Clin. Otolarjngol, 1992, 17, 491-496
Culture of human middle ear mucosal explants; mucin production J.HILL*, D . A . H U T T O N , G.G.R.GREEN, J.P.BIRCHALL* & J.P.PEARSON Department of Physiological Sciences, The Medical School, University of Newcastle upon Tyne and *Department of ENT, Freeman Hospital, Freeman Road, Newcastle upon Tyne, U K Accepted for publication 10 June 1992
HILL J . , HUTTON D . A . , G R E E N G . G . R . , B I R C H A L L J . P . & P E A R S O N J . P .
(1992) Clin.Otolaryngol. 17, 4 9 1 4 9 6
Culture of human middle ear mucosal explants; mucin production Middle ear mucosal biopsies could be maintained in culture for up to 7 days, the longest time attempted in this study. Mucin biosynthesis and secretion were measured by incorporation of ‘‘C-glucosamine. Three peaks of radioactivity were present when the dialysed medium was chromatographed. Peak I which accounted for about 10% of the total radioactivity had properties characteristic of mucin. The other two peaks were not characteristic of mucins. Labelled macromolecules excluded on Sepharose 2B were also present in the tissue. Autoradiography of the explants showed that the labelled glucosamine was concentrated in the epithelial layer. Morphometry demonstrated that 1-2% of the epithelial cell volume consisted of goblet cells. The proportionate incorporation of radioactivity into macromolecules increased with increasing epithelial cell volume. This system will allow assessment of factors implicated in the pathogenesis of otitis media with effusion and the study of the action of pharmacological agents on biosynthesis and secretion. Keywords middle ear mucosa mucin biosynthesis
Middle ear epithelium is a modified respiratory epithelium’ and although there are many papers on the culture of explants and cells from human and animal respiratory epithelium, e.g.’l3 there are few studies directly using middle ear epithelium.4 The difficulty in obtaining middle ear tissue has lead some workers to use nasal epithelium as a substitute.’ However, as metaplastic changes in the middle ear mucosa have been well documented in otitis media with effusion,6 it is necessary to study middle ear mucosa directly. Therefore any changes in biosynthetic and secretory properties that occur when the epithelium changes from normal to diseased can be quantified. Mucus glycoproteins or mucins are the most important molecules in determining the viscoelastic properties of mucus secretions’.* and mucins are present in middle ear effusions in otitis media with effusion (OME) at between 8 and 25% of the non-dialysable solids. Mucins are highly glycosylated proteins which are linked together by disulphide bridges to form large macromolecular complexes.’ The carbohydrate Correspondence: Dr J.P.Pearson, Department of Physiological Sciences, The Medical School, Framlington Place, Newcastle upon Tyne NE2 4HH, UK.
human
chains are packed closely together in extensive glycosylated regions of the protein core. This close packing makes it impossible for proteolytic enzymes to gain access to the protein core. Therefore proteolysis of mucin secretions destroys contaminating protein, proteoglycans and serum glycoproteins leaving a large molecular size mucin fragment. Treatment of a mixed secretion with proteolytic enzymes is therefore one method of determining the mucin content. Mucins also have a characteristic buoyant density 1.40-1.55 g/ml’O different from nucleic acid and proteoglycans which have higher buoyant densities and protein and lipids which have lower ones. The carbohydrate side chains are rich in N-acetyl glucosamine (17.1 f3.8% by weight of middle ear mucin”) and therefore I4C labelled glucosamine can be used to label newly synthesized mucin, allowing measurement of mucin production. In this paper we describe the maintenance of human middle ear mucosal explants in culture of up to 7 days and describe methods of measuring mucin production by these explants. The integrity and viability of the mucosa were assessed histologically and autoradiography was used to assess the distribution of radioisotope. The number of mucin
49 1
492 J.Hill et al.
secreting cells in the middle ear mucosa could be expected to vary according to the anatomical site of the biopsy' and the degree of metaplasia in individual samples. It was therefore necessary to subject the biopsies to morphometric analysis in order to relate the number of mucin secreting cells to m u c h secretion in an individual sample and allow comparison of mucin secretion between samples.
Methods COLLECTION A N D M A I N T E N A N C E OF E X P L A N T S
Mucosal biopsies were removed with micro-cupped forceps from patients undergoing middle ear surgery. Biopsies were taken through myringotomy incisions in patients undergoing grommet insertion and from patients undergoing myringoplasty or stapedectomy. The biopsies were transferred to the laboratory in 5 m l of tissue culture medium. The biopsies were then placed on Nucleopore filters, and floated on 3 ml of medium (RPMI 1640) containing 10% fetal calf serum, 2 m M glutamate, 1OO/ v/v penicillin and streptomycin and 2.5 mM fungizone. The explants were incubated at 37°C in 5% CO, and air for up to 7 days.
HISTOLOGICAL TECHNIQUES
Explants removed at intervals of 1-7 days were preserved in 10% formal saline solution. Paraffin sections 4 pm thick were stained with either haematoxylin and eosin (H and E). periodic acid Schiffs, or alcian blue.
I4cL A B E L L I N G STUDIES The medium was removed after 48-h incubation of the biopsies and replaced by 3 ml of medium containing I4C glucosamine (Amersham International UK) final concentration of 74 kBq/ml. The medium was collected after a further 24 h and stored frozen until used for further analysis. The tissue was either preserved in formal saline for histological or autoradiographic analysis, or homogenized in 0.067 M phosphate buffer containing a proteolytic inhibitor cocktail" to allow analysis of intracellular labelled macromolecules.
