conjunctival transdifferentiation induced by systemic...

Conjunctival Transdifferentiation Induced by SystemicVitamin A Deficiency in Vascularized Rabbit Corneas

Scheffer C. G. Tseng,* Mahmood Farazdaghi,t and Agarha A. Rider:}:

Conjunctival transdifferentiation, the process in which conjunctival epithelium transforms into acornea-like epithelium with the loss of goblet cells during the healing of a total corneal epithelialdefect, can be retarded or reversed by corneal neovascularization. We have previously shown that thisprocess normally occurring on non-vascularized corneas can be retarded or reversed by topical retin-oids, suggesting that vitamin A may be one of the factors from blood circulation which is responsiblefor modulating transdifferentiation. Herein, we have examined the effect of systemic vitamin A defi-ciency on vascularized corneas starting 4 months after epithelial denudation, and compared thisdeficient group with their vascularized and non-vascularized controls. Mean serum retinol level Gxg/dl)(n = 4) measured by HPLC was gradually reduced from 83 of the controls to 20 in a 10 monthfollow-up. Topographical analysis disclosed a centrifugal loss of goblet cell density with time. Histol-ogy showed complete transdifferentiation in vascularized areas at 9 months, initiated by the loss ofmucin contents from receding zones first noted at 2 months. Using impression cytology, all corneaswere not keratinized and all conjunctivas maintained a normal goblet cell density at 10 months. Theseresults indicate that conjunctival epithelium on corneal surface is more sensitive to the decrease ofserum vitamin A levels than that on conjunctiva, and support tti« hypothesis that the relative vitamin Adeficiency on vascularized corneas can also result in the conjunctival transdifferentiation. InvestOphthalmol Vis Sci 28:1497-1504, 1987

When a total corneal epithelial defect is createdextending 2 to 3 mm beyond limbus, the epithelialsource for wound healing is derived from the sur-rounding conjunctiva. During the healing process,the fate of the migrating conjunctival epithelium onthe denuded corneal surface will be determined bythe presence or absence of corneal neovasculariza-tion. In its absence, the conjunctiva-like epitheliumwill undergo serial stages of morphological transfor-mation into a cornea-like epithelium with the loss ofgoblet cells,1"3 a process now termed conjunctivaltransdifferentiation. However, if corneal vasculariza-tion is present during the healing period1'34 or intro-duced afterwards,5'6 conjunctival transdifferentiationis either inhibited1'3"5 or reversed,6 as evidenced bythe persistence of goblet cells. These investigations

From the *Department of Ophthalmology, Bascom Palmer EyeInstitute, University of Miami, Miami, Florida, the tlmmunologyUnit, Wilmer Ophthalmological Institute, and the ^Department ofBiochemistry, School of Hygiene and Public Health, the JohnsHopkins Medical Institution, Baltimore, Maryland.

Presented in part at the ARVO meeting, May 6-10, 1985, atSarasota, Florida.

Supported by grant EY-05656 (SCGT) from the National EyeInstitute.

Submitted for publication: December 5, 1986.Reprint requests: Scheffer C. G. Tseng, MD, PhD, P.O. Box

016880, Bascom Palmer Eye Institute, Miami, FL 33101.

indicate that corneal vascularization may be a causa-tive factor in modulating conjunctival transdifferen-tiation.

We speculated that the persistence of goblet cellson the corneal surface in this model may be sup-ported by some modulating factor(s) from the bloodcirculation.7 We further hypothesized that vitamin Aor retinoids may be among the factors that can inhibitconjunctival transdifferentiation by maintaininggoblet cell differentiation, based on the following in-formation about vitamin A. First, it has been shownthat under normal conditions, the conjunctival andscleral blood vessels at the limbal region are the majorsource of vitamin A for the cornea when the distribu-tion of retinol-binding protein is analyzed,8 or whenthe blood-borne horseradish peroxidase tracer is stud-ied.9 Second, retinoids are essential for epithelialgrowth and differentiation.1011 Numerous studies,both in vivo and in vitro, have demonstrated thatdeficiency of this vitamin can convert a secretory epi-thelium to a squamous epithelium (squamous meta-plasia), and a vitamin excess can convert a stratifiedsquamous epithelium to a secretory epithelium(mucous metaplasia).10" Because of the relative defi-ciency of vitamin A in the normal avascular corneacompared to that of vascularized conjunctiva, con-junctival transdifferentiation occurs.

