Quantitation of Histochemical Staining of Salivary Gland Mucin Using Image Analysis in Cats and Dogs Mahmut Sozmen, Peter J

Home , Fuchsine

Quantitation of histochemical staining of salivary gland mucin using image analysis in cats and dogs Mahmut Sozmen, Peter J. Brown, Peter J. Cripps

To cite this version:

Mahmut Sozmen, Peter J. Brown, Peter J. Cripps. Quantitation of histochemical staining of salivary gland mucin using image analysis in cats and dogs. Veterinary Research, BioMed Central, 1999, 30 (1), pp.99-108. hal-00902560

HAL Id: hal-00902560 https://hal.archives-ouvertes.fr/hal-00902560 Submitted on 1 Jan 1999

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Original article

Quantitation of histochemical staining of salivary gland mucin using image analysis in cats and dogs

Mahmut Sozmena Peter J. Browna Peter J.b Cripps

a Comparative Pathology Laboratory, University of Bristol, School of Veterinary Science, Langford, Bristol BS40 5DU, UK b Farm Animal Studies, University Veterinary Field Station, Leahurst, Neston, South Wirral, L64 7TE, UK

(Received I October 1997; accepted 17 September 1998)

Abstract - Two different, computer-based, image analysis methods were employed to complement subjective assessment of patterns of mucin staining (by periodic acid Schiff/alcian blue, aldehyde fuchsin/alcian blue and potassium hydroxide-alcian blue/phenylhydrazine hydrocholoride) in digitised images of sections of major and minor salivary glands from cats and dogs. Image analysis based on red, green and blue (RGB) staining was not suitable for quantitation of histochemical staining of mucin in salivary glands. Image analysis based on hue, saturation and lightness (HSL) allowed quantitative assessment of staining by different stain components and of mixed staining, and enabled comparison of staining of different glands in dogs and cats. Quantitative analysis based on HSL allowed differentiation of differences in staining patterns of major and minor cat and dog salivary glands, and between the species; such differences would have been impossible to distinguish on subjective grounds alone. Quantitative assessment of normal salivary gland histochemistry allows comparison with staining patterns in disease. © Inra/Elsevier, Paris. morphometry / image analysis / cat / dog / salivary gland

Résumé - Quantification par analyse d’image de la coloration histochimique des mucines des glandes salivaires de chien et de chat. Deux méthodes différentes d’analyse d’image sur ordinateur ont été utilisées pour compléter l’appréciation subjective des modes de coloration des mucines (par acide périodique de Schift7bleu alcian, fuchsine aldéhyde/bleu alcian et hydroxyde de potassium-bleu alcian/hydrochlorure de phénylhydrazine) à partir d’images digitalisées de coupes des glandes salivaires majeures et mineures provenant de chiens et de chats. L’analyse d’image fondée sur la coloration bleu, vert et rouge (r-ed, green and blue: RGB) n’a pas permis de quantifier la coloration histochimique des glandes salivaires. L’analyse d’image fondée sur la teinte, la saturation et la brillance (hue, .saturation and lightne.ss : HSL) a permis d’une part de quantifier le marquage par les différents composants des colorations et par leur mélange et d’autre part de comparer le marquage des diverses

* Correspondence and reprints Tel.: (44) Ol 17 928 9525; fax: (44) 01 17 928 9588; e-mail [email protected] glandes chez les chiens et les chats. L’analyse d’image fondée sur la méthode HSL a permis la dis- tinction de différents modes de coloration des glandes salivaires majeures et mineures chez le chien et le chat et entre espèces ; de telles distinctions n’auraient pas pu être obtenues à partir des seules don- nées subjectives. L’appréciation quantitative de l’histochimie de glandes salivaires normales per- met la comparaison avec les colorations obtenues lors de maladies. © Inra/Elsevier, Paris. morphométrie / analyse d’image / chat / chien / glande salivaire

