|Molecular Vision 2006;
Received 8 July 2005 | Accepted 7 April 2006 | Published 10 May 2006
Systemic investigation of keratoepithelin deposits in TGFBI/BIGH3-related corneal dystrophy
Ilhem El Kochairi,1
Igor Letovanec,2 Sylvie Uffer,3 Francis L. Munier,3
Pascal Chaubert,2 Daniel F. Schorderet1,3
1Institut de Recherche en Ophtalmologie, Sion, Switzerland; 2Institut de Pathologie, Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland; 3Jules Gonin Eye Hospital, Lausanne, Switzerland
Correspondence to: Ilhem El Kochairi, Institut de Recherche en Ophtalmologie, 64, Avenue Grand Champsec, CH-1950, Sion, Switzerland; Phone: +41 (0)27 205 79 00; FAX: +41 (0)27 205 79 01; email: firstname.lastname@example.org
Purpose: To investigate the location and tissue-specificity of the pathologic keratoepithelin (KE) deposition in a patient with a keratoepithelinopathy (KEP), TGFBI/BIGH3-related corneal dystrophy.
Methods: An autopsy was performed in a patient with lattice type I corneal dystrophy (LCDI) after authorization was obtained from the family. Mutation screening in TGFBI/BIGH3 was done on the patient several years ago. Eighteen different tissues or organs, including brain, heart, lung, kidney, liver, lymph nodes, spleen, aorta, esophagus, bone marrow, urinary bladder (including a papillary urothelial carcinoma), samples of a metastatic squamous cell carcinoma, adrenal gland, parathyroid gland, muscle, prostate, and cornea were investigated, and sections from the tissues were labeled with KE2 rabbit TGFBI/BIGH3 antiserum.
Results: The patient, diagnosed with LCDI and Alzheimer's disease, died at 79 years of age from a complicated chronic obstructive lung disease. Mutation analysis showed the classical Arg124Cys mutation in exon 4 of TGFBI/BIGH3, associated with LCDI. Except for the cornea, immunostaining with KE2 antisera did not reveal any deposits in any of the 17 other organs analyzed.
Conclusions: Pathologic deposits caused by KE accumulation were only observed in the cornea and in no other tissue or organ in this patient. These results suggest a cornea-specific mechanism in the aggregation of KE. Further studies need to be done to investigate whether the degradation of mutated KE generates cornea-specific fragments that aggregate or whether the clearing of normal fragments is different in affected corneas, which then leads to aggregation.
Corneal dystrophies are inherited diseases of the cornea characterized by the progressive accumulation of deposits in the cornea . Eight years ago, we reported that allelic mutations in one single gene, the TGFBI/BIGH3 gene located on 5q31 and coding for a 683 amino acid secreted protein called keratoepithelin (KE), were responsible for many of these diseases . KE is a member of the extracellular matrix and is ubiquitously found, but not in the brain . We showed in mouse embryos that expression starts as early as 10.5 days after conception with high expression in all the mesoderms . Until now, all the pathogenic mutations reported on the TGFBI/BIGH3 gene were linked to corneal dystrophies and were not related to other diseases.
KE is expressed in various organs [5,6] and is also present in blood serum . However, clinical examination of patients affected with 5q31-linked corneal dystrophies has never revealed KE deposits anywhere else than in the cornea. Schmitt-Bernard et al.  investigated the skin of a patient carrying LCDI. Although KE is abundantly expressed in the skin, they did not find amyloid deposits in this tissue, even after carefully checking with electron microscopy. Furthermore, Klintworth  has also added to this statement that despite the deposition of amyloid within the cornea in numerous inherited diseases, the corneal tissue is not affected in many genetically determined forms of amyloidosis that affect other tissues. Corneal deposits in TGFBI/BIGH3-related corneal dystrophies seem to be a tissue-specific local process. Although it has regularly been mentioned that deposits are only found in the corneas of affected patients [10-15], no systematic search for deposits elsewhere in the body has been performed and no autopsy of a patient with TGFBI/BIGH3-related corneal dystrophy has been reported.
