Dynamics of protein and fluid secretion from the major salivary glands of rat: relevance of research findings to clinically observed defective secretion in cystic fibrosis.

Although there are indications of a defect in secretion of protein from exocrine cells in cystic fibrosis (CF), this remains an aspect of CF research that has not been adequately addressed. Using salivary glands of rat as model systems, and following the effects of parasympathetic and sympathetic autonomic nerve stimulation on these glands, we demonstrate the existence of three separate pathways through which secretion of protein can be evoked from serous and mucous exocrine cells. These pathways allow the secretion of proteins from the intracellular compartments in a constitutive, intermediate or regulated manner. The primary aspects of secretory profile including concentration and the degree of hydration of secreted material differ greatly between the pathways, are cell type specific, and presumably are a direct consequence of controlled changes in the levels of second messengers induced upon stimulation of these cells. As previously published reports suggest that only the beta-adrenergic regulated pathway is affected by CF, differences between the pathways in their secretory profiles may influence the development of lung disease, through disparate disturbances in the secretion of protein and fluid from serous and mucous cells of the submucosal glands that line the bronchiolar tree in humans. We gratefully acknowledge support from The Wellcome Trust and from The European Union Biomed II Programme.

Ultrastructural changes in exocrine tissues of a DeltaF-508 CFTR mouse model.

Cystic fibrosis (CF) is characterized by abnormal secretion from epithelial cells. We wanted to detect changes in the ultrastructural characteristics of cells within a number of exocrine tissues, including the colon, submandibular and parotid salivary glands of DeltaF-508 CFTR animals. Therefore, in the present study a DeltaF-508 CFTR mouse model was compared to control, by applying conventional and complex carbohydrates staining techniques to tissue sections at the electron microscope level. The colon of DeltaF-508 CFTR mice contained thick mucous secretions that harbored many bacteria, along with cytoplasmic fragments and leukocytes. Leukocytes were also seen to infiltrate the cytoplasm of goblet cells. Tissues were taken before, 10 min after isoprenaline, and 30 min after a further injection of methacholine. In the submandibular gland, there is limited secretory activity after isoprenaline treatment, and this increases further with methacholine treatment. Depletion of the secretory granules of acinar cells is observed, following the combined isoprenaline and methacholine treatment, but no significant changes in granule numbers occurred in granular tubule cells. Glycogen, abundant before treatment, is reduced within 10 min of isoprenaline treatment and is completely exhausted by 30 min, especially in the convoluted granular tubule cells. A few secretory granules in acinar and in granular tubule cells of the DeltaF-508 CFTR submandibular glands displayed two electron densities. The secretory responses of the parotid gland cells were similar to those in submandibular gland cells, except that in these DeltaF-508 CFTR cells, secretory granules appeared more polymorphic in structure than those found in control animals.

Sulphation of the salivary mucin MG1 (MUC-5B) is not correlated to the degree of its sialylation and is unaffected by cystic fibrosis.

Defective acidification of intracellular organelles, particularly the trans-Golgi network, has been proposed to explain the decreased sialylation and increased sulphation of secreted proteins in cystic fibrosis (CF). To test this hypothesis we compared expression of sulphate and sialic acid on three salivary mucins namely MG1 (MUC-5B), MG2 (MUC-7) and GL. Proteins in whole mouth saliva (WMS) from four individuals were separated by fast protein liquid chromatography (FPLC) on a Superdex 200 column and the partially purified mucins slot-blotted and assayed for sulphate content by staining with Alcian Blue. Sulphation varied with the individual and with the mucin: MG1 was the most sulphated and contributed almost the entire sulphate content of WMS. These results allowed us to test small volumes of WMS from 20 CF patients and age- and sex-matched controls for estimates of sulphate content on MG1. Wherever possible sulphate on MG1 was also visualised by staining washed SDS-PAGE gels with Alcian Blue at pH 1.0. To assess the sialic acid content of salivary mucins, electroblots of SDS-PAGE gels were probed with labelled Triticum vulgaris agglutinin. In summary, our results show MG1 to be the main sulphated protein in whole mouth saliva and there are large differences in the expression of sulphate and of sialic acid on this mucin, both in control and CF groups. CF led neither a decrease in sialylation nor an increase in sulphation and direct comparisons of sialic acid content with sulphate in MG1 failed to reveal any obvious link between the two in health or in disease. Our data thus do not support the hypothesis of defective acidification as the underlying cause of altered glycosylation in CF, but point instead to inter-individual differences in expression/functioning of terminal glycosyltransferases for published observations. We thank the European Union Biomed II programme for support.

Altered sialyl- and fucosyl-linkage on mucins in cystic fibrosis patients promotes formation of the sialyl-Lewis X determinant on salivary MUC-5B and MUC-7.

Destruction of the lungs as a consequence of recurrent infections with microorganisms such as Pseudomonas aeruginosa remains the underlying cause of most morbidity and mortality in cystic fibrosis (CF). We have hypothesized that changes in the glycosylation of key tracheal mucins such as MUC5B and MUC7 might increase the risk of pulmonary disease in CF patients. However, in preference to sputum we have examined the sugar-chains on these mucins in saliva because in the latter not only can the glycoproteins be collected from controls, but they are essentially free from modifications made following bacterial infection in disease. Proteins in ductal or whole-mouth saliva from 20 CF patients with the Delta F-508 CFTR mutation and age-and sex-matched controls were separated by SDS-PAGE and blotted onto nitrocellulose and then probed with labelled lectins of known specificity. Linkage of terminal sialic acid on the blotted mucins was determined using Sambucus nigra agglutinin (detects the 2-->6 linkage) and Maackia amurensis agglutinin (the 2-->3 linkage). Fucose was detected by Ulex europaeus agglutinin-1 (1-->2 linkage) and Aleuria aurantia agglutinin (1-->3 linkage). We found that each mucin shows a characteristic glycosylation pattern and in controls most of the sialic acid is 2-->6 linked on MG1 (MUC 5B) and 2-->3 linked on MG2