Inflammatory mediator release on conjunctival provocation of allergic subjects with allergen David Proud, PhD,* James Sweet, MD,** Paul Stein, PhD,*** Russell A. Settipane, MD, ** Anne Kagey-Sobotka, PhD,* Mitchell H. Friedlaender, MD,*** and Lawrence M. Lichtenstein, MD, PhD* Baltimore, Md., and La Jolla, Calif. To evaluate the role of injammatory mediators in the pathogenesis of the ocular allergic response, 23 subjects with positive histories of allergies to either cat dander or ragweed pollen and positive skin tests to the appropriate allergen extract were recruited and were subjected to conjunctival provocation. The tear duct of the left eye of each subject was blocked with a collagen plug while the right eye was left unplugged. In all cases, the eye was initially provoked with saline and subsequently with the appropriate allergen extract. Nonallergic subjects, or allergic subjects provoked with nonrelevant allergen, were used as control subjects. After each provocation, symptoms were recorded. and tears were collected with preweighed strips of filter paper (Schirmer strip). Each strip was placed into a tared tube containing fluid appropriate for the optimal preservation of the mediator to be measured. It was therefore possible to calculate the weight of tears collected and to express mediator levels per milliliter of tears. All allergic subjects demonstrated a positive symptomatic response to allergen challenge, whereas the control subjects remained asymptomatic. Blockage of the tear duct did not significantly alter the response. For allergic subjects, the levels of histamine, kinins, prostaglandin D,, albumin, and TAME-esterase activity were all signi$cantly (p < 0.005 in each case) greater after allergen challenge than after saline challenge. Furthermore, levels of each of these mediators after allergen challenge (expressed as increases above levels after saline provocation) were significantly greater for allergic subjects than for control subjects ( p < 0.005 in each case). Thus, the clinical response to conjunctival provocation with allergen is associated with increases in the levels of inflammatory mediators in tears. ( J ALLERGY CLIN IUMJNOL 1990;85: 896-905. ) The eye is a common target organ of the allergic response and is readily accessible to study. Despite this accessibility, however, there have been relatively few studies of the pathogenesis of the ocular allergic response, and the allergic manifestations of ocular hay From the *Division of Clinical Immunology, The Johns Hopkins University School of Medicine, Baltimore, Md., and **Division of Allergy and Immunology, ***Department of Molecular and Experimental Medicine, and Division of Ophthalmology, Scripps Clinic and Research Foundation, La Jolla, Calif. Supported by National Institutes of Health Grants HL 32272, EY 02502, AI 07290, and AI 08270. Received for publication July 26, 1989. Revised Dec. 8, 1989. Accepted for publication Dec. 11, 1989. Reprint requests: David Prot~I, PhD, The Johns Hopkins Asthma and Allergy Center, Clinical Immunology Unit office 2, 301 Bayview Boulevard, Baltimore, MD 21224. Publication No. 005 from The Johns Hopkins Asthma and Allergy Center. Dr. Lawrence M. Lichtenstein is the recipient of a Pfizer Biomedical Research Award. 111/18973 666 Abbreviations used PGD2, FGE,, PGF: F’rostaglandin D,, E,, F, HPLC: High-performance liquid chro- matography RIA: Radioimmunoassay LTC,, LTD.,, LTE.,: Leukotrienes C,, D,, E, fever have been monitored largely by subjective mea- sufes. Although Bisgaard et al.’ have observed in- creased levels of immunoreactive leukotrienes in tear fluids after conjunctival provocation, and it has been reported that histamine levels in tear fluids are not elevated in ocular hay fever,2 there has been no sys- tematic attempt to relate the clinical response to al- lergen to the generation of inflammatory mediators. In other tissues, including the lung,3” nose,7-10 and ~kin,“-‘~ the ability to determine the levels of media- tors generated during the allergic response has not only provided a more objective means of monitoring
`
`APOTEX EX1054
`
`Page 1
`
`

`