The exact role of KE in the pathogenesis of the 5q31-linked corneal dystrophies is still unclear. Korvatska et al.  did present several hypotheses which could explain the origin of the corneal deposits in these inherited corneal dystrophies. One is the possible role of cornea-specific enzymes in the catabolism of normal and mutated KE. Alternatively, the local conditions observed in the cornea may favor the spontaneous aggregation of mutated KE peptides. Schmitt-Bernard et al.  have also presented a similar hypothesis of an alteration in the degradation of the mutated TGFBI/BIGH3 protein and a participation of local factors to explain KE amyloidogenesis.
It has already been shown that TGFBI/BIGH3 transcripts and KE are intensively synthesized in the corneal epithelium of the adult eye and that primary corneal fibroblasts established from normal and 5q31-affected corneas secreted KE in the culture medium . Recently, it was also shown that KE itself constituted most, if not all, of the content of the deposits, as opposed to a recruitment of other amyloidogenic peptides or proteins [16,17]. Finally, these authors showed that abnormal KE isoforms specifically observed in diseased corneas were not produced by cultured fibroblasts from these same corneas.
We present the first report of an autopsy of a patient affected with LCDI due to an Arg124Cys mutation. We did not observe deposits other than in the cornea. This suggests a corneal-specific protein processing that may explain why the cornea is the only site where pathologic deposits are found [10-15].
An autopsy was performed after informed consent was obtained from the family. After careful macroscopic examination, the heart, lung, brain, and samples from liver, spleen, pancreas, kidney, esophagus, aorta, bone marrow, and lymph nodes were routinely processed and fixed in a 6% buffered formaldehyde solution. After dehydration and embedding in paraffin, 4 μm sections were stained with hematoxylin and eosin (H&E) and Congo red.
KE2 antiserum and anti-amyloid P component (anti-P) were used as described previously in the literature . Preimmune serum was also used to detect nonspecific staining and improve specificity. Samples of brain, heart, lung, kidney, liver, lymph nodes, spleen, aorta, oesophagus, bone marrow, urinary bladder (including a papillary urothelial carcinoma), adrenal gland, parathyroid gland, muscle, prostate, and samples of a metastatic squamous cell carcinoma were selected for immunohistochemical analysis. Samples of cornea obtained after keratoplasty of our patient were selected as a positive control.
All selected samples were immunostained using the peroxidase-anti-peroxidase (PAP) technique of Sternberger et al.  Tissue sections (4 μm thick) were mounted on Superfrost® Plus slides (Milian SA, Geneva, Switzerland) deparaffinized in xylol, taken through to absolute alcohol, blocked for endogenous peroxidase with 0.1% hydrogen peroxidase in methanol for 45 min, and rehydrated through graded alcohols. Sections were incubated with KE2 antibody and preimmune serum. They were all first digested in 0.05% Pronase (Dako, Glostrup, Denmark) for 15 min at room temperature. To reduce nonspecific primary antibody binding, the slides were incubated for 10 min with normal goat serum (PelFreez Biologicals, Rogers, AR) diluted 1:30 in Tris-buffered saline (TBS). The primary antibody KE2 (1:50), preimmune serum and anti-P component (1:40; Dako), secondary antibodies, goat anti-rabbit immunoglobulins (1:100) and PAP complex (1:400; Sternberger Monoclonals, Baltimore, MD) were diluted in TBS containing 2 mg/ml skim milk. All incubations were for 30 min at room temperature, followed by washing in TBS.
The patient died at 79 years of age of bronchopneumonia and respiratory failure. He was suffering from mild Alzheimer's disease. An older sister has the disease as well . Both individuals were part of a large Swiss family affected with LCDI. The patient was diagnosed when he was an adult (about 50 years ago) and molecular analysis, performed in 1996, revealed an Arg124Cys mutation in exon 4 of the TGFBI/BIGH3 gene . His sister was not affected and did not carry the mutation.
A poorly differentiated squamous cell carcinoma of the inguinal fold with widespread metastases (inguinal, paraaortic, lung hilar lymph node, and multiple lung metastases) was diagnosed. Lung examination revealed also bronchopneumonia and emphysema with subsequent right heart hypertrophy. A noninvasive papillary urothelial carcinoma was also diagnosed. Grading was not possible due to autolysis artefacts. For the purpose of this study, all H&E stained slides were carefully examined and no unusual eosinophil deposits were observed.