To test the above hypothesis, we have previously

1497

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933134/ on 04/24/2018

1498 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / Seprember 1987 Vol. 28

demonstrated that topical retinoids, ie etretinate and13-cis retinoic acid, can inhibit conjunctival transdif-ferentiation in non-vascularized corneas by main-taining goblet cell differentiation.12 Recently, we alsodemonstrated that topical application of all-trans ret-inoic acid or increased all-trans retinol levels in tearsand aqueous by a model of immunogenic uveitis canreverse the transdifferentiated cornea-like epitheliumin the non-vascularized corneas to a conjunctiva-likeepithelium with the reappearance of goblet cells.13 Inthe present study, we further tested this hypothesis invascularized corneas by inducing conjunctival trans-differentiation, which would otherwise not occur invascularized conditions, using the model of systemicvitamin A deficiency.

Materials and Methods

Rabbit Model of Conjunctival Transdifferentiation

A rabbit model of conjunctival transdifferentiationwas created in New Zealand albino rabbits, bodyweight 2-3 kg, in a manner similar to a previousmethod3 using n-heptanol originally described byCintron et al.14 Corneas that healed within 6 dayswere excluded because the resultant epithelium mayhave originated from corneal remnants. On the 10thday after wounding, the remaining corneas werescreened for epithelial defects and corneal vasculari-zation. Those corneas healed without epithelial de-fects could then be classified into the following twomajor groups: the non-vascularized group with ves-sels extending less than 2 mm into the cornea, and thevascularized group with vessels extending more than2 mm into the cornea and involving more than threequadrants. The occurrence of these two corneal typeswas haphazard, unpredictable, and possibly related tothe severity of the chemical debridement. Corneasthat had not healed completely by day 10 were re-screened on day 12. A total of 38 rabbits were used inthis experiment, which gave 43 non-vascularized cor-neas, 24 vascularized corneas, and 9 corneas whichwere excluded due to vascularization occurring onlyin part of the cornea. Our study conformed to theARVO Resolution on the Use of Animals in Re-search.

Induction of Vitamin A Deficiency

Four months after re-epithelialization, 18 rabbitswith 24 vascularized corneas were subjected to thevitamin A-depleted diet (Test diet No. 7722, TekladTest Diets, Madison, WI), originally described byVan Horn et al in the rabbit xerophthalmia model.15

Three to four corneas were obtained from the experi-mental group at 2, 4, 6, 7, 9, and 10 months after the

special diet by sacrificing the rabbits with an overdoseof intravenous phenobarbital.

Tissue Preparation For Morphological Study andGoblet Cell Analysis

After sacrifice, we removed the cornea with a 1 to 2mm scleral rim using a razor blade and scissors. Thecorneal button was bisected through the vertical me-ridian into nasal and temporal halves. One half wasassigned to routine histological study and the other togoblet cell analysis. Histological study was performedusing 6 ixm tissue sections stained with periodic acid-Schiff (PAS) reagent for goblet cells, and counter-stained with hematoxylin and eosin (H-E). Gobletcell analysis was performed as previously described.3

In brief, flat-mount preparations were made on eachcorneal half and then subjected to fixation and se-quential stainings with 0.5% Alcian blue (pH 2.5) andPAS reagents. The topographical density of gobletcells was counted at X400 magnification at each 1mm zone from the center to the limbus. The meanvalue of the three meridians of the three pies of eachcorneal half was designated as the value of that cor-neal specimen. The mean value and its standard errorof three or four of such specimens was reported herefor comparison at any given time of sacrifice, in a waysimilar to that of our previous report.13