`VOLUME 85 NUMBER 5 Mediator generation in ocular allergy 897 the allergic reaction but has also increased our un- derstanding of the pathogenesis of the allergic re- sponse in each tissue. Furthermore, the measurement of inflammatory mediators during pharmacologic in- tervention studies in these tissues has provided useful information regarding the mechanisms of actions of various antiallergic drugs.‘5x l6 One particularly im- portant observation to arise from studies of mediator generation during allergic events is that tissue-specific differences are observed. For example, while FGD, is generated concomitantly with histamine during the immediate allergic response in the upper airways,’ this eicosanoid is not detectable until considerably later after antigen provocation in the skin and does not display the same time course of generation and per- sistence as histamine. I4 Such tissue-specific responses may result, in part, from the demonstrable heteroge- neity of mast cells from different tissues”-*’ and clearly emphasize the importance of examining each target organ of interest, rather than trying to extrap- olate from one tissue to another. To evaluate the role of inflammatory mediators in the pathogenesis of the ocular response to allergen, we have therefore monitored the levels of histamine, kinins, PGD2, sulfidopeptide leukotrienes, TAME- esterase activity, and albumin in recovered tear fluids from allergic and nonallergic subjects after conjunc- tival provocation with saline and, subsequently, with an appropriate allergen. We now report that a positive symptomatic response to allergen provocation is as- sociated with significant increases in the concentra- tions of all of these parameters (with the exception of leukotienes) in tears. METHODS Subjects Healthy male and female volunteers between the ages of 18 and 55 years were recruited. No subjects with vernal or contact lens conjunctivitis were used. Allergic subjects were defined as individuals having a positive clinical history of allergic symptoms (not necessarily conjunctival) to an an- tigen to which they had a positive intradermal skin test (>6 mm wheal to 0.5 allergy units per milliliter or less of allergen extract). In the present study, 14 subjects (four male and 10 female subjects) were allergic to cat dander, and nine subjects (three male and six female subjects) were allergic to ragweed. The control population consisted of 19 nonallergic subjects (five male and 14 female subjects with no clinical history of allergic symptoms and negative skin tests to the allergens used) and three of the cat-allergic subjec’? who were exposed to a nonrelevant antigen (rag- weed). Several subjects were challenged on more than one occasion In such instances, challenges were separated by a minimum of 1 week to avoid the possibility of priming. All allergic subjects were asymptomatic and were not re- ceiving any medications at the time of challenge. All sub- jects gave informed consent before participation in the study, and the study protocol was approved by the Human Subjects Committee of the Scripps Clinic and Rese‘arch Foundabon. Challenge procedure Ragweed-pollen extract was obtained from Hollister-Stier Laboratories (Spokane, Wash.), and cat-hair and dander- allergenic extract were obtained from ALK America, Inc. (Milford, Conn.). On entering the laboratory, all subjects were examined to establish that no visible ocular symptoms were present at the time of challenge. The protocol then involved applying a collagen plug to block the tear duct in the left eye. A 40 p+l drop of the normal saline diluent used for the allergen extracts was then administered to each eye. After a brief pause (approximately 5 minutes), symptoms were evaluated by the investigator, and tears were collected by placing a preweighed strip of filter paper (Schimier strip) into the inferior fomix. After again determining that subjects were asymptomatic, a 40 l.~l drop of allergen extract was then administered into the eye. The concentration of allergen extract used was usually 30,000 to SO.000 times that re- quired to elicit a threshold skin response. Thus, the actual dose of allergen administered to the eye was tn the range of 500 to 1000 allergy units. In allergic subjects. symptoms usually developed within minutes after allergen challenge. Symptoms were graded by the investigator. and tears were collected as described above. Symptoms assessed were hy- peremia, chemosis, and itching. Examinations mcluded slit lamp evaluations. Each symptom was graded on a scale of 0 to 4 (i.e., the maximum possible score was 12) with 0 indicating no symptoms and 1, 2, 3, and 4 representing mild, moderate, moderately severe, and severe symptoms, respectively. Since each filter paper strip was used for the measurement of no more than two mediators (see below). several individuals were challenged on more than I day to allow the collection of sufficient data for each mediator. In a limited number of instances, an allergic subject was chal- lenged with the protocol described earlier, but three separate filter paper strips were superimposed and then placed in the inferior fomix. The strips were then separated and placed into