Corneal samples obtained after keratoplasty of our patient, known for hereditary corneal dystrophy (LCDI), several years before his death. These samples were used as a positive control and showed eosinophil/amorphous fusiform deposits upon H&E staining in the anterior part of the corneal stroma (Figure 1A). The amyloid nature of the deposits was confirmed by Congo red staining (Figure 1A) and immunostaining using nonspecific antisera to amyloid P component (not shown) which are nonspecific to disease type. The same deposits showed intense staining with the specific KE2 antibody (Figure 1D). Little nonspecific staining was present as shown in Figure 1C after incubation with preimmune serum. All selected and immunostained autopsy samples were carefully examined and no KE2-positive pathological deposits were present in any of the selected organs (Figure 2, Figure 3) or tumor samples (Figure 4). The staining of the samples incubated with either specific KE2 antibody or the preimmune serum is similar. Except for the cornea, immunostaining with KE2 antiserum and anti-P as well as staining with Congo red did not reveal any deposit in all the other tissues.
The expression of TGFBI/BIGH3 has been demonstrated in different tissues by several groups. In the present study, deposition of KE reactive material was investigated in samples from 16 different organs and also from a squamous cell and a papillary urothelial carcinoma. Except for the cornea, immunostaining with KE2 antiserum did not reveal any deposit in all the other tissues. Samples were not available for transcript or protein quantification and we therefore cannot say whether the expression of TGFBI/BIGH3 was normal or increased in these tissues. However, Billings et al.  have shown that immunohistochemical analysis of human lung localized Big-h3 to the vascular and airway extracellular matrix (ECM), and particularly to the septal tips of alveolar ducts and alveoli, and that TGFBI/BIGH3 protein is produced in lung by bronchial smooth muscle cells. They also mentioned that no abnormal findings relative to the lung have been reported in patients with TGFBI/BIGH3 related corneal dystrophy. The same authors reported a positive immunostaining of the matrix and urothelium in intact bladder tissue with anti-TGFBI/BIGH3 antibodies and high amounts of the TGFBI/BIGH3 protein secreted by cells obtained from human bladder tissue. In addition, TGFBI/BIGH3 has been found upregulated in many cancers such as colorectal tumors  and esophageal adenocarcinoma . TGFBI/BIGH3 mRNA expression was found to be 32.4 fold increased in pancreatic cancer in comparison to normal control tissues . Finally, renal expression of TGFBI/BIGH3 has also been shown in the rat .
Except in the cornea, no pathogenic mutation of TGFBI/BIGH3 has been related to a disease in any other tissue. KE is a widely expressed protein in the ECM and may play an important role in organogenesis and in the maintenance of tissue architecture. Analysis of the cDNA deduced amino acid sequence of KE identified within the second and fourth internal fasciclin-like repeat domains shows two conserved sequences (Asn-Lys-Asp-Ile-Leu and Glu-Pro-Asp-Ile-Met, respectively) known to mediate cell adhesion . In addition, the developmental expression pattern of TGFBI/BIGH3 in the mouse embryo shows that TGFBI/BIGH3 transcripts are abundantly expressed in the mesenchyme of most organs . However, all of the mutations reported so far in TGFBI/BIGH3 are exclusively related to different forms of inherited corneal dystrophy . Our patient was suffering from LCDI, characterized by a progressive accumulation of amyloid corneal deposits.
Corneal amyloidogenesis in TGFBI/BIGH3 related corneal dystrophy is likely a tissue and protein specific process . Systemic amyloidoses are often the result of a chronic inflammatory process , while corneal amyloid deposits in TGFBI/BIGH3 related corneal dystrophy preceed the development of recurrent corneal erosions, rather than developing secondarily. This has been strongly suggested in a case reported by Aldave et al. , where a TGFBI/BIGH3 related corneal dystrophy patient had subepithelial opacification in his left eye without any history of corneal erosion in this eye. Billings et al.  have reported that proinflammatory cytokine, interleukin 1β, which upregulates a number of proinflammatory genes, had no effect on TGFBI/BIGH3 expression levels. The mechanism leading to corneal amyloidosis is still unclear.