AUTORADIOGRAPHY
Autoradiographs were prepared by dipping the histological slides in liquid photographic emulsion (Ilford (35). Following exposure in the dark for 2 weeks the slides were developed and counterstained with H and E. Controls were taken using explants not exposed to I4C labelled glucosamine.
MORPHOMETRY
Serial sections 4 pm thick were made from the explant after standard fixation. The sections were stained with alcian blue to visualize the mucus containing goblet cells. The whole explant was completely sectioned to give between 25 and 50 sections and every fifth section was measured to obtain ( I ) the total cross-sectional area; (2) the area occupied by surface epithelium; and (3) the area occupied by goblet cells. The assumption was made that the explants approximated in shape to a sphere. The formula $rr3 (using the section with the largest cross-sectional area) produces values for explant volume of epithelium and volume occupied by goblet cells. Measurements were made on a semi-automatic image analysis system, MOP videoplan using Kontron software. BIOCHEMICAL ANALYSES
Free label was removed from the medium and the tissue homogenate by exhaustive dialysis against distilled water at 4°C. The labelled macromolecules ( M , > 104) recovered were characterized in terms of their hydrodynamic size, buoyant density and susceptibility to various enzymes. Hydrodynamic size, before and after enzymatic digestion, was analysed by size exclusion chromatography on a Sephadex G I 5 0 column (80 x 1.5 cm) and by chromatography on a Sepharose CL-2B column (150 x 1.5 cm) eluted with 0.2 M NaCl containing 0.2% w/v NaN,. The buoyant density of radiolabelled material was measured after centrifugation a t 40000 r.p.m. a t 4°C for 48 h in a CsCl density gradient (starting density 1.42 g/ml)." Labelled macromolecules were digested with either papain (Sigma Chemical Company, Limited, London, UK) 0.08-0.125 mg of papain per ml of sample in 0.067 M sodium phosphate buffer pH 6.7 containing 5 mM cysteine HCI and 5 mM NazEDTA for 18-72 h a t 60"C;'0 chondroitinase ABC (Sigma Chemical Company Limited, UK), 3.2pg chondroitinase per ml of sample in 0.1 M sodium acetate/O.l M Tris hydroxymethylaminomethane pH 7.3 for 4 h at 37"C;I2 or heparinase I11 (Sigma Chemical Company Limited, London, UK), 0.01 U/ml of sample in 0.01 M sodium phosphate buffer pH 7.0 for 16 h at 25"C.13 The products of digestion were then chromatographed by gel exclusion chromatography. The activity of the enzymes was confirmed by parallel digestions of specific substrates.
Results TISSUE M A I N T E N A N C E
Thirty-six explants were maintained in support medium for over 72 h. Evidence of viability was demonstrated by histological staining. Figure 1 shows a H & E stain of a biopsy after 72 h in medium. The explant shows a continuous
Culture of mucosal explants 493
Figure 1. Section of middle ear explant epithelium after 72 h in culture. H & E x 150.
epithelium with intact cells and clearly defined nuclei. The explants were viable up to at least 7 days, the longest time attempted in this study. Figure 2 shows a biopsy maintained in support medium for 6 days, stained with alcian blue. Again the epithelial layer is intact and the cells appear normal with evidence of alcian blue staining material in the epithelial cells, presumably mucin. Five out of 36 of the explants did not survive in the medium: these samples were taken from chronically discharging ears and had bacterial overgrowth identified as Pseudomonns aeruginosa. No evidence of bacterial infection was found in the other 31 explants. AUTORADIOGRAPHY
In order to determine whether the cells in the explant had taken up the I4C label and, if so, in which cells it was concentrated. 4 explants were studied. The explants were supported in medium for 48 h, then bathed in I4C glucosamine containing medium for a further 24 h, washed, and prepared for autoradiography. Figure 3 shows a representative section countcrstained with H & E. As can be seen, the
Figure 2. Section of middle ear explant epithelium after 6 days in culture. Alcian blue. x 300.
I4C label has been taken up and concentrated in the epithelial layer where the mucus secreting cells are found. IDENTITY OF LABELLED MACROMOLECULES
Chromatography of the labelled macromolecules recovered from the medium on Sephadex G. 150 showed the presence of three peaks of radioactivity (Figure 4). Peak I eluting at the void volume (Vo) of the column contained 10f2% (X+ 1 s.e.m.) (n = 4) of the labelled material. Two other peaks were present, peak I1 (Kav, 0.22) contained 76*4% ( n = 4) and peak 111 (Kav, 0.67) contained 14+5% (n = 4) of the labelled material. The excluded peak, peak I, was also excluded on Sepharose 4B gel filtration. When the dialysed
494 J.Hill et al.
3000
m:
r
0
R
20
40
60
80
Fraction number
Figure 4. Gel filtration on Sephadex GI50 of culture medium after dialysis to remove 0 , unbound radioactive label and the medium after 18 h digestion with papain. The V o and Yt were determined using blue dextran and methyl orange respectively.
+,
Figure 3. Autoradiograph counterstained with H 8c E of middle ear explant after 48 h in culture followed by a further 24 h in the presence of I4C-glucosamine.The small black dots indicate the position of the radioactive label. x 420.
medium was subjected to papain digestion for 18 h peak I was unaffected; however, peak I1 was almost entirely removed and peak Ill near the total column volume was now the major peak (Figure 4). None of the peaks present on GI50 gel filtration was affected by digestion with chrondroitinase ABC or heparinase 111. The tissue homogenate also contained some large molecular size labelled material which was excluded on Sepharose CL-2B. When the medium was subjected to CsCl equilibrium density gradient centrifugation (starting density 1.42 g/ml) the major portion of the labelled material separated into the low buoyant density fractions 1-4, with densities below 1.40 g/nil. There was however a second band of labelled material at a higher buoyant density 1.45-1 3 5 g/ml, pcaking at I .52 g/ml. This density is chardcteristic for a mucin (Figure 5). When peaks 1-111 from gel filtration were subjected to CsCl gradients, peak 11 separated to the top of the gradient with a buoyant density below 1.4 g/ml and the majority of the material was in fraction 1 with a density of about 1.3 g/ml. Peak 111 banded at a buoyant density of 1.42 g/ml at the bottom of the range for a mucin and peak I had a buoyant density of about 1.50 g/ml characteristic of a mucin.
