Analysis of Human Corneal Igg by Isoelectric Focusing

Analysis of Human Corneal IgG by Isoelectric Focusing

J. Clifford Woldrep,* Robin L. Noe,f and R. Doyle Stulringf

Parameters which regulate the localization and retention of IgG within the corneal stroma are complex and poorly understood. Although multiple factors are involved, electrostatic interactions between IgG and anionic corneal tissue components, ie, proteoglycans (PG) and glycosaminoglycans (GAG) may regulate the distribution of antibodies within the corneal stroma. Isoelectric focusing (IEF) and blotting analysis of IgG revealed a restricted pi profile for both central and peripheral regions of the normal cornea. Similar analysis of pathological corneas from keratoplasty specimens in Fuchs' dystrophy and keratoconus reveal a variable IEF profile. In the majority of keratoplasty specimens from patients with corneal edema or graft rejection, there was generally little or no IgG detectable. These results suggest that in edematous corneas where there is altered PG/GAG in the stroma and modified fluid dynamics, there is a concomitant loss of IgG. These findings may have implications for immunologic surveillance and protection of the avascular cornea. Invest Ophthalmol Vis Sci 29:1538-1543, 1988 The humoral immune system plays an important the soluble plasma proteins through ionic interac- role in mediating immunologic surveillance and pro- tions. The PGs and GAGs have long been known to tection of the avascular cornea. Immunoglobin G serve important physiological functions within the (IgG) is the major component of this humoral net- cornea regulating the collagen fiber organization,9 work which is responsible for diverse functional and pH,10 and swelling properties.9" Borcherding et al biological properties within the eye.1 The source of have demonstrated that the concentration of acidic corneal IgG is thought to be diffusion from the limbal GAGs differs from the central to the limbal regions of vessels.2"4 Factors that regulate the diffusion of IgG the cornea.9 Since the different GAGs vary with re- into the corneal stroma are extremely important spect to their numbers of negatively charged sulfate since there is normally no IgG produced locally groups, a subsequent electrostatic gradient within the within the quiescent eye.1"3 The localization and re- corneal stroma might, therefore, indirectly determine tention of IgG in the tissues, including the corneal which species of IgG gain access to the central and stroma, is dependent, in part, on the plasma IgG peripheral regions. Experiments were designed to test concentration, catabolic rate, and diffusion coeffi- this hypothesis through isoelectric focusing and anal- cient.5 Interactions between IgG and tissue compo- ysis of IgG isolated from normal human corneas ob- nents also regulate the tissue distribution of circulat- tained at autopsy and diseased corneas removed at ing IgG.6 Electrostatic factors may be paramount in keratoplasty. regulating the corneal distribution of plasma proteins. This is suggested by studies demonstrating the differential distribution of positively charged IgG and Materials and Methods negatively charged albumin within the corneal Ocular Tissues stroma.7 In the corneal stroma, the acidic proteogly- gans (PG) and the glycosaminoglycans (GAG) may Normal human globes and corneas were obtained function as fixed, anionically charged tissue ele- within 24 h of death from the Georgia Lions Eye ments8 which indirectly regulate the distribution of Bank at Emory University Eye Center (Atlanta, GA) and the Lions Eyes of Texas Eye Bank at the Cullen Eye Institute of Baylor College of Medicine (Hous- From the *Baylor College of Medicine, Center for Biotechnol- ton, TX). The corneas and aqueous humor were re- ogy, Division of Vision Research, The Woodlands, Texas, and the moved and frozen at — 80 °C prior to analysis. Kera- fEmory University Eye Center, Department of Ophthalmology, Atlanta, Georgia. toplasty specimens, aqueous humor and serum were Supported by grants EY-07154, 5T-32 EY-07092, and obtained at the time of surgery and were frozen EY-05097 from the National Eye Institute, Bethesda, Maryland. within three hours. Just prior to analysis, 8 mm nor- Submitted for publication: February 8, 1988; accepted April 18, mal corneal buttons were thawed and the central 4 1988. mm was separated from the remaining peripheral 4 Reprint requests: J. Clifford Waldrep, PhD, Baylor College of Medicine, The Center for Biotechnology, Division of Vision Re- mm ring using a trephine. Keratoplasty specimens search, 4000 Research Forest Drive, The Woodlands, TX 77381. consisted of one-half of an 8 mm button which was