different tubes containing the appropriate solutions for each mediator assay. In this manner it wds possible to demonstrate that the entire spectrum of mediators being assayed were all generated in the same sample of tear fluid. Sample processing Each preweighed filter paper strip was placed immedi- ately into a preweighed tube containing 300 p.1 of the ap- propriate collection fluid for the mediators to be measured. Samples for histamine and Tahoe-esterase activity were collected into isotonic saline. Samples for kinins and al- bumin were collected into saline containing 40 mmol/L of ethylenediaminetetraacetic acid, whereas samples far PGD, and immunoreactive leukotrienes were collected into 95% ethanol. In each case, the tubes were weighed again im- mediately after placement of the filter strip into the tube. In this manner the weight of collected tear fluid could be calculated for each paper strip. This method provided re- liable estimates of weights unless very low volumes (<:S 111) of tear fluid were collected. Each collection tube was
`
`Page 2
`
`

`
`898 Proud et al. J. ALLERGY CLIN. IMMUNOL. MAY 1990 loo, LEFTEYE(PLUGGED) V. ALLIiRGEN CHALLENGE loo RIGHT EYE (UNPLUGGED) 90 80 70 60 I P CHALLENGE FIG. 1. Volume of tear fluids collected after challenge of collagen-plugged and unplugged eyes from the same allergic subjects. Tear volumes were determined by weight subtraction. Horizontal bars represent medians. Tear volumes were not different after saline challenge of the two eyes, and allergen challenge resulted in significant increases in tear weights in both eyes (p < 0.02 in unplugged eyes and p < 0.005 in plugged eyes). These allergen-induced increases above saline challenge were not different between the two eyes. then stored at -80” C until assay. When samples were thawed, fluids for the determination of histamine, TAME- esterase activity, kinins, and albumin were aliquoted and assayed at appropriate dilutions. The samples for arachi- donic acid metabolites were thawed, and the strips were removed. The samples were then centrifuged at 10,000 g for 5 minutes, and the supematants were removed and con- centrated to dryness with a vacuum concentrator (Savant Instruments, Inc., Hicksville, N.Y.). Each sample was then reconstituted with 0.5 ml of phosphate-buffered saline and assayed immediately for immunoreactive PGD, and LTC,. Mediator assays Histamine was measured in 50 p.1 aliquots of fluid with a radioenzymatic assay sensitive to 50 pg/m12’, 22 This assay uses partially purified histamine N-methyl transferase from rat kidney and S-adenosyl-C-[‘HI-methyl-methionine (New England Nuclear/DuPont, Boston, Mass.) to convert histamine to [3H]-methylhistamine, which is then isolated from other radiolabeled materials and quantified by liquid scintillation counting. A standard curve ranging from 50 pglml to 10,000 pg/ml was used in each experiment, and unknown values were determined by reading from the linear portion of the standard curve. The identity of measured histamine was confirmed by also assaying selected samples in a fluorometric assay” and by demonstrating that treatment with diamine oxidase destroyed the measured material. Kinins were assayed with a competitive RIA sensitive to 20 pglml of bradykinin8 The antibody used does not dis- criminate, on a molar basis, between bradykinin, lysylbra- dykinin, and methionyllysylbradykinin. The only other ma- terials known to cross-react in the assay are the human kininogens, which display ~2% cross-reaction on a molar basis. The assay has inter- and intra-assay coefficients of variation that am both <5%. The identity of the kinins present in tear fluids after antigen provocation was deter- mined by HPLC, as described previously.* In brief, pooled tear fluids were extracted by passage over SepPak Cl8 cartridges (Waters Associates, Milford, Mass.). After car- tridges were washed with 0.05% trifluoroacetic acid per 0.025% triethylamine, pH 2.5 (solution A), kinins were eluted from the cartridge with 50% solution A per 50% solution B (solution B: 80% acetonitrile in solution A) and concentrated to 50 to 100 ~1 with a vacuum concentrator. Recovery of [‘HI-bradykinin with this method was 90%. Samples were centrifuged at 10,OOfl g for 2 minutes to remove any particulate material before application to the HPLC column. HPLC separation was achieved with a 3.9 by 300 mm p, Bondapak C,, column (Waters Associates) that was eluted at a flow rate of 1 ml/mm with a mobile phase consisting of 76% solution A and 24% solution B. Absorbance of standard bradykinin, lysylbradykinin, and methionyllysylbradykinin was monitored at 210 nm. Frac- tions were collected and assayed