, eye, heart, nervous system, and gastrointestinal tract. 1 Early diagnosis of WG is difficult, but if diagnosed and treated promptly with immunosuppressive therapy and corticosteroids, the prognosis is much improved. Up to 90% of patients have circulating anti-neutrophil cytoplasmic antibodies. 2 The presence of these antibodies is commonly used as a diagnostic marker for this condition. 3  
Ophthalmic involvement may be present in up to 58% of WG cases. Moreover, in some cases ocular manifestations may be the major symptom or presenting feature of the disease. Although circulating anti-neutrophil cytoplasmic antibodies (usually anti–proteinase 3 antibodies) are a sensitive and specific marker for WG-associated scleritis 4 and are used in the early evaluation of patients, not all WG patients carry anti–proteinase 3 antibodies, particularly in early stages or limited presentation of the disease. In such cases, correct diagnosis and treatment may be delayed. Additional disease markers would, therefore, be a useful in the diagnostic procedure, helping to differentiate between ophthalmic manifestations of WG and other corneal inflammatory conditions with an autoimmune background. 
We have previously demonstrated the presence of autoantibodies to a corneal protein of 66 kDa (BCEA-A) in WG with and without peripheral ulcerative keratitis (PUK). 5 In this study, we have purified BCEA-A and used molecular techniques to identify it as cytokeratin 3. 
Methods
Methods of securing human and animal tissues complied with the National Institutes of Health Guidelines on the Care and Use of Animals in Research, the Declaration of Helsinki, and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 
DE-52 Purification of BCEA-A
Corneal epithelium was scraped from the central region of bovine corneas. Three hundred milligrams of tissue collected from 20 eyes was homogenized in 5 ml of 1 M NaCl, 40 mM Tris/HCl, pH 7.8, containing 1μ l 10 mM phenylmethylsulfonyl fluoride. The homogenized extract then was centrifuged at 20,000g for 30 minutes. The supernatant was dialyzed against distilled water, followed by 20 mM Tris/HCl, pH 8.0, for 48 hours. The concentration of the protein was estimated by spectrophotometer at a wavelength of 280 nm. The integrity of the corneal extract was assessed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and Coomassie Blue staining to ensure that the proteins were not degraded. The soluble corneal extract was dialyzed into 0.05 M NaCl, 20 mM Tris/HCl, pH 7.8, and diluted to a protein concentration of 29 mg/ml in the same buffer. Five milliliters of the extract was loaded on to a DE-52 column. Unbound proteins were eluted from the column with 0.05 M NaCl, 20 mM Tris/HCl, pH 7.8. The resulting fraction was dialyzed into 20 mM Tris/HCl, pH 7.8, analyzed by SDS-PAGE, and immunoblotted with patient sera containing antibodies to BCEA-A. 
SDS-PAGE and Immunoblotting
Protein from tissue extracts or antigen-enriched protein preparations were separated by SDS-PAGE and immunoblotted by standard techniques. 5  
Production of Rabbit Polyclonal Antibodies to BCEA-A
A BCEA-A–enriched extract was separated by SDS-PAGE and electroblotted onto a pure nitrocellulose membrane. The membrane was stained with Ponceau S to identify the relevant band. The identified protein band was excised from the membrane and used to raise polyclonal antibodies as described previously. 5  
RNA Extraction
Bovine corneal epithelium was powdered in liquid nitrogen and taken up in RNAzolB (Biogenesis, Poole, UK). Total RNA was purified according to the manufacturer’s instructions. 
cDNA Library Preparation and Screening
A cDNA expression library was made from bovine corneal epithelial mRNA in the Uni-ZAP-XR lambda vector (Stratagene, Amsterdam, The Netherlands) according to the manufacturers instructions. Briefly, cDNA was synthesized by reverse transcription using a polyT primer containing a XhoI restriction site at the 3′ end. EcoRI adapters were ligated to the two blunt termini, followed by a double digestion with XhoI/EcoRI. The resulting product was directionally cloned into the EcoRI/XhoI site of Uni-ZAP-XR vector. The resulting library contained 2 × 108 recombinant clones. The library was transfected into 200 μl of Escherichia coli XL-1 MRF′ cells, and 5 NZY agar plates, each containing 50,000 clones, were prepared for immunoscreening with the rabbit polyclonal BCEA-A antibodies. The antisera was diluted 1:2000 and preincubated with 12 mg/ml E. coli lysate prior to the library screening. Immunoscreening was carried out according to the Stratagene protocol. Positive clones were purified by secondary and tertiary screening. 
DNA Sequencing
