the effect of thermal burns on the release of collagenase from

The effect of thermal burns on the releaseof collagenase from corneas of

vitamin A-deficient and control ratsWen Lin Seng, Joseph A. Glogowski, George Wolf, Michael B. Berman,

Kenneth R. Kenyon, and Timothy C. Kiorpes

A mild trauma in the form of a thermal burn was applied to corneas of vitamin A-deficient ratsand their pair-fed controls. The control corneas routinely showed rapid re-epithelializationwithout stromal changes. The corneas of deficient rats recovered more slowly, frequentlyexhibiting stromal edema, leukoma, and sometimes ulceration. Because collagenase is thoughtto initiate collagen destruction in corneal ulceration, the relationships among vitamin A status,severity of trauma, and collagenase levels were determined. Mild thermal burns were found tocause corneas from less severely deficient rats to ulcerate rarely but to release increased levels ofcollagenase, mainly on the first day of culture, as in the case of nonburned, severely deficientrats. Comparable burns of corneas of pair-fed control rats resulted in no idceration and in verylittle collagenase release. Severe burns of either pair-fed control or normal rat corneas causedulceration and collagenase release, but collagenase activity was maximal on the second and.third days of culture. Differences in vitamin A status at time of burning gave rise to differentpatterns of collagenase. By following the development of the vitamin deficiency, it was deter-mined that little active collagenase is released after mild burns of corneas in animals in thepre -weight plateau stage but that much more active enzyme is released when animals are inweight plateau or 5% weight loss stages. Studies of the effect of recovery from vitamin Adeficiency on the response to mild thermal burn indicated that the longer the interval betweenfeeding vitamin A and the burn, the lower the postburn level of collagenase in the day 1medium. Thus it would appear that restitution of vitamin A status decreases the level of activecollagenase after the mild thermal burn. The system developed here can be used to study thebiochemical basis for ulceration in vitamin A deficiency, and the possibility exists that theulceration characteristic of keratomalacia in people can he initiated by an environmentaltrauma.

Key words: cornea, vitamin A deficiency, vitamin A, rat, keratomalacia,collagenase, collagen, trauma, thermal burn, corneal ulceration,

polymorphonuclear neutrophils

From the Department of Nutrition and Food Science,Massachusetts Institute of Technology, Cambridge,and the Eye Research Institute of Retina Foundationand the Department of Ophthalmology, HarvardMedical School (M. B. B., K. R. K., and T. C. K.),Boston, Mass.

This work was supported in part by Research grants EY515 (G. W.) and EY 02146 (M. B. B.) of the NationalEye Institute, NIH, and the Adelaide Breed BayrdFoundation (G. W.); by Academic Investigator AwardEY 156 from the National Eye Institute, NIH, the

William R. Milton Fund of Harvard University, and agrant-in-aid from Fight-For-Sight, Inc., New York,N.Y. (K.R.K.); and by Biomedical Research Supportgrant PHS 5SO7RR05527 from the National Eye Insti-tute and the Massachusetts Lions Eye Research Fund,Inc. (M. B. B., K. R. K., and T. C. K.).

Submitted for publication July 18, 1979.Reprint requests: G. Wolf, Department of Nutrition and

Food Science, Massachusetts Institute of Technology,Cambridge, Mass. 02139.

0146-0404/80/121461 + 10$01.00/0 © 1980 Assoc. for Res. in Vis. and Ophthal., Inc. 1461

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933089/ on 03/25/2018

1462 Seng et al.Invest. Ophthalmol. Vis. Set.

December 1980

Keratomalacia is the rapid corneal ulcer-ation that occurs in children who are severelydeficient in vitamin A. It has been estimatedthat tens of thousands of new cases of thisdisorder occur each year in malnourishedchildren. Although the exact sequence ofevents leading to tissue dissolution has notbeen elucidated, it would seem that as incases of ulceration after chemical burns,thermal burns, or infection, the degradationof fibrillar corneal collagen is initiated by theenzyme collagenase. Indeed, Pirie et al.1

have recently described the release in cul-ture of collagenolytic activity from vitaminA-deficient rat corneas.