at appropriate dilutions to confirm the imniunoreactivity of eluted materials. To ensure that there was no carryover of standard kinins into tear-fluid samples, the sample loop was thoroughly washed with ace- to&rile after each run, and a negative control run was performed before administration of tear samples. For all tear samples the identity of kinins present was determined by immunoassay of recovered fractions and comparison to the retention times of authentic kinins. Enzymes that have arginine esterase activity were assayed according to the method of Imanari, et al. ,% which is based on the liberation of tritiated methanol from the synthetic substrate )H-TAME. In brief, duplicate 40 p,l aliquots of appropriately diluted samples are incubated with 10 p,l of 0.2 mol/L of Tris buffer, pH 8.0, and 10 pl of 3H-TAME for 60 minutes at room temperature. The reaction is then terminated, and tritiated methanol is partitioned into an or- ganic scintillation fluid and quantied by liquid scintillation counting. Results are recorded as counts per minute above background (corrected for dilution). Although the enzymes responsible for the TAME-e&erase activity in tears have yet to be identified, during the immediate allergic reaction in the upper airways, TAME-esterase activity is largely a reflection of the activities of a plasma kallikrein/a,- macroglobulin complex and of mast cell trypmse with an additional, smaller, contribution by tissue kallikrein.“~Z6 The albumin content of recovered tears was measured by a specific RIA sensitive to 1 ng/ ml.*’ Immunoreactive sul-
`
`Page 3
`
`

`
`VOLUME 85 NUMBER 5 Mediator generation in ocular d~ergy ~!BS fidopeptide leukotrienes were measured with a previously described IUA.28 The antibody used was raised against LTC.,, and cross-reactivity with LTD, and LTE, was 68% and 25%, respectively. The sensitivity limit of the assay is 200 pg/ml of LTC,. PGDz was measured with an RIA capable of detecting 50 pg/ml of PGD,.Z9 The antibody used dem- onstrates < 1% cross-reaction with PGEl, PC&, thrombox- ane B,, PGF,,, and 6-keto-PGF,,. Statistical analysis All data were corrected to be expressed as concentrations in original tear fluids (i.e., not diluted by the collection fluid) except TAME-esterase activity, which was expressed as counts per minute generated by a 40 ~.rl aliquot of tears. Comparisons of mediator values after saline versus allergen challenges in each group of subjects were performed non- parametrically with the Wilcoxon matched-pairs sign-ranks test. This test was also used for comparing weights of tears, baseline mediator levels, and allergen-induced increases in mediators (above saline challenge) between plugged and unplugged eyes. The degree of association between media- tor levels and symptoms was evaluated with Spearman’s rank correlations. Across-group analysis (allergic subjects versus control subjects) of allergen-induced increases in me- diators (above saline challenge) was also performed non- parametrically with the Mann-Whitney U test. Statistical significance was presumed to be achieved for values of p < 0.05, Statistical analysis was performed on a Macin- tosh computer (Apple Computer, Inc., Cupertino, Calif.) with Statview software (Brainpower, Inc., Calabasas, Calif.). RESULTS In our initial analysis of these data obtained in the present study, we evaluated the effects on conjunctival provocation of the use of collagen plugs to block the tear duct. Our purpose was to determine if quantitative recoveries of tear fluids and mediators could be ob- tained without occluding the tear duct. We analyzed the effects of collagen plugging on the weights of tear fluids recovered in all challenge experiments (n = 37 for allergic subjects and n = 26 for control subjects). In control subjects there was no significant difference in tear weights after saline challenge in plugged versus unplugged eyes (median tear weights, 16.7 mg and 14.4 mg, respectively). Furthermore, there was no significant increase in tear weights in either eye after allergen challenge of control subjects, and the post- challenge tear weights were not different between plugged and unplugged eyes (median, 14.8 mg and 12.3 mg, respectively). The tear weight data for al- lergic subjects are illustrated in Fig. 1. Again, no significant differences were observed in tear weights after saline provocation of plugged and unplugged eyes (median, 9.7 mg and 10.1 mg, respectively). Significant increases in tear weights were noted, how- ever, after allergen challenge of both plugged and o-1 F CONTROLS ALLEWfCS FtG. 2. Distribution of symptom scorea after albrgen chal- lenge of allergic subjects. Symptom