Measurement of Serum Retinol Levels by HPLC

In order to monitor the progress of vitamin A de-pletion, serum level of all-trans retinol (retinol) wasmeasured by a HPLC method. Aliquots of 0.2 ml ofserum were mixed with an equal volume of 125 ret-inol equivalent (RE)/dl of retinyl acetate in absoluteethanol which was used as an internal standard. Afteraddition of 1.0 ml hexane, the mixture was agitatedon a Vortex Mixer for 60 sec. It was then centrifugedat 2000 rpm for 15 min. The supernatant was re-moved and evaporated at room temperature in astream of nitrogen. The residue was then dissolved in0.2 ml absolute ethanol, of which an aliquot of 30 iAwas injected into a Varian High Performance LiquidChromatography (HPLC) apparatus (Sunnyvale,CA), which was a Model 5000 equipped with a Var-ichrom U-Vis detector set at 328 nm. The reversephase chromatography was run using a Waters C-18,10 M, 3.9 X 30 cm column at a flow rate of 2.5 ml/minof methanol:water (96:4). This method was similar tothat reported previously by Driskell et al.16 The con-centration of serum retinol was calculated using theinternal standard method by comparing the peakheights. Duplicate runs were made on each sample.Before use, the samples were covered by aluminumfoil and stored at -20°C. The sample preparation


No. 9 CONJUNCTIVAL TRANSDIFFERENTIATION AND VITAMIN A DEFICIENCY / Tseng er ol.

was proceeded under yellow fluorescent light. All re-agents used here were of HPLC grade.

Impression Cytology

Ten months after vitamin A depletion, impressioncytology was performed on both corneal and con-junctival surface using our previous method17 to de-termine if there was any change of squamous meta-plasia on the ocular surface as a result of vitamin Adeficiency.

Results

Four months after total removal of corneal andlimbal epithelia, conjunctival epithelium on the non-vascularized corneas transdifferentiated into a cor-nea-like epithelium with total disappearance of gobletcells. In contrast, vascularized corneas still main-tained a conjunctiva-like epithelium containing gob-let cells (results not shown). These findings were con-sistent with those of our previous reports.31213 Itshould be noted that the conjunctiva-like epitheliumwould persist indefinitely without spontaneous trans-differentiation on the vascularized corneas, since vas-cularized controls maintained these conjunctival fea-tures for more than 2 yr after wounding in a separatesurvey (unpublished results). Figure 1 shows the find-ing of numerous goblet cells on such a vascularizedcontrol cornea 20 months after debridement.

Rabbits with vascularized cornea(s) were thenplaced on a vitamin A-depleted diet. The dietary ef-fect on the serum retinol concentration was moni-tored monthly by HPLC in a total of 10 month fol-low-ups except at the 3rd and 8th month after vita-min A deficiency. As shown in Figure 2, mean serumretinol levels gradually decreased from 83 /ug retinolequivalent (RE)/dl of the pre-dietary control to 20 ugRE/dl at 7 months and then remained stationarythereafter until the end of 10 months of vitamin Adeficiency.

The density and distribution of goblet cells werethen analyzed on the vascularized corneas during this10 month period of study. The results are summa-rized in the Figures 3 and 4. The goblet cell density,measured as number of cells per mm2 at 2 months ofvitamin A deficiency, was high and distributedthroughout the entire vascularized corneas from thecenter to the mid-periphery and the periphery of thecornea (Fig. 4, upper panel). As shown in Figure 3, atup to 4 months goblet cells were still found from thelimbus (L) to the center (C) of the cornea. Thereafter,one began to see the gradual centrifugal receding ofgoblet cells from the central part of the cornea to theperiphery (Figs. 3, 4, middle panel), and finally at theend of 10 months, goblet cells could only be found in

Fig. 1. Flat-mount preparation of the vascularized control corneaobtained 20 months after debridement. The picture was taken 4mm from limbus at X160 (Alcian blue and PAS). Note the gobletcells on the corneal surface.

the peripheral 1 to 2 mm from limbus (Figs. 3, 4,lower panel). Accompanied with the centrifugal lossof goblet cells was the morphological transformationof the conjunctiva-like epithelium containing gobletcells to a cornea-like one with increased cellular strat-ification which was first noted at 9 months (Fig. 4,upper through lower panels).