It is unlikely that vitamin A deficiencyalone causes corneal ulceration. Rather, it isour hypothesis that tissue changes specific tothe deficiency can predispose the cornea toulceration. This predisposition could occureither by the lowering of the threshold forthe initiation of the ulcerative process, or byrendering ineffective the control mechanismsthat normally prevent tissue destruction, orby both. In this view, vitamin A deficiency isa necessary but not a sufficient component ofkeratomalacia. The events leading to ulcer-ation and destruction of the cornea wouldthen have to be initiated by some externalinsult, be it physical or chemical or the resultof infection.

In this communication, we compare vita-min A-deficient and pair-fed control rats intheir response to mild thermal burn of thecornea. The differences observed, both clini-cally and biochemically, support the sugges-tion that the vitamin A-deficient cornea is atrisk with respect to ulceration.

Materials and methods

Rats. Weanling Holtzmann albino rats were puton a vitamin A—deficient diet2 for 3 to 4 weeks andthen paired by weight with normal control ratswhich were fed the same diet with a vitamin Asupplement (200 /xg of retinyl acetate per week).After they had been paired, the weights of thecontrol and deficient rats were recorded. Theamount of food eaten by the deficient rat on oneday was given to its paired control rat on the nextday.

For the vitamin A recovery experiments, ratswere kept on a vitamin A-deficient diet until they

developed deficiency symptoms (5% to 15% weightloss). Then retinyl acetate (150 //.g in 0.2 ml ofcottonseed oil) was given intragastrically to thedeficient rats at different time intervals either be-fore or after the thermal trauma.

Thermal burns. Rats without eye lesions wereselected from animals having lost 5% to 15%weight after maintenance on the vitamin A -deficient diet. Animals were anesthetized by in-traperitoneal injection of 0.5 ml of chloral hydratesolution per 100 gm body weight (3.6% solutionfor deficient; 7% for control rats). The eyes werethen treated with 1 drop each of 0.5% pro-paracaine hydrochloride (E. R. Squibb, Bur-lington, Mass.). After application, the eyes werewashed with saline and blotted dry. A thermalprobe (Thermokeratophore; Frigitronics, Inc.,Shelton, Conn.) of 2 mm diameter was used at130° C to burn their corneas. An instant touch ofthe probe to the central cornea produced a "mildburn." The bum produced by applying the probeto the cornea twice at the same spot with increasedpressure and 1 sec duration each time was definedas a "severe burn."

Clinical and histopathologic observations. Ani-mals were examined at intervals by slit lamp.When ready, they were sacrificed, and their eyeswere enucleated, placed in 10% neutral-bufferedformalin, and prepared by standard techniques forparaffin-embedded sections. These were stainedwith hematoxylin and eosin, and serial sectionswere taken at four stepped intervals across theentire cornea. Representative sections were exam-ined and photographed with a Zeiss photomicro-scope III.

Organ culture methods. Generally, unlessotherwise stated, burned rats were sacrificed 3days after burning. The reason for choosing thistime interval was that Kenyon et al.1 had foundulceration of rabbit corneas, accompanied by col-lagenase release, to begin 3 days after a severethermal burn. Corneas were cut into small piecesand cultured in Dulbeccos modified EaglesMedium with NaHCO3 (0.2%), penicillin (110U/ml), and streptomycin (100 /^g/ml). Each cul-ture dish containing six corneas and 1 ml ofmedium was placed in a sterile box and gassedwith 5% CO2 + 95% O2 at 37° C. The mediumwas changed every day for 5 days, and the har-vested medium was stored at —20° C. The sterilityof cultures was confirmed by the absence of bac-terial growth on nutrient agar plates subsequent toinoculation with media harvests.

Preparation of collagenase. The stored mediawere thawed and concentrated by ultrafiltration to0.5 ml (filter PTGS, cutoff point 104 mol. wt.; Mil-


Volume 19Number 12 Corneal collagenase in vitamin A deficiency 1463

lipore Corp., Bedford, Mass.), for purposes ofassay. Concentrated media were then dialyzedovernight against 0.025M Tris HCl buffer, pH 7.8(4° C), containing 0.005M CaCl, and 0.1M NaCl,and then lyophilized. The lyophilized enzyme wasthen redissolved in 50 fx\ of double-distilled ordeionized distilled water and was assayed.