1538

Downloaded from iovs.arvojournals.org on 10/02/2021 No. 10 IEF OF CORNEAL IgG / Woldrep er ol 1539

trephined into a central 4 mm hemicircle and a pe- dilution of mouse monoclonal antibody to human ripheral portion. IgG (MAB 066, Chemicon International, El Se- gundo, CA) for 60 min. The blots were washed for 15 Isoelectric Focusing (IEF) min with three changes of Tris-Saline and incubated with a 1:100 biotin-conjugated horse anti-mouse IgG Gels were prepared for IEF as described by Hoff- (VectaStain kit). After washing for 15 min with three man and Jump, on GelBond Agarose Film (FMC changes of Tris-Saline, the blots were incubated in a Corporation, Rockland, ME) by an open casting 1:200 dilution of the biotin-avidin ABC peroxidase technique.12 Using pyrogen-free water, 0.64 ml of pi complex (Vectastain kit, Vector Laboratories, Bur- 3-10 ampholytes (FMC) was added to 7 ml of lingame, CA) for 30 min. The blots were then washed 60-65°C melted FMC isogel agarose (1% W/v final for 15 min in Tris-Saline with three changes and de- concentration) containing 1.0 gram sorbitol and 1.0 veloped for 5 min in chloro-1-naphthol (Sigma, St. ml of glycerol. Some gels were supplemented with pi Louis, MO), 30 mg in 10 ml methanol plus 40 ml 8.0-10.5 ampholytes to increase resolution of basic 0.05 M Tris Buffer, pH 7.6, and 5-^1 of 30% H O . proteins. The final volume was adjusted to 10 ml, and 2 2 The developed blots were washed in dH O and stored the gels were poured under a heatlamp onto pre- 2 in the dark to dry, overnight. All antibody dilutions warmed, GelBond film on a level casting platform. were made in BGEN. All incubations and washing After solidification, the gels were cured for a mini- procedures were carried out on a rocker platform. mum of 1 hr at 4°C in a humidified chamber. Sam- The pi range of the serum and corneal IgG was deter- ples were added using applicator wicks (FMC pi 3-10 mined from the migration distances of the pi stan- standards, serum, or aqueous humor) or corneal dards. pieces were applied directly to the gel for direct tissue isoelectric focusing.13 A pH gradient was generated across the gel using a 1.0 N NaOH catholyte solution Results (pH 13.0) and a 0.5 N acetic acid anolyte solution (pH 2.6). An initial current of 1.0 W was used for 10 As an extension of previous quantitative and quali- min to establish the pH gradient. Isoelectric focusing tative analyses of immunoglobulins eluted from was initiated for 30 min and the sample applicators human corneas, we sought to further characterize and corneal pieces removed from the gel surface. Iso- corneal IgG using direct tissue isoelectric focusing electric focusing was continued for an additional 30 (IEF) and blotting onto nitrocellulose. Our assays re- min (600 volt/hours). Water accumulation on the gel quired as little as a 2 X 4 mm piece of corneal tissue for IgG analysis. Initial IEF analysis of normal surface was removed by periodic blotting with filter human corneal IgG revealed that the pi of both cen- paper. The isoelectric point (pi) or the pH at which tral and peripheral corneal IgG is restricted when there is no net mobility of a protein in an electric field compared to the pi of normal human serum IgG was determined using defined pi standards. (Figs. 1, 2). The pis of IgG species within the normal cornea is distributed within a narrower pH range than Blotting that of serum. The average lower limit detected for The IEF gels were rinsed for 60 seconds in distilled central corneal IgG was 5.78 ± 0.11 and the average water with agitation. The IEF gels were then blotted for peripheral corneal IgG was 5.60 ± 0.10 (Table 1). by direct application of pieces of nitrocellulose equili- These lower limit pi values are significantly different brated in 0.05 M Tris-Saline, pH 7.6. A wet piece of from the corresponding lower serum IgG pi (5.27 filter paper was placed onto the surface of the nitro- ± 0.27) (Table 1). The upper pi limit for central cor- cellulose, and the gel was then inverted onto ten neal IgG was 8.73 ±0.14 and for peripheral corneal layers of absorbent paper towels. A glass sheet was IgG was 8.95 ± 0.11 (Table 1). These pi values were then placed against the back of the GelBond surface not significantly different from the corresponding and a 1 kg weight applied for 10 min. Transfer of the upper serum IgG pi (9.4 ± 0.10) (Table 1). Qualita- focussed proteins occurred by diffusion. The nitro- tive IEF profiles of aqueous humor were not obtained cellulose blot was then removed and cut for staining since only trace amounts were detectable on the ni- and analysis. The Standards were stained with amido trocellulose blots due to extensive fractionation of the black or india ink. IgG within the pH gradient. The blots were analyzed for IgG distribution by the IEF and IgG blotting analyses were carried out biotin-avidin ABC peroxidase technique. The blots using pathological human corneas obtained at the were first blocked for 30 min with a solution of 3% time of penetrating keratoplasty. In five patients with BSA, 0.25% gelatin, 5 mM EDTA, and 0.05% NP-40 Fuchs' corneal dystrophy, although more variable, (BGEN). Next, the blots were incubated in a 1:100 the pi of the central and peripheral IgG was not sig-