Pure clones were grown up, and phage DNA was extracted using phage DNA extraction columns according to the manufacturer’s instructions (Qiagen, Crawley, UK). The size of inserts was measured on agarose gels after polymerase chain reaction amplification, using primers specific to the T3 (5′AAT TAA CCC TCA CTA AAG GG3′) and T7 (5′GTA ATA CGA CTC ACT ATA GGG C3′) binding sites. Samples were amplified for 40 cycles of 45 seconds at 94°C, 45 seconds at 60°C, and 80 seconds at 72°C. PCR products were cloned into the TA cloning vector (Invitrogen, Groningen, The Netherlands) according to manufacturer instructions. Plasmids containing inserts of the correct size were purified using Qiagen columns. Inserts were sequenced directly by cycle sequencing (BIG dye primer kit; Perkin Elmer, Warrington, UK) using M13 forward (5′GTA AAA CGA CGG CCA G3′) and M13 reverse (5′CAG GAA ACA GCT ATG AC3′) primers. Sequenced samples were run on a Perkin Elmer 377 automated DNA sequencer. Sequences obtained were used in DNA homology searches of the GenBank DNA databases using FastA. 
Tissue Localization of BCEA-A by Immunofluorescence
Fresh human eyes that were unsuitable for corneal transplant were obtained from the Manchester Eye Bank and stored in a moist chamber with a balanced salt solution until the sample was processed. A corneoscleral disc was removed, using a trephine and scissors. The disc was then cut into blocks of approximately 2 × 10 mm, ensuring that the limbus was included. Other human tissues including esophagus, oral mucosa, liver, skin, lungs, and kidney were obtained postmortem, frozen, and cut into 5-μm sections. The tissue was snap frozen in liquid nitrogen–cooled isopentane and mounted in OCT compound. BCEA-A expression was localized in these tissues as described previously. 5  
Results
Tissue Localization of BCEA-A
Frozen sections of normal human corneas were used as the substrate in indirect immunofluorescence. The rabbit anti–BCEA-A antibodies detected antigen in human corneal epithelium showing a cytoplasmic binding pattern (Fig. 1) . There was no binding to the underlying sections of cornea, i.e., Bowman’s membrane, stroma, Descemet’s membrane, or endothelium. 
Frozen sections of normal human tissues also were used as a substrate for rabbit anti–BCEA-A antibody binding. The selected tissues included kidney, lung, liver, and skin, which are often involved in WG. The other tissues, intestine and esophagus, have a squamous epithelial layer similar to the epithelial layer of the eye. The anti–BCEA-A antibody bound only the epidermal layer of the skin, showing a cytoplasmic binding pattern, similar to the binding seen in the corneal epithelium (Fig. 1)
Identification of BCEA-A
A bovine corneal epithelial Uni-ZAP-XR lambda expression library was produced and 1 × 105 recombinant clones were screened with rabbit anti–BCEA-A antibodies. One clone (A10) was identified as positive. The rabbit antibodies were not purified for their antigen specificity; therefore, it was possible that the antibodies, which recognized clone A10, were different from those that recognize BCEA-A. Antibodies bound to A10 were eluted from the tertiary screening filter and used to probe a western blot of BCEA-A–enriched protein extract. The eluted antibodies bound to a 66-kDa band. 
The 550-bp insert in clone A10 was sequenced, and a DNA homology search revealed strong homology with rabbit cytokeratin 3 and human keratin 2 (Table 1) . The DNA sequence of A10 was translated into a putative amino acid sequence, and protein homology searches were performed. The results from these searches confirmed the strong homology between clone A10 and basic cytokeratins 3 and 2 (Table 1 , Fig. 2 ). 
To confirm the identity of BCEA-A as keratin, a human epidermal keratin extract (ICN) containing both basic and acidic keratins was probed with both human sera and rabbit anti–BCEA-A antibodies. The keratin preparation was resolved by SDS-PAGE along with a BCEA-A–enriched protein extract. Immunoblotting with rabbit anti–BCEA-A antibodies and human serum known to carry anti–BCEA-A antibodies revealed binding to a band slightly above 66 kDa (Fig. 3)
Discussion
We previously reported the association of antibodies to a basic 66-kDa corneal epithelial–derived antigen (BCEA-A) and PUK-associated WG. 5 To further characterize this antigen, rabbit polyclonal antibodies were raised and used to localize CEA-A by indirect immunofluorescence, which demonstrated that BCEA-A is expressed in the epithelium of human cornea and in human skin epidermis. A bovine corneal epithelial cDNA library was screened with the rabbit anti–BCEA-A antibodies and a positive clone was isolated, purified, and sequenced. The clone showed 85% DNA homology with a rabbit corneal–specific cytokeratin. This as well as other indirect evidence indicates that BCEA-A is cytokeratin 3. 
