Conjunctival Goblet Cell Frequency after Alkali Injury is not Accurately Reflected by Aqueous Tear Mucin Content

Transcription

1 Conjunctival Goblet Cell Frequency after Alkali Injury is not Accurately Reflected by Aqueous Tear Mucin Content Judith Friend, Timothy Kiorpes, and Richard A. Thoft Goblet cell counts have been used to evaluate the suitability of conjunctiva as a source of ocular surface epithelial cells. However, since tear mucin content can be determined without tissue excision, it seemed that the concentration of those compounds might be a useful indicator of conjunctival vitality. To test the extent to which aqueous tear composition reflects conjunctival goblet cell frequency, goblet cell frequency and aqueous tear mucin content were measured after alkali injury in rabbits. Mild alkali injury (0.1 N NaOH for 30 sec) caused a transient but substantial decrease in goblet cells (to 25% of normal at day 7) with a return to normal by six weeks. Tear mucin content was decreased to a lesser degree, from a normal value of 6.4 ± 0.47 nmol oligosaccharide per nl (n = 10) to a minimal value of 4.7 ± 0.64 (n = 7) (73% of normal) at day 7, returning to normal 4 weeks after injury. Thus, the direction of the change was the same, but the magnitudes were quite different. These results suggest that conjunctival goblet cell frequency is not accurately reflected by aqueous tear mucin content, and, therefore, that tear mucin content cannot be used directly as an indicator of conjunctival health. Invest Ophthalmol Vis Sci 24: , 1983 A plethora of reports has documented the effect of alkali burns on various aspects of corneal epithelial and stromal metabolism and morphology in rabbits, 1 " 7 but relatively little attention has been paid to the effects of such burns on the conjunctiva. 8 Recent observations suggest that peripheral corneal and/or conjunctival epithelium may play an important role in corneal epithelial maintenance and integrity, implying that changes in the conjunctiva may eventually become manifest on the corneal surface. In injured eyes, for example, where the corneal and limbal epithelia are completely destroyed, there is no question that the conjunctiva resurfaces the cornea. Even in the absence of total epithelial loss, there is evidence that there is an ongoing centripetal motion of cells from the periphery towards the center of the cornea, as evidenced by central movement of epithelial dots in postkeratoplasty eyes, 9 by attenuation of a central corneal epithelium isolated from the peripheral epi- From the Eye Research Institute of the Retina Foundation and the Department of Ophthalmology, Massachusetts Eye and Ear Infirmary and Harvard Medical School, Boston, Massachusetts. Supported in part by Research Grants EY 01830, EY 03061, Eye Research Core Services Support P30 EY from the National Eye Institute, National Institutes of Health, and the Lions Eye Research Fund, Inc. Submitted for publication June 16, Reprint requests: Judith Friend, Eye Research Institute, 20 Staniford Street, Boston, MA thelium by a glue ring, 10 and by replacement of donor epithelium by host epithelium in rabbit keratoplasties." Although it is not proven, the implication of those studies is that conjunctival epithelium may be involved in corneal epithelial maintenance even in normal eyes. It is also important to understand the effects of injury on the conjunctiva if one proposes to use such conjunctiva to resurface damaged corneas, since such conjunctiva serves as the source of healthy epithelial surface cells. Conjunctival transplantation is now used in the treatment of several kinds of ocular surface disease. 12 " 14 Currently, its use is limited to autografts in patients with unilateral disease because it is assumed that only undamaged conjunctiva can be used for grafting. Verification of the usefulness of previously injured conjunctiva for corneal resurfacing might increase the number of patients for whom conjunctival grafting is possible. Previous studies have suggested that goblet cell frequency is a useful indicator for the health of the conjunctiva The presence of goblet cells may reflect a degree of differentiation or maturation of the conjunctiva which in turn reflects the presence of healthy tissue. However, it is not always possible or desirable to take conjunctival biopsies. Since tear mucin is believed to be secreted primarily by conjunctival goblet cells, 17 tear mucin content might be a useful assay to determine conjunctival differentiation /83/0500/612/$ 1.15 Association for Research in Vision and Ophthalmology 612

