density and distribution of canine conjunctival goblet cells

Density and Distribution of ConineConjunctivol Goblet Cells

Cecil P. Moore,* Norman J. Wilsmaat Erik V. Nordheim4 Lynda J. Majors,§ and Linda L. Collier"

Conjunctival goblet cells (GCs) were quantitated to establish baseline values for density and distribu-tion of these cells in healthy canine eyes. From each of 18 sites, tissue was collected, sectioned at 2 fim,and stained with periodic acid Schiff stain. Within each sampling site, 500 epithelial cells (GCs,squamous, polygonal, and basal epithelial cells) were counted and the ratio of GCs to total epithelialcells was computed as an index of goblet cell density or goblet cell index (GCI). A heterogenousdistribution of canine conjunctival goblets cells was demonstrated. Lower nasal fornix (LNf) andadjacent sites, lower middle fornix (LMf) and lower nasal tarsal (LNt), had the highest mean densitiesof goblet cells. In contrast, GCs were essentially absent from the upper and lower bulbar areas.Remaining sites had intermediate GCIs. Sex differences in GCIs were noted for LNf and LNt sites.Mean tear film breakup times (BUTs) were determined, and, for normal beagle dogs, were 19.38(±4.80 sees) OS and 19.96 (±5.01 sees) OD. The similarities between canine and human conjunctivalgoblet cell distributions support the use of the dog for studying the conjunctival mucous system. InvestOphthalmol Vis Sci 28:1925-1932, 1987

Conjunctival epithelial goblet cells are the sourceof an undissolved gel-like mucin which forms thedeepest of three tear film layers.1"3 Goblet cell mucin(gel mucin) covers the glycocalyces of conjunctivaland corneal epithelial cells and forms a multipurpose"mucociliary blanket" over the ocular surface.45 Gelmucin serves as a corneal wetting agent by forming ahydrophilic interface between the aqueous tear layerand the hydrophobic corneal epithelium and fills insurface irregularities of the epithelial cells therebyproviding an optically smooth corneal surface.4"6

Goblet cell mucin protects the ocular surface by trap-ping particulate debris and bacteria, and contributesto local immunity by providing a medium for theadherence of immunoglobulins (IgA) and microbici-dal lysozymes.4"7

From the *Department of Veterinary Medicine and Surgery,College of Veterinary Medicine, University of Missouri, Columbia,Missouri; the tDepartment of Comparative Biosciences, School ofVeterinary Medicine, the ^Department of Statistics, College ofLetters and Science and the ^Department of Forestry, College ofAgricultural and Life Sciences, and the §Department of SurgicalSciences, School of Veterinary Medicine, University of Wisconsin,Madison, Wisconsin; the "Department of Pathology, College ofVeterinary Medicine, University of Missouri, Columbia, Missouri.

Supported by grants from the Graduate School, University ofWisconsin (140708) and COR Grant, University of Missouri(1352).

Submitted for publication: August 26, 1986.Reprint requests: Cecil P. Moore, W-104 Veterinary Medicine,

College of Veterinary Medicine, University of Missouri, Columbia,MO 65211.

Ocular diseases associated with quantitative andqualitative deficiencies of the tear gel mucin layer arecharacterized by instability of the tear film with sub-sequent drying of the ocular surface.8'9 Conjunctivalinflammation, squamous metaplasia, ocular pain,corneal ulceration, scarring and opacification, andblindness have been associated with these diseases.Spontaneously occurring GC deficiency with asso-ciated mucin insufficiency has recently been reportedin dogs.10

Tear film breakup time (BUT), tear mucin mea-surement, and GC quantitation are methods used toassess presence and functional capacity of preoculargell mucin.""13 Lack of standardization and variablesrelated to performance of the procedure have limitedthe usefulness of tear BUT as a diagnostic and re-search tool.1415 Mucin has been quantified by mea-suring hexosamine and 0-linked oligosaccharides.1216