Characterization of collagenolytic activity. Di-alyzed harvests from thermally burned ulceratingrat corneas were incubated with 0,1% calf tail ten-don collagen at 23° C. For the determination ofcleavage products of collagen, reaction mixtureswere electrophoresed in 6% or 10% polyacryl-amide gels containing 0.1% sodium dodecyl sul-fate.4

Collagenase assay. Collagenase activity was de-termined in a capillary "heat gel" system whichemploys undenatured (native) calf tail tendon col-lagen as substrate.5 This system discriminates ac-tive collagenase from other proteases that can de-grade denatured collagen (gelatin). Collagenaseactivity was determined from the mean of threereplicate assays incubated for 100 hr. Rates of col-lagen gel lysis were demonstrated to be pro-portional to enzyme concentration, over a range ofapproximately 0.05 to 0.25 mm of lysis per hour(unpublished data), as had been shown previouslyto hold for the enzyme from rabbit corneas.5 Dataare expressed as millimeters of lysis per hour percornea.

Collagen solttbilization. The amount of hy-droxyproline in the medium as a marker for solu-bilized collagen was determined by the method ofRojkind and Gonzales,6 after centrifugation of har-vested media.

Results

Clinical and histological observationsNonburned animals. In nonburned pair-

fed control animals, the corneal epitheliumremained intact and lustrous (Fig. 1,A) incontrast to vitamin A-deficient animals(weight loss 20%), which showed varying de-grees of corneal xerosis, with punctate kera-topathy evidenced by methylene blue stain-ing of the irregular epithelium (Fig. 1,J5),These deficient animals were never noted toshow gross lesions such as stromal edema orulceration. Severely deficient rats (>20%weight loss), on the other hand, frequentlyshowed eye lesions such as stromal edema orleukoma and, rarely, ulceration.

By light microscopy, nondeficient controlsdemonstrated entirely unremarkable mor-

Fig. 1. A, Pair-fed control animal shows intactcorneal epithelium with normal luster (note cor-neal light reflex). B, Vitamin A-deficient animal(10% weight loss) demonstrates corneal xerosis asirregular epithelial surface (note blurred lightreflex) with diffuse superficial punctate kerato-pathy, as indicated by diffuse methylene bluestaining (circled). C, Severely vitamin A-deficientanimal (20% weight loss) 3 days after thermal burnexhibits diffuse corneal xerosis (as in B), plus cen-tral stromal ulceration with perforation.

phology (Fig. 2, A), and deficient animals (5%to 15% weight loss) showed only minimal su-perficial epithelial keratinization with nostromal abnormality (Fig. 2, B). Severe vi-


1464 Seng et al. Invest, Ophthaimol. Vis. Sci.December 1980

Fig. 2. For legend, see facing page.



tamin A deficiency, however, resulted inmarked epithelial surface keratinization,granulosa cell layer formation, stromal infil-tration by inflammatory cells, and peripheralstromal neovascularization (Fig. 2, C and D).

Thermally burned animals. In response tomild thermal burns, nondeficient controlsshowed rapid epithelial recovery, did notshow significant stromal opacification, andnever ulcerated. After the burn, deficientanimals (5% to 15% weight loss) often exhib-ited stromal edema and leukoma but rarelyulceration. In contrast, with severe deficien-cy a mild burn resulted in rapid stromal ul-ceration and frequent perforation (Fig. 1, C).Severe thermal burns resulted in stromal ul-ceration in both normal and vitaminA-deficient animals. Histologically, no in-flammatory response was noted in corneas ofmildy burned, nondeficient animals (Fig. 2,E). In contrast, deficient animals 3 days aftermild thermal burn usually showed intact epi-thelium and nonulcerating stroma with mod-erate acute inflammatory infiltrate (Fig. 2, F).In the severely deficient animal, looselyadherent epithelium often terminated atmargins of ulcerating stromas that wereheavily infiltrated by acute inflammatorycells (Fig. 2, G and H).

Characterization and assay of collageno-lytic activity. Electrophoresis of the reaction

products of collagen with media harvestsfrom both mildly burned deficient and se-verely burned normal corneas demonstratedthat collagen had been cleaved intrahelicallyinto approximately %-length and V4-lengthproducts (Fig. 3). As described by Pirie etal.,1 the initial cleavage products appear tohave been degraded further, presumably byother proteases in the media.