The analysis of KE structure may help in understanding this corneal amyloidogenesis process. It has been shown that corneal deposits are constituted by the mutant KE protein itself [16,17]. KE is an extracellular matrix protein containing an NH2-terminal secretory signal peptide, four 140 amino acid repeats with internal homology, and an arg-gly-asp (RGD) motif at the COOH-terminus . The RGD motif is found in many extracellular matrix proteins modulating cell adhesion and serves as a ligand recognition sequence for several integrins [27-29]. Most of the mutations in TGFBI/BIGH3 occur at the two arginine codons Arg124 and Arg555 [2,30]. Schmitt-Bernard et al.  have assessed the influence of the structure of mutant TGFBI/BIGH3 protein related peptides centered on codon 124 on amyloid formation. They have shown that the particular amino acid occupying position 124, the presence of valine-valine at positions 112-113, and the existence of disulfide and hydrogen bonding, are of utmost importance in determining the tendency of TGFBI/BIGH3 protein related peptides to adopt a β-pleated sheet configuration and to form amyloid. They showed that of all the peptides analyzed, and on the contrary to the native Arg110-131 peptide (amino-acid sequence between codons 110 and 131), Cys110-131 is the peptide with the greatest tendency to form amyloid fibrils . Since cysteine has the capacity to form disulfide bonds, it participates in the appearance of aggregates and predisposes to the formation of amyloid fibrils. Besides disulfide bonds, the same authors suggest that other factors participate to amyloidogenesis. Hydrophobicity is an important feature in the formation of β-pleated insoluble aggregates. The native Val112-Val113 hydrophobic sequence has been implicated in amyloid fibril formation from the TGFBI/BIGH3 protein derived peptides mutated at codon 124 .
As suggested in a previous report , a corneal specific protein-protein interaction may explain the aggregation of abnormal KE isoforms that can also be produced by a corneal specific protease. Takacs et al.  showed that the 42 kDa protein variant found in LCDI corneas starts at residue 26 (Ala) whereas the NH2-terminal amino acid of the non mutated KE is Gly24. An altered processing of the mutated KE protein could then explain the amyloid corneal deposition. It could be interesting to identify the KE isoforms that may be present in other organs. If these isoforms are different from those identified in the cornea, the investigations should be oriented toward a corneal specific protease as a likely causal element of 5q31 corneal dystrophies. Stix et al.  confirmed very recently that amyloid deposits in LCDI corneas contain proteolytic fragments of KE. They proposed the hypothesis of an unstable globular state of the mutated KE that leads to unfolded polypeptide chains favoring proteolytic cleavage and intermolecular association leading to amyloid formation. Proteolysis has also been presented as a mechanism for clearing the pathologic corneal protein aggregates. Whether secretion of the mutated protein or an inefficient clearing process is the causal mechanism of the corneal deposit is still unknown. As an ECM protein, KE interacts with other matrix proteins, such as collagen type I, II, IV, and VI [33-35], in order to maintain the corneal tissue architecture which is essential for corneal transparency. But the interaction between mutant KE and the other ECM proteins does not seem to be disrupted . Aberrant interactions between mutant KE or its degradation products and ECM proteins or other corneal components may also explain KE related amyloidogenesis. Since deposits in TGFBI/BIGH3 related corneal dystrophy are localized mainly in the center of the cornea as opposed to BIGH3 being expressed in both the central and peripheral cornea, identification of a corneal component that is present mainly in the central cornea could help understanding KE aggregation.
In summary, our study confirms that deposits are only present in the cornea of patients affected with LCDI and probably in no other tissue, including tumoral tissues. This suggests that specific conditions may exist in the cornea. Such conditions could be responsible for the production of abnormal KE through a cornea specific catabolic pathway. Alternatively, cornea-specific physical or biochemical conditions may trigger the aggregation of abnormal KE products. Further investigations should focus on corneal specific components in order to identify the pathological mechanism of the 5q31 related corneal dystrophies.
This research was supported by the Fondation pour la Recherche sur les Maladies Hereditaires and the Swiss National Science Foundation grant number 31-66713. This data was presented at the 2005 ARVO meeting.
1. Munier F, Schorderet D, Uffer S. Dystrophies héréditaires de la cornée. In: Oeil et Génétique. Dufier JL, Kaplan J, editors. Issy-les-Moulineaux (France): Masson; 2005. p. 139-58.
2. Munier FL, Korvatska E, Djemai A, Le Paslier D, Zografos L, Pescia G, Schorderet DF. Kerato-epithelin mutations in four 5q31-linked corneal dystrophies. Nat Genet 1997; 15:247-51.