The clone A10, identified by antibody binding, showed 85% DNA homology and 87.9% amino acid homology with rabbit cytokeratin 3 (K3). The homology of the A10 clone also was compared with other members of cytokeratin family and with cytokeratins across species. The result revealed 78% amino acid homology with human skin keratin 2 (K2), and 84% homology with human K3. Neither the DNA nor the amino acid sequence of bovine K3 is available on databases. However, human and rabbit K3 sequences showed 80% amino acid homology to one another. Therefore, it is likely that the homology of 87.9% between A10 clone and rabbit K3 is due to the species differences between rabbit and bovine K3. 
K3 is a member of the intermediate filament (IF) superfamily of proteins. Keratins are the major structural components of the cytoskeleton that are exclusively expressed by epithelial cell types throughout the body. K3 is a basic component in a pair with acidic cytokeratin 12 (K12); both proteins are cornea specific, and bovine K3 weighs 66 kDa. 6 These characteristics make K3 a good candidate for BCEA-A, which is corneal epithelial–specific, basic, and 66 kDa. 
To further investigate the possible identity of BCEA-A as K3, a preparation of human skin epidermal keratins was immunoblotted with rabbit anti–BCEA-A antibodies and with human serum positive to BCEA-A. This preparation was chosen, in the absence of purified corneal epithelial keratin, because human K2 and K3 share 80% identity, therefore, polyclonal antibodies to K3 are likely to cross-react with K2. The experiment revealed a very strong binding of rabbit anti–BCEA-A antibodies to a spectrum of keratins present in skin epidermal preparation. It also showed pronounced binding of human serum antibodies to a band positioned slightly higher than 66 kDa (human K2 is 67 kDa) in the same preparation. Both types of antibody also recognized a 66-kDa band in both nonpurified extract and BCEA-A–enriched extract. This result strongly suggests that the original antigen against which anti–BCEA-A antibodies were raised is K3. 
To further confirm the identity of BCEA-A as K3, immunohistochemistry was used to localize the antigen within human skin and cornea. In a study of K3 expression in rabbit cornea, using the highly specific monoclonal antibody AE5, it has been demonstrated that K3 is localized to corneal epithelium and the suprabasal layer of the limbus. 7 The localization of BCEA-A using indirect immunofluorescence on frozen corneal sections showed that rabbit and human antibodies recognize an antigen in both the corneal and limbal part of the epithelium, in keeping with the expected pattern for K3. The cross-reactivity of anti BCEA-A antibodies with skin epidermal keratins has been confirmed by indirect immunofluorescence (IIF) on frozen human skin sections. The antibody binding on these sections was limited to the epidermal layer above the basal layer of the human skin, where K2 (the skin equivalent of K3) is localized. 8 Moreover, the homology search demonstrated that K2 has the highest amino acid and DNA homology to K3 and clone A10, in comparison with other human keratins. This implies that K2 is a likely target for cross-reacting anti–BCEA-A antibodies, which may explain the lack of binding of anti–BCEA-A antibodies to other tissues, such as intestine and esophagus, with different type of epithelium containing basic type keratins of much lower homology to K3, e.g., K4 (esophageal equivalent of K3), which shows only 68% homology with clone A10 and K3. The above results strongly suggest that the target antigen for anti–BCEA-A antibodies in the cornea is K3. 
Autoantibodies against constituents of the cytoskeleton have been reported and investigated by a number of groups. The best characterized are anti-filaggrin antibodies, 9 which are specific for RA and often appear before clinical onset of the disease. 10 However, to the date there have been no reports of anti-cytoskeleton antibodies present in primary systemic vasculitis. 
We previously have demonstrated that a high percentage of WG patients, especially those with eye complications, have antibodies to a corneal protein (BCEA-A). In this study we identified this protein as cytokeratin 3. Identification of disease-specific markers provides a tool for investigation into disease pathology and etiology and may also provide new diagnostic or prognostic markers. Further investigation will be required to define the role of anti-K3 antibodies in disease pathology and their suitability as diagnostic markers for WG and PUK. 
Anti-K3 antibodies are present in a large proportion of WG patients, in particular those with eye complications. Even though it is unclear what the role or source of these antigens is, they may prove a valuable diagnostic marker in WG and PUK. 
 