Because specimens collected for biochemical studiesare aqueous tear samples, the value of these measure-ments has been questioned.1718 Morphologic study ofthe conjunctiva and GC counts provide an indirectmeasure of mucous production. Recent findings in-dicate that determination of GC density is a moresensitive indicator of ocular surface health thanmethods which measure aqueous tear mucin contentdirectly.1718 Accurate interpretation of GC quantitiesusing impression cytology depends upon priorknowledge of GC frequencies at accessible samplingsites.19"21

In domestic and laboratory animals considerablespecies variation has been noted in the location and

1925

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933133/ on 04/12/2018

1926 INVESTIGATIVE OPHTHALMOLOGY b VISUAL SCIENCE / December 1987 Vol. 28

Materials and Methods

Animals

Fig. 1. Incision of the lateral aspect of the conjunctival sac. Smallblunt tissue scissors are used to extend the incision from the cutane-ous lateral canthus to the lateral limbus. Note that the lateral longposterior ciliary artery is used as a limbal marker to define the lineof incision. An incision of the medial aspect of the conjunctival sacwas similarly performed using the medial long posterior ciliaryartery as a marker at the medial limbus.

density of conjunctival GCs.22"25 Variation in GCnumbers has been demonstrated at different sites inwhole mounts of human conjunctiva.26 The objectiveof this investigation was to quantify GCs at specificsites within the conjunctiva of normal dogs. The hy-pothesis tested was that GCs are disturbed homogen-eously throughout normal canine conjunctiva. Thisstudy was performed to provide essential informationfor accurate interpretation of biopsy specimens fromdiseased canine eyes and to permit interspecies com-parisons of GC frequencies and locations. Tear filmbreakup time (BUT), a noninvasive method of as-sessing preocular mucin function, was also investi-gated in dogs.

Fig. 2. Removal of the cornea and conjunctiva in toto. A scleralgroove is placed 1 mm posterior to the limbus and a 3 mm full-thickness incision is made into the anterior chamber. The cornea isthen separated from the sclera circumferentiatly with scissors asshown.

Twenty-four healthy young adult laboratory beagledogs (10 males and 14 females) between 9-18 monthsof age were studied and eyes of each animal werescreened for normality using a hand-held slit-lampbiomicroscope and a binocular indirect ophthalmo-scope. Schirmer tear tests and tear film breakup testswere performed 5 days prior to tissue collection andeuthanasia. All animals used in this study were caredfor and treated in accordance with the ARVO Reso-lution on the Use of Animals in Research.

Schirmer Tear Test

A Schirmer tear test (STT) strip was inserted intothe inferior conjunctival culdesac of each eye of eachdog and the tear levels were recorded after 60 sec-onds.

Tear Film Breakup Test

The dogs were sedated with xylazine and ketamineto minimize ocular movements. Tear film breakuptime (BUT) was recorded following instillation of twodrops of fluorescein dye into the conjunctival sac ofeach eye. After allowing two to three blinks to distrib-ute the fluorescein, the eyelids were held open man-ually. A 2 mm slit beam light with cobalt blue filterand X16 magnification on a stationary slit-lamp bio-microscope were used for the tear BUT tests. Timewas measured from the last blink to the appearance ofthe first spot of fluorescein disruption.

Tissue Collection, Processing and Sampling

After 5 days, each animal was anesthetized, theeyelids were sutured together with 4-0 silk, the peri-ocular area was aseptically prepared and the orbitswere exenerated. The orbital contents, extraocularmuscles, and periocular skin were excised and dis-carded leaving the closed conjunctival sac attached tothe globe. Subconjunctival connective tissues wereseparated from the episclera with a round-tipped sur-gical blade. The conjunctiva was reflected to thelimbus for the entire 360° limbal circumference.

This method was developed and used for in totocollection of conjunctiva and cornea from each eyeand is illustrated in Figures 1-3. This procedure al-lowed consistent division of the conjunctivae of alleyes into dorsal and ventral halves. Following tissuecollection the flat mount preparation was immersedin 10% buffered formalin and stored until tissueswere processed histologically. For each half of each


No. 12 CONJUNCTIVAL GODLET CELL DENSITY AND DISTRIBUTION / Moore er ol. 1927

conjunctival preparation, three horizontal zones(bulbar, fornix, and tarsal) were identified anatomi-cally and each zone represented approximately equalareas. Vertical zones were established by dividing thetissue into three approximately equal areas identifiedas nasal, middle and temporal zones. By combininghorizontal and vertical designations, nine sites weredefined for upper and lower conjunctiva and, there-fore, a total of 18 sampling sites were identified pereye. Designations and notations for the 18 sites aregiven in Figure 4.

