`Conjunctival Epithelia
`
`Tsutomu Inatomi,* Sandra Spurr-Michaud,* Ann S. Tisdale,* Qian Zhan*
`Sandy T. Feldman,\ and Ilene K. Gipson*
`
`Purpose. To determine whether human conjunctival epithelium expresses any of the human
`mucin genes designated MUC2 through MUC7.
`Method. Northern blot analysis was performed using total RNA isolated from surgically re-
`moved conjunctival tissues. Complementary DNA or oligonucleotides to the tandem repeat
`region of each mucin gene were labeled and hybridized to conjunctival RNA. In situ hybridiza-
`tion also was performed to determine the distribution of mucin mRNA.
`
`Results. Only MUC4 and MUC5 probes hybridized to conjunctival RNA by Northern blot
`analysis. Both probes bound in a polydispersed pattern, which is characteristic of mucin genes.
`Using in situ hybridization, MUC4 mRNA was detected in the cells of the stratified conjunctival
`epithelium, whereas MUC5 mRNA expression was limited to goblet cells. MUC4 or MUC5
`probes did not hybridize to sections of corneal epithelium.
`
`Conclusions. The mucins MUC4 and MUC5 are expressed by the human conjunctiva. These
`mucins may play an important role in forming the tear-film layer at the air and ocular surface
`epithelium interface. Invest Ophthalmol Vis Sci. 1996;37:1684-1692.
`
`1 he surfaces of the corneal and conjunctival epithe-
`lia are covered by tear film, which has a variety of
`components, including mucus. Mucus constitutes the
`innermost layer of the tear film and interfaces with
`the apical surface of the epithelium. 1 This relationship
`of mucus to the apical surface of epithelium is present
`along all other mucosal surfaces, including those of
`the gastrointestinal, tracheobronchial, and reproduc-
`tive tracts. The most prominent components of the
`mucus layer are mucins. Mucins are glycoproteins in
`which at least 50% of mass is O-linked carbohydrate. 2
`Mucins are notoriously difficult glycoproteins to study;
`their high carbohydrate content impedes many con-
`ventional assays.
`Although the mucins are essential for the integrity
`of the tear film, in which they form a viscoelastic gel, 3
`the characteristics of mucins on the ocular surface
`are not well known. Several studies 4"7 have detailed
`
`From *Schepens Eye Research Institute and the Department of Ophthalmology,
`Harvard Medical School, Boston; and the "[Department of Ophthalmology,
`University of California, San Diego.
`Supported by National Institutes of Health research grant R37-EY03306 (IKG) and
`l/y The Massachusetts Lions Eye Research Fund.
`Submitted for publication October 18, 1995; revised February I, 1996; accepted
`March 7, 1996.
`Proprietary interest catebtny: N.
`Reprint requests: Ilene K. Cipson, Schepeiu Eye Research Institute, 20 Staniford
`Street, Boston, MA, 02114.
`
`1684
`
`approximate molecular weight, carbohydrate content,
`and amino acid composition of glycoproteins and
`plasma proteins of ocular mucus. These studies have
`provided the initial data suggesting the presence of a
`heterogeneous mixture of glycoproteins in the tear
`film.
`Generally, the mucins on the ocular surface have
`been considered to be derived from the goblet cells
`in the conjunctiva. Histologically, vesicles packed with
`periodic acid-Schiff or Alcian blue staining material
`can be detected inside the conjunctival goblet cells. 8
`These staining data suggest the presence of acidic and
`neutral carbohydrates in mucins. It has been proposed
`that the stratified squamous cells of the conjunctiva
`provide a second source of mucus for the tear film.' 110
`These cells were shown to contain vesicles that bind
`dyes with affinity for carbohydrate, and, by electron
`microscopy, the vesicles appeared to be emptying into
`the tear film. In support of that hypothesis, the con-
`junctival apical cell vesicles, as well as vesicles in apical
`cells of corneal epithelium, bind a monoclonal anti-
`body (H185) to mucin-like glycoproteins." Thus, con-
`junctival and corneal stratified squamous epithelia
`may provide a source of mucins for the ocular surface.