1. INTRODUCTION measure the relative percentages of different classes of mucosubstances in normal major Combinations of histochemical stains and minor salivary glands in cats and dogs. allow simultaneous demonstration of dif- The histochemical methods used allowed ferent types of mucins. Examples of these mucins to be recognised within three major methods are periodic acid Schiff/alcian blue categories within the classification of Spicer (PAS/AB); high iron diamine/alcian blue et al. [19]: neutral mucins, sulphomucins (HID/AB): PAS stains neutral mucins; and sialomucins. Two different analyses alcianophilia is attributable to the presence were available within the image analysis of carboxyl groups; HID selectively stains package employed; these were based on sulphated mucosubstances. A combined threshold extraction imagery of the three aldehyde fuchsin/alcian blue (AF/AB) tech- primary colours - red, green and blue nique provides an alternative to the high (RGB), and threshold imagery based on hue, iron diamine-alcian blue [17].] . saturation and lightness (HSL). Previous studies of salivary gland histochemistry of cats and dogs have been qualitative and subjective [ 10, 12, 20]. Such an 2. MATERIALS AND METHODS approach is valuable and plays a major role in investigation of normal tissues and in disease diagnosis. In other situations, estimation 2.1. Animals and tissues of size and tissue distribution of particular cell is and components important morpho- Six samples of normal tissue from the pala- metrical has increased in recent analysis tine, parotid, sublingual, submandibular, tongue years. Most morphometrical analyses have and zygomatic glands were removed during the utilised the stereological point-counting course of post-mortem examination from a total technique for assessing proportional vol- of 13 dogs [16] and ten cats; in the cat samples stained with AB/PAS it was to consider umes of salivary constituents; in people, possible the serous von Ebner’s from have included an assessment of the glands separately they the more extensive mixed In all of glands. cases, replacement salivary gland parenchyma the cause of death or euthanasia was not related by fibrous and fatty tissue as an effect of to salivary gland disease and the glands were ageing [3, 14] or to detect morphometrical macroscopically nonnal. Details of the animals changes to justify a diagnosis of Sjogren’s are given in table I. studies in syndrome [1]. Morphometrical The tissues were fixed in 10 % neutral animal are few No salivary glands [4, 13]. buffered formol saline for 1-2 days at room tem- quantitative data are available to show the perature before being routinely processed and proportional amounts of different mucins embedded in polywax at 60 °C. Sections were secreted by cat and dog salivary glands. cut at 5 pm thickness and mounted on slides. Representative sections were stained with haema- This study was undertaken to determine toxylin and eosin to confirm their normal struc- the distribution of different mucins and to ture. 2.2. Mucosubstance histochemical 2) aldehyde fuchsin/alcian blue (AF/AB) [18]: procedures aldehyde fuchsin solution for 20 min; alcian blue (I %) solution (pH 2.5) for 5 min; De-waxed sections were stained, at room tem- 3) potassium hydroxide/AB/periodic acid perature, with: phenylhydrasine Schiff (KOH/AB/PAPS) [11]: 1) alcian blue/periodic acid Schiff (AB/PAS) [9]: 0.5 % KOH in 70 °lo (v/v) aqueous ethanol for 15 min; alcian blue (I %) in 0.1 M HCL, 1 % alcian blue (pH 2.5) solution for 5 min; pH 1.0 for 30 min; I % (w/v) periodic acid I % periodic acid for 5 min; rinsed and for 2 h; 0.5 % aqueous phenylhydrasine exposed to Schiff’s reagent for IS min; hydrocholoride for 2 h; basic fuchsin Schiff Mayer’s haematoxylin counterstain; for 4 h. 2.3. Morphometrical analysis Two different areas (each representing a tissue section area of about 0.640 mm2) were mea- sured from each section combina- Digitised images of stained sections were cap- (gland/stain tion). Areas were selected to include secretory tured using a Videovue capture card (Video Associates Labs Inc.; Austin, Texas, USA) tissue and associated ducts but to exclude larger blood vessels and interlobular/interlobar con- installed in a 486, DX266 computer linked to a nective tissue. Results were as median microscope (Nikon microphot-fxa.) with a one expressed stained area. chip colour video camera. Images were obtained percentage as 10 x 40, full colour (24 bit).tif files. Data were analysed using Minitab (version 9.2; 1993) and differences were tested The were on a by using captured images investigated non-parametric (Kruskal Wallis) statistical anal- 6100/60 Power Macintosh computer using a yses, group means were compared using Mann- semiautomatic image analysis program (Optilab Whitney rank sum test. A null hypothesis of no 2.5; ME Electronics, UK). were Reading, Images difference was used and a critical P = in terms of RGB probability analysed following thresholding 0.05 was assumed. and HSL.