Downloaded from iovs.arvojournals.org on 10/02/2021 1540 INVESTIGATIVE OPHTHALMOLOGY 6 VISUAL SCIENCE / October 1988 Vol. 29

one of the two edema patients with normal coraeal IgG, there was asymetrical, localized edema sur- rounded by normal tissue. In the 17 IgG deficient corneas, there was no apparent correlation between IgG content, and duration of edema, severity of edema or pathology. A similar absence of corneal IgG was detected in one of two patients grafted for alkali burns (Table 1). It is of interest to note that both of these corneas were vascularized. Analysis of aqueous humor from patients with graft rejection, corneal edema or alkali

-9.5

Fig. 1. IEF blot of normal human serum IgG with multiple bands at different isoelectric points within the pH gradient. The pi range for this normal was 5.5-9.5. The gel range was pH 3.0-10.0.

niflcantly different from the serum IgG or the normal corneal IgG profile (Table 1). The IEF and IgG blotting analysis of keratoconus corneas also apparently mirrored that of the serum, although the lower pi limit for the serum, central and peripheral cornea tended to be elevated with marked variability (Table 1). The pi of IgG present in keratoconus corneas was not quantitatively different from that of normal cor- -5.8 neas or those with Fuchs* dystrophy. In contrast, corneal specimens from patients with graft failure or corneal edema revealed an entirely different pattern (Fig. 3). Fifty-seven percent (4/7) of failed corneal grafts had little or no detectable IgG within the central cornea, 43% (3/7) had little or no detectable IgG in the peripheral cornea (Table 1). One of the rejected corneas that was vascularized dis- played an IEF profile like that of serum. Another was vascularized but had reduced amounts of corneal IgG. Most striking was the finding that there was little or no detectable IgG in 92% (17/19) of central corneas and 84% (16/19) of peripheral corneas from patients who were undergoing penetrating keratoplasty for corneal edema (Table 1). All of the 17 who had little or no detectable corneal IgG had relatively normal serum IgG profiles, indi- Fig. 2. IEF blot of normal human corneal IgG after direct tissue cating that the absence of corneal IgG was probably isoelectric focusing. The pi range of the central (C) cornea was due to abnormalities of the cornea rather than a de- 6.0-9.1; the peripheral (P) cornea was 5.8-9.3; and the limbal (L) fective humoral immune system in the patients. In cornea was 5.8-9.5. The gel range was pH 3.0-10.5.