`A major advance in understanding mucin struc-
`ture, function, and heterogeneous character has come
`
`Investigative Ophthalmology & Visual Science, July 1996, Vol. 37, No. 8
`Copyright © Association for Research in Vision and Ophthalmology
`
`APOTEX 1021, pg. 1
`
`
`
`1685
`
`Chromosomal
`Mapping
`
`Amino Acids in
`Tandem Repeat
`
`References
`
`20
`23
`17
`16
`8
`29
`169
`23
`13/41
`
`15, 16
`18, 19
`24
`25
`19, 20
`21
`27
`23
`26
`
`Iq21q24
`Ilpl5
`
`73I
`
`lpl5
`Ilpl5
`Ilpl5
`. 4
`12
`
`Mucin Gene Expression in Conjunctiva
`
`TABLE l. Human Mucin Genes
`
`Designation
`
`Type if Sequence Verified
`
`cDNA Clone Source
`
`MUC1
`MUC2
`MUC3
`MUC4
`MUC5AC
`MUC5B
`MUC6
`MUC7
`MUC8
`
`Membrane-spanning
`Gel-forming-secretory
`
`Gel-forming-secretory
`
`Soluble monomer-secretory
`
`Mammary/pancreatic tumor
`Intestine
`Intestine
`Trachea
`Trachea
`Trachea
`Stomach
`Salivary gland
`Trachea
`
`from the recent cloning of mucin genes.2'12 M To date,
`nine human mucin genes have been cloned; they have
`been designated MUC1 through MUC8 (two MUC5
`genes have been described), in order of their discov-
`ery (Table 1). Besides the fact that the biochemical
`property of a high percentage of their mass (>50%)
`is O-linked oligosaccharides, each mucin gene pro-
`vides a unique amino acid sequence that is repeated
`tandemly in the protein backbone. The number and
`sequence of amino acids of each mucin tandem repeat
`unit is different for each mucin gene (Table I).1'1 The
`only common feature in the repetitive sequences are
`that they contain a high percentage of serine and/or
`threonine residues,1'1 which provide the O-glycosyla-
`tion sites on the protein. The number of amino acids
`in the tandem repeats of the eight mucin genes cloned
`to date varies from 8 to 169,M but the number of
`tandem repeats per mucin also can vary in individuals
`as a result of genetic polymorphism.1'
`Functionally, mucins have been subdivided into
`two types, transmembrane mucins and secretory mu-
`cins.2 Only one transmembrane mucin has been
`cloned. This mucin, MUC1, is found on the surfaces
`of various epithelial cells and carcinomas.15"' Re-
`cently, we reported17 that MUC1 mucin is expressed
`by corneal and conjunctival epithelial cells but not by
`goblet cells. We hypothesize that this transmembrane
`mucin plays a role in tear film spread and association
`of the mucus gel on the epithelial surface. Its anti-
`adhesion property may prevent adhesion of foreign
`debris to the ocular surface.
`Of the remaining mucins, MUC2 to MUC8, the
`best characterized are MUC2, MUC5, and MUC7.
`Complete cDNA sequence is available for MUC2 and
`MUC7. The MUC2 cDNA contains two tandem repeat
`domains and four cysteine-rich D domains that are
`homologous to von Willebrand factor.18 l9 These cys-
`teine-rich D domains appear to be necessary for the
`disulfide linking of mucin molecules, which leads to
`polymerization into gels. Initially, partial sequences of
`a number of MUC5 cDNAs were cloned.2021 They were
`placed into three groups, MUC5A, MUC5B, and
`MUC5C. All localize to chromosome Ilpl5. Recent
`
`data suggest that MUC5A and MUC5C are at precisely
`the same gene locus.22 Thus, MUC5A and MUC5C are
`part of the same gene and have been designated
`MUC5 or MUC5AC. MUC5B appears to be distinct
`from MUC5AC. Although the full length of MUC5
`cDNA is unknown, MUC5 has cysteine-rich regions
`with a high degree of homology to MUC2 cysteine-
`rich domains. The presence of this conserved domain
`of MUC5 suggests that MUC5, like MUC2, may be a
`gel-forming mucin. The second secretory mucin for
`which full-length cDNA sequence is known was cloned
`from the salivary gland and designated MUC7.2H In
`contrast to other larger secretory mucins, the full-
`length cDNA of MUC7 is 2350 bp. This small mucin,
`which exists as a monomer, lacks the cysteine-rich do-
`main found in MUC2 and MUC5.2:'
`The other mucin genes MUC3, MUC4, MUC6,
`and MUC8 are not as well characterized.1'1 MUC3 was
`cloned from a small intestine cDNA library and is ex-
`pressed in small intestine and colon.2'1 MUC4 and
`MUC8 were cloned from a tracheobronchial cDNA
`library; however, only tandem repeat sequences are
`available.2'21' Even though various tissues express
`MUC4, this mucin
`is relatively uncharacterized.
`MUC6, cloned from a stomach cDNA library, has the
`longest tandem repeat sequence reported.27
`It is unknown whether ocular surface epithelium
`expresses any of the cloned, so-called secretory mu-
`cins. To understand better the characteristics of the
`ocular surface mucins, we determined the expression
`of MUC2 through MUC7 by Northern blot analysis of
`conjunctival RNA and then, by in situ hybridization,
`determined the cellular origin of the expressed mu-
`cins.