Control corneas given mild burns showedno lesions 3 days after burning. Upon cultur-ing of these corneas, low levels of collagenasewere detected on days 2 and 3 (Fig. 4, A).Much more collagenase was detected in cul-tures of corneas from vitamin A-deficientrats (weight loss 10.6% ± 0.65 S.E.M.) afterthe mild thermal burn. The active colla-genase level was greatest on day 1 of culture(Fig. 4, A). When nonburned corneas werecultured from control (pair-fed) and from vi-tamin A-deficient rats without eye lesions(weight loss 20.7% ± 5.4 S.E.M.), no de-tectable collagenase was released within thelimits of sensitivity of the assay,5 even after 5days of culture.

Nonburned corneas of severely deficientrats (weight loss > 20%; stromal edema andleukoma in corneas) released considerableactive collagenase into the medium whenplaced in culture. Collagenase levels werefound to be highest on the first day of culture

Fig. 2. A, Pair-fed control animal exhibits normal corneal cytoarchitecture with respect touniform, five-layered, nonkeratinizing epithelium, avascular stroma devoid of inflammatorycells, and unremarkable Descement's membrane and endothelium. (xlOO.) B, VitaminA-deficient animal (10% weight loss) shows only minimal keratinization of superficial epithelialcells. (X250.) C, Severe vitamin A deficiency (20% weight loss) results in marked epithelialsurface keratinization, granulosa cell layer formation, stromal inflammatory infiltration, andstromal neovascularization (peripherally). (X250.) D, Severely deficient animals showinfiltration of central stroma by acute inflammatory cells, predominantly PMNs. (X400.) E,Pair-fed control animal 3 days after mild thermal burn has re-established intact corneal epithe-lial layer. The stroma is intact but acellular, since neither keratocytes nor inflammatory cellshave repopulated the central stroma. (x250.) F, Deficient animal (10% weight loss) 3 days aftermild thermal burn shows intact epithelium and nonulcerating stroma which is, however,moderately infiltrated by acute inflammatory cells. (X150.) G, Severely deficient animal 3 daysafter mild thermal burn demonstrates extensive central epithelial defect (left) and nonadherentmarginal epithelium (right), with ulcerating stroma heavily infiltrated by acute inflammatorycells. (xlOO.) H, Severely deficient animal 3 days after mild thermal burn exhibits stromalulcer (left) with epithelium (right) terminating abruptly at margin of stromal defect. Thestromal lamellae appear diffusely disorganized and fragmented, and they are infiltrated byPMNs. (X250.)


1466 Seng et at. Invest. Ophthalmol. Vis, Set.December 1980

2 3 4

Enzyme rMixture \

P

a2

Fig. 3. Cleavage of collagen by media concentrates. Gel electrophoresis4 of the denaturedproducts of hydrolysis of collagen in solution. Concentrated culture media were incubated withcalf tail tendon collagen in solution at 23° for times indicated below and then acidified to pH 3.5by acetic acid. Incubation has resulted in degradation of the /3-dirners (/3U) /3i2) to %-lengthdimers (j3A

n, /3A12); of single a-chains (au a2) to %-length a-chains (aA

u ocA2). Left panel,

Media from mildly burned vitamin A-deficient rat corneas. (1) Calf tail tendon collagen incu-bated 72 hr in buffer;.(2) combined day 1 to day 4 culture concentrate; (3) calf tail collagen and 20£il of day 1 culture concentrate incubated 17 hr; (4) calf tail collagen and 150 /xl of day 4 cultureconcentrate incubated 17 hr. The one-fourth fragments are not resolved by the 6% gel usedand migrate with the buffer front. (3n, Intact dimer of a^ f}l2, intact divner of otx and a2; /3'V,%-length of /3 n ; /3A

12, %-length of/312; aY intact a,; a2, intact a2; aAu %-length a,; aA

2,%-length a2. Right panel, Media from severely thermally burned ulcerating rat corneas incu-bated for 4 days (3). Fragments with sizes between aA], and aA

2 as well as fragments smallerthan aA

2 are produced. Cleavage has also formed the V^-length fragment aB2. The aB, fragment

is nearly completely missing, indicating its continued degradation to small-molecular-weightfragments. Shown for comparison are /3 and a components of undegraded collagen (1) and %-and Vi-length collagen reaction products from incubation with rabbit corneal fibroblast colla-genase (2) used as markers. Acrylamide concentration, 10%.4 Notation is same as left panelexcept that /3 and /SA not resolved into /3 n and /3l2, nor 0A into /3Au and /3A

t2; aB1; Vi-length a,;

aB2, ^-length a2.