One eye was used from each of five randomly se-lected males and five randomly selected females. Oneanimal from each sex group was randomly selected tocontribute both eyes to the study. This resulted in theinitial analysis of 12 eyes (six males-three left, threeright; six females-three left, three right). For each ofthe 18 sites per eye a 6 mm diameter core sample wascollected with a disposable cutaneous biopsy punch.

All Sampling Sites(Left Eye)

Upper(U)

Fig. 3. A corneal/conjunctival flat mount preparation (right eye).Following separation of the cornea from the sclera, sutures areremoved from the eyelid margins and the conjunctival sac isopened. By grasping the eyelid margins the tissues are carefullymounted on cardboard (conjunctival mucosa facing up) and at-tached at each corner with 4-0 silk suture. Note that the third eyelidhas been incised medially and that its larger portion remains at-tached to the lower one-half of the specimen. The entire flat mount,as shown, is prepared for placement in fixative solution.

Nasal (N)

I UNt

\""

V

\ UNbL J

Middle (M)p f • ~ r-

UMt

UMf

UMb

L 1

Temporal IT)

UTf /i

, --JUTb I

tarsal(t )

formxff )

buibaHb)

LNb

LNf

LNt

LMb

LMf

LMt

LTb

LTf

LTt

bulbar (b)

fornix <f)

tarsaHl)

>

Nasal (M) Middle(M) Temporal (T)

Lower (L)

Fig. 4. Conjunctival sampling sites for left eye. Notations foreach site are: UNt-upper nasal tarsal; UMt-upper middle tarsal;UTt-upper temporal tarsal; UNf-upper nasal fornix; UMf-uppermiddle fornix; UTf-upper temporal fornix; UNb-upper nasal bul-bar; UMb-upper middle bulbar; UTb-upper temporal bulbar;LNb-lower nasal bulbar; LMb-lower middle bulbar; LTb-lowertemporal bulbar; LNf-lower nasal fornix; LMf-lower middle fomix;LTf-lower temporal fornix; LNt-!ower nasal tarsal; LMt-lowermiddle tarsal; and LTt-lower temporal tarsal.

Samples were cut in half perpendicular to the con-junctival surface with randomly oriented planes ofincision. The hemisections were placed side-by-side,cut face down, embedded in glycol methacrylate, andsectioned at 2 nm thickness with a LKB Historangemicrotome (LKB Industries, Inc., Gaithersburg,


1928 INVESTIGATIVE OPHTHALMOLOGY b VISUAL SCIENCE / December 1987 Vol. 28

MD). Two serial sections (four specimens) were takenfrom each site. Sections were mounted on glass slidesand stained with periodic acid SchifF (PAS) and he-matoxylin reagents.

Goblet Cell Index

Goblet cell indices was determined by counting GCprofiles versus all epithelial cell profiles, ie, GCs plussquamous, polygonal, and basal cells. Goblet cellswere identified by the presence of PAS-positive intra-cellular material. Crescent-shaped nuclei associatedwith PAS-positive material were considered nuclei ofGCs and were not included in the remaining epithe-lial cell counts. The ratio of GC profiles to total epi-thelial cell profiles for each site was computed as anindex of GC density or goblet cell index (GCI).

A single linear section was selected from each sitefor epithelial cell quantitation. This selection wasbased on absence of technical artifacts. For each site,500 total epithelial cells were counted using the X40objective and a standard light microscope. Five fociof 100 cells/focus were selected along the length ofeach linear section in the following manner: Cells intwo focused high power fields, one at each end of thelinear specimen, were counted just inside left andright tissue margins. Cells in a third high power fieldwere counted at the center of the specimen and thefourth and fifth foci were located midway betweenthe center and each end of the section. Counting wasinitiated at the left-hand side of a high power fieldand proceeded from left to right until a total of 100epithelial cell profiles, including GCs, had beencounted. To validate this sampling scheme, the entirepopulation of cells was counted for some specimens(in addition to counting the five designated foci) andsimilar proportions were determined.