Figure 1.
 
Detection of BCEA-A in noncorneal tissues. Several human tissues, including cornea (A, C), kidney and skin (B, D), lung, esophagus, intestine, and liver, were screened with rabbit anti–BCEA-A sera. BCEA-A was only detected in the cornea (A) and skin (C) sections. No positive staining was detected when the secondary antibodies were used alone (B, D). e, epithelium; s, stroma; ed, epidermis; m, dermis.
Figure 1.
 
Detection of BCEA-A in noncorneal tissues. Several human tissues, including cornea (A, C), kidney and skin (B, D), lung, esophagus, intestine, and liver, were screened with rabbit anti–BCEA-A sera. BCEA-A was only detected in the cornea (A) and skin (C) sections. No positive staining was detected when the secondary antibodies were used alone (B, D). e, epithelium; s, stroma; ed, epidermis; m, dermis.
Table 1.
 
Homology between the cDNA and Putative Amino Acid Sequence of Clone A10 and Basic Keratin Sequences
Table 1.
 
Homology between the cDNA and Putative Amino Acid Sequence of Clone A10 and Basic Keratin Sequences
Comparison DNA Amino Acid
Position (bp) Homology (%) Position (aa) Homology (%)
Clone A10 vs rabbit K3* , † 432–670, 1600–1769 85 116–238 85
Clone A10 vs human K3, ‡ 118–253 84
Clone A10 vs human K2, § , ∥ 372–907 79 109–236 78
Human K3, ‡ vs rabbit K3, † 80
Figure 2.
 
Amino acid sequence alignment of clone A10 with rabbit and human keratin. The putative amino acid sequence of clone A10 compared to rabbit and human keratin 3 and human keratin 2. Areas of nonhomology are shaded.
Figure 2.
 
Amino acid sequence alignment of clone A10 with rabbit and human keratin. The putative amino acid sequence of clone A10 compared to rabbit and human keratin 3 and human keratin 2. Areas of nonhomology are shaded.
Figure 3.
 
Binding of rabbit and human anti-BCEA-A antibodies to human skin epidermal keratin. BCEA-A enriched protein extract (1) and human epidermal keratins (ICN) (2) were separated by SDS-PAGE, blotted, and probed with rabbit anti–BCEA-A antibodies (a) and patient sera (b). m, moleculer weight markers.
Figure 3.
 