2 No. 5 ALKALI INJURY, GOBLET CELLS, AND TEAR MUCIN / Friend er ol. 613 The work to be reported here, therefore, measured conjunctival goblet cell frequency and aqueous tear mucin, glucose and protein content after alkali injury in rabbits to (1) evaluate the response of the conjunctiva to chemical injury, and (2) evaluate the extent to which aqueous tear composition reflects conjunctival goblet cell content and, therefore, whether it is a useful indicator of conjunctival differentiation. Materials and Methods Animal Preparation and Tissue Sampling Alkali bums: Rabbits, weighing kg, were used for all experiments. Anesthesia was induced by intramuscular chlorpromazine hydrochloride-ketamine hydrochloride plus ether inhalation and topical proparacaine. The eyelids were opened and a Grade I (Hughes' classification) 18 burn produced by putting 0.1 N sodium hydroxide onto the cornea and into the cul-desac. The eyelids were held open and moved slightly to ensure the alkali covered the entire ocular surface epithelium. After 30 sec, the alkali was washed out of the eye with a 1-min wash of 0.9% sodium chloride. In some cases, the nictitating was removed, and the alkali was applied 1 week later after the eye was healed and noninflamed. Burns were made in a total of 66 eyes in 47 rabbits. Topical antibiotic ointment was applied immediately after the burn. The eyes were examined with a flashlight and topical antibiotic ointment was applied daily until the epithelial defect healed. Thereafter, the animals were examined withflashlightor slit lamp two to four times weekly. At the end of the experiments, rabbits were killed by an overdose of intravenous sodium pentobarbital followed by air embolism. Tear samples: At 1 day and 1, 4, and 6 weeks after burns or in unburned normal eyes, aqueous tear samples were collected from the lower cul-de-sac of unanesthetized rabbit eyes using a 10-/ul capillary pipette. No stimulators of tear flow were used. The 2-10 /ill of fluid collected within 1-2 min were stored in the sealed capillary pipettes at -80 C pending analysis. Conjunctival biopsies for goblet cell analysis: Rabbits were anesthetized as described above, except no topical proparacaine was used. One day and 1, 4, and 6 weeks after burns, or in unburned normal eyes, 2-3 mm diameter, approximately round, samples of conjunctiva near the limbus were excised using fine forceps and corneal scissors. Thus, four samples were taken from the four quadrants of the conjunctiva (upper and lower nasal, and upper and lower temporal) in each eye over the 6-week period. Samples for goblet cell counts were also taken deep in the fornices, to confirm that the epithelium in that region was also injured by the alkali. Goblet cell counts in those regions were similar to those found around the limbus after injury. Analyses Goblet cellfrequency: The conjunctival biopsy samples were fixed in 10% buffered formalin, embedded in paraffin, and sectioned. Seven-micron sections were stained with hematoxylin-esoin, or with the periodic acid-schiff (PAS) reaction to demonstrate goblet cells. Goblet cells were counted in PAS stained sections. The frequency was expressed as the number of goblet cells per 100 epithelial cells. A minimum of 200 epithelial cells was counted for each value. Tears Tear mucin and glucose: The amount of mucinlike glycoprotein in tears was determined by quantitation of the alkali-labile oligosaccharides in tears. 19 This was done by specifically labeling N-acetylgalactosamine (the linkage sugar of mucin-like glycoprotein) with tritium, then determining the amount of those sugars present in tear samples by counting the radioactivity and comparing it to appropriate standards. Briefly the method was as follows. Alkali treatment of mucin-containing samples with tritiated borohydride produced a tritium-labeled linkage amino sugar alcohol attached to oligosaccharide. Subsequent acid hydrolysis separated the labeled amino sugar alcohol from the oligosaccharide. The amino sugar alcohol was then isolated by binding to an AG 50W-X8 cation exchange column followed by elution with 1 M hydrochloric acid, and