2.3.1. based on RGB Image analysis 3.RESULTS

RGB threshold analysis was performed using an automatic function of the analysis software 3.1. assessment of mucin package as well as by manipulation of individual Subjective red (R), green (G) and blue (B) values. histochemical staining characteristics 2.3.2. Image analysis based on HSL Three general types of epithelial mucosubstances were identified the histo- Images were subjected to threshold analysis in by terms of H, S and L. H values were chosen which chemical techniques employed; in broad distinguished staining with each individual stain, terms, they were recognised as belonging and combined staining, which correlated with to three major categories: neutral mucins, that observed on subjective assessment. The same sulphomucins and sialomucins. values were used for all sections stained by the same stain combination and are presented in AB/PAS: neutral mucins were identified table II. by their positive (red) PAS reaction. Acidic mucins, stained by alcian blue alone, Differences in staining pattern between appeared sky blue. Those mucins stained the various cat and dog salivary glands are by both stains varied in colour from reddish shown in tables V-VII. There were few dif- purple to royal blue, depending on the ferences between cat and dog glands with degree of alcianophilia. AB/PAS staining (table V); there were more differences between cat and AF/AB: were stained var- dog glands sulphomucins stained with AF/AB and ious shades of acidic mucins (table VI) purple; KOH/AB/PAPS (table VII). appeared blue. Mixed staining exhibited various shades of purple-bluish, depending on the of and sul- proportions carboxylated 4. DISCUSSION phated mucins. KOH/AB/PAPS: the results were simi- In the present study, three general types lar to those given by the AF/AB method of epithelial mucosubstances were identi- except that sulphomucins were stained aqua- fied; in broad terms, they were recognised as marine instead of reddish or bluish purple. neutral mucins, acidic sulphated mucins and acidic mucins to the Differences sialylated according among histochemically classification of and and demonstrable mucins were both Leppi Spicer 161 salivary et al. who mucins qualitative and quantitative. Only minor dif- Spicer [19], categorised into four histochemical groups. Type I ferences in staining characteristics were mucins are characterised noted in the same obtained from dif- by PAS-positiv- glands and with cationic ferent subjects. Subjective assessment of ity negligible staining II mucins are and the distributions and relative proportions of dyes. Type carboxylated also react with PAS but, in addition, bind different types of mucin in the different sali- alcian blue; they fail to react with the HID vary glands are summarised in table III. reagents. Types III and IV mucins are PAS- positive, alcianophilic and stain positively with HID reagents, with differences in the 3.2. Morphometrical analysis intensity of cationic dye binding. The significance of the different types of 3.2.1. Image analysis based on RGB secretions in different salivary epithelial cells remains unknown. Menghi et al. [8]] No combination of RGB thresholds was observed that changes of diet do not affect found which selectively identified staining the type of secretion by postnatal rabbits of different components or mixed staining. while sialic acid content increased with growth. Differences in chemical compo- 3.2.2. Image analysis based on HSL nents (amylase) from rat parotid saliva have been shown to vary in response to changes in diet and the of diet to masti- It was possible to select H values which relationship distinguished staining with each individual cation [5]. In the present investigation, some differences in characteristics were stain, and combined staining, which corre- staining lated with that observed subjectively noted in the same glands obtained from different variations in (table III). There were marked differences subjects; staining may between the proportions of different glands be indicative of different stages in secretory stained by the various stain components, cycles [2]. and combinations, in cats and dogs. The The distribution of different staining pat- results (median values ± SD) are presented terns in cats and dogs reported in the present in table IV. Statistical analyses of the results study are in general agreement with results are presented in tables V-VII. of other workers !10, 12, 15, 20]. Some minor differences were noted on the basis of employed in the present study indicated that morphometrical analysis. For example, cat sulphated mucous secretion predominated in submandibular and sublingual gland are both submandibular and sublingual glands both mixed (serous and mucous) glands; (69.3 and 22.3 % KOH/AB-positive stain- Tachi [20] indicated that mucous cells pre- ing with KOH/AB/PAPS, respectively). dominate in the submandibular gland while The difference between those figures and serous cells are more numerous in the sub- the findings of Tachi [20] may be that what, lingual glands. The quantitative approach on simple microscopic observation, appeared to be solely sialomucin was, on The mucous acinar cells present in cat and analysis, mixed with sulphated mucin. In a dog minor salivary glands are a mixed pop- similar way, Reifel and Travill [12] found ulation mainly producing sulphated muco- that all dog parotid gland acini were substances, which are histochemically sim- alcianophilic (with AB/PAS) indicating the ilar to comparable cells in the sublingual presence of mainly acidic carbohydrates. and submandibular glands. On the histochemical basis of the the aci- In this investigation, no purely alcianophilic present study, nar cells of all minor are staining was detected and PAS-positive salivary glands staining dominated (41.8 %). This discrep- best categorised as mucous, since the cells contain neutral and acidic ancy might reflect small amounts of PAS- mucopolysac- positivity in the cytoplasm resulting in charides with mainly sulphomucin and sialomucin. alcian blue-positive staining being recognised as mixed staining. The quantity of In many glands in the present study, the PAS positive secretory granules is affected staining patterns included staining with each by the phase of the secretory cycle; the component stain alone, as well as combined secretory granules are concentrated during staining; morphometric analysis allowed a phase of maturation, after enzyme syn- further discrimination of the mucin stain- thesis and before discharge of secretion [2]. ing patterns in different glands. Such a method may also be of value in assessment [2] Donath K., Die Sialadenose der Parotis. Ultra- of salivary gland tissue in disease. Changes strukturelle, klinische und experimentelle Befunde zur Sekretionspathologie, Fischer, Stuttgart, 1976. in histochemically demonstrable mucins have been demonstrated in diseased sali- [3] Drummond J., Chisholm D.M., A qualitative and of the human labial sali- Lev and quantitative study ageing vary glands by Spicer [7]. vary glands, Arch. Oral Biol. 29 (1984) 151-155. [4] Hagelqvist E., Ahlner B.H., Lind M.G., Mor- phology and histochemistry of rabbit sub- ACKNOWLEDGEMENT mandibular glands, Acta Otolaryngol. Suppl. 480 (1991) 1-17.