`
`MATERIALS AND METHODS
`
`Tissue Samples
`
`All investigations followed the tenets of the Declara-
`tion of Helsinki, and informed consent and full institu-
`tional review board approval were obtained. Small seg-
`ments (approximately 1 x2 mm) of forniceal con-
`
`APOTEX 1021, pg. 2
`
`
`
`1686
`
`Investigative Ophthalmology & Visual Science, July 1996, Vol. 37, No. 8
`
`TABLE 2. Mucin cDNA and Oligonucleotide Probes Corresponding
`to Tandem Repeat Region
`
`Mucin Gene
`
`cDNA/oligo
`
`Designations
`
`Length (bp)
`
`References
`
`MUC2
`
`MUC3
`MUC4
`MUC5
`MUC6
`MUC7
`
`cDNA
`cDNA
`cDNA
`oligo
`cDNA
`cDNA
`oligo
`
`SMUC41
`HAM-1
`SIB124
`MUC4oligo
`4F
`MUC6.2T
`MUC7oligo
`
`836
`90
`387
`48
`494
`1014
`45
`
`bp = base pair; oligo = oligonucleotide; cDNA = complementary DNA.
`
`18
`32
`24
`25
`31
`27
`23
`
`junctiva were excised at the time of cataract surgery.
`All conjunctiva appeared normal at the time of preop-
`erative examination. None of the patients from whom
`samples were obtained were receiving chronic topical
`eye medications. Tissues for Northern blot analysis
`were frozen in liquid nitrogen immediately after re-
`moval and stored at — 70°C until RNA isolation. Four
`pieces of conjunctiva (three women and one man, 56
`to 77 years of age) were used to obtain sufficient RNA
`for analysis. Conjunctival tissues for in situ hybridiza-
`tion were fixed immediately after removal and pro-
`cessed as described below. Six samples were obtained;
`they included tissues from three women (70 to 75
`years of age) and three men (21, 67, and 70 years
`of age). Corneas with adjacent limbus and adjacent
`conjunctiva used for in situ hybridization were ob-
`tained from eye banks and were fixed less than 24
`hours after death.
`
`Isolation of RNA and Northern Blot Analysis
`Total RNA was isolated with an RNA Isolation Kit
`(Stratagene, La Jolla, CA) using an acid guanidinium
`thiocyanate phenol-chloroform
`single-step extrac-
`tion.28 Conjunctival RNA, isolated from two biopsies
`of forniceal conjunctiva, were pooled
`to obtain
`enough for one lane on each Northern blot. The pool-
`ing of several samples decreases the possibility of a
`false-negative result caused by tissue variation. Total
`RNA was isolated from cultured human corneal epi-
`thelium as described previously. 17 Total RNA from hu-
`man trachea, stomach, salivary gland, and small intes-
`tine, purchased from Clontech (Palo Alto, CA), were
`used for positive controls; a BT-20 breast carcinoma
`cell RNA, cultured rabbit corneal fibroblast RNA, and
`human umbilical vein endothelial cell RNA were used
`as negative controls.
`RNA samples of 10 (xg were separated on 1% aga-
`rose gels in the presence of formaldehyde. The integ-
`rity and amount of RNA loaded on each gel was deter-
`mined by staining with ethidium bromide. The RNA
`was then transferred to Gene Screen Plus membrane
`(DuPont, Boston, MA).
`The membranes were prehybridized in a solution
`
`of 50% formamide, 5 X saline-sodium phosphate-
`EDTA buffer, 5 X Denhardt's solution, 1% sodium
`dodecyl sulfate (SDS), 10% dextran sulfate, and 100
`//g/ml denatured salmon sperm DNA for 2 hours at
`42°C. Hybridization was performed in the same hy-
`bridization solution with 2 X 106 cpm/ml of 32P-la-
`beled probe at 42°C overnight. After hybridization,
`the membranes that were hybridized with cDNA
`probes were washed using high-stringency conditions;
`two washes with 2 X saline-sodium citrate (SSC) buffer
`for 15 minutes at room temperature, two washes with
`2 X SSC, 1% SDS at 55°C for 20 minutes, followed by
`two washes with 0.1 X SSC, 0.1% SDS at 55°C for 20
`minutes. For the oligonucleotide probes, high-strin-
`gency washes were decreased to 1 X SSC, 0.1% SDS
`at 42°C. After autoradiography, each membrane was
`stripped and reprobed three times.