Downloaded from iovs.arvojournals.org on 10/02/2021 No. 10 IEF OF CORNEAL IgG / Woldrep er ol 1541

Table 1. IEF of IgG in normal and abnormal corneas Range of IgG pi values (lower to upper limits) in: Diagnosis Serum Central cornea Peripheral cornea Normal 5.27 ± 0.27 5.78 ±0.1 If 5.60 ±0.1 Of to 9.40 ±0.10 (3/3)* to 8.73 ±0.14(3/3)* to 8.95 ±0.11(3/3)$ Fuchs' dystrophy 5.86 ±0.33 5.68 ± 0.67 5.68 ± 0.67 to 9.38 ± 0.24 (5/5) to 9.18 ±0.22 (5/5) to 9.26 ±0.19 (5/5) Keratoconus 6.13 ± 1.28 6.10 ±0.90 6.27 ± 1.00 to 9.40 ± 0.56 (6/6) to 9.30 ± 0.69 (6/6) to 9.40 ± 0.60 (6/6) Graft failure 6.04 ±0.21 Deficient (4/7) Deficient (3/7) to 9.36 ± 0.26 (7/7) 6.70 ± 0.52 6.48 ± 0.46 to 9.13 ± 0.29 (3/7) to 8.98 ± 0.25 (4/7) Cornea] edema 5.94 ± 0.33 Deficient (16/19) Deficient (16/19) to 9.15 ±0.21 (19/19) Deficient (1/19) 4.80-8.60 (1/19)§ 5.15 ±0.49 5.10 ±0.42 to 9.15 ±0.21 (2/19) to 9.05 ±0.21 (2/19) Alkali Burn 6.25 ±0.21 Deficient (1/2) Deficient (1/2) to 8.95 ± 0.64 (2/2) 6.20-9.30 (1/2) 6.20-9.40(1/2)

* Number of specimens/number of patients. t Not different from serum IgG pi. f Different from serum IgG pi at P < 0.01 by student t-test. § 25% of normal serum value.

burns generally revealed only trace or nondetectable amounts of IgG.

Discussion Immunoglobulins (Ig) with antibody activity are the mediators of the humoral immune system. These bifunctional glycoproteins are capable of binding for- eign antigens and initiating a variety of biological activities such as complement fixation, cell receptor in- teraction and regulation of catabolic processes. In humans, at least five distinct classes of Ig have been identified with different size, charge, amino acid composition and carbohydrate content.14 Each Ig class is a complex group of heterogeneous molecules. This is clearly demonstrated by the characteristic electrophoretic migration pattern of the IgG extend- ing from the alpha to the gamma fractions of normal serum. Of the Ig classes, IgG is the major Ig in both the serum and the tissues. IgG is extremely heterogeneous with respect to biological activity and exhibits the most diverse electrophoretic migration pattern.14 Analysis of human serum IgG has demonstrated that there are multiple species with pis from the anionic range of 5.5 to the cathodic range of 9.5.15 The parameters that regulate the egress of IgG from the circulation and its ultimate localization within the tissues are complex and poorly understood. These factors are particularly important within the avascu- Fig. 3. IEF blot of corneal IgG in edema (left) and graft failure lar cornea which normally contains no plasma cells (right). The serum (S) pi range of this edema patient was 6.0-9.0. for the local production of IgG.1"3 It is generally be- There was minimal IgG within the central (C) cornea and none in lieved that corneal IgG is derived from the circulation the peripheral (P) cornea. The serum (S) of this graft failure patient 2 4 was 6.0-9.1. There was a trace of IgG within the central (C) cornea through diffusion from the limbal vessels. " The and none in the peripheral (P) cornea. The gel range was pH rates at which different IgG molecules traverse from 3.0-10.0.