Because three sites (lower nasal fornix (LNf), lowermiddle fornix (LMf) and lower nasal tarsal (LNt) ofthe initial 12 eyes studied showed a difference be-tween male and female eyes, expanded studies wereperformed on these sites. Additional samples (LNf,n = 11; LMf, n = 12; and LNt, n = 19) were pro-cessed and GCIs were determined for these threesites. Some samples were not usable because of arti-facts and were not included in the study. Therefore,numbers were unequal for the three sites in the ex-panded study.

Statistical Analysis

Schirmer tear test (STT) and tear film breakuptime (BUT) data were analyzed statistically for differ-ences between left and right eyes using a two-tailedpaired t-test. Sex differences for STT and tear BUTwere tested with an independent two-sample t-test

after averaging measurements across left and righteyes. Sex differences in GCI for each site were testedwith one way analysis of variance (ANOVA). Multi-ple comparisons of mean GCIs for sites were madeusing ANOVA least squares differences (LSD). Theassumption of equal variance for the GCI valuesacross sites was tested with Levene's test.27

Results

Schirmer Tear Tests and Tear Film Breakup Times

The mean Schirmer tear test (STT) value was 23.42mm/min (SD 6.25 mm/min) for left eyes (OS) and24.29 mm/min (SD 5.58 mm/min) for right eyes(OD). Tear film breakup times (BUTs) were 19.38(SD 4.80 seconds) OS and 19.96 (SD 5.01 seconds)OD. No significant differences were found betweenright and left eyes for STT (0.5 > P > 0.4) or BUTs (P> 0.5). Differences were not significant between eyesof males and females for STT (P > 0.5) or for tearBUTs (P> 0.5).

Goblet Cell Index

Summary data for goblet cell indices (GCIs) for all18 sites from the initial 12 eyes studied (six males andsix females) are recorded in Table 1. Marked differ-ences in GCIs were noted among sites. Highest GCIswere found in two adjacent sites, LNf and LMf, in thelower fornix region of the conjunctiva (Table 1, Fig-ures 5 and 6). The GCI for LNt, a site on the nasalarea of the lower eyelid adjacent to LNf, was highestof the ten intermediate sites. Considering combinedmale and female data for the 12 sites with nonnegli-gible numbers of GCs, the two sites with the highestmean GCIs, LNf (0.299) and LMf (0.296), had theleast variability, ie, smallest standard deviations (LNf0.043; LMf 0.044).

For the 12 original eyes, sex differences in GCIs foreach site were tested statistically using one-wayANOVA for which no significant differences werefound for 15 of the 18 sites. Significant differenceswere discovered for the three highest GCI sites, LNf,LMf, and LNt (LNf P < 0.001; LMf P = 0.050; LNt P= 0.007). Goblet cell indices were similarly low ornegligible for all six bulbar sites for both males andfemales.

To investigate further possible sex differences inGCIs for the three highest sites, additional male andfemale samples were processed (see Materials andMethods), for LNf (n =11), LMf (n = 12), and LNt (n= 19) sites. Of the additional tissues studied, someanimals were represented by two eyes while otheranimals were represented by one eye. After increasingthe sample size, the mean GCIs for male and female


No. 12 CONJUNCTIVAL GODLET CELL DENSITY AND DISTRIBUTION / Moore er ol. 1929

samples at the LMf site (Table 2) were not differentstatistically (P = 0.930). Two sites, LNf and LNt,remained significantly different (LNf P = 0.001; LNtP = 0.040).