`
`cDNA and Probe Preparation
`Because tandem repeat sequences are specific for
`each mucin gene 14 and probes corresponding to this
`region may hybridize to a number of the repeats of
`these sequences per mRNA, we chose to use cDNAs
`or oligonucleotides corresponding to tandem re-
`peat regions as probes for Northern blot analysis
`and in situ hybridization (Table 2). The sequence
`of cDNAs for MUC2, MUC3, and MUC5 were de-
`scribed in the references cited in Table 2. MUC6.2T
`cDNA clone contains two repeats of MUC6 tandem
`repeat sequence. Because cDNA probes for MUC4
`and MUC7 were unavailable, the most frequently
`found repetitive sequences of each tandem repeat
`were selected for oligonucleotide probes. MUC4
`oligo corresponds to antisense sequence of the 48-
`bp fragment of the 69-bp tandem repeat (5'-GTC-
`GGTGACAGGAAGAGGGGTGGCGTGACCTGT -
`GGATGCTGAGGAAGT-3'), and MUC7 oligo corre-
`sponds to the antisense sequence of a 45-bp tandem
`repeat ( 5' - GGTGTGGGTGGGGCAGCTGTGGTC-
`TCTGGTGGAGCTGAGGAAGAT-3'). These oligo-
`nucleotides, purified by reverse-phase chromatogra-
`phy, were purchased from BioServe (Laurel, MD)
`and labeled with y- 32P adenosine triphosphate by
`
`APOTEX 1021, pg. 3
`
`
`
`Mucin Gene Expression in Conjunctiva
`
`1687
`
`A
`
`12
`
`3
`
`4
`
`5 6 B 12
`
`3 4
`
`5 6
`
`12
`
`3 4 5 6
`
`5.5-
`
`2.8-
`
`1.9-
`1.G-
`
`2.8-
`
`1.9-
`1.6-
`
`7.4-
`5.5-
`
`2.8-
`
`1.9-
`1.6-
`
`7.4-
`
`1.9-
`1.6-
`
`2
`
`3
`
`5 6
`
`12
`
`3 4
`
`7.4-
`
`2.8-
`
`1.9-
`1,6-
`
`12
`
`3 4
`
`5
`
`28S
`
`18S
`
`H
`
`28S
`
`18S
`
`FIGUREI. Northern blot anal-
`ysis demonstrating tissue-spe-
`cific expression of mucin
`mRNA in conjunctiva and
`other tissues. RNA samples of
`10 fxg were separated on 1%
`agarose gels in the presence
`of formaldehyde. Two mem-
`branes, blots 1 and 2, were
`prepared, respectively. Blot 1
`was hybridized serially with
`probes for MUC2 (A), MUC4
`(C), and MUC6 (E), and blot
`2 was hybridized with probes
`for MUC3 (B), MUC5 (D),
`and MUC7
`(F). Ethidium
`bromide staining of each gel
`show the integrity of RNA
`blot 1 (G) and blot 2 (H).
`Note that probes for only
`MUC4 and MUC5 bound to
`the conjunctival RNA. RNA
`ladders were used as size
`markers: lane 1 = conjunc-
`tiva; lane 2 = stomach; lane
`3 = small intestine; lane 4 =
`trachea;
`lane 5 = salivary
`gland; lane 6 = human um-
`bilical vein endothelial cell
`HUVEC.
`
`T4 polynucleotide kinase (BioLabs, Richmond, CA).
`cDNA probes were labeled with a-:^P deoxycytidine
`triphosphate by Random Primer Plus kit (DuPont).
`Labeled probes were purified using a Sephadex-G25
`or Sephadex-G50 column (Boehringer-Mannheim,
`Indianapolis, IN) before Northern blot analysis.
`
`In Situ Hybridization
`Because MUC4 and MUC5 mRNA were detected in
`conjunctival RNA by Northern blot analysis, their dis-
`tribution was determined by in situ hybridization. Be-
`cause recent ARVO abstracts*1'™ suggest that MUC2
`is present at the ocular surface and because a faint
`band was detected for MUC7 by Northern blot analysis
`with extended exposure, in situ hybridization of
`MUC2 and MUC7 mRNA also was performed. Tech-
`
`niques were as previously described.17 Oligonucleo-
`tide probes to MUC4 and MUC7 were labeled with 3r>S-
`dATP by terminal deoxynucleotidyl transferase (Gibco
`BRL, Gaithersburg, MD). Antisense and sense ribo-
`probes, generated from the MUC5 cDNA (designated
`F4)3' and from the MUC2 cDNA (designated HAM-
`1),V2 were labeled with 35S-uridine triphosphate by T3
`or T7 RNA polymerase (Boehringer Mannheim, India-
`napolis, IN).