(Fig. 4, B). Thus less severely vitamin A—deficient rats without eye lesions (Fig. 4, A)showed a pattern of collagenase after burningclosely similar to that of nonburned severelydeficient corneas (Fig. 4, B).

Severe thermal burns of corneas from con-trol animals caused ulceration in almost allanimals and the presence of active colla-genase in the culture media. As opposed tothe deficient corneas, active collagenase wasmaximal on day 2 of culture and continuedthrough day 3 (Fig. 4, A).

To study corneal collagen solubilization asan index of actual stromal destruction, de-terminations were made of hydroxyproline-

containing fragments in daily harvest ofmedia. Inspection of hydroxyproline profiles(Fig. 5, A) shows that more collagen was sol-ubilized on day 1 in culture of burneddeficient corneas than on day 1 in cultures ofburned or nonburned pair-fed control cor-neas. Thereafter, daily solubilization of cor-neal collagen appeared comparable in cultureof burned deficient and pair-fed controls.Nonburned pair-fed controls showed muchless collagen solubilization. Despite the factthat collagenase showed a peak at day 1 inculture of burned deficient corneas, the peakof collagen solubilization by such corneas wason day 2 of culture. When the first harvest



Days

Fig. 4. Collagenase activities in cultures of corneas. Culture media were harvested daily for 5days and assayed for collagenase. A,—•— Mildly burned vitamin A-deficient corneas. Eachpoint is the mean of five separate experiments, with a total of 17 culture dishes (five to sixcorneas per dish). Bars represent S.E.M.---A---, Severely burned, pair-fed control corneas.Each point is the mean of two experiments, with a total of eight culture dishes, (five to sixcorneas per dish). Bars represent the range. • , Mildly burned, pair-fed control cor-neas. Numbers of corneas same as for severely burned control corneas. B, Nonburned, se-verely vitamin A-deficient corneas. Each point represents the mean of three experiments;bars give S.E.M. Culture dishes (five to six corneas per dish) used in each experiment werefive, four, and two.

was made after 2 days of culture, the initialhydroxyproline levels were much higher thanin cultures of burned and nonburned pair-fedcontrols (Fig. 5, B). Thereafter, daily sol-ubilization of collagen was comparable in allthree types of cultures. These data are inter-preted to mean that there is a lag in collagendegradation after introduction of the corneasto culture, so that even though collagenaselevels are higher initially in burned deficientcultures, solubilized collagen does not reflectthe initial difference in collagenase levels.

Collagenase during development and re-covery from vitamin A deficiency. The re-sponse to mild burns is correlated with theextent of progressive vitamin A deficiency(Fig. 6, A). At the stage when there was stillsufficient vitamin available to support growth(preplateau) a small amount of collagenasewas detected on days 2 to 5 of culture. Whengrowth ceased (plateau stage), presumably

after a decline in tissue vitamin A, the mildburn caused the appearance of collagenaseactivity on days 1 to 2 of culture. The amountof collagenase detected increased with ad-vancing vitamin A deficiency. Conversely,upon intragastric intubation of one dose ofretinyl acetate (150 /w-g), complete recoveryoccurred from the effects of deficiency on thecornea, provided the vitamin was given 48 hrbefore the mild burn; a very low level of col-lagenase was detectable on day 1 of culture,and a small amount was detected on days 2 to5 (Fig. 6, B). If the vitamin was given 24 hrbefore the burn, some collagenase was de-tected on day 1 of culture, and more was de-tected if the vitamin was given immediatelyafter or 24 hr after burning.

Discussion

The results of the current studies suggestthat corneas from vitamin A-deficient rats


1468 Seng et al.Invest. Ophthalmol. Vis. Sci.

December 1980

20

16

12

8

4

- A ]

-/

--

V

\ ~"\

I /

_

1

-

t

Fig. 5. Release of hydroxyproline (OH-Pro) (mi-crograms per cornea) into the media of cultures ofrat corneas. —•—, Mildly burned, vitaminA-deficient corneas. Each point is the mean ofthree separate experiments, with a total of threeculture dishes (four to six corneas per dish). Barsrepresent S.E.M.---•---, Mildly burned, pair-fedcontrol corneas. Numbers of corneas same as fordeficient • • • • • • • , Nonburned, pair-fed controlcorneas. A, Culture media harvested daily and as-sayed for hydroxyproline.6 B, Culture media ini-tially harvested 2 days after culturing, then daily,and assayed for hydroxyproline.