The changes of goblet cells were then deliberatelyexamined at the receding zone. It became clear thatgoblet cells followed sequential steps of cellular de-generation before they disappeared. These degenerat-ing goblet cells were quantitated in the same manneras mentioned above from C to L of the cornea, theresults of which are presented in Figure 5. The de-tailed features of the sequential degenerative changesare illustrated in Figure 6. The first degenerativechange was noted at 2 months after vitamin A defi-ciency using flat-mount preparations, on which one

2 - ao

8 1

A-DEPLETED DIET

Fig. 2. Serum retinol concentrations (̂ g retinol equivalent/dl)measured by HPLC. In the 10 month course of vitamin A defi-ciency, serum retinol concentrations were measured monthly fromfour samples of which the mean and its standard error (extensionbar) was plotted for comparison.


1500 INVESTIGATIVE OPHTHALMOLOGY b VISUAL SCIENCE / Seprember 1987 Vol. 28

7500 r 4 m

6 m

5 4 3 2 I

DISTANCE FROM LIMBUS (mm)

L 0.5

Fig. 3. Topographical distribution of goblet cell density. Gobletcell density (cells per mm2) was measured at each 1 mm zone fromthe center (C) to the limbus (L) of the cornea, and expressed asmean and standard error (extension bar) at 4, 6, 7, 9, and 10months (n = 3 or 4). Note the centrifugal loss from C to L withtime.

could see sporadic emptying of central mucin contentin some goblet cells (Fig. 6A). Following this, at 4months, the remaining mucin further disintegrated,and some goblet cells became smaller in size, con-taining bluish mucin, and some formed halos as aresult of the loss of goblet cells in the center of a groupof goblet cells (Fig. 6B). Since the bluish colorationsignifies the acidic nature of the mucin,18 acidicmucin was used here. These degenerating goblet cellsoccurred sporadically and were distributed through-out the entire vascularized cornea at 2 months andincreased in density at 4 months; the density at thelatter time is displayed in Figure 5. The distributioncurve indicated that the peak density was at the 2 mmzone from limbus, or 4 mm from the central corneawhere the goblet cells could still be found. As thevitamin A depletion further progressed to 6 months,the degenerating goblet cells at the receding zone ex-hibited coalescence of the halo formation as well asacidic mucin rings, resulting in a total loss of goblet

cells (Fig. 6C). At this time, the density of the degen-erating goblet cells with halo formation receded cen-trifugally, and the peak density moved to 1 mm fromlimbus, or, interestingly, 4 mm from the goblet cellleading edge, which was 5 mm from the limbus (Fig.5). At 7 months or later, this density curve of degen-erating goblet cells further receded centrifugally andthe peak density could only be found at the limbus,and the goblet cell leading edge was at 3-4 mm fromthe limbus (Fig. 5). The centrifugal receding of thedegenerating goblet cells was accompanied by the dis-appearance of goblet cells from the central cornea tothe leading edge, which was characterized by the co-alescence of residual acidic coloration into a clusteror acinus pattern without any staining distributedalong the remaining underlining vasculatures (Fig.6D). The results of the changes of the density curvesof degenerating goblet cells from 4 to 9 months alongwith the centrifugal shift of the peak density zone,which was always at 3-4 mm from the receding edge,strongly indicated that the conjunctival transdiffer-entiation was under the influence of a centrifugalgradient difference from the limbus to the center ofthe cornea.

It should be noted that we did not note any kera-tinization of the corneal surface (Fig. 4), nor anysquamous metaplasia of the conjunctival surfaceusing impression cytology (results not shown) even atthe end of the 10 month period of study.