acoma are reported on a scale of 0 to 12. The median is represented by a horizontal bar. unplugged eyes (p < 0.005 for collagen-plugged eyes and p < 0.02 in unplugged eyes). Although postallergen-challenge tear weights were statistically different in the two eyes (median, 18.1 mg in phtgged eyes and 14.3 mg in unplugged eyes;p -C O.oOS), the magnitude of the difference was modest, particularly in view of the inherent inaccuracies of the method for determining tear weight and appears unlikely to be of biologic significance. Moreover, when a comparison was made of the allergen-induced increases in tear weights above increases recorded for sahne chalIenge in plugged and unplugged eyes, no significant effects of collagen plugging were observed. Next, a comparison of mediator levels in plugged versus unplugged eyes of the same subjects was per- formed. No sign&ant differences were observed in the levels of baseline mediators in ~1~~~ versus unplugged eyes. Similarly, aIIergen-induced mediator levels were analyzed in terms of mediator generation above that induced by saline provocation of the same subjects (n = 12 to 14 for each mediator in aIlergic subjects and n = 8 for each mediator in control sub- jects). No significant differences for unplugged eyes were observed for any m either allergic or control subjects. T&en in its entirety, we interpreted these data to indicate that collagen plugging of the tear duct had littIe or no effect on the response to allergen provocation. A~~~~~, for all subsequent analyses, symptom score and mediator data from plugged and unplugged eyes were com- bined. Administration of sahne or allergen to the con- junctiva of nonallergic individuals Med to induce any detectable symptoms. Similarly, admi n of a nonrelevant allergen to the eyes of allergic subjects failed to elicit a clinical response. By contrast, while administration of saline pm&iced no safe ef- fect in allergic subjects, provocation with a relevant
`
`Page 4
`
`

`
`900 Proud et al. J. ALLERGY CLIN. IMMUNOL. MAY 1990 loo0 CONTROLS ALLERGICS SAirNE ALLERGEN * SALINE ALLERGEN CHALLENGE CHALLENGE CHALLENGE CHALLENGE FIG. 3. Histamine levels in tear fluids from allergic and control subjects. Median levels are demonstrated by horizontal bars. Histamine levels increased significantly after allergen chal- lenge of allergic subjects (p < 0.001) but not control subjects. Histamine levels after saline challenge were also significantly elevated in allergic subjects compared to control subjects 1000 ZONTROLS 10 SALINE ALLE-RGEN - CHALLENGE CHALLENGE ALLERGICS SALINE ALLERGEN CHALLENGE CHALLENGE FIG. 4. PGD2 levels in tear fluids from allergic and control subjects. Median levels are illustrated by horizontalbars. Levels after saline challenge are not different for allergic and control subjects. Allergen challenge induced a significant increase in PGD2 levels in allergic subjects (p < 0.005) but not in control subjects. allergen led to itching, chemosis, and hyperemia in all subjects tested. Although there was some variation in the responses observed (Fig. 2), the median com- posite symptom score of 9 was consistent with a mod- erately severe clinical reaction. While histamine levels (Fig. 3) were not signifi- cantly affected after allergen challenge of control sub- jects when levels were compared to levels measured after saline provocation (median, 13.6 rig/ml versus 6.9 rig/ml, respectively), significant elevations (p < 0.001) were observed for allergic subjects (median, 60 rig/ml after allergen versus 15.6 rig/ml after sa- line). Interestingly, the concentration of histamine de- tected after saline challenge of allergic subjects was also significantly elevated (p = 0.048) compared to that of control subjects. This was the only mediator for which saline challenge induced a significant dif- ference between allergic and nonallergic subjects. PGD, is the major prostanoid produced by human mast cells.30 In light of the increases observed for histamine after allergen challenge of allergic subjects, we also examined the levels of this newly generated mediator in tear fluids (Fig. 4). No effect of allergen provocation was observed in control subjects (median, 7.0 rig/ml after allergen versus 6.2 rig/ml after sa- line), but a significant effect was observed for allergic subjects (median, 12.3 rig/ml after allergen versus 9.6 rig/ml after saline; p < 0.005). In contrast to PGD*, immunoreactive leukotrienes were not signif- icantly increased after challenge of either control (me- dian, 5.3 rig/ml after allergen challenge versus 4.6 rig/ml after saline) or allergic (median, 12.7 rig/ml after allergen versus 7.3 rig/ml after saline; p = 0.073) subjects (Fig. 5). Changes in vascular permeability were also ob- served in response to allergen provocation (Fig. 6). Although a slight increase in the concentration of al- bumin relative to saline provocation was noted in con- trol subjects (median, 1.4 mg/ml versus 0.6 mg/ml; p < 0.05), perhaps signifying some irritant effect of the challenge procedure, the increases in allergic sub- jects were significantly (p < 0.001) greater (median, 7.3 mg/ml after allergen versus 0.5 mg/ml after sa- line; p < 0.001). TAME-esterase activity (Fig. 7),