are at increased risk of ulceration as a resultof trauma. This was demonstrated by the dif-ferential clinical and biochemical responsesto the same mild burn by the deficient andthe control corneas. The burns caused nocorneal lesions in pair-fed controls, and littleor no collagenase was detected in media fromsuch corneas. Control corneas could becaused to ulcerate but only by severe burns,and the pattern of release of collagenase dif-fered from that of the burned vitamin A-deficient corneas. We were also able toconfirm the observation of Pirie et al.1 thatnonburned corneas from severely vitaminA-deficient rats (>20% weight loss; grosseye lesions) released active collagenase intothe medium. The collagenase was detectedmainly on the first day of culture.

The difference in patterns of collagenaseactivity (Fig. 4), mainly on day 1 for thedeficient (burned and nonburned) and days 2and 3 for the burned pair-fed controls, is ofsignificance in that it might bear upon themechanism of ulceration in the vitamin A-deficient cornea. The recovery experiment inparticular (Fig. 6, B) shows that on days 2 to 5the levels of detectable collagenase presentin the media are almost identical for corneasfrom animals whose day 1 collagenase valuesdo differ as a function of vitamin A status.Moreover, collagenase values for days 2 to 5are comparable, whether vitamin A was orwas not restored to the deficient animals. It isclear, then, that the day 1 media reflect thevitamin A status. Collagenase levels on sub-sequent days of culture are thought to reflectthe results of the tissue injury involved in thescissoring and mincing procedures used toprepare corneas for culture.

The hydroxyproline data (Fig. 5) are also insupport of the interpretation that day 1 mediareflect the vitamin A status. They suggestthat the greatest daily solubilization of cor-neal collagen in vitro in culture of burnedvitamin A-deficient corneas occurs duringthe early part of the culture period. Thereaf-ter collagen solubilization occurs at slowerrates in all types of cultures. It is to be notedthat although pair-fed controls do not showpersistent corneal pathology after thermalburns in vivo, they do demonstrate solubili-zation of corneal collagen in organ culturesomewhat above that of nonburned controls(Fig. 5). If one considers the hydroxyprolineprofile (Fig. 5, B) in nonburned controls asreflecting merely the tissue mincing and cul-turing and subtracts this from the burnedcontrol profile, it becomes clear that nearlyall the hydroxyproline in the burned controlis due to tissue preparation. Despite the factthat there is a relationship between colla-genase, vitamin A status, and hydroxyprolinerelease, our results show that hydroxyprolinesolubilization in culture is not correlated withcollagenase levels and most likely reflects theactivity of other proteases as well.

The present studies do not provide an an-swer to the question why nonburned se-verely vitamin A-deficient corneas some-



Fig. 6. A, Collagenase activities in cultures of corneas from mildly burned pre—weight plateau,plateau, and vitamin A-deficient rats. Culture media were harvested daily for 5 days andassayed for collagenase. —-•---, Preplateau rats; —A—, plateau rats; —o—, deficient rats (5%to 10% weight loss, no eye lesions). Each experiment was done with the combined media fromtwo culture dishes (six corneas per dish). B, Collagenase activities as a function of recoveryfrom vitamin A deficiency. Corneas were removed from mildly burned deficient rats (5% to15% weight loss, no eye lesions) at various times after vitamin A administration and culturedfor 5 days with daily change of medium. Collagenase activities were determined. — • — ,Vitamin A supplementation 48 hr before burning (exp. 1); . . . • . . . , vitamin A supplementa-tion 24 hr before burning (exp. 2); --O--, vitamin A supplementation 24 hr after burning (exp.3); -.-.A-.-., vitamin A supplementation immediately upon burning (exp. 4);—A—, no vitaminA supplementation (exp. 5). The numbers of culture dishes used for each experiment weretwo, two, three, three, two, respectively, with four to six corneas per dish. For experiments 1,2, and 5, the points represent means of two separate experiments, and the bars show the range;for experiments 3 and 4, points represent the means of three separate experiments, with barsrepresenting S.E.M. Combined collagenase values for experiments 1 and 2 (day 1) are statisti-cally significantly different from combined values of experiments 3, 4, and 5 (day 1) (p < 0.01).