the amount of sugar quantitated by counting the radioactivity on a scintillation counter. Free tear glucose was converted to sorbitol and labeled in this reduction procedure. Labeled sorbitol did not bind to the AG 5OW-X8 column, but was eluted in water wash. Its amount was also determined by scintillation counting. Counting for radioactivity was done in a toluene-triton X-100 scintillation cocktail using a Beckman LS 8100 counter. Correction for the efficiency of counting was based on the H number that the Beckman LS 8100 scintillation counter determines automatically for each sample. The amount of free tear glucose and of N-acetylgalactosamine was determined by comparison to glucose standards. The amount of alkali-labile oligosaccharide, measured as N-acetylgalactosamine, was

3 614 INVESTIGATIVE OPHTHALMOLOGY 6 VISUAL SCIENCE / May 1980 Vol. 24 Fig. 1. Conjunctiva after alkali injury. One day after a 30-sec 0.1 N NaOH alkali burn, epithelium was absent from portions of the conjunctiva (A). In other areas, it was present as a single cell layer, with no goblet cells (B). At one day, the subepithelial tissue was edematous (A and B). One week after a 30-sec 0.1 N NaOH alkali burn, the epithelium covered the conjunctiva] surface, was of normal thickness, but had far fewer goblet cells than normal conjunct!val epithelium (C). Six weeks after the alkali burn, the epithelium covering the conjunctiva was normal both in thickness and in goblet cell content, and the subconjunctival edema had subsided (D). (See also Fig. 3) (Epi = epithelium, Gc = goblet cell, bv = blood vessel) (periodic acid-schiff reaction) (XI03). 1D 4 '

4 No. 5 ALKALI INJURY, GODLET CELLS, AND TEAR MUCIN / Friend er ol. 615 Distribution of Goblet Cells around the Rabbit upper Limbus inflammatory cells into the conjunctiva. Over the next weeks, the epithelium was re-established, and the inflammation subsided (Fig. 1). temporal 7.8*0.7(9) (12) (8) (9) lower nasal Fig. 2. Distribution of goblet cells around the rabbit limbus. Goblet cell frequency was the same in the upper and lower nasal and upper and lower temporal quadrants of the conjunctiva surrounding the limbus. used as an estimate of the mucin content of the tears. That is expressed as nmol oligosaccharide per ^1 tears. The free tear glucose content of tears is expressed as mg per 100 ml. Tear protein: Tear protein content was measured in l-/il tear samples using the Coomassie Brilliant Blue procedure as modified for micro samples. 20 ' 21 Results Appearance of Cornea and Conjunctiva The clinical and histologic appearance of the conjunctival epithelium after Hughes' Grade 1 alkali injuries has been documented previously. 818 One day following injury, there was a significant, but usually not total, loss of epithelium, edema, and invasion of Goblet Cell Frequency The frequency of goblet cells in conjunctival epithelium near the limbus was the same in the upper and lower nasal, and upper and lower temporal quadrants (Fig. 2). Alkali burns caused a transient decrease in the number of goblet cells in limbal conjunctiva relative to the number of epithelial cells from 7.7% to less than 3% within 24 hours. Within 4-6 weeks after injury, goblet cell frequency returned to normal (Figs. 1, 3). Similarly at 1 day and 1 week, there were reduced numbers of goblet cells in conjunctiva taken from the fornices. Tear Analysis Removal of the nictitating significantly decreased mucin-oligosaccharide and protein content of the tears (Table 1), but had no effect on glucose content. Alkali burns caused 3- to 12-fold, albeit transient, increases in tear glucose content, but the values were normal 4 weeks after injury. After 6 weeks, the glucose level in eyes with nictitating s was below normal, while that in eyes without such s was slightly elevated (Table 2). Tear protein levels were increased one day after injury in those eyes without nictitating s, but returned promptly to normal and remained nor- Conjunctival Goblet Cells After Total 3Os Alkali Burn (iisem(n)) O 0.1N NaOH 0.1N NaOH, without nictitating (4) Fig. 3. Conjunctival goblet cells after 30-sec 0.1 N NaOH alkali burn. Goblet cell frequency in the conjunctiva was significantly below normal for 1 week after alkali injury, but was normal by 4-6 weeks after alkali burn. t «(«g»