Binding of rabbit and human anti-BCEA-A antibodies to human skin epidermal keratin. BCEA-A enriched protein extract (1) and human epidermal keratins (ICN) (2) were separated by SDS-PAGE, blotted, and probed with rabbit anti–BCEA-A antibodies (a) and patient sera (b). m, moleculer weight markers.
Leavitt RY, Fauci AS, Bloch DA, et al. The America College of Rheumatology 1990 Criteria for the Classification of Wegener’s Granulomatosis. Arthr Rheum. 1990;33:1101–1107.
Kallenberg CGM, Mulder AHL, Tervaert JWC. Antineutrophil cytoplasmic antibodies—a still-growing class of autoantibodies in inflammatory disorders. Am J Med. 1992;93:675–682. [CrossRef] [PubMed]
Jayne DR, Gaskin G, Pusey CD, Lockwood CM. ANCA and predictiotory relapse in systemic vasculitis. Mon J Assoc Physicians. 1995;88:127–133.
Jager MJ, Vos A, Pasmans S, Hoekzema R, Broersma L, Vandergaag R. Circulating cornea-specific antibodies in corneal disease and cornea transplantation. Graefes Archive Clin Exp Ophthalmol. 1994;232:82–86. [CrossRef]
Reynolds I, John SL, Tullo AB, et al. Characterization of two corneal epithelium derived antigens associated with vasculitis. Invest Ophthlmol Vis Sci. 1998;39:2594–2601.
Cooper D, Sun TT. Monoclonal-antibody analysis of bovine epithelial keratins—specific pairs as defined by coexpression. J Biol Chem. 1986;261:4646–4654. [PubMed]
Schermer A, Galvin S, Sun TT. Differentiation-related expression of a major 64k corneal keratin in vivo and in culture suggests limbal location of corneal epithelial stem-cells. J Cell Biol. 1986;103:49–62. [CrossRef] [PubMed]
Sun TT, Eichner R, Nelson WG, et al. Keratin classes—molecular markers for different types of epithelial differentiation. J Invest Dermatol. 1983;81:S109–S115. [CrossRef]
Girbal E, Sebbag M, Gomesdaudrix V, Simon M, Vincent C, Serre G. Characterization of the rat esophagus epithelium antigens defined by the so-called antikeratin antibodies, specific for rheumatoid-arthritis. Ann Rheum Dis. 1993;52:749–757. [CrossRef] [PubMed]
Cordonnier C, Meyer O, Palazzo E, et al. Diagnostic-value of anti-ra33 antibody, antikeratin antibody, antiperinuclear factor and antinuclear antibody in early rheumatoid-arthritis—comparison with rheumatoid-factor. Br J Rheumatol. 1996;35:620–624. [CrossRef] [PubMed]
Figure 1.
 
Detection of BCEA-A in noncorneal tissues. Several human tissues, including cornea (A, C), kidney and skin (B, D), lung, esophagus, intestine, and liver, were screened with rabbit anti–BCEA-A sera. BCEA-A was only detected in the cornea (A) and skin (C) sections. No positive staining was detected when the secondary antibodies were used alone (B, D). e, epithelium; s, stroma; ed, epidermis; m, dermis.
Figure 1.
 
Detection of BCEA-A in noncorneal tissues. Several human tissues, including cornea (A, C), kidney and skin (B, D), lung, esophagus, intestine, and liver, were screened with rabbit anti–BCEA-A sera. BCEA-A was only detected in the cornea (A) and skin (C) sections. No positive staining was detected when the secondary antibodies were used alone (B, D). e, epithelium; s, stroma; ed, epidermis; m, dermis.
Figure 2.
 
Amino acid sequence alignment of clone A10 with rabbit and human keratin. The putative amino acid sequence of clone A10 compared to rabbit and human keratin 3 and human keratin 2. Areas of nonhomology are shaded.
Figure 2.
 
Amino acid sequence alignment of clone A10 with rabbit and human keratin. The putative amino acid sequence of clone A10 compared to rabbit and human keratin 3 and human keratin 2. Areas of nonhomology are shaded.
Figure 3.
 
Binding of rabbit and human anti-BCEA-A antibodies to human skin epidermal keratin. BCEA-A enriched protein extract (1) and human epidermal keratins (ICN) (2) were separated by SDS-PAGE, blotted, and probed with rabbit anti–BCEA-A antibodies (a) and patient sera (b). m, moleculer weight markers.
Figure 3.
 
Binding of rabbit and human anti-BCEA-A antibodies to human skin epidermal keratin. BCEA-A enriched protein extract (1) and human epidermal keratins (ICN) (2) were separated by SDS-PAGE, blotted, and probed with rabbit anti–BCEA-A antibodies (a) and patient sera (b). m, moleculer weight markers.
Table 1.
 
Homology between the cDNA and Putative Amino Acid Sequence of Clone A10 and Basic Keratin Sequences
Table 1.
 
Homology between the cDNA and Putative Amino Acid Sequence of Clone A10 and Basic Keratin Sequences
Comparison DNA Amino Acid
Position (bp) Homology (%) Position (aa) Homology (%)
Clone A10 vs rabbit K3* , † 432–670, 1600–1769 85 116–238 85
Clone A10 vs human K3, ‡ 118–253 84
Clone A10 vs human K2, § , ∥ 372–907 79 109–236 78
Human K3, ‡ vs rabbit K3, † 80
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×