MORPHOMETRY
The estimated epithelial cell volume of the explants varied from 7.15 to 323 pm3 and the estimated goblet cell volume
ranged from 0.133 to 2.65 pm’. Therefore the fraction of the epithelial cell volume which was goblet cells was 1-2%. The total counts remaining in the medium after dialysis increased with increasing epithelial cell volume and the percentage incorporation of labelled glucosamine into macromolecules 1 s.e.m.) (n = 3) again increasing with was 0.9*0.35% (.i* increasing epithelial cell volume in the explant. As this study was aimed at measuring mucin production the non-dialysable counts in the medium were related to goblet cell volume. The non-dialysable count per pm’ of goblet cell volume was 3.5 f 1.2 x 10’ d.p.m. (n = 3). However, most of this activity was lost after exhaustive digestion with papain with 2.Of 1.2 x lo4 d.p.m./pm’ of goblet cell volume (Xf 1 s.e.m.) (n = 3), that is, bctween 3.9 and 6.9% of the non-dialysable counts, remaining. Resistance to papain digestion is a characteristic of the highly glycosylated regions of mucins.
Discussion The method of tissue maintenance used here was able to maintain the mucosa in a viable state for up to 7 days as assessed by histological staining which showed the presence of normal cells, some containing alcian blue staining material, presumably mucin. Previous studies with human middle ear mucosa’ were set up only to measure ciliary activity. A semi-solid agar medium was used which would make assessment of mucin production difficult. Guinea-pig ’~ middle ear mucosa has been maintained in fluid ~ u l t u r ebut these workers also confined their study to assessing ciliary activity.
Culture of mucosal explants 495
r
600
500
E:
400
4
v
0
f
300
\ I
l
/
l
In this study we floated the explants o n inert Nucleopore filters allowing the transfer of nutrients from the medium and the secreted molecules to pass back into the medium. Explants rather than monolayers were used in this study because they contained fully differentiated mucus secreting cells and were easy to maintain whereas respiratory epithelial cells need highly defined culturc conditions to express mucus synthesis and secretion activities.” Mucins have three biochemical characteristics which allow their identification in the medium and tissue. Firstly, their large hydrodynamic size will allow them to be separated from most contaminants by gel filtration. Secondly, the multiple densely packed carbohydrate side chains render large sections of the protein core (at least 50%) of the molecule resistant to the action of proteolytic enzymesI6 so proteolysis of mucins results in large hydrodynamic sized species remaining. Finally, the buoyant density of mucins is characteristically between I .40 and 1.55 g/ml. The autoradiography studies demonstrated that the 14C label has been taken up by the explants and is concentrated in the epithelial layer where it is presumably being actively incorporated into secreted macromolecules. Gel filtration of the dialysed medium on Sephadex GI50 showed a peak eluting at the void volume, where a mucin would elute. However, this is only a minor peak accounting for about 10% of the total medium radioactivity. The other two peaks are presumably non-mucin macromolecules. This is supported by the fact that peak I1 is destroyed by proteolysis with papain. The identity of the molecules in peaks I1 and 111 is unknown but as with peak I they did not contain
glycosaminoglycans or proteoglycans as shown by the lack of any effect of chondroitinase ABC or heparinase 111. A similar distribution of label has been shown in medium from human mucin secreting cells in c u l t ~ r e ’in~ that the large excluded mucin molecules contain less of the total radioactivity than the included molecules o n gel filtration. Others working with a human colonic epithelial cell line have shown that most of the material secreted by these cells is excluded on gel filtration, characteristic of a mucin.” However, in the later study the column was eluted with 4~ guanidinium chloride (GuHCI) and Kim” has shown that the culture medium from hamster tracheal surface epithelial cells labelled with 3H-glucosamine contains virtually all the radioactivity in the Vo of a Sepharose 4B column eluted with 4 M GuHCI. Subsequent analysis showed that this excluded peak containcd non-mucin proteins and glycoproteins linked to the m u c h via hydrophobic interactions. These smaller size contaminants could be separated from the mucin by treatment with SDS or heating. After this treatment, up to 60% of the radioactivity eluted in the included volume of the column. These results suggest that elution of media containing mucin secretions on gel filtration in the presence of GuHCl may lead to an overestimate of the m u c h content. In our studies only about 10% of the non-dialysable medium radioactivity was mucin as assessed by exclusion on Sephadex (3150. We used a gel filtration medium with a smaller pore size than that used by other workers to exclude mucins, i.e. Sephadex GI 50 compared to Sepharose 4B”,l9 because our previous studies” showed that the proteolytically digested middle ear mucin was not excluded on Sepharose 4B. That the G 150 excluded peak I was characteristic of a mucin was confirmed by CsCl equilibrium density centrifugation,’”,” peak I having a buoyant density of about 1.5 g/ml. Neither of the other two peaks on gel filtration had hydrodynamic sizes or buoyant densities characteristic of a highly glycosylated mucin. After exhaustive digestion of the medium with papain (48-72 h) only 3.9-6.9% of the radioactivity remained; this compares favourably with about 10% of the non-dialysable medium label excluded on gel filtration. The appearance of a large peak 111 on gel filtration after papain digestion was because digestion was incomplete since incubation was for 18 h only. Peak I11 was lost when digestion was exhaustive, i.e. 48-72 h. Therefore, only peak I was mucin. These studies demonstrate that middle ear mucosal explants can be maintained in organ culture and incorporate ‘‘C-glucosamine into glycosylated molecules secreted by the cells including macromolecules with properties characteristic of mucins. However, only about 10% of the non-dialysable radioactivity is in mucins. This method of in-vitro culture of mucosal explants will permit the study of the effects of drugs and endogenous agents on mucin production and thus elucidate some of the aetiological factors in the pathogenesis of glue ear.