5 616 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / May 1983 Vol. 24 Table 1. Rabbit tear glucose, protein, and mucin content* Tear component concentration Glucose (mg/100 ml) Protein (MSS/M') Mucin oligosaccharide (nmol//il) ±6 (10) ± 1 CO) 6.4 ± 0.47 (10) One week after removal of nictitating 18 ±4 (21) 11 ± 1 (21) 3.5 ±0.24(21) Pf <0.01 <0.01 * Values are given as averages ± the standard errors of the mean with the number of measurements in parentheses. f P values is given if P < 0.05 using the Student's t-test. mal for the duration of the study (Table 3). In eyes that had retained the nititating, the tear protein remained normal throughout the 6 weeks after injury (Table 3). Mucin concentration (nmol oligosacchride per /A tears) in eyes with nictitating s was decreased initially by sodium hydroxide burns, but the decrease was only statistically significant for the first week (Table 4). Even at its lowest point, the mucin concentration in the tears of the alkali burned rabbits was approximately 73% of the normal value. Eyes without nictating s showed no decrease in tear mucin content when the values were compared to that for unburned eyes without nictitating s. Discussion Tear glucose and protein content recover promptly from the alkali injury. The increases in those sub : stances seen immediately after injury probably reflect only increased leakage through the large epithelial defect, which is present early. Within a few weeks, the glucose levels return to the expected low values. 22 There is a large variability in the glucose values perhaps because this method measures all neutral sugars, Table 2. Tear glucose content after alkali injury (mg/100 ml*) After sodium hydroxide burn Day 1 Day 7 Day 28 Day 42 With nictitating 28 ± 6.3 (10) 90 ±9 11 ± 8 20 ±9 10 ± 3 (8)t (7) (4) (3)t Without nictitating 18 ± 219 ± 42 ± 25 ± 40 ± 4.4 (21) (18)t (17)t (8) (7)t * Values are expressed as averages ± the standard errors of the mean with the number of determinations in parentheses. t When compared to the corresponding normal value using the Student's t-test, the value is significantly different at the 5% level. Table 3. Tear (^ig//il tears*) After sodium hydroxide burn Day 1 Day 7 Day 28 Day 42 protein after alkali injury With nictitating 17 ± 1 (10) 19 ± 3 (8) 16 ± 1 (7) 23 ± 3 (4) 16 ± 1 (4) Without nictitating 11 ± 1 (21) 29 ± 3(18)f 19 ± 1 (17)t 15 ±8(18) 17 ±2(7) * Values are given as averages ± the standard errors of the mean with the number of determinations in parentheses. f When compared to the corresponding normal value using the Student's t-test, the value is significantly different at the 5% level. not just glucose. The tear protein values are also normal 23 within a few weeks of the alkali injury. In rabbits, unlike humans, 17 the goblet cell frequency is uniform around the lumbus (Fig. 2), permitting a comparison of goblet cell frequency in the same animal over a period of weeks after alkali burn. By so doing, we have demonstrated a decrease followed by a prompt recovery of histologically identifiable goblet cells after alkali injury (Fig. 3). Conjunctival goblet cell frequency has been shown to be very sensitive to other abnormal conditions. For example, the number of these cells relative to the bulk of epithelial cells is severely decreased when conjunctiva is incubated in organ culture, 16 when conjunctiva migrates across the cornea, 10 ' 24 and in a variety of pathologic conditions, eg, ocular pemphigoid, 25 ' 26 and Sjogren's syndrome. 25 Some have also shown a decrease in goblet cells acutely following chemical injury in humans, 25 but this is apparently not the case long after injury. 27 In all these conditions, it is, of course, difficult to say whether the goblet cells are, in fact, absent, or whether they are simply unrecognizable because of a lack of PAS, positive material. Table 4. Tear mucin concentration after alkali injury (oligosaccharide-nmol/^1*) After sodium hydroxide burn Day 1 Day 7 Day 28 Day 42 With nictitating 6.4 ± 0.47 (10) 4.9 ± 0.72 (8) 4.7 ± 0.64 (7)f 6.7 ± 1.8 (4) 5.8 ± 0.55 (3) Without nictitating 3.5 ±0.24(21) 5.1 ± 0.27(18)f 5.1 ±0.24(17)f 4.6 ± 0.58 (8) 4.4 ± 0.59 (7) * Values are expressed as averages ± the standard errors of the mean with the number of determinations in parentheses. f When compared to the corresponding normal value using the Student's t-test, the value is significantly different at the 5% level.