Downloaded from iovs.arvojournals.org on 10/02/2021 1542 INVESTIGATIVE OPHTHALMOLOGY b VISUAL SCIENCE / October 1988 Vol. 29

the limbal vessels through the corneal stroma is de- specimens revealed a relative increase in keratan sul- pendent on individual diffusion coefficients.16 The fate with a decrease in chondroitin sulfate.21 The diffusion coefficient may, however, be partially regu- presence of elevated amounts of keratan sulfate in the lated in the eye by electrostatic interactions. Highly central cornea9 might explain our observed IgG pro- charged, cationic IgG species readily egress from the file in keratoconus specimens. The relatively normal circulation and deposit within ocular tissues.6 Addi- IgG profile in Fuchs' dystrophy is more difficult to tional factors, such as the state of hydration can also interpret since some of the patients had mild stroma regulate the diffusion of molecules into the corneal edema with thickened Descemet's membrane. A pos- stroma.16 The fluid dynamics of the edematous cor- sible explanation for the demonstration of IgG in nea may also play an important role in determining these corneas might be altered corneal biochemistry the character of corneal proteins. The flow of fluid in which there is an abundance of oxytalan fibers from the aqueous humor through the epithelium present.22 These fibers are highly charged, enriched might alter corneal proteins by "wash out" or by ac- with cysteine, glutamate and aspartate.23 Thus Fuchs' celerated diffusion.16 It is, however, unlikely that this corneas might maintain different types of fixed nega- mechanism alone could account for the absence of tive charges capable of interacting with IgG. IgG in edematous corneas since there were patients Analysis of corneal IgG has provided insights into with corneal edema (Fuchs' dystrophy, graft failure the dynamic relationship between corneal physiol- and alkali burn) who had detectable IgG within nor- ogy, biochemistry and immunology. Our studies sug- mal range. gest that edema causes changes corneal immunoglob- Since PG and GAG have been implicated in main- ulin distribution. Patients with corneal edema may be taining the swelling properties of the corneal at increased risk for the development of infectious stroma,9,11,17-19 i:t is likely that they also have marked k?ratitis because of low levels of intracorneal immu- effects on the corneal IgG distribution. In addition, noglobulins. The occurrence of ruptured bullae, in PG and GAG provide fixedanioni c sites contributing combination with the loss of corneal IgG might make significantly to the pH of the corneal stroma.8 It is these patients more susceptible to certain types of likely that there is an intimate association between infection. The same might hold true for virtually any the anionic corneal PG and GAG, fluid dynamics episode which alters the PG and GAG and resultant and proteins which indirectly regulates the distribu- electrostatic interactions either throughout the cor- tion of IgG within the cornea. Indeed, there is a re- neal stroma or within localized regions. Further ex- stricted IgG IEF profile demonstrated in the normal periments are in progress to resolve these issues. central and peripheral cornea compared to serum. Key words: cornea, immune-globulin gamma (IgG), isoelec- Our results with pathological corneas suggest that the tric focusing, corneal edema, keratoplasty biochemical changes responsible for edema also regulate the retention of IgG within the corneal stroma. Previous observations have demonstrated that there Acknowledgments 1118 is loss of GAG in edematous corneas. Similar al- The authors wish to acknowledge the expert technical terations in GAG levels have been reported for re- assistance of Jayne Schulte. The authors also wish to ac- jected corneal transplants.18 If one considers the cor- knowledge both the Lion's Eyes of Texas and the Georgia neal GAG as fixednegativ e charges, somewhat analo- Lion's Eye Banks for their assistance in these studies, and April King and Candace Wendling for their expert secre- gous to those on a DEAE anionic exchange resin and tarial assistance. capable of binding positively charged proteins, ie, IgG, then the loss of corneal GAG would likely result in a corresponding loss of corneal IgG. This explana- References tion is consistent with our observations in edematous 1. Allansmith MR: The Eye and Immunology. St. Louis, The and rejected corneas. C.V. Mosby Company, 1982. IgG profiles of pathological corneas from patients 2. Allansmith M, Newman L, and Whitney C: The distribution of with keratoconus and Fuchs' dystrophy were not immunoglobulin in the rabbit eye. Arch Ophthalmol 86:60, 1971. grossly abnormal and were more difficult to interpret. 3. Allansmith M, Whitney C, McClellan BH, and Newman LP: None of the corneas from keratoconus patients were Immunoglobulins in the human eye: Location, type, and edematous. Previous studies of PG and GAG in these amount. Arch Ophthalmol 89:36, 1973. pathological corneas demonstrated that there are ab- 4. Henkind P, Hansen RJ, and Szalay J: Ocular Circulation. In normalities, however, these are less pronounced. Tis- Biomedical Foundations of Ophthalmology, Vol. 2, Chapt. 5, Duane TD and Jaeger EA, editors. Philadelphia, Harper and sue culture studies of keratoconus stromal cells dem- Row Publishers, 1983, pp. 1-54. onstrate normal synthesis levels of sulfated GAG in 5. Spiegelberg HL: Biological activities of immunoglobulins of 20 vitro. Analysis of GAG in keratoconus keratoplasty different classes and subclasses. Adv Immunol 22:259, 1974.