Considering all data, mean GCIs for each of the 18sites were compared using an ANOVA least-signifi-cant-difference (LSD) method of multiple compari-sons. Results of the LSD multiple comparisons (P< 0.05) with combined male and female data aregiven in Table 3. If an animal had both right and lefteyes represented, each eye was considered indepen-dently. Those sites included in the same group werestatistically indistinguishable. Sites with the highestGCIs, LNf (0.300) and LMf (0.290), were statisticallythe same and were designated as group A. The sixbulbar regions had few or essentially no GCs (GCI< 0.035) and formed a distinct group designated F.The remaining 10 sites comprised four intermediategroups (B, C, D, and E) with GCIs ranging from LNt(0.141) to LNt (0.242). For the 10 sites in the inter-mediate groups, considerable overlapping of meanGCIs was observed (Table 3). Because generally simi-lar patterns were noted for males and females ana-lyzed separately, pooling of data across sexes was jus-tified despite differences between sexes noted for twosites.

Because the bulbar sites had negligible numbers ofGCs, they were excluded from the analysis and theANOVA LSD multiple comparisons were recom-puted for the remaining 12 sites. Although someminor changes in the overlapping of groups were ob-

Table 1. Goblet cell indices (GCIs) for 18 sitesof initial 12 eyes studied

GCIs of CANINE CONJUNCTIVA

upper(u)

N M T

Site

LNfLMfLNtLTfUTfLMtUMfUNtUNfUMtUTtLTtUNbUMbUTbLNbLMbLTb

Combinedmale andfemale(n =

Mean

0.2990.2960.2290.2280.2280.2230.2140.2140.2080.1900.1780.1410.0330.0220.0160.0100.0040.001

12)

S.D.

0.0430.0440.0810.0700.0620.0650.0600.08!0.0820.0770.0720.0800.0410.0450.0210.0140.0110.002

Female(n-

Mean

0.3350.3200.2850.2360.2420.2370.2020.2440.2150.2120.1940.1660.0370.0340.0110.0120.0010.000

= 6)

S.D.

0.0150.0390.0410.0470.0490.0550.0700.0670.0880.0330.0850.0600.0480.0620.0120.0150.0010.001

Male(n-

Mean

0.2630.2710.1720.2210.2150.2100.2260.1830.2000.1670.1610.1150.0290.0100.0200.0080.0070.00!

~-6)

S.D.

0.0270.0370.0710.0920.0740.0760.0520.0880.0830.1030.0600.0930.0370.0150.0280.0130.0160.002

ioh GC O«f»»ityIt 0 290-0 3O0)

lnt*rm«diala GC Dantity(GCI. 0 141 0 242)

I—I Low GC DensityI I (GCIl 0 0 0 1 - 0 033)

lower ( I )

Fig. 5. Location of high, intermediate and low density GC areas(left eye). Note that the two sites with the highest goblet cell indices(GCIs) are adjacent in the lower conjunclival fornix (LNf and LMfsites). Goblet cells are essentially absent from all six bulbar sites.The remaining ten tarsal and fornix sites have intermediate GCdensities.

served, results were similar to those shown in Table 3in that the highest GCI sites (LNf and LMf) remainedstatistically indistinguishable. The third highest GCIsite (LNt) was no longer significantly different fromthe second highest GCI site (LMf).

Discussion

In canine and human conjunctiva GCs are absent,or present in extremely low numbers, on the perilim-bal bulbar surface and are present in highest numbersin the lower nasal fornix, lower middle fornix, andlower nasal palpebral sites. Although little is knownabout factors that influence normal conjunctival GCdensity and distribution, the degree of conjunctivalhydration has been proposed as a significant exoge-nous factor.26 We concur that surface hydration is animportant determinant in GC location and density.However, contrary to Kessing's idea of surface hy-


1930 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / December 1987 Vol. 28(G

CI)

X9TJC

1

gobl

e

0.310.3

0.290.280.270.260.250.240.230.220.210.2

0.190.180.170.160.150.140.130.120.11

Means and Standard Errors for GCIs(12 sample sites)

Fig. 6. Goblet cell indices(GCIs) for 12 conjunctiva!sites. Note that the highestGCIs and smallest standarderrors (SEs) occur for LNfand LMf sites. The remain-ing intermediate densitysites have greater variabilityin mean GCIs as noted bygreater SEs (ventricle bars).The lines connecting themean values identify sitesfor each horizontal area, ie,tarsal or fornix, from upperand lower conjunctivalhalves. Mean GCIs for bul-bar sites, which were negli-gible, are not shown.