Discussion

The rabbit model of conjunctival transdifferentia-tion is an intriguing and ideal system to study thecellular differentiation of conjunctival epithelium.This process was first described by Friedenwald1 andsubsequently thoroughly studied by several other in-vestigators, including Thoft, Friend et al,2'519"21 Liuet al,6 and us.3'71213 These pathophysiological studiesof this experimental model have given us insights intothe ocular surface changes caused by a total loss ofcorneal and limbal epithelia. This information ishelpful in understanding such clinical problems aschemical burns, in which the first histopathologicaldescription of a similar finding of conjunctival trans-differentiation was given by Maumenee and Scholz.4

Our major interest in this model has been the mo-dulating mechanism by which conjunctival transdif-ferentiation occurs on non-vascularized corneas.Based on previous observations that conjunctivaltransdifferentiation can be either inhibited14'5 or re-versed6 by corneal neovascularization, we speculatedthat the modulating factor(s) might be coming fromthe blood circulation and should meet the following


No. 9 CONJUNCTIVAL TRANSDIFFERENTIATION AND VITAMIN A DEFICIENCY / Tseng er ol. 1501

Fig. 4. Histological sections of the vascularized corneas collected at 2 (upper panel), 6 (middle panel), and 9 (lower panel) months aftervitamin A deficiency. The photographs were taken at the central 1 mm (left panel), mid-peripheral 3 mm (middle panel), and peripheral 5 mm(right panel) from the center of the cornea. Note the goblet cells in the overlying epithelium (some indicated by arrows), and the cross-sectionsof the blood vessels, (some marked by asterisks). Note also the centrifugal loss of goblet cells and gradual replacement by a cornea-likeepithelium with time (all original magnifications X160, H-E and PAS).

two criteria: (1) be present in systemic vascular circu-lation in a high concentration, but relatively low ornegligible in extravascular tissues; and (2) possess theability of modulating goblet cell differentiation. Be-cause vitamin A or retinoids have the above twoproperties, we hypothesized that vitamin A might beone of the candidate factors.7 Our previous stud-ies31213 together with the present report have sup-ported this hypothesis, which is summarized in Fig-ure 7.

Conjunctival epithelium growing on the denudedcorneal surface can exhibit one of the following twomorphologies: either conjunctiva-like containinggoblet cells or cornea-like without goblet cells butwith increased cellular stratification. The transfor-mation of original conjunctiva-like epithelium to acornea-like one is termed transdifferentiation, andoccurs as a result of relative local vitamin A defi-

400

300

200

100

L 0.5

DISTANCE FROM LIMBUS (mm)

Fig. 5. Topographical distribution of degenerating goblet celldensity. Degenerating goblet cells per mm2 were measured at each Imm zone from the center (C) to the limbus (L) of the cornea andexpressed as mean and standard error (extension bar) at 4,6, 7, and9 months (n = 3 or 4). For each curve, note the peak density was 3to 4 mm from the receding edge and its centrifugal shift with time.


15Q2 INVESTIGATIVE OPHTHALMOLOGY b VISUAL SCIENCE / September 1987 Vol. 28

Fig. 6. Flat-mount prepa-rations of the vascularizedcorneas collected at 2 (A), 4(B), 6 (C), and 9 (D) monthsafter vitamin A deficiency.A and B were taken at themid-peripheral zone; C andD at the area where gobletcells entirely disappeared.Note the sporadic emptyingof central mucin in somegoblet cells (A), (arrow),which progressed into cen-tral holes, (arrow), and fur-ther disintegrated intosmaller-sized goblet cells,indicated by arrow heads(B). Total disappearance ofgoblet cells produced the re-sidual colorless negativehalos arranged in a clusterfashion shown by arrows (C,D). Note also that the un-derlying vasculature be-came clear at the later stage(D) (all original magnifica-tions XI60 except A X400,Alcian blue and PAS).