6 No. 5 ALKALI INJURY, GODLET CELLS, AND TEAR MUCIN / Friend er ol. 617 Recovery of goblet cells as shown in the present experiments has not been demonstrated previously in an experimental system. Clinically, recovery of goblet cells has been seen within 2-4 weeks after restoration of vitamin A levels in deficient patients. 28 " 30 However, the mechanism in those cases would seem to be very different from that seen in our experiments. In vitamin A deficiency, the epithelium is abnormal, with keratinization, but there is no significant decrease in the number of cells. 30 In that case, the goblet cell deficiency observed seems to be due to abnormal differentiation of the ocular surface cells, probably secondary to vitamin A deficiency. In the case of the experimental alkali injury reported here, on the other hand, reappearance of the goblet cells occurs only shortly after restoration of the epithelial cell mass implying that goblet cell appearance may be an advanced stage of ocular surface differentiation. Despite the partial loss of goblet cells seen after the mild chemical injury to the conjunctiva, the tear mucin concentration remains at nearly normal levels throughout the experiment. Although there is correlation between goblet cell frequency and tear mucin content, substantial amounts of mucin are present even when goblet cell frequency is low (Fig. 4). This suggests that these aqueous tear samples do not include the undissolved mucin present on the ocular surface. Examination of human conjunctival mucin using a permeable (Millipore R ) shows there is a significant amount of mucin on the ocular surface which can be collected on filters and stained with the PAS reaction 3132 indicating that this mucin is not dissolved in the aqueous tears. If the bulk of this undissolved mucin is derived from goblet cells, and only a small amount of mucin is dissolved in the aqueous tears, the relative independence of aqueous tear mucin and conjunctival goblet cell frequency is explicable. The relative contributions to ocular surface mucin from the lacrimal gland and mucin production by nongoblet cells 3334 remain undefined. A lack of strict correlation between goblet cell frequency and aqueous tear mucin has also been shown in clinical experiments where disease conditions associated with markedly decreased or absent goblet cells (eg, ocular pemphigoid, keratoconjunctivitis sicca, and Stevens-Johnson syndrome) still had at least 60% of the normal tear mucin The role of tear mucin in the structure and stability of the tear film has been described 36 and even small mucin decreases may adversely affect tear stability, especially if they affect the mucus layer of the tear film. 37 However, Holly et al 37 were not able to demonstrate abnormalities in aqueous tear component surface activity sufficient to account for tear film instability in patients with ocular pemphigoid or mild keratocon Goblet cells/100 epithelial cells Fig. 4. Tear mucin concentration vs. conjunctival goblet cell frequency. The linear correlation coefficient and y-intercept were 0.70 and 4.3 nmol/^1, respectively, for eyes with the nictitating s ( ), and 0.88 and 5.4 nmol/^1 for eyes without the nictitating s (O). The slopes were 0.21 for eyes with nictitating s ( ) and 0.19 for eyes without (O). (Based on the data presented in Table 4 and Fig. 3). junctivitis, although they did demonstrate such abnormalities in tears from patients with Stevens-Johnson syndrome. 37 Thus, the role of aqueous tear mucin in tear film stability remains unclear. These studies suggest that, at least in some diseases, there may be primary abnormalities of the ocular surface epithelium, not just problems related to tear film abnormalities. This report documents, for the first time, experimental confirmation of the fact that aqueous tear mucin content does not reflect accurately conjunctival goblet cell frequency, and, therefore, that aqueous tear mucin content is not a useful indicator of conjunctival differentiation or health. Key words: ocular surface epithelium, conjunctiva, alkali burn, goblet cells, tear mucin Acknowledgments The authors thank Patricia Pearson for the histologic preparations, and Shigeru Kinoshita, MD, for helpful discussions. References 1. Schmidt-Martens FW, Hennighausen U, and Bremer HJ: Aqueous humour changes after alkali burns. Trans OphthalmolSoc UK 97:715, Newsome DA and Gross J: Regulation of corneal collagenase production: stimulation of serially passaged stroma cells by blood mononuclear cells. Cell 16:895, Kenyon KR, Berman M, Rose J, and Gage J: Prevention of stromal ulceration in the alkali-burned rabbit cornea by gluedon contact lens. Evidence for the role of polymorphonuclear leukocytes in collagen degradation. Invest Ophthalmol Vis Sci 18:570, Matsuda H and Smelser GK: Epithelium and stroma in alkaliburned corneas. Arch Ophthalmol 89:396, 1973.