UNt UMt LTTt UNf UMf UTf LNf LMf LTf LNt LMt LTt( 1 2 ) ( T n ( 1 2 ) U 2 > ( U ) U 2 > < " > < 2 « > < " > < 3 1 > < " > < " >

sample site

dration, we believe that the bulbar conjunctiva ispredisposed to exposure and drying and, therefore,represents a relatively poorly hydrated conjunctivalarea. Furthermore, we contend that gravitation ofaqueous tear into the lower conjunctival sac, forma-tion of the lacrimal lake, and accumulation of tear atthe medial canthus results in maximal hydration ofthe lower nasal fornix, lower nasal palpebral con-junctiva, and plica semilunaris (in man). Therefore,we propose that a direct relationship exists betweenthe degree of surface hydration and the number ofconjunctival GCs.

Such a direct relationship is supported by Ralph'sfindings that the mean GC count among patientswith keratitis sicca was significantly less than similarcounts performed on normal subjects.28 Also sup-porting the notion that a high degree of tear fluidhydration is essential for GC health is evidence thatnormal human tears directly stimulate secretorycells.29 The relationship between aqueous tear fluidand the propagation and longevity of conjunctivalGCs merits further investigation.

Although a previous topographical study has pro-vided clinically useful information,26 the methodused was not directly applicable to the clinical studyof spontaneous conjunctival disease. We have devel-oped a new method, which is less disruptive to theconjunctival mucosa, for harvesting and preparingthe conjunctiva in toto. This method allows collec-tion Of samples from multiple sites for microscopicalstudies of conjunctiva. Because 18 different samplingsites may be consistently identified using this newmethod, predictable sampling of multiple sites fromone eye or sampling of comparable sites from differ-ent eyes is possible.

Because of the elasticity of conjunctiva and thenonlinearity of processed tissue, we believe propor-tion indexing methods provide a more consistentmeans of quantitating GC profiles histologically thanlinear counting methods. Because of the inherentvariability in the number of epithelial layers presentat various sites within the conjunctiva, we furtherbelieve that expressing the ratio of GC profiles/100total epithelial cell profiles more accurately reflects

Table 2. Combined summary data for initial and expanded GCIs for sites LNf, LMf, and LNt

Site

LNfLMfLNt

No.

232431

Combined male and female

Mean

0.3000.2900.242

S.D.

0.0380.0470.074

No.

121219

Female

Mean

0.3220.2900.263

S.D.

0.0260.0520.062

No.

111212

Male

Mean

0.2760.2910.207

S.D.

0.0340.0420.082


No. 12 CONJUNCTIVA!. GODLET CELL DENSITY AND DISTRIBUTION / Moore er ol. 1931

actual volumetric relationships than alternative in-dexing methods, such as GCs/100 basal epithelialcells.

Since, in the present study, the two areas withhighest GCIs (LNf and LMf) were also characterizedby the least variability in GC numbers, these are rec-ommended as the most appropriate biopsy sites forquantitating canine conjunctival GCs in experimen-tally-induced or spontaneously-occurring ocular sur-face diseases. In the live animal, the lower conjuncti-val fornix is easily exposed and readily accessible forsampling. Furthermore, the elasticity of fornicealconjunctiva allows the tissue to be easily grasped andelevated for incisional biopsy.

Because this study has established that extremelylow GCIs occur at all bulbar areas in normal canineeyes, cellulose acetate (CA) impression methods ap-plied to bulbar conjunctiva would seem to be inap-propriate for quantitating GCs in diseased canineeyes. Furthermore, since it is technically difficult toinsert CA strips onto fornix or palpebral conjunctivain unanesthetized animals, the CA method wouldappear to be a less advisable sampling procedure forGC quantitation in the dog than histological exami-nation of conjunctiva.