ciency. This relative deficiency state can in turn occureither on normally non-vascularized cornea in whichblood-borne vitamin A is relatively lower than innormally vascularized conjunctiva,3 or on vascular-ized corneas in which systemic vitamin A deficiencyis induced (this report). The notion of the relativelocal vitamin A deficiency is verified in this report bythe finding that serum retinol levels were subnormalbut not undetectable. As a matter of fact, the firstchange of goblet cell degeneration was noted at 2months and became pronounced from 4 to 7 months,during which time serum retinol levels remained inthe range of 20 to 60 Mg RE/dl. Even when transdif-ferentiation was completed at 9 months, its level wasstill around 20 fig RE/dl. At that time, goblet celldensity over conjunctiva remained unchanged andcornea showed no keratinization, indicating that thevitamin A-deficient state of the ocular surface was arelative one. In other words, under the above twoconditions, the conjunctival epithelium growing onthe denuded corneal surface will lose goblet cell dif-ferentiation earlier than that growing on the normalconjunctival area.

The ocular surface changes of this vitamin A defi-ciency model should not be confused with those origi-nally described by Van Horn et al,15 though the samediet was used. Severely depleted xerophthalmia canbe induced in young rabbits with a body weight lessthan 0.5 kg, as originally reported by Van Horn et

all5>22 and recently reproduced by us,23 since the vita-min A storage level was low in young animals. In thisreport, vitamin A depletion was created 4 monthslater on rabbits with a starting body weight of 2-3 kg,which had a high vitamin A storage level and thusmade the clinical xerophthalmia inapparent.

According to this hypothesis, one can imagine thattransdifferentiation can also be inhibited or reversedby increasing a local supply of vitamin A to the cor-neal surface. Indeed, we have shown that this wasachievable by either making the corneas vascular-ized3 or by applying topical retinoids onto the non-vascularized corneas'213 (Fig. 7). All the evidencepoints to a close relationship between local vitamin Asupply and goblet cell differentiation, and the latter isvery sensitive to the former.

Previous studies by Rask et al8 and Raviola9 indi-cate that the major supply of vitamin A to the epithe-lium on the corneal surface is from blood circulation.This notion is supported by all of our previous exper-iments31213 and the present study. They all show thatthe loss of goblet cells in conjunctival transdifferen-tiation is always centrifugal and that their regenera-tion in the reversal process is always centripetal. Re-cently, Ubels et al demonstrated free retinol in rabbittears24 and lacrimal gland fluid25 in the concentrationof 6.9 Mg/dl and 6.2 jtg/dl respectively. The facts thatthese values are ten times less than that of serum andthat the goblet cell changes actually follow a radial


No. 9 CONJUNCTIVAL TRANSDIFFERENTIATION AND VITAMIN A DEFICIENCY / Tseng er ol. 1503

Non-vascularized

Corneas

Vascularized Corneasunder Vitamin A Deficiency

Relative LocalVitamin A Deficiency

TRANSDIFFERENTIATION

CONJUNCTIVA-LIKEEPITHELIUM

ICORNEA-LIKE

EPITHELIUM

REVERSAL OFTRANSDIFFERENTIATION

Increased LocalVitamin A level

Non-vascularizedCorneas + Topical

Retinoids

^Vascularized Corneas

Fig. 7. Summary of the hypothesis explaining the modulatingmechanism of conjunctival transdifferentiation. The process oftransdifferentiation is thought to result from relative local vitaminA deficiency, which has been tested experimentally by either exam-ining non-vascularized corneas3 or by inducing vitamin A defi-ciency in vascularized ones (present report). On the other hand, thereversal of transdifferentiation occurs as a result of increased localvitamin A level to the corneas, either by applying topical retinoidsto non-vascularized corneas1213 or by examining vascularizedones.3

gradient difference, as shown in Figures 3 and 4, in-dicate that the major supply of vitamin A to cornea isfrom limbal vessels and not from tears.