times ulcerated and why thermal burns ofless severely vitamin A-deficient corneassometimes also caused ulceration. In keepingwith the hypothesis that trauma of some sortis necessary to initiate the events leading toulceration, it is possible that some severelyvitamin A-deficient animals have sufferedtrauma. This may have caused the release ofdestructive mediators which are successfullyregulated in well-nourished animals. Ulcer-ation in the present studies is correlated withthe presence in the ulcer region of polymor-phonuclear neutrophils (PMNs). Although

their exact roles are not understood, PMNshave been implicated in ulceration from di-verse causes. Indeed, in a study of ulcer-ations in vitamin A-deficient rat corneas,Pirie et al.1 observed that higher levels ofcollagenolytic activity were correlated withthe appearance of PMNs in the vitaminA-deficient corneas. Possibly the observedinability of the burned corneal epithelium toheal in vitamin A-deficient animals results inthe continuous infiltration of the cornealstroma by PMNs. A role of persistent epithe-lial defects in attracting PMNs into the cor-


1470 Seng et al. Invest. Ophthalmol. Vis, Sci.December 1980

nea has been discussed elsewhere.7 PMNs docontain hydrolases that can attack collagen;and it would seem that PMN-derived elastaseand cathepsin,8 in addition to collagenase,might all contribute to the collagen destruc-tion of corneal ulceration. In this regard,hydrolase-containing vesicles might be morelabile in vitamin A deficiency,9 so that hy-drolases more readily reach the extracellularmatrix. Some PMN hydrolases, so released,could then facilitate PMN infiltration of thestroma.7

Although PMNs appear to participate incorneal ulceration, roles for other cell typesare not ruled out. Indeed, rabbit cornealfibroblasts in cell culture produce a colla-genase with the same chromatographic andactivation properties as that obtained fromorgan cultures of corneas that are ulceratingafter alkali burns.

We have shown in this report that a mildthermal trauma inflicted on corneas, even atthe earliest stage of deficiency (Fig. 6,A),caused corneal lesions and release of activecollagenase whereas a trauma of the samemagnitude to control corneas caused no le-sions and resulted in little detectable colla-genase release. Thus trauma can initiate bio-chemical events in early stages of vitamin Adeficiency that may ultimately lead to stromalulceration. The possibility exists, then, thatthe very erratic and usually unilateral ap-pearance of ulceration characteristic of kera-

tomalacia, both in rat and man, may also beinitiated by an environmental trauma.

We are grateful to Dr. Peter F. Davison, BostonBiomedical Research Institute, for performing the elec-trophoresis of collagen cleavage products produced bycollagen obtained from vitamin A-deficient corneas.

R E F E R E N C E S

1. Pirie A, Werb Z, and Burleigh MC: Collagenase andother proteinases in the cornea of the retinol-deficient rat. Br J Nutr 34:297, 1975.

2. Sneider WD and Wolf G: Evidence that vitamin A isnot required for the biosynthesis of ovalbumin inchicks. J Nutr 106:1515, 1976.

3. Kenyon KR, Berman M, Conn H, Henriques A,Gage J, and Rose J: Ulceration and repair in rabbitcornea following thermal injury. INVEST OPHTHAL-MOL VIS SCI 17(ARVO Suppl.):252, 1977.

4. Laemmli UK: Cleavage of structural proteins duringthe assembly of the head of bacteriophage tissue.Nature 227:680, 1970.

5. Berman MB, Manabe R, and Davison PF: Tissuecollagenase: a simplified, semiquantitative enzymeassay. Anal Biochem 54:522, 1973.

6. Rojkind M and Gonzalez E: An improved methodfor determining specific radioactivities of proline-14C and hydroxyproline-14C in collagen and noncol-lagenous proteins. Anal Biochem 57:1, 1974.

7. Berman M: Regulation of collagenase. Therapeuticconsiderations. Trans Ophthalmol Soc UK 98:397,1978.

8. Burleigh MC: Degradation of collagen by non-specific proteinases. In Proteinases in MammalianCells and Tissues, Barrett AJ, editor. New York,1977, North-Holland Publishing Co., pp. 285-310.

9. Sudhabaran PR and Kurup PA: Vitamin A and lyso-somal stability in rat liver. J Nutr 104:1466, 1974.


the effect of thermal burns on the release of collagenase from

Documents