496 J,Hill et al.
Acknowledgements We thank the Medical Rescarch Council and Catherine Cookson for support and the staff of ENT Newcastle for providing the mucosal samples.
References SAD^ J. (1966) Middle ear mucosa. Arch. Otoluryngol. 78, 137-143 ClENG P.W.. SHERMANJ.M., BOATT.F. & BRUCEM. (1981) Quantitation of radiolabelled mucous glycoproteins secreted by tracheal explants. Anal. Biochem. 117, 301-306 KIMK.C., REARICK J.I., NETTERSHEIM P. & JETTENA.M. (1985) Biochemical characterization of mucous glycoproteins synthesised and secreted by hamster tracheal epithelial cells in primary culture. J . B i d . Chenr. 260, 4021-4027 S.C. & GROTEJ.J. (1989) VAN BLITTERSWIJK C.A., HESSELING The effect of calcium on terminal differentiation, proliferation and morphology of cholesteatoma matrix meatal epidermis and middle ear epithelium. In Chalesteaioma and Mastoid Surgery. pp. 77-81, Kugler & Ghedinin Publications, Amsterdam DRUKER I., W ~ I S M AZ.N & SAD^ J . (1976) Tissue culture of human adult adenoids and middle ear mucosa. Ann. O l d . Rhinol. Lur.vngol. 85, 3277333 SAUEJ. (1966) Pathology arid pathogenesis of secretory otitis media. Arch. Otoluryngol. Head Neck Surg. 84, 297-305 S.B. (1984) BELLA.E.. ALLENA., MORRISE.R. & ROSS-MURPHY Functional interactions of gastric mucus glycoprotein. Int. J . Riol. Mucromul. 6, 309-3 I5 SELLERSL.A., ALLENA,. MORRISE.R. & ROSS-MURPHY S.B. ( I 983) Rheological studies on pig gastrointestinal mucous secretions. Biochem. Soc. Trans. 11, 763 -764. ALLENA. (1981) Structure and function of gastrointestinal mucus. In Physiology of the Gasiroinresrinal Tract. Volume 1, pp. 617-639. Raven Press, New York 10 FITZGERALI> J.E., GREENG.G.R., STAFFORDF.W., BIRCHALI. J.P. & PEARSONJ.P. (1987) Characterization of human middle
ear mucus glycoprotein in chronic secretory otitis media. Clin. Chim. Acta 169, 281 -298 11 STARKEY B.J., SNARYD. & ALLENA. (1974) Characterisation of gastric mucoproteins isolated by equilibrium density gradient centrifugation. Biochem. J . 141, 633-639 12 IIASCALLV.C. & HEINECARUD. (1974) . . Olieosaccharide competitors of the proteoglycan-hyaluronic acid interaction. J . Biol. Chem. 249, 4242-4249 13 MCLEANM.W. (1987) Action of heparinases I1 and 111 on bovine kidney Heparin sulphate. In Proceedings of the 9th International S,ymposium on Glycoconjrrgates. B34 A. Lerouge, France 14 OHASHI Y., NAKAI Y., KOSHIMOH. & ESAKIY. (1986) Ciliary activity in the in vitro tubotympanum. Arch. Otoluryngol. 243, 317-319 15 Wu R., MARTINW.R., ROBINSONC.B., ST GEORGEJ.A., PLOPPER C.G., KURI.AND G., LASTJ.A., CROSSC.E., MCDONALD R.J. & BOUCHERR. (1990) Expression of mucin synthesis and secretion in human tracheobronchial epithelial cells grown in culture. A m . J . Respir. Cell Mol. Bid. 3, 467-478 16 PEARSON J.P. & ALLENA. (1982) Reduction by mercaptoethanol and proteolysis of the non-glycosylated peptide region of pig gastric mucus glycoprotein. In Mucus in Health and Diseuse. Volume 2. pp. 152-153. Plenum Press, New York 17 CORFIELD A.P., CLAMPJ.R., CASEYA.D. & PARASKEVA C. (1990) Characterization of a sialic acid rich mucus glycoprotcin secreted by a premalignant human colorectal adenoma cell line. Inr. J . Cuncer. 46, 1059-1065 18 KIMK.C. (1991) Mucin-like glycoproteins secreted from cultured hamster tracheal surface epithelial cells: their hydrophobic nature and amino acid composition. Exp. Lung Res. 17, 65-76 19 Wu R., PLOPPER C.G. & CIENZ P. (1991) Mucin-like glycoprotein secreted by cultured hamster tracheal epithelial cells. Biochem. J . 277, 713-718 20 ALLEN A. (1989) Gastrointestinal mucus. In Handbook of Physiology USA. pp. 359-382. American Physiological Society 21 NEUTRAM.R. & FORSTNER J.F. (1987) Gastrointestinal mucus: synthesis, secretion and function. In Physiology of’ rhe Gastrointestinal Tract. Second edition. pp. 975- 1009. Raven Press, New York
Culture of human middle ear mucosal explants; mucin production J.HILL*, D . A . H U T T O N , G.G.R.GREEN, J.P.BIRCHALL* & J.P.PEARSON Department of Physiological Sciences, The Medical School, University of Newcastle upon Tyne and *Department of ENT, Freeman Hospital, Freeman Road, Newcastle upon Tyne, U K Accepted for publication 10 June 1992
HILL J . , HUTTON D . A . , G R E E N G . G . R . , B I R C H A L L J . P . & P E A R S O N J . P .