3. Ferguson JW, Mikesh MF, Wheeler EF, LeBaron RG. Developmental expression patterns of Beta-ig (betaIG-H3) and its function as a cell adhesion protein. Mech Dev 2003; 120:851-64.
4. Schorderet DF, Menasche M, Morand S, Bonnel S, Buchillier V, Marchant D, Auderset K, Bonny C, Abitbol M, Munier FL. Genomic characterization and embryonic expression of the mouse Bigh3 (Tgfbi) gene. Biochem Biophys Res Commun 2000; 274:267-74.
5. Skonier J, Neubauer M, Madisen L, Bennett K, Plowman GD, Purchio AF. cDNA cloning and sequence analysis of beta ig-h3, a novel gene induced in a human adenocarcinoma cell line after treatment with transforming growth factor-beta. DNA Cell Biol 1992; 11:511-22.
6. Escribano J, Hernando N, Ghosh S, Crabb J, Coca-Prados M. cDNA from human ocular ciliary epithelium homologous to beta ig-h3 is preferentially expressed as an extracellular protein in the corneal epithelium. J Cell Physiol 1994; 160:511-21.
7. Klintworth GK, Valnickova Z, Enghild JJ. Accumulation of beta ig-h3 gene product in corneas with granular dystrophy. Am J Pathol 1998; 152:743-8.
8. Schmitt-Bernard CF, Chavanieu A, Herrada G, Subra G, Arnaud B, Demaille JG, Calas B, Argiles A. BIGH3 (TGFBI) Arg124 mutations influence the amyloid conversion of related peptides in vitro. Eur J Biochem 2002; 269:5149-56.
9. Klintworth GK. The molecular genetics of the corneal dystrophies--current status. Front Biosci 2003; 8:d687-713.
10. Ridgway AE, Moller HU. Genetics of granular dystrophy. Ophthalmology 1992; 99:1753.
11. Klintworth GK. Lattice corneal dystrophy. An inherited variety of amyloidosis restricted to the cornea. Am J Pathol 1967; 50:371-99.
12. Holland EJ, Daya SM, Stone EM, Folberg R, Dobler AA, Cameron JD, Doughman DJ. Avellino corneal dystrophy. Clinical manifestations and natural history. Ophthalmology 1992; 99:1564-8.
13. Hall P. Reis-Bucklers dystrophy. Arch Ophthalmol 1974; 91:170-3.
14. Rodrigues MM, Krachmer JH. Recent advances in corneal stromal dystrophies. Cornea 1988; 7:19-29.
15. Miller CA, Krachmer JH. Epithelial and stromal dystrophies. The Cornea. Kaufman HE, Barron BA, McDonald MB, Waltman SR, editors. New York: Churchill Livingstone; 1988. p. 44-70.
16. Korvatska E, Henry H, Mashima Y, Yamada M, Bachmann C, Munier FL, Schorderet DF. Amyloid and non-amyloid forms of 5q31-linked corneal dystrophy resulting from kerato-epithelin mutations at Arg-124 are associated with abnormal turnover of the protein. J Biol Chem 2000; 275:11465-9.
17. Korvatska E, Munier FL, Chaubert P, Wang MX, Mashima Y, Yamada M, Uffer S, Zografos L, Schorderet DF. On the role of kerato-epithelin in the pathogenesis of 5q31-linked corneal dystrophies. Invest Ophthalmol Vis Sci 1999; 40:2213-9.
18. Sternberger LA, Hardy PH Jr, Cuculis JJ, Meyer HG. The unlabeled antibody enzyme method of immunohistochemistry: preparation and properties of soluble antigen-antibody complex (horseradish peroxidase-antihorseradish peroxidase) and its use in identification of spirochetes. J Histochem Cytochem 1970; 18:315-33.
19. Berg L, Hughes CP, Coben LA, Danziger WL, Martin RL, Knesevich J. Mild senile dementia of Alzheimer type: research diagnostic criteria, recruitment, and description of a study population. J Neurol Neurosurg Psychiatry 1982; 45:962-8.
20. Billings PC, Herrick DJ, Howard PS, Kucich U, Engelsberg BN, Rosenbloom J. Expression of betaig-h3 by human bronchial smooth muscle cells: localization To the extracellular matrix and nucleus. Am J Respir Cell Mol Biol 2000; 22:352-9.