`Tissues prepared by two different methods were
`available. In one, frozen conjunctival sections were
`mounted on gelatin-coated slides and fixed with 4%
`paraformaldehyde. In the second, sections from 4%
`paraformaldehyde-nxed, paraffin-embedded
`tissues
`were mounted on gelatin-coated slides. Proteinase K
`treatment and acetylation were performed before hy-
`
`APOTEX 1021, pg. 4
`
`
`
`1688
`
`Investigative Ophthalmology & Visual Science, July 1996, Vol. 37, No. 8
`
`A
`
`12
`
`3 4 5
`
`1
`
`2
`
`3
`
`45
`
`
`
`u
`
`1
`
`2
`
`3 4
`
`5
`
`28S-
`
`18S
`
`7.4
`
`5.5
`
`2.8
`
`1.9
`1.6
`
`7.4-
`
`5.5-
`
`2.8-
`
`1.9-
`1.6-
`
`FIGURE 2. Northern blot anal-
`ysis demonstrated lack of ex-
`pression of MUC4 (A) and
`MUC5 (B) mRNA in cul-
`tured human corneal epithe-
`lium. RNA samples of 10 jig
`were separated on a 1% aga-
`rose gel in the presence of
`formaldehyde. Ethidium bro-
`mide staining shows the in-
`tegrity of RNA (C). Lane 1 =
`cultured human corneal epi-
`thelium; lane 2 = trachea;
`lane 3 = small intestine; lane
`4 = BT-20 breast cell carci-
`noma; lane 5 = cultured rab-
`bit corneal fibroblast.
`
`bridization. Hybridization was carried out in a solution
`of 50% formamide, 10% dextran sulfate, 0.3 M NaCl,
`1 X Denhardt's solution, 10 mM Tris, pH 8, 5 mM
`EDTA, 0.5 mg/ml tRNA, and 10 mM dithiothreitol.
`Sections were hybridized at 42°C for MUC4 and MUC7
`oligoprobes and at 57°C for MUC 2 and MUC5 ribo-
`probes. After washing, autoradiography was per-
`formed, and the sections were counterstained with
`hematoxylin and eosin. To verify the cellular distribu-
`tion of MUC4 mRNA, colorimetric, nonradioactive in
`situ hybridization was performed. MUC4 oligoprobes
`were labeled with digoxigenin (DIG)-dUTP (Boeh-
`ringer Mannheim) by
`terminal deoxynucleotidyl
`transferase. Sections were pretreated according to the
`protocol used with the 35S probe. MUC4 hybridization
`was performed at 37°C in a solution containing 50%
`formamide, 10% dextran sulfate, 20 //g/ml tRNA.
`After hybridization, washes were performed for 30
`minutes serially in each of the following: 2 X SSC at
`room temperature, 1 X SSC at room temperature, 0.5
`X SSC at 37°C, and 0.5 X SSC at room temperature.
`MUC4-hybridized probes were detected by DIG detec-
`tion kit (Boehringer Mannheim) according to the
`manufacturer's instruction.
`
`RESULTS
`
`Northern Blot Analysis
`To examine the expression of secretory mucin genes
`MUC2 through MUC7, Northern blot analysis (Fig. 1)
`was performed with probes to each MUC tandem re-
`peat region (Table 2). MUC2 and MUC3 mRNA were
`detected only in small intestine RNA and not in other
`RNA samples (Figs. 1A, IB). Positive signals of MUC2,
`MUC3, MUC4, MUC5, and MUC6 showed the typical
`smeared, polydispersed pattern previously reported as
`characteristic of these mucin genes. 18'24'25'27
`
`MUC4 mRNA expression was observed in con-
`junctiva as well as in trachea; the latter was used as
`positive control (Fig. 1C). Although the binding to
`conjunctiva is less intense than to trachea, both signals
`in conjunctiva and trachea show the typical polydis-
`persed pattern. In the MUC5 Northern blot (Fig. ID),
`an intense signal was detected in the positive controls
`of trachea and stomach, similar to that reported by
`Toribara et al. 27 The same polydispersed intense bind-
`ing was detected in conjunctiva (Fig. ID, lane 1).
`MUC6 mRNA expression was detected only in
`stomach RNA, as reported by Toribara et al.27 No bind-
`ing was observed in the conjunctival sample. Intense
`MUC7 mRNA signal, approximately 2.4 kbp, was de-
`tected in the salivary gland only (Fig. IF). A faint band
`of the same size appeared in conjunctiva only after an
`extended 4-day exposure of the autoradiogram (data
`not shown). The integrity and amount of RNA loaded
`on each gel was verified by staining with ethidium
`bromide, as shown in Figures 1G and 1H.
`To determine whether MUC4 or MUC5 mRNA ex-
`pression was limited to the conjunctival region of the
`ocular surface epidtelia, Northern blot analysis using
`total RNA from cultured human corneal epithelium,
`known to express MUC1,17 was performed. MUC4 and
`MUC5 probes bound to positive control RNA from tra-
`chea (Fig. 2, lanes A2 and B2) in the same fashion as
`shown in Figure 1C, lane 4, and Figure ID, lane 4. Figure
`2, lane A3, shows weak binding of MUC4 probe to small
`intestine RNA as previously reported14 and as shown in
`Figure 1C, lane 3. Neither MUC4 nor MUC5 mRNA was
`detected in cultured corneal epithelial RNA (Fig. 2). To
`verify these results, in situ hybridization was performed
`on corneal tissues (see next section).