(1992) Clin.Otolaryngol. 17, 4 9 1 4 9 6
Culture of human middle ear mucosal explants; mucin production Middle ear mucosal biopsies could be maintained in culture for up to 7 days, the longest time attempted in this study. Mucin biosynthesis and secretion were measured by incorporation of ‘‘C-glucosamine. Three peaks of radioactivity were present when the dialysed medium was chromatographed. Peak I which accounted for about 10% of the total radioactivity had properties characteristic of mucin. The other two peaks were not characteristic of mucins. Labelled macromolecules excluded on Sepharose 2B were also present in the tissue. Autoradiography of the explants showed that the labelled glucosamine was concentrated in the epithelial layer. Morphometry demonstrated that 1-2% of the epithelial cell volume consisted of goblet cells. The proportionate incorporation of radioactivity into macromolecules increased with increasing epithelial cell volume. This system will allow assessment of factors implicated in the pathogenesis of otitis media with effusion and the study of the action of pharmacological agents on biosynthesis and secretion. Keywords middle ear mucosa mucin biosynthesis
Middle ear epithelium is a modified respiratory epithelium’ and although there are many papers on the culture of explants and cells from human and animal respiratory epithelium, e.g.’l3 there are few studies directly using middle ear epithelium.4 The difficulty in obtaining middle ear tissue has lead some workers to use nasal epithelium as a substitute.’ However, as metaplastic changes in the middle ear mucosa have been well documented in otitis media with effusion,6 it is necessary to study middle ear mucosa directly. Therefore any changes in biosynthetic and secretory properties that occur when the epithelium changes from normal to diseased can be quantified. Mucus glycoproteins or mucins are the most important molecules in determining the viscoelastic properties of mucus secretions’.* and mucins are present in middle ear effusions in otitis media with effusion (OME) at between 8 and 25% of the non-dialysable solids. Mucins are highly glycosylated proteins which are linked together by disulphide bridges to form large macromolecular complexes.’ The carbohydrate Correspondence: Dr J.P.Pearson, Department of Physiological Sciences, The Medical School, Framlington Place, Newcastle upon Tyne NE2 4HH, UK.
human
chains are packed closely together in extensive glycosylated regions of the protein core. This close packing makes it impossible for proteolytic enzymes to gain access to the protein core. Therefore proteolysis of mucin secretions destroys contaminating protein, proteoglycans and serum glycoproteins leaving a large molecular size mucin fragment. Treatment of a mixed secretion with proteolytic enzymes is therefore one method of determining the mucin content. Mucins also have a characteristic buoyant density 1.40-1.55 g/ml’O different from nucleic acid and proteoglycans which have higher buoyant densities and protein and lipids which have lower ones. The carbohydrate side chains are rich in N-acetyl glucosamine (17.1 f3.8% by weight of middle ear mucin”) and therefore I4C labelled glucosamine can be used to label newly synthesized mucin, allowing measurement of mucin production. In this paper we describe the maintenance of human middle ear mucosal explants in culture of up to 7 days and describe methods of measuring mucin production by these explants. The integrity and viability of the mucosa were assessed histologically and autoradiography was used to assess the distribution of radioisotope. The number of mucin
49 1
492 J.Hill et al.
secreting cells in the middle ear mucosa could be expected to vary according to the anatomical site of the biopsy' and the degree of metaplasia in individual samples. It was therefore necessary to subject the biopsies to morphometric analysis in order to relate the number of mucin secreting cells to m u c h secretion in an individual sample and allow comparison of mucin secretion between samples.
Methods COLLECTION A N D M A I N T E N A N C E OF E X P L A N T S
Mucosal biopsies were removed with micro-cupped forceps from patients undergoing middle ear surgery. Biopsies were taken through myringotomy incisions in patients undergoing grommet insertion and from patients undergoing myringoplasty or stapedectomy. The biopsies were transferred to the laboratory in 5 m l of tissue culture medium. The biopsies were then placed on Nucleopore filters, and floated on 3 ml of medium (RPMI 1640) containing 10% fetal calf serum, 2 m M glutamate, 1OO/ v/v penicillin and streptomycin and 2.5 mM fungizone. The explants were incubated at 37°C in 5% CO, and air for up to 7 days.
HISTOLOGICAL TECHNIQUES
Explants removed at intervals of 1-7 days were preserved in 10% formal saline solution. Paraffin sections 4 pm thick were stained with either haematoxylin and eosin (H and E). periodic acid Schiffs, or alcian blue.
I4cL A B E L L I N G STUDIES The medium was removed after 48-h incubation of the biopsies and replaced by 3 ml of medium containing I4C glucosamine (Amersham International UK) final concentration of 74 kBq/ml. The medium was collected after a further 24 h and stored frozen until used for further analysis. The tissue was either preserved in formal saline for histological or autoradiographic analysis, or homogenized in 0.067 M phosphate buffer containing a proteolytic inhibitor cocktail" to allow analysis of intracellular labelled macromolecules.