Downloaded from iovs.arvojournals.org on 10/02/2021 No. 10 IEF OF CORNEAL IgG / Woldrep er ol 1543

6. Waldrep JC: Uveal IgG distribution: regulation by electrostatic into fractions by isoelectric focusing. Protides of the Biological interactions. Curr Eye Res 6:897, 1987. Fluids 17:449, 1969. 7. Allansmith MR and McClellan BH: Immunoglobulins in the 16. Allansmith M, deRamus A, and Maurice D: The dynamics of human cornea. Am J Ophthalmol 80:123, 1975. IgG in the cornea. Invest Ophthalmol Vis Sci 18:947, 1979. 17. Donn A: Cornea and sclera. Arch Ophthalmol 75:261, 1966. 8. Hedbys BO: The role of polysaccharides in corneal swelling. 18. Hart GW: Corneal proteoglycans. In Cell Biology of the Eye, ExpEyeRes 1:81, 1961. McDevitt DS, editor. New York, Academic Press, 1982, pp. 9. Borcherding MS, Blacik LJ, Sittig RA, Bizzell JW, Breen M, 1-52. and Weinstein HG: Proteoglycans and collagen fibre organiza- 19. Elliott GF, Goodfellow JM, and Woolgar AE: Swelling studies tion in human corneascleral tissue. Exp Eye Res 21:59, 1975. of bovine corneal stroma without bounding membranes. J 10. Mishima S and Hedbys BO: Physiology of the cornea. Int Physiol 298:453, 1980. Ophthalmol Clinics 8:527, 1968. 20. Yue BYJT, Baum JL, and Silbert JE: The synthesis of glycos- 11. Anseth A: Studies on corneal polysaccharides: V. Changes in aminoglycans by cultures of corneal cells from patients with corneal glycosaminoglycans in transient stromal edema. Exp keratoconus. J Clin Invest 63:545, 1979. Eye Res 8:297, 1969. 21. Praus R and Goldman JN: Glycosaminoglycans in human corneal buttons removed at keratoplasty. Ophthalmic Res 12. Hoffman WL and Jump AA: Rapid blotting of IgG to nitro- 2:223, 1971. cellulose with minimal IgM contamination. J Immunol Meth 22. Alexander RA, Grierson I, and Garner A: Oxytalan fibers in 76:263, 1985. Fuch's endothelial dystrophy. Arch Ophthalmol 99:1622, 13. Saravis CA, O'Brien M, and Zamcheck N: Direct tissue iso- 1981. electric focusing in agarose. J Immunol Meth 29:97, 1979. 23. Goldfisher S, ColtofT-Schiller B, and Goldfisher M: Microfi- 14. Roitt IM, Brostoff J, and Male DK: Immunology. St. Louis, brils, elastic anchoring components of the extracellular matrix, TheC.V. Mosby Company, 1985. are associated with fibronectin in the zonule of Zinn and aorta. 15. Howard A and Virella G: The separation of pooled human IgG Tissue Cell 17:441, 1985.

Downloaded from iovs.arvojournals.org on 10/02/2021