Tear film breakup time (BUT), another noninva-sive technique, has been regarded as a clinically use-ful procedure for assessing the stability of the tear filmin man."'30 In the present study tear BUTs are re-ported for the first time in normal dogs. In this study,a combination of tranquilizer and dissociative anes-thetic agent was administered prior to performingBUTs. Anesthetic and preanesthetic drugs have beenshown to reduce aqueous tear secretions and could,thereby, influence BUT. In addition, this protocolmay be undesirable in animals with spontaneous dis-ease. Although tear BUT may be a useful screeningprocedure, morphological study of conjunctivalbiopsy specimens with GC quantitation appears to bea more definitive method for evaluating productionof preocular gel mucin in the dog.

Although the influence of sex on GC numbers isnot totally clear from this study, a significant sex ef-fect was noted at two sites. Based on recent reports ofendocrine influences on lacrimal tissues and con-junctiva, it is not surprising that differences in GCIsbetween sexes might be seen.31-32 Since the sites whichshowed sex differences were among sites with highestGC densities, it is possible that greater variabilitywithin intermediate GCI sites accounted for theirlack of significant sex differences. Additional studiesare necessary to further clarify the influence of sex onconjunctival GC densities.

The present study has demonstrated quantitativelya heterogeneous distribution of conjunctival GCs in

Table 3. Results of analysis of variance (ANOVA)multiple comparison least significant difference(LSD) for all sites*

LSD grouping(P<

AA

BB CB CB CB CB CB C

C

0.05)

DDDDDD E

E

FFFFFF

Mean

0.3000.2900.2420.2280.2280.2230.2140.2140.2080.1900.1780.1410.0330.0220.0160.0100.0040.001

N

2324

31121212121212121212

121212121212

Site

LNfLMf

LNtLTfUTfLMtUMfUNtUNfUMtUTtLTt

UNbUMbUTbLNbLMb

* Mean GCIs for initial and expanded data for all 18 sites were used inthese statistical comparisons.

young adult beagle dogs. The highest densities of GCwere noted in the lower nasal fornix and adjacentsites (lower nasal tarsal and lower middle fornix) and,with respect to relative densities of GCs, canine con-junctiva closely resembles human conjunctiva. In thedog, intermediate densities of conjunctival GCs werefound in the remaining fornix and tarsal areas andthe prelimbal bulbar conjunctiva contained essen-tially no GCs. The spontaneous occurrence of caninetear deficient disease and the similarities which existbetween canine and human conjunctivae support theuse of the dog as a model species for studying theconjunctival mucous system.

Key words: canine conjunctival GC, goblet cell index,preocular mucin, conjunctival biopsy sites, goblet cell den-sity and distribution

Acknowledgments

The authors thank Ms. Linda Ostrander for her experttechnical assistance and Ms. Gail Ribble for preparing theillustrations.

References

1. Milder B: The lacrimal apparatus. In Adler's Physiology of theEye, Moses RA, editor. St. Louis, C.V. Mosby, 1981, pp.16-26.

2. Holly FJ: Tear film physiology. Am J Optom Physiol Optics57:252, 1980.

3. Moore JC and Tiffany JM: Human ocular mucus: Origins andpreliminary characteristics. Exp Eye Res 33:203, 1981.

4. Nichols BA, Chiapinoo ML, and Dawson CR: Demonstration


1932 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / December 1987 Vol. 28

of the mucous layer of the tear film by electron microscopy.Invest Ophthalmol Vis Sci 26:264, 1985.

5. Nichols B, Dawson CR, and Togni B: Surface features of theconjunctiva and cornea. Invest Ophthalmol Vis Sci 24:570,1983.

6. Holly FJ and Lemp MA: Wettability and wetting of cornealepithelium. Exp Eye Res 11:239, 1971.

7. Lemp MA, Holly FJ, Iwata S, and Dohlman CH: The precor-neal tear film: I. Factors in spreading and maintaining a con-tinuous tear film over the corneal surface. Arch Ophthalmol83:89, 1970.

8. Lemp MA, Dohlman CH, Kuwabara T, Holly FJ, and CarrollJM: Dry eye secondary to mucus deficiency. Trans Am Oph-thalmol Otolaryngol 75:1223, 1971.

9. Lemp MA: Surfacing abnormalities. Int Ophthalmol Clin13:191, 1973.

10. Moore CP and Collier LL: Mucin deficient external eye diseasein dogs. ACVO Proc, San Francisco, California, American Col-lege of Veterinary Ophthalmologists, 1985, pp. 152-167.