AUTORADIOGRAPHY
Autoradiographs were prepared by dipping the histological slides in liquid photographic emulsion (Ilford (35). Following exposure in the dark for 2 weeks the slides were developed and counterstained with H and E. Controls were taken using explants not exposed to I4C labelled glucosamine.
MORPHOMETRY
Serial sections 4 pm thick were made from the explant after standard fixation. The sections were stained with alcian blue to visualize the mucus containing goblet cells. The whole explant was completely sectioned to give between 25 and 50 sections and every fifth section was measured to obtain ( I ) the total cross-sectional area; (2) the area occupied by surface epithelium; and (3) the area occupied by goblet cells. The assumption was made that the explants approximated in shape to a sphere. The formula $rr3 (using the section with the largest cross-sectional area) produces values for explant volume of epithelium and volume occupied by goblet cells. Measurements were made on a semi-automatic image analysis system, MOP videoplan using Kontron software. BIOCHEMICAL ANALYSES
Free label was removed from the medium and the tissue homogenate by exhaustive dialysis against distilled water at 4°C. The labelled macromolecules ( M , > 104) recovered were characterized in terms of their hydrodynamic size, buoyant density and susceptibility to various enzymes. Hydrodynamic size, before and after enzymatic digestion, was analysed by size exclusion chromatography on a Sephadex G I 5 0 column (80 x 1.5 cm) and by chromatography on a Sepharose CL-2B column (150 x 1.5 cm) eluted with 0.2 M NaCl containing 0.2% w/v NaN,. The buoyant density of radiolabelled material was measured after centrifugation a t 40000 r.p.m. a t 4°C for 48 h in a CsCl density gradient (starting density 1.42 g/ml)." Labelled macromolecules were digested with either papain (Sigma Chemical Company, Limited, London, UK) 0.08-0.125 mg of papain per ml of sample in 0.067 M sodium phosphate buffer pH 6.7 containing 5 mM cysteine HCI and 5 mM NazEDTA for 18-72 h a t 60"C;'0 chondroitinase ABC (Sigma Chemical Company Limited, UK), 3.2pg chondroitinase per ml of sample in 0.1 M sodium acetate/O.l M Tris hydroxymethylaminomethane pH 7.3 for 4 h at 37"C;I2 or heparinase I11 (Sigma Chemical Company Limited, London, UK), 0.01 U/ml of sample in 0.01 M sodium phosphate buffer pH 7.0 for 16 h at 25"C.13 The products of digestion were then chromatographed by gel exclusion chromatography. The activity of the enzymes was confirmed by parallel digestions of specific substrates.
Results TISSUE M A I N T E N A N C E
Thirty-six explants were maintained in support medium for over 72 h. Evidence of viability was demonstrated by histological staining. Figure 1 shows a H & E stain of a biopsy after 72 h in medium. The explant shows a continuous
Culture of mucosal explants 493
Figure 1. Section of middle ear explant epithelium after 72 h in culture. H & E x 150.
epithelium with intact cells and clearly defined nuclei. The explants were viable up to at least 7 days, the longest time attempted in this study. Figure 2 shows a biopsy maintained in support medium for 6 days, stained with alcian blue. Again the epithelial layer is intact and the cells appear normal with evidence of alcian blue staining material in the epithelial cells, presumably mucin. Five out of 36 of the explants did not survive in the medium: these samples were taken from chronically discharging ears and had bacterial overgrowth identified as Pseudomonns aeruginosa. No evidence of bacterial infection was found in the other 31 explants. AUTORADIOGRAPHY
In order to determine whether the cells in the explant had taken up the I4C label and, if so, in which cells it was concentrated. 4 explants were studied. The explants were supported in medium for 48 h, then bathed in I4C glucosamine containing medium for a further 24 h, washed, and prepared for autoradiography. Figure 3 shows a representative section countcrstained with H & E. As can be seen, the
Figure 2. Section of middle ear explant epithelium after 6 days in culture. Alcian blue. x 300.
I4C label has been taken up and concentrated in the epithelial layer where the mucus secreting cells are found. IDENTITY OF LABELLED MACROMOLECULES
Chromatography of the labelled macromolecules recovered from the medium on Sephadex G. 150 showed the presence of three peaks of radioactivity (Figure 4). Peak I eluting at the void volume (Vo) of the column contained 10f2% (X+ 1 s.e.m.) (n = 4) of the labelled material. Two other peaks were present, peak I1 (Kav, 0.22) contained 76*4% ( n = 4) and peak 111 (Kav, 0.67) contained 14+5% (n = 4) of the labelled material. The excluded peak, peak I, was also excluded on Sepharose 4B gel filtration. When the dialysed
494 J.Hill et al.
3000
m:
r
0
R
20
40
60
80
Fraction number
Figure 4. Gel filtration on Sephadex GI50 of culture medium after dialysis to remove 0 , unbound radioactive label and the medium after 18 h digestion with papain. The V o and Yt were determined using blue dextran and methyl orange respectively.
+,
Figure 3. Autoradiograph counterstained with H 8c E of middle ear explant after 48 h in culture followed by a further 24 h in the presence of I4C-glucosamine.The small black dots indicate the position of the radioactive label. x 420.
medium was subjected to papain digestion for 18 h peak I was unaffected; however, peak I1 was almost entirely removed and peak Ill near the total column volume was now the major peak (Figure 4). None of the peaks present on GI50 gel filtration was affected by digestion with chrondroitinase ABC or heparinase 111. The tissue homogenate also contained some large molecular size labelled material which was excluded on Sepharose CL-2B. When the medium was subjected to CsCl equilibrium density gradient centrifugation (starting density 1.42 g/ml) the major portion of the labelled material separated into the low buoyant density fractions 1-4, with densities below 1.40 g/nil. There was however a second band of labelled material at a higher buoyant density 1.45-1 3 5 g/ml, pcaking at I .52 g/ml. This density is chardcteristic for a mucin (Figure 5). When peaks 1-111 from gel filtration were subjected to CsCl gradients, peak 11 separated to the top of the gradient with a buoyant density below 1.4 g/ml and the majority of the material was in fraction 1 with a density of about 1.3 g/ml. Peak 111 banded at a buoyant density of 1.42 g/ml at the bottom of the range for a mucin and peak I had a buoyant density of about 1.50 g/ml characteristic of a mucin.