7 618 INVESTIGATIVE OPHTHALMOLOGY b VISUAL SCIENCE / Moy 1983 Vol Pfister RR and Burstein N: The alkali burned cornea. I. Epithelial and stromal repair. Exp Eye Res 23:519, Pfister RR, Nicolaro ML, and Paterson CA: Sodium citrate reduces the incidence of cornea] ulcerations and perforations in extreme alkali-burned eyes acetylcysteine and ascorbate have no favorable effect. Invest Ophthalmol Vis Sci 21:486, Schmidt-Martens FW and Bremer HJ: Alteration nutritiver Metabolite in Hornhautstroma des Kaninchenauges in der fru- "hen Phase nach experimenteller Alkaliveratzung. Albrecht von Graefes Arch Klin Exp Ophthalmol 213:207, Schmidt-Martens FW, Busch, L, and Klaas D: Friihe Veranderungen der Conjunctiva nach Alkali-Veratzung. Albrecht von Graefes Arch KJin Exp Ophthalmol 201:135, Kaye DB: Epithelial response in penetrating keratoplasty. Am J Ophthalmol 89:381, Shapiro MS, Friend J, and Thoft RA: Corneal re-epithelialization from the conjunctiva. Invest Ophthalmol Vis Sci 21:135, Kinoshita S, Friend J, and Thoft RA: Sex chromatin of donor corneal epithelium in rabbits. Invest Ophthalmol Vis Sci 21:434, Thoft RA: Conjunctival transplantation. Arch Ophthalmol 95:1425, Thoft RA: Conjunctival transplantation as an alternative to keratoplasty. Ophthalmology 86:1084, Thoft RA: Indications for conjunctival transplantation. Ophthalmology 89:335, Herman WK, Lindstrom RL, and Doughman DJ: Conjunctival autograft transplantation for unilateral ocular surface disease. Presented at the 12th Corneal Research Conference, Boston, MA, Thoft RA and Friend J: Ocular surface evaluation. Doc Ophthalmol Proc Ser 20:201, Kessing SV: Mucus gland system of the conjunctiva. A quantitative normal anatomical study. Acta Ophthalmol Suppl 95, Hughes WF Jr: Alkali burns of the eye. II. Clinical and pathologic course. Arch Ophthalmol 36:189, Kiorpes TC, Friend J and Thoft RA: Measurement of ocular surface mucus after mobilization with N-acetylcysteine (AcCys). ARVO Abstracts. Invest Ophthalmol Vis Sci 18(Suppl):142, Bradford MM: A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72:248, Bio-Rad Protein Assay Instruction Manual, Bio-Rad Laboratories, Richmond, CA, Reim M, Lax F, Lichte H, and Turss R: Steady state levels of glucose in the different layers of the cornea, aqueous humor, blood and tears in vivo. Ophthalmologica 154:39, Dartt DA, Knox I, Palau A, and Botelho SY: Proteins in fluids from individual orbital glands and in tears. Invest Ophthalmol Vis Sci 19:1342, Friedenwald JS: Growth pressure and metaplasia of conjunctival and corneal epithelium. Doc Ophthalmol 5-6:184, Ralph RA: Conjunctival goblet cell density in normal subjects and in dry eye syndromes. Invest Ophthalmol 14:299, Kinoshita S, Kiorpes TC, and Thoft RA: Goblet cell density is a better indicator of ocular surface disease than is tear mucin. Arch Ophthalmol, in press. 27. Faulkner WJ, Kenyon KR, Rowsey JJ, Hanninen L, Gipson I, and Gaylor J: Chemical burns of the human ocular surface: clinicopathologic studies of 14 cases. ARVO Abstracts. Invest Ophthalmol Vis Sci 20(suppl):8, Sullivan WR, McCulley JP, and Dohlman CH: Return of goblet cells after vitamin A therapy in xerosis of the conjunctiva. Am J Ophthalmol 75:720, Sommer A, Green WR, and Kenyon KR: Clinical-histopathologic correlations of vitamin A responsive and non-responsive Bitot's spots. Arch Ophthalmol 100:953, Thoft RA: Dietary deficiencies. In The Cornea Book, Smolin, G and Thoft, RA, editors. Boston, Little, Brown & Co., (in press). 31. Adams AD: The morphology of human conjunctival mucus. Arch Ophthalmol 97:730, Egbert PR, Lauber S, and Maurice DM: A simple conjunctival biopsy. Am J Ophthalmol 84:798, Srinivasan BD, Jakobiec FA, Iwamoto T, and DeVoe G: Giant papillary conjunctivitis with ocular prostheses. Arch Ophthalmol 97:892, Greiner JV, Kenyon KR, Henriquez AS, Korb DR, Weidman TA, and Allansmith MR: Mucus secretory vesicles in conjunctival epithelial cells of wearers of contact lenses. Arch Ophthalmol 98:1843, Dohlman CH, Friend J, Kalevar V, Yagoda D, and Balazs E: The glucoprotein (mucus) content of tears from normals and dry eye patients. Exp Eye Res 22:359, Holly FJ and Lemp MA: Wettability and wetting of corneal epithelium. Exp Eye Res 11:239, Holly FJ, Patten JT, and Dohlman CH: Surface activity determination of aqueous tear components in dry eye patients and normals. Exp Eye Res 24:479, 1977.