The hypothesis that transdifferentiation is modu-lated by local vitamin A supply via blood vessels is ofclinical significance. Several ocular surface disordersare known to bear a common manifestation of squa-mous metaplasia of the ocular surface epithelia,7 apathological process including loss of goblet cells,mucin deficiency, increased cellular stratification,and keratinization.17 Since conjunctival goblet celldifferentiation is so sensitive to the local vitamin Alevel, we have speculated that some of these disordersmight have resulted from a local deficiency. For ex-ample, the scar or fibrosis in the subconjunctival pro-pria, which is found at the chronic cicatricial stage ofsome of these diseases, may well be a barrier to localvitamin A supply to the overlying epithelium. Basedon this concept, we have advocated the use of topicalretinoic acid in treating some of these disorders.26"28

The preliminary results were encouraging and sup-ported the above hypothesis, further indicating aclose resemblance between the pathological processof squamous metaplasia and the experimental modelof conjunctival transdifferentiation.

As stated above, conjunctival transdifferentiationinvolves morphological transformation of the con-junctiva-like epithelium to a cornea-like one, which isindeed a process of cellular differentiation. In thepresent report, the findings that acidic mucin came todominate neutral mucin together with the shrinkageof goblet cell size occurring in an early stage of vita-min A deficiency are similar to our earlier study ofthe transdifferentiation on non-vascularized cor-neas,3 and to our previous finding on mucin changesin xerophthalmic rats.29 These results indicate thatbiochemical changes of conjunctival mucin mightoccur far ahead of those of goblet cell density. Wetherefore speculate that one of the major actionmechanisms for vitamin A modulation of conjuncti-val goblet cell differentiation is through the control ofmucin expression. Recently, we have purified andpartially characterized rabbit ocular mucin30 and de-veloped monoclonal antibodies to it (see the com-panion papers in this issue). In the future, we hope tofurther explore the molecular controlling mechanismby which vitamin A modulates conjunctival transdif-ferentiation.

Key words: conjunctival transdifferentiation, goblet cells,retinoids, vascularization, vitamin A deficiency

References

1. Friedenwald JS: Growth pressure and metaplasia of conjuncti-val and corneal epithelium. Doc Ophthalmol 5-6:184, 1951.

2. Shapiro MS, Friend J, and Thoft RA: Corneal re-epithelializa-tion from the conjunctiva. Invest Ophthalmol Vis Sci 21:135,1981.

3. Tseng SCG, Hirst LW, Farazdaghi M, and Green WR: Gobletcell density and vascularization during conjunctival transdif-ferentiation. Invest Ophthalmol Vis Sci 25:1168, 1984.

4. Maumenee AE and Scholz RO: The histopathology of theocular lesions produced by the sulfur and nitrogen mustards.Bull Johns Hopkins Hosp 82:121, 1948.

5. Thoft RA, Friend J, and Murphy HS: Ocular surface epithe-lium and corneal vascularization in rabbits: I. The role ofwounding. Invest Ophthalmol Vis Sci 18:85, 1979.

6. Liu SH, Tagawa Y, Prendergast RA, Franklin RM, and Silver-stein AM: Secretory component of IgA: A marker for differen-tiation of ocular epithelium. Invest Ophthalmol Vis Sci20:100, 1981.

7. Tseng SCG, Hirst LW, Maumenee AE, Kenyon KR, Sun T-T,and Green WR: Possible mechanisms for the loss of goblet cellsin mucin-deficient disorders. Ophthalmology 91:545, 1984.Rask L, Geijer C, Bill A, and Peterson PA: Vitamin A supply ofthe cornea. Exp Eye Res 31:201, 1980.Raviola G: Conjunctival and episcleral blood vessels are per-meable to blood-borne horseradish peroxidase. Invest Oph-thalmol Vis Sci 24:725, 1983.Logan WS: Vitamin A and keratinization. Arch Dermatol105:748, 1972.Elias PM and Williams ML: Retinoids, cancer, and the skin.Arch Dermatol 117:160, 1981.