MORPHOMETRY
The estimated epithelial cell volume of the explants varied from 7.15 to 323 pm3 and the estimated goblet cell volume
ranged from 0.133 to 2.65 pm’. Therefore the fraction of the epithelial cell volume which was goblet cells was 1-2%. The total counts remaining in the medium after dialysis increased with increasing epithelial cell volume and the percentage incorporation of labelled glucosamine into macromolecules 1 s.e.m.) (n = 3) again increasing with was 0.9*0.35% (.i* increasing epithelial cell volume in the explant. As this study was aimed at measuring mucin production the non-dialysable counts in the medium were related to goblet cell volume. The non-dialysable count per pm’ of goblet cell volume was 3.5 f 1.2 x 10’ d.p.m. (n = 3). However, most of this activity was lost after exhaustive digestion with papain with 2.Of 1.2 x lo4 d.p.m./pm’ of goblet cell volume (Xf 1 s.e.m.) (n = 3), that is, bctween 3.9 and 6.9% of the non-dialysable counts, remaining. Resistance to papain digestion is a characteristic of the highly glycosylated regions of mucins.
Discussion The method of tissue maintenance used here was able to maintain the mucosa in a viable state for up to 7 days as assessed by histological staining which showed the presence of normal cells, some containing alcian blue staining material, presumably mucin. Previous studies with human middle ear mucosa’ were set up only to measure ciliary activity. A semi-solid agar medium was used which would make assessment of mucin production difficult. Guinea-pig ’~ middle ear mucosa has been maintained in fluid ~ u l t u r ebut these workers also confined their study to assessing ciliary activity.
Culture of mucosal explants 495
r
600
500
E:
400
4
v
0
f
300
\ I
l
/
l
In this study we floated the explants o n inert Nucleopore filters allowing the transfer of nutrients from the medium and the secreted molecules to pass back into the medium. Explants rather than monolayers were used in this study because they contained fully differentiated mucus secreting cells and were easy to maintain whereas respiratory epithelial cells need highly defined culturc conditions to express mucus synthesis and secretion activities.” Mucins have three biochemical characteristics which allow their identification in the medium and tissue. Firstly, their large hydrodynamic size will allow them to be separated from most contaminants by gel filtration. Secondly, the multiple densely packed carbohydrate side chains render large sections of the protein core (at least 50%) of the molecule resistant to the action of proteolytic enzymesI6 so proteolysis of mucins results in large hydrodynamic sized species remaining. Finally, the buoyant density of mucins is characteristically between I .40 and 1.55 g/ml. The autoradiography studies demonstrated that the 14C label has been taken up by the explants and is concentrated in the epithelial layer where it is presumably being actively incorporated into secreted macromolecules. Gel filtration of the dialysed medium on Sephadex GI50 showed a peak eluting at the void volume, where a mucin would elute. However, this is only a minor peak accounting for about 10% of the total medium radioactivity. The other two peaks are presumably non-mucin macromolecules. This is supported by the fact that peak I1 is destroyed by proteolysis with papain. The identity of the molecules in peaks I1 and 111 is unknown but as with peak I they did not contain
glycosaminoglycans or proteoglycans as shown by the lack of any effect of chondroitinase ABC or heparinase 111. A similar distribution of label has been shown in medium from human mucin secreting cells in c u l t ~ r e ’in~ that the large excluded mucin molecules contain less of the total radioactivity than the included molecules o n gel filtration. Others working with a human colonic epithelial cell line have shown that most of the material secreted by these cells is excluded on gel filtration, characteristic of a mucin.” However, in the later study the column was eluted with 4~ guanidinium chloride (GuHCI) and Kim” has shown that the culture medium from hamster tracheal surface epithelial cells labelled with 3H-glucosamine contains virtually all the radioactivity in the Vo of a Sepharose 4B column eluted with 4 M GuHCI. Subsequent analysis showed that this excluded peak containcd non-mucin proteins and glycoproteins linked to the m u c h via hydrophobic interactions. These smaller size contaminants could be separated from the mucin by treatment with SDS or heating. After this treatment, up to 60% of the radioactivity eluted in the included volume of the column. These results suggest that elution of media containing mucin secretions on gel filtration in the presence of GuHCl may lead to an overestimate of the m u c h content. In our studies only about 10% of the non-dialysable medium radioactivity was mucin as assessed by exclusion on Sephadex (3150. We used a gel filtration medium with a smaller pore size than that used by other workers to exclude mucins, i.e. Sephadex GI 50 compared to Sepharose 4B”,l9 because our previous studies” showed that the proteolytically digested middle ear mucin was not excluded on Sepharose 4B. That the G 150 excluded peak I was characteristic of a mucin was confirmed by CsCl equilibrium density centrifugation,’”,” peak I having a buoyant density of about 1.5 g/ml. Neither of the other two peaks on gel filtration had hydrodynamic sizes or buoyant densities characteristic of a highly glycosylated mucin. After exhaustive digestion of the medium with papain (48-72 h) only 3.9-6.9% of the radioactivity remained; this compares favourably with about 10% of the non-dialysable medium label excluded on gel filtration. The appearance of a large peak 111 on gel filtration after papain digestion was because digestion was incomplete since incubation was for 18 h only. Peak I11 was lost when digestion was exhaustive, i.e. 48-72 h. Therefore, only peak I was mucin. These studies demonstrate that middle ear mucosal explants can be maintained in organ culture and incorporate ‘‘C-glucosamine into glycosylated molecules secreted by the cells including macromolecules with properties characteristic of mucins. However, only about 10% of the non-dialysable radioactivity is in mucins. This method of in-vitro culture of mucosal explants will permit the study of the effects of drugs and endogenous agents on mucin production and thus elucidate some of the aetiological factors in the pathogenesis of glue ear.