[5] Johnson D.A., L.M., Effect of food con- Mahmut Sozmen was supported by a grant Sreebny and starvation on the diurnal of from the Turkish Educational Council. sistency cycle Higher the rat parotid gland, Arch. Oral Biol. 16, (1971) 177-185. [6] Leppi T.J., Spicer S.S., The histochemistry of REFERENCES mucins in certain primate salivary glands, Am. J. Anat. 118 (1966) 833-859. I1 i De Wilde P.C.M., Baak J.P.A., Slootweg P.J., [7] Lev R., Spicer S.S., A histochemical comparison Hene R.J., Kater L., Morphometry in the diag- of human epithelial mucins in normal and in nosis of Sjogren’s syndrome, Anal. Quant. Cytol. hypersecretory states including pancreatic cystic Histol. 8 (1986) 49-55. fibrosis, Am. J. Pathol. 46 (1965) 23-48. [81 Menghi G., Vitaioli L., Bondi A.M., Materazzi ! 14] Scott J., Quantitative age changes in the histo- G., Glycoconjugates in salivary glands of rats logical structure of human submandibular sali- during postnatal development, Anat. Anz. 149 vary glands, Arch. Oral Biol. 22 ( 1977) 221-227. (1981)226-231. [ 15! Shackleford J.M., Klapper C.E., Structure and [91 Mowry R.W., Alcian blue techniques for the his- carbohydrate histochemistry of mammalian sali- tochemical study of acidic carbohydrates, J. His- vary glands, Am. J. Anat. III l (1962) 25-47. tochem. Cytochem. 16 ( 1956) 530-539. [ 16J Sozmen M., Brown P.J., Eveson J.W., Cytokeratin [10] Munhoz C.O.G., Histochemical classification of immunostaining in normal dog major and minor acini and ducts of parotid glands from artio- salivary glands, Vet. Res. 29 ( 1998) 457-466. dactyles, carnivores and rodents, Acta Histochem. [)7) Spicer S.S., Diamine methods for differentiating 39 (1971 ) 302-317. mucosubstances histochemically, J. Histochem. [ I 1 ! Reid P.E., Owen D.A., Ramey C.W., Dunn W.L., Cytochem. 13 (1965) 21 I-234. Clay M.G., Jones E.A., Histochemical procedures [ 18J Spicer S.S., Meyer D.B., Histochemical differ- for the simultaneous visualization of sialic acids entiation of acid mucopolysaccharides by mean of its side chain 0- acyl variants and 0- sulphate combined aldehyde fuchsin-alcian blue staining, ester, Histochem. J. 17 ( 1985) 113-117. Am. J. Clin. Pathol. 33 ( 1960) 453-460. [12! Reifel C.W., Travill A.A., Structure and carbo- ! 19J Spicer S.S., Leppi T.J., Stoward P.J., Suggestions hydrate histochemistry of postnatal canine salivary for a histochemical terminology of carbohydrate- glands, Am. J. Anat. 134 ( 1972) 377-393. rich tissue components, J. Histochem. Cytochem. [ 13) Sashima M., Age-related changes of rat sub- 13 (1965) 559-603. mandibular glands: a morphometric and ultra- 120] Tachi T., Histological and histochemical investi- structural study. J. Oral Pathol. 15 ( 1986) gations on the major salivary glands of the cat, 507-512. Aichi Gakuin J. Dent. Sci. 10 (1972) 201-225.