`
`Page 5
`
`

`
`VOLUME 85 NUMBER 5 Mediator generation in ocular .~~l!lergy 901 SALINE SALINE CHALLENGE CHALLENGE C#&!!!!& E v FIG. 5. lmmunoreactive leukotriene levels in tearfiuidsfrom allergic and control subjects. Median levels are illustrated by horizontalbars. Concentrations after saline challenge were not different between control and allergic subjects. No significant increases were observed for either group in response to allergen challenge. .l SALINE ALLERGEN ALLERGEN CHALLENGE CHALLENGE CHALLENGE FIG. 6. Albumin levels in tear fluids from allergic and control subjects. Median levels are iilus- trated by horizontal bars. Levels after saline challenge were not different between allergic and control subjects. Allergen challenge induced significant increases in albumin concentrations in both control subjects (p < 0.05) and allergic subjects (p < 0.001). Allergen-induced increases above saline challenge levels were significantly greater for allergic than for control subjects (p < 0.001). which, during allergic events in the upper airways, represents largely the activity of a plasma kalli- krein/q macroglobulin complex, was also strikingly increased during symptomatic responses to allergen challenge in allergic subjects (median, 1008 x lo3 cpm after allergen challenge versus 58 x lo3 cpm after saline; p < O.OOl), whereas modest increases in control subjects (median, 121 x lo3 cpm after aller- gen versus 34 X 10’ cpm after saline) were not sig- nificant. Similarly, the concentrations of kinins, po- tent vasoactive peptides, were dramatically elevated (Fig. 8) in tears during the response to allergen oc- curring in allergic subjects (median, 212 rig/ml after allergen versus 9.9 rig/ml after saline; p < O.OOl), whereas no such effects were observed after challenge of control subjects (median, 17.4 ng / ml after allergen versus 12.9 ng/ ml after saline). HPLC analysis of the kinins generated during the ocular allergic response revealed that a mixture of bradykinin and lysylbra- dykinin was produced (Fig. 9). To examine the degree of association between me- diator levels and the clinical response to allergen chal- lenge, we performed Spear-man’s rank cormlatio~ be- tween symptom scores and the levels of each me&ator in tear fluids after both saline and allergen provo- cation of allergic subjects. Significant correlations (p < 0.001 in each case) were observed for hista- mine (rs = 0.76), kinins (rs = 0.71), albumin (r> = 0.77), and TAME$esterase activity (f, = 0.82), but not for immunoreactive leukotrienes (q = 0.13) or PGD, (rs = 0.23). Finally, to demonstrate statistically that the allergen-induced increases in mediafor concer&rations in allergic subjects were greater than increases ob- served in control subjects, across-groq analysis of the allergen-induced elevations in mediator concemra- tions (expressed as increases above levels after saline challenge) was performed. This analysis n&rated significant increases in allergic subjects compared to control subjects for histamine, FGD, (p K 0.005 in
`
`Page 6
`
`

`
`902 Proud et al. J. ALLERGY CLIN. IMMUNOL. MAY 1990 low0 $ : ALLERGICS 5 i: loo0 so- WE: q * 100 ZL $3 F 10 54 2 1 SALINE ALLERGEN SALIkE ALLERGEN CHALLENGE CHALLENGE CHALLENGE CHALLENGE FIG. 7. Levels of TAME-esterase activity in tear fluids from allergic and control subjects. Median levels are illustrated by horizontal bars. Levels after saline challenge were not different between allergic and control subjects. Allergen challenge induced significant increases in TAME-esterase activity in allergic subjects (p < 0.001) but not in control subjects. CONTROLS FIG. 8. Kinin levels in tear fluids from allergic and control subjects. Median levels are illustrated by horizontal bars. Levels after saline challenge were not different between allergic and control subjects. Allergen challenge induced significant increases in kinin levels in allergic subjects (p c 0.001) but not in control subjects. each case), kinins, albumin, and TAME-esterase ac- tivity (p < 0.001 in each case), but not for immu- noreactive leukotrienes . Thus, increased generation of inflammatory mediators is associated with the symp- tomatic response to ocular challenge with allergen. DISCUSSION The pathogenesis of allergic reactions involves the interaction of a specific allergen with IgE antibodies bound to receptors on mast cells and basophils. Bridg- ing of the IgE receptors on these cells results in a complex sequence of biochemical events that leads to cellular activation, arachidonic acid metabolism, and the release of preformed mediators, such as histamine and trypta~e.