496 J,Hill et al.
Acknowledgements We thank the Medical Rescarch Council and Catherine Cookson for support and the staff of ENT Newcastle for providing the mucosal samples.
References SAD^ J. (1966) Middle ear mucosa. Arch. Otoluryngol. 78, 137-143 ClENG P.W.. SHERMANJ.M., BOATT.F. & BRUCEM. (1981) Quantitation of radiolabelled mucous glycoproteins secreted by tracheal explants. Anal. Biochem. 117, 301-306 KIMK.C., REARICK J.I., NETTERSHEIM P. & JETTENA.M. (1985) Biochemical characterization of mucous glycoproteins synthesised and secreted by hamster tracheal epithelial cells in primary culture. J . B i d . Chenr. 260, 4021-4027 S.C. & GROTEJ.J. (1989) VAN BLITTERSWIJK C.A., HESSELING The effect of calcium on terminal differentiation, proliferation and morphology of cholesteatoma matrix meatal epidermis and middle ear epithelium. In Chalesteaioma and Mastoid Surgery. pp. 77-81, Kugler & Ghedinin Publications, Amsterdam DRUKER I., W ~ I S M AZ.N & SAD^ J . (1976) Tissue culture of human adult adenoids and middle ear mucosa. Ann. O l d . Rhinol. Lur.vngol. 85, 3277333 SAUEJ. (1966) Pathology arid pathogenesis of secretory otitis media. Arch. Otoluryngol. Head Neck Surg. 84, 297-305 S.B. (1984) BELLA.E.. ALLENA., MORRISE.R. & ROSS-MURPHY Functional interactions of gastric mucus glycoprotein. Int. J . Riol. Mucromul. 6, 309-3 I5 SELLERSL.A., ALLENA,. MORRISE.R. & ROSS-MURPHY S.B. ( I 983) Rheological studies on pig gastrointestinal mucous secretions. Biochem. Soc. Trans. 11, 763 -764. ALLENA. (1981) Structure and function of gastrointestinal mucus. In Physiology of the Gasiroinresrinal Tract. Volume 1, pp. 617-639. Raven Press, New York 10 FITZGERALI> J.E., GREENG.G.R., STAFFORDF.W., BIRCHALI. J.P. & PEARSONJ.P. (1987) Characterization of human middle
ear mucus glycoprotein in chronic secretory otitis media. Clin. Chim. Acta 169, 281 -298 11 STARKEY B.J., SNARYD. & ALLENA. (1974) Characterisation of gastric mucoproteins isolated by equilibrium density gradient centrifugation. Biochem. J . 141, 633-639 12 IIASCALLV.C. & HEINECARUD. (1974) . . Olieosaccharide competitors of the proteoglycan-hyaluronic acid interaction. J . Biol. Chem. 249, 4242-4249 13 MCLEANM.W. (1987) Action of heparinases I1 and 111 on bovine kidney Heparin sulphate. In Proceedings of the 9th International S,ymposium on Glycoconjrrgates. B34 A. Lerouge, France 14 OHASHI Y., NAKAI Y., KOSHIMOH. & ESAKIY. (1986) Ciliary activity in the in vitro tubotympanum. Arch. Otoluryngol. 243, 317-319 15 Wu R., MARTINW.R., ROBINSONC.B., ST GEORGEJ.A., PLOPPER C.G., KURI.AND G., LASTJ.A., CROSSC.E., MCDONALD R.J. & BOUCHERR. (1990) Expression of mucin synthesis and secretion in human tracheobronchial epithelial cells grown in culture. A m . J . Respir. Cell Mol. Bid. 3, 467-478 16 PEARSON J.P. & ALLENA. (1982) Reduction by mercaptoethanol and proteolysis of the non-glycosylated peptide region of pig gastric mucus glycoprotein. In Mucus in Health and Diseuse. Volume 2. pp. 152-153. Plenum Press, New York 17 CORFIELD A.P., CLAMPJ.R., CASEYA.D. & PARASKEVA C. (1990) Characterization of a sialic acid rich mucus glycoprotcin secreted by a premalignant human colorectal adenoma cell line. Inr. J . Cuncer. 46, 1059-1065 18 KIMK.C. (1991) Mucin-like glycoproteins secreted from cultured hamster tracheal surface epithelial cells: their hydrophobic nature and amino acid composition. Exp. Lung Res. 17, 65-76 19 Wu R., PLOPPER C.G. & CIENZ P. (1991) Mucin-like glycoprotein secreted by cultured hamster tracheal epithelial cells. Biochem. J . 277, 713-718 20 ALLEN A. (1989) Gastrointestinal mucus. In Handbook of Physiology USA. pp. 359-382. American Physiological Society 21 NEUTRAM.R. & FORSTNER J.F. (1987) Gastrointestinal mucus: synthesis, secretion and function. In Physiology of’ rhe Gastrointestinal Tract. Second edition. pp. 975- 1009. Raven Press, New York