11. Lemp MA: Breakup of the tear film. Int Ophthalmol Clin13:97, 1973.

12. Dohlman CH, Friend J, Kalevar V, Yagoda D, and Balazs EA:The glycoprotein (mucus) content of tears from normals anddry eye patients. Exp Eye Res 22:359, 1976.

13. Egbert PR, Lauber S, and Maurice DM: A simple conjunctivalbiopsy. Am J Ophthalmol 84:798, 1977.

14. Lemp MA and Hamill JR: Factors affecting tear film breakupin normal eyes. Arch Ophthalmol 89:103, 1973.

15. Vanley GT, Irving LH, and Gregg TH: Interpretation of tearfilm breakup. Arch Ophthalmol 95:445, 1977.

16. Kiorpes TC, Friend J, and Thoft RA: Measurement of ocularsurface mucus after mobilization with N-acetylcysteine.ARVO Abstracts. Invest Ophthalmol Vis Sci 18(Suppl):142,1979.

17. KJnoshita S, Kiorpes TC, Friend J, and Thoft RA: Goblet celldensity in ocular surface disease: A better indicator than tearmucin. Arch Ophthalmol 101:1284, 1983.

18. Friend J, Kiorpes T, and Thoft RA: Conjunctival goblet cellfrequency after alkali injury is not accurately reflected by

aqueous tear mucin content. Invest Ophthalmol Vis Sci24:612, 1983.

19. Greiner JV, Herriques AS, Covington HI, Weidman TA, andAllansmith MR: Goblet cells of the human conjunctiva. ArchOphthalmol 99:2190, 1981.

20. Nelson JD, Havener VR, and Cameron JD: Cellulose acetateimpressions of the ocular surface: Dry eye states. Arch Oph-thalmol 101:1869, 1983.

21. Nelson JD and Wright JC: Conjunctival goblet cell densities inocular surface disease. Arch Ophthalmol 102:1049, 1984.

22. Riis CR: The Normal Canine Conjunctiva. MS Thesis, CornellUniversity, Ithaca, New York, 1976, pp. 45-78.

23. Latkovic S and Nilsson SEG: The ultrastructure of the normalconjunctival epithelium of the guinea pig: I-IV. Acta Oph-thalmologica 57:106,123,305,321, 1979.

24. Rohen JW and Steuhl P: Specialized cell types and their re-gional distribution in the conjunctival epithelium of the cyno-molgus monkey. Graefes Arch Clin Ophthalmol 218:59, 1982.

25. Weyrauch KD: The conjunctival epithelium in the domesticruminant: I. Light microscopic investigations. Zeitschrift furMikroskopisch-Anatomische Forschung 97:565, 1983.

26. Kessing SV: Mucous gland system of the conjunctiva: Aquantitative normal anatomic study. Acta Ophthalmol95(Suppl):9, 1968.

27. Snedecor GW and Cochran WC: Levene's test of homogeneityof variance. In Statistical Methods, 7th edition. Ames, Iowa,The Iowa State University Press, 1980, pp. 253-254.

28. Ralph RA: Conjunctival goblet cell density in normal subjectsand in dry eye syndromes. Invest Ophthalmol 14:299, 1975.

29. Franklin M and Band BG: Mucus stimulating factor in tears.Invest Ophthalmol Vis Sci 19:430, 1980.

30. Lamberts DW: Dry eyes. In The Cornea, Smolin G and ThoftRA, editors. Boston, Little Brown, 1983, pp. 293-314.

31. Jacobs M, Buxton D, Dramer P, Lubkin V, Dunn M, Herp A,Weinstein B, Southern AL, and Perry H: The effect of oopho-rectomy on the rabbit lacrimal system. ARVO Abstracts. In-vest Ophthalmol Vis Sci 27(Suppl):25, 1986.

32. Sullivan DA and Allansmith MR: Hormonal modulation oftear volume in the rat. Exp Eye Res 42:131, 1986.


density and distribution of canine conjunctival goblet cells

Documents