`In Situ Hybridization
`The stratified conjunctival epithelium showed positive
`binding of the 35S-labeled MUC4 oligoprobe (Figs. 3A,
`
`APOTEX 1021, pg. 5
`
`
`
`Mucin Gene Expression in Conjunctiva
`
`1689
`
`FIGURE 3. MUC4 mRNA dis-
`tribution in conjunctiva us-
`ing in situ hybridization with
`HlS-labeled MUC4 antisense
`oligoprobe: (A) dark field;
`(B) bright
`field. MUC4
`mRNA distribution
`in con-
`junctiva using in situ hybrid-
`ization with MS-labeled MUC4
`sense oligoprobe: (C) dark
`field; (D) bright field. The
`conjunctival epithelium shows
`hybridization with
`the anti-
`sense probe. Goblet cells do
`not appear to be labeled pref-
`erentially {arrows, AJB). Digox-
`igenin-labeled probe (E, anti-
`sense probe; F, sense probe),
`which allows better resolution
`of cell borders, also demon-
`strated convincing hybridiza-
`tion of probes to the conjunc-
`tival epithelium, but hybridiza-
`tion to goblet cells was not
`obvious (arrows, E). In a sec-
`tion in which corneal, limbal,
`and conjunctival epithelium
`were present, hybridization of
`a'S probe,
`the
`indicating
`MUC4 mRNA expression, dis-
`appeared at die limbus (G).
`No expression was detected in
`either corneal (small arrow) or
`limbal
`(large arrow) epithe-
`lium. Bar = 50 //m (A to D);
`20 //m (E,F); 100 pm (G).
`
`\
`
`#
`* ,
`
`3B). The sense sequence of the same region of MUC4,
`used as a negative control probe, did not show any
`binding (Figs. 3C, 3D). To increase the resolution of
`the cellular distribution of MUC4 mRNA and to verify
`the tr'S data, DIG-labeled oligoprobe was used for in
`situ hybridization (Fig. 3E). The DIG data were consis-
`tent with the 35S results; the stratified conjunctival epi-
`thelium showed MUC4 mRNA expression; DIG-la-
`beled sense probe did not show any binding (Fig. 3F).
`Using tissue sections that included cornea and
`conjunctiva, expression of MUC4 mRNA in cornea
`was determined using MS-labeled MUC4 oligoprobe
`(Fig. 3G). MUC4 mRNA was detected in conjunctival
`epithelium. However, MUC4 mRNA expression disap-
`peared near the limbal region. No expression was ob-
`served in the corneal epithelium.
`Localization of MUC5 mRNA was determined
`
`with antisense riboprobe generated from the MUC5
`cDNA (Figs. 4A, 4B). Goblet cells showed intense sig-
`nal, but no signal was detected in the stratified epithe-
`lial cells. Some of the goblet cells did not show bind-
`ing. Sense probe transcribed from the MUC5 cDNA
`did not show any binding (Fig. 4C). Using tissue sec-
`tions that included cornea and conjunctiva, expres-
`sion of MUC5 mRNA in cornea also was determined.
`MUC5 mRNA was not detected in the corneal epithe-
`lium (Figs. 4D, 4E) of tissue sections in which goblet
`cells of the conjunctival epithelium were labeled with
`the MUC5 probe (Fig. 4F).
`MUC7 and MUC2 mRNA expression was exam-
`ined using in situ hybridization with MUC7 oligoprobe
`or HAM-1 riboprobe to MUC2, respectively. No MUC7
`binding was observed, even after an extensive 4-week
`exposure time (data not shown). No binding of MUC2
`
`APOTEX 1021, pg. 6
`
`
`
`1690
`
`A
`
`• . *
`
`Investigative Ophthalmology & Visual Science, July 1996, Vol. 37, No. 8
`
`FIGURE 4. MUC5 mRNA dis-
`tribution in conjunctiva us-
`ing in situ hybridization with
`35S-labeled MUC5 antisense
`riboprobe (A, low magnifica-
`tion), (B, high magnifica-
`tion), and sense riboprobe
`(C). Low magnification (A)
`shows overall pattern of in-
`tense signal in scattered cells.
`At high magnification (B),
`antisense probe bound to
`the goblet cells
`intensely
`(arrows). The control sense
`probe showed no binding
`over controls (C). In sections
`in which cornea, limbus, and
`conjunctiva were present,
`MUC5 mRNA was not de-
`tected in corneal epithelium
`(D, dark field; E, bright
`field), but conjunctival gob-
`let cells showed positive bind-
`ing (arrow, F), Bar =100 fim
`(A); 50 /jm (B toF).