~’ The eye is a frequent target of inflam- mation during allergic reactions, with most ocular al- lergy affecting the conjunctiva, the mucous membrane of the eye.32v 33 This presumably relates to the fact that, whereas many structures in the eye, such as the cornea, lens, iris, and retina, are virtually devoid of mast cells, these cells are present in a high density (5000 cells per cubic millimeter) in the human con- junctiva.% Conjunctival provocation therefore pro- vides a relevant model to study the ocular allergic response. Our current data represent the first comprehensive, controlled study of the profile of inflammatory me- diators generated during the ocular allergic response. A major advantage of the system described here, com- pared to models used in other tissues, is the ability to determine the actual volume of secretions produced and thereby to express the concentrations of mediators produced per milliliter of tear fluid. This is of im- portance, since the criticism has been raised in other model systems that increased mediator levels in lavage Auids during the allergic reaction could be explained by increases in the volumes of secretions recovered by lavage rather than by increased concentrations of mediators in secretions. Our comparison of tear col- lection using the Schirmer filter paper strips in collagen-plugged versus unplugged eyes suggests that collagen plugging of the tear duct had little, if any, effect on the recovery of tear fluids from the eyes or on mediator concentrations, and therefore, plugging of the tear duct does not appear necessary to obtain reproducible data. The incus in the volume of ff uid
`
`Page 7
`
`

`
`VOLUME 85 NUMBER 5 Mediator generation in ocular allergy 303 recovered after allergen challenge of allergic, but not normal, subjects are consistent with a specific tearing response during the allergic reaction. Thus, increases in mediator concentrations in tear fluid together with an increased volume of tears would, clearly, result in a greatly exaggerated increase in the total level of a mediator generated during the allergic response. The generation of histamine and PGDl during the allergic response is consistent with mast cell activation in the conjuntiva. Although it has previously been reported that histamine levels in tears are not increased during ocular hay fever,* this study compared levels in symptomatic individuals to levels in normal control subjects. Our present data clearly demonstrate that there is a considerable range of histamine levels in both subject populations and that it is possible that differences may not be detected if limited numbers of subjects are used. In contrast, by use of each indi- vidual as their own control, as in the present study, such individual variations are avoided. The median concentrations of histamine (5 X 10m7 mol/L) and PGD, (5 x lo-* mol/L) in tear fluids during the al- lergic response are clearly sufficient to initiate inflam- matory events. A surprising finding of the current study was the lack of a significant increase in the levels of immu- noreactive leukotrienes in tears during the allergic re- sponse, particularly since this is the only mediator previously reported to be increased after conjunctival provocation.’ The same antibody was used in both studies; therefore, different patterns of cross-reaction can not explain the different results. Similar levels of immunoreactive leukotrienes were detected after al- lergen challenge in both studies, but Bisgaard et al.’ detected much lower levels in control eyes than were detected in the present study. The reason for this dis- crepancy is not clear, but further studies, with more sophisticated techniques to identify cross-reacting ma- terials, like authentic leukotrienes, are needed to clarify the question of leukotriene generation during ocular allergic reactions. In addition to the direct release of mast cell-derived mediators, such as histamine and PGD2, it is clear that other mediators are generated during the allergic response in the eye. Kinins are potent vasoactive pep- tides that have been demonstrated to be generated during allergic events in the airways