21. Buckhaults P, Rago C, St Croix B, Romans KE, Saha S, Zhang L, Vogelstein B, Kinzler KW. Secreted and cell surface genes expressed in benign and malignant colorectal tumors. Cancer Res 2001; 61:6996-7001.
22. Hourihan RN, O'Sullivan GC, Morgan JG. Transcriptional gene expression profiles of oesophageal adenocarcinoma and normal oesophageal tissues. Anticancer Res 2003; 23:161-5.
23. Schneider D, Kleeff J, Berberat PO, Zhu Z, Korc M, Friess H, Buchler MW. Induction and expression of betaig-h3 in pancreatic cancer cells. Biochim Biophys Acta 2002; 1588:1-6.
24. Gilbert RE, Wilkinson-Berka JL, Johnson DW, Cox A, Soulis T, Wu LL, Kelly DJ, Jerums G, Pollock CA, Cooper ME. Renal expression of transforming growth factor-beta inducible gene-h3 (beta ig-h3) in normal and diabetic rats. Kidney Int 1998; 54:1052-62.
25. Buxbaum JN. The systemic amyloidoses. Curr Opin Rheumatol 2004; 16:67-75.
26. Aldave AJ, Rayner SA, King JA, Affeldt JA, Yellore VS. A unique corneal dystrophy of Bowman's layer and stroma associated with the Gly623Asp mutation in the transforming growth factor beta-induced (TGFBI) gene. Ophthalmology 2005; 112:1017-22.
27. Hautanen A, Gailit J, Mann DM, Ruoslahti E. Effects of modifications of the RGD sequence and its context on recognition by the fibronectin receptor. J Biol Chem 1989; 264:1437-42.
28. Kim JE, Jeong HW, Nam JO, Lee BH, Choi JY, Park RW, Park JY, Kim IS. Identification of motifs in the fasciclin domains of the transforming growth factor-beta-induced matrix protein betaig-h3 that interact with the alphavbeta5 integrin. J Biol Chem 2002; 277:46159-65.
29. Nam JO, Kim JE, Jeong HW, Lee SJ, Lee BH, Choi JY, Park RW, Park JY, Kim IS. Identification of the alphavbeta3 integrin-interacting motif of betaig-h3 and its anti-angiogenic effect. J Biol Chem 2003; 278:25902-9.
30. Kinoshita S, Adachi W, Sotozono C, Nishida K, Yokoi N, Quantock AJ, Okubo K. Characteristics of the human ocular surface epithelium. Prog Retin Eye Res 2001; 20:639-73.
31. Takacs L, Boross P, Tozser J, Modis L Jr, Toth G, Berta A. Transforming growth factor-beta induced protein, betaIG-H3, is present in degraded form and altered localization in lattice corneal dystrophy type I. Exp Eye Res 1998; 66:739-45.
32. Stix B, Leber M, Bingemer P, Gross C, Ruschoff J, Fandrich M, Schorderet DF, Vorwerk CK, Zacharias M, Roessner A, Rocken C. Hereditary lattice corneal dystrophy is associated with corneal amyloid deposits enclosing C-terminal fragments of keratoepithelin. Invest Ophthalmol Vis Sci 2005; 46:1133-9.
33. Hashimoto K, Noshiro M, Ohno S, Kawamoto T, Satakeda H, Akagawa Y, Nakashima K, Okimura A, Ishida H, Okamoto T, Pan H, Shen M, Yan W, Kato Y. Characterization of a cartilage-derived 66-kDa protein (RGD-CAP/beta ig-h3) that binds to collagen. Biochim Biophys Acta 1997; 1355:303-14.
34. Rawe IM, Zhan Q, Burrows R, Bennett K, Cintron C. Beta-ig. Molecular cloning and in situ hybridization in corneal tissues. Invest Ophthalmol Vis Sci 1997; 38:893-900.
35. Hanssen E, Reinboth B, Gibson MA. Covalent and non-covalent interactions of betaig-h3 with collagen VI. Beta ig-h3 is covalently attached to the amino-terminal region of collagen VI in tissue microfibrils. J Biol Chem 2003; 278:24334-41.
36. Kim JE, Park RW, Choi JY, Bae YC, Kim KS, Joo CK, Kim IS. Molecular properties of wild-type and mutant betaIG-H3 proteins. Invest Ophthalmol Vis Sci 2002; 43:656-61.