`
`was observed in conjunctiva, whereas control sections
`of intestinal goblet cells showed intense binding (data
`not shown).
`
`DISCUSSION
`From the combined results of Northern blot analysis
`and in situ hybridization, we conclude that MUC4 and
`MUC5 mRNA are expressed in the human conjunc-
`tiva. Screening of MUC2 through MUC7 by Northern
`blot analysis showed a positive signal only with MUC4
`and MUC5. Further investigation by in situ hybridiza-
`tion demonstrated a different pattern of MUC4 and
`MUC5 mRNA distribution in conjunctival epithelium.
`The data suggest that the stratified epithelium, which
`expresses MUC4 mRNA, and the goblet cells, which
`express MUC5 mRNA, contribute to the production
`of ocular mucus. Recently, we reported17 the expres-
`sion of the transmembrane mucin, MUC1, by corneal
`and conjunctival stratified epithelia, but not goblet
`cells. To date, three of the cloned mucin genes are
`known to be expressed by the ocular surface epithe-
`lium and may contribute to the tear mucus layer. It
`is interesting tfiat each of these three mucins has a
`different pattern of mRNA expression in the ocular
`surface epithelium and that both classes of the cloned
`mucins (transmembrane and secretory mucins) are
`expressed.
`Expression of the secretory mucin MUC5 mRNA
`is limited to the goblet cells of the conjunctiva. On
`secretion, MUC5 may be the gel-forming mucin that
`provides the tear film mucus gel. The demonstrated
`
`presence of cysteine-rich domains in the MUC5 pro-
`tein backbone22 indicates that cross-links form be-
`tween individual MUC5 molecules by disulfide bonds,
`thus bringing about mucus gel formation.1'21'1 Some
`of the goblet cells at the conjunctival surface were not
`labeled with MUC5 probe. This result may reflect a
`difference in the differentiation stage of the labeled
`and unlabeled cells; perhaps the terminally differenti-
`ated goblet cells have translated all their MUC5
`mRNA. Studies of goblet cell-specific gene expression,
`such as that seen with MUC5, may provide further
`insight into the differentiation of goblet cells. Another
`possibility for lack of binding by some cells is that the
`goblet cell population may be heterogeneous. Some
`goblet cells may synthesize a different mucin than
`MUC5.
`The current study demonstrated convincing hy-
`bridization of MUC4 mRNA to conjunctival epithe-
`lium. MUC4 has been classified as a secretory mucin,14
`but, because only the tandem repeat sequence of
`MUC4 cDNA has been cloned,25 its detailed structure
`and function are still unclear. More cloning data are
`necessary for complete characterization as to whether
`MUC4 is truly a secretory type. Its designation as a
`secretory type mucin probably relates to its demon-
`strated expression by simple epithelia whose function
`is secretion. Expression of MUC4 mRNA by the stra-
`tified conjunctival epithelium is surprising because ex-
`pression of MUG4 had been demonstrated in only
`simple secretory-absorptive epithelia of bronchial, co-
`Ionic, and endocervical mucosa.33 It has been sug-
`gested that nongoblet cells of the conjunctiva provide
`
`APOTEX 1021, pg. 7
`
`
`
`Mucin Gene Expression in Conjunctiva
`
`a second source of mucin for the tear film. The sugges-
`tion was based on the presence of periodic acid-Schiff-
`positive vesicles inside the conjunctival epithelial cyto-
`plasm.'"0 Our data on the expression of MUC4 mRNA
`by nongoblet cell conjunctival epithelium is consistent
`with this suggestion. However, further research is re-
`quired to identify the contents of these secretory vesi-
`cles as MUC4.
`To determine whether MUC4 and MUC5 mRNA
`expression in the ocular surface is limited to the con-
`junctiva, Northern blot analysis was used to investigate
`the expression of these mucin mRNAs by cultured
`human corneal epithelium. Because it is difficult to
`obtain enough RNA for Northern blot analysis from
`in vivo human corneal epithelial samples, RNA from
`cultured human corneal epithelium, which produces
`MUC1 mRNA,17 was used. The results of these studies
`showed no expression of MUC4 or MUC5 mRNA by
`the cultured human corneal epithelium. In situ hy-
`bridization verified this result. Neither MUC4 nor
`MUC5 mRNA was detected in corneal epithelium in
`sections that had positive binding in the conjunctiva.
`It is noteworthy that expression of MUC4 mRNA
`ended near the limbal region. Thus, in the ocular
`surface epithelium, MUC4 and MUC5 may be specific
`to the conjunctiva.