12. Tseng SCG, Hirst LW, Farazdaghi M, and Green WR: Inhibi-

8.

9.

10

11


1504 INVESTIGATIVE OPHTHALMOLOGY b VISUAL SCIENCE / September 1987 Vol. 28

tion of conjunctival transdifferentiation by topical retinoids.Invest Ophthalmol Vis Sci 28:538, 1987.

13. Farazdaghi M, Liu SH, Rider AA, Lutty GA, Prendergast RA,and Tseng SCGT: Reversal of conjunctival transdifferentiationby retinoids. ARVO Abstracts. Invest Ophthalmol Vis Sci25(Suppl):76, 1984.

14. Cintron C, Hassinger L, Kublin CL, and Friend J: A simplemethod for the removal of rabbit corneal epithelium utilizingn-heptanol. Ophthalmic Res 11:90, 1979.

15. Van Horn DL, Schutten WH, Hyndiuk RA, and Kurz P: Xe-rophthalmia in vitamin A-deficient rabbits: Clinical and ultra-structural alterations in the cornea. Invest Ophthalmol Vis Sci19:1067, 1980.

16. Driskell WJ, Neese JW, Bryant CC, and Bashor MM: Measure-ment of vitamin A and vitamin E in human serum by high-per-formance liquid chromatography. J Chromatogr 231:439,1982.

17. Tseng SCG: Staging of conjunctival squamous metaplasia byimpression cytology. Ophthalmology 92:728, 1985.

18. Sheehan DC and Hrapchak BB: In Theory and Practice ofHistotechnology. St. Louis, C.V. Mosby, 1980, p. 159.

19. Thoft RA and Friend J: Biochemical transformation of regen-erating ocular surface epithelium. Invest Ophthalmol Vis Sci16:14, 1977.

20. Friend J and Thoft RA: Functional competence of regenerat-ing ocular surface epithelium. Invest Ophthalmol Vis Sci17:134, 1978.

21. Kinoshita S, Friend J. Kiorpes TC, and Thoft RA: Keratin-likeproteins in corneal and conjunctival epithelium are different.Invest Ophthalmol Vis Sci 24:577, 1983.

22. Hatchell DL and Sommer A: Detection of ocular surface ab-normalities in experimental vitamin A deficiency. Arch Oph-thalmol 102:1389, 1984.

23. Farazdaghi M, Tseng SCG, and Prendergast RA: Corneal en-dothelial changes in rabbit xerophthalmia. ARVO Abstracts.Invest Ophthalmol Vis Sci 27(Suppl):175, 1986.

24. Ubels JL and MacRae SM: Vitamin A is present as retinol inthe tears of humans and rabbits. Curr Eye Res 3:815, 1984.

25. Ubels JL, Foley KM, and Rismondo V: Retinol secretion bythe lacrimal gland. Invest Ophthalmol Vis Sci 27:1261, 1986.

26. Tseng SCG: Topical retinoid treatment for dry eye disorders.Trans Ophthalmol Soc UK 104:489, 1985.

27. Tseng SCG, Maumenee AE, Stark WJ, Maumenee IH, JensenAD, Green WR, and Kenyon KR: Topical retinoid treatmentfor various dry-eye disorders. Ophthalmology 92:717, 1985.

28. Tseng SCG: Topical tretinoin treatment for severe dry-eye dis-orders. J Am Acad Dermatol 15:860, 1986.

29. Kenyon KR, Huang AJW, Tseng SCG, and Gipson IK: Mor-phogenesis of conjunctival goblet cells in the rat. ARVO Ab-stracts. Invest Ophthalmol Vis Sci 26(Suppl):174, 1985.

30. Tseng SCG, Huang AJW, and Sutter DJ: Purification andcharacterization of rabbit ocular mucin. Invest OphthalmolVis Sci 28:1473, 1987.


conjunctival transdifferentiation induced by systemic...

Documents