`Two recent abstracts report use of antibodies de-
`scribed as 4F1 and 3A229 and an oligonucleotide probe
`to a region of the tandem repeat of MUC230 to assay
`for tear presence and human conjunctival expression,
`respectively. The abstracts report MUC2 in tears and
`MUC2 mRNA in conjunctival tissue. Using several
`methods, we could not corroborate the abstract data.
`As described above, in our hands, repeated Northern
`blot analysis using a cDNA probe (SMUC-41), which
`corresponds to an 836-bp region of several tandem
`repeats of the MUC2 molecule, did not show hybrid-
`ization to conjunctival mRNA but did show hybridiza-
`tion to small intestine mRNA. In addition, in situ hy-
`bridization using a 90-bp HAM-1 MUC2 probe to the
`tandem repeat region did not hybridize to conjuncti-
`val epithelium but did hybridize to small intestine. To
`extend these studies to tissues in which optimal RNA
`preservation was possible, in situ hybridization and
`Northern blot analysis using rat tissue and a 48-bp
`oligonucleotide probe to the tandem repeat of the rat
`MUC2 homologue34 were carried out. We could not
`detect MUC2 expression in rat conjunctiva by either
`method (unpublished data, 1995). One possible ex-
`planation for the discrepancy between our results and
`those of the two abstracts is tissue variation. Upregula-
`tion of MUC2 expression has been reported in in-
`flamed or hyperplastic respiratory tract epithelia.35 It
`would be interesting to determine whether a similar
`alteration of mucin expression occurs in the ocular
`surface in conditions of inflammation.
`Although integrity of RNA was confirmed by
`
`1691
`
`ethidium bromide staining before Northern blot anal-
`ysis, the positive signals of each mucin showed the
`typical polydispersed pattern. This polydispersity of
`mucin mRNA has been reported as a characteristic
`of mucin genes12 and has been discussed in several
`articles.1213 It may reflect the rapid turnover or insta-
`bility of mucin mRNA resulting from the extremely
`large size of mucin genes. Further investigation, how-
`ever, is required to clarify the reason for the polydis-
`persed pattern of mucin mRNA on Northern blots.
`The relationship between goblet cell density and
`mucus alteration in ocular surface disease has been
`well studied.36'37 It is obvious that the loss of goblet
`cells causes mucus deficiency. However, in some cases,
`decrease of tear breakup time does not correlate with
`the loss of goblet cells.38 Perhaps in some of these
`cases there is a deficiency in mucin production in
`the stratified conjunctival and corneal epithelia. The
`relationship between ocular disease and the alteration
`of MUC4, MUC5, or both, remains to be studied.
`In conclusion, we have demonstrated the expres-
`sion of MUC4 and MUC5 mucin mRNA in conjunc-
`tiva. These two mucins may play an important role in
`forming the gelled mucus layer at the air-ocular sur-
`face epithelium interface. This brings to three the
`number of cloned mucins expressed by the ocular
`surface epithelium, but their functions remain to be
`determined experimentally. It also remains to be de-
`termined whether the ocular surface expresses addi-
`tional uncharacterized or unique mucins.
`
`Key Words
`conjunctiva, gene expression, goblet cell, mucin (MUC)
`
`Acknowledgments
`
`The authors thank Dr. Neil Toribara for providing MUC2,
`MUC3, MUC5, and MUC6 cDNA, Dr. Carol Basbaum for
`HAM-1 cDNA, Dr. Enrico Cagliero for HUVEC mRNA, and
`Gale Unger for editorial assistance.
`
`References
`
`1. Nichols BA, Chiappino ML, Dawson CR. Demonstra-
`tion of the mucous layer of the tear film by electron
`microscopy. Invest Ophthalmol Vis Sri. 1985; 26:464-
`473.
`2. Strous GJ, Dekker J. Mucin-type glycoproteins. Crit Rev
`Biochem Mol Biol. 1992; 27:57-92.
`3. Holly F, Lemp M. Tear physiology and dry eyes. Surv
`Ophthalmol. 1977; 22:69-87.
`4. Iwata S, Kabasawa I. Fractionation and chemical prop-
`erties of tear mucoids. Exp Eye Res. 1971; 12:360-367.
`5. Moore JC, Tiffany JM. Human ocular mucus: Origins
`and preliminary characterisation. Exp Eye Res.
`1979; 29:291-301.
`6. Moore JC, Tiffany JM. Human ocular mucus: Chemi-
`cal studies. Exp Eye Res. 1981;33:203-212.
`7. Chao CC, Vergnes JP, Brown SI. Fractionation and
`partial characterization of macromolecular compo-
`
`APOTEX 1021, pg. 8
`
`
`
`1692
`
`Investigative Ophthalmology & Visual Science, July 1996, Vol. 37, No. 8
`
`nents from human ocular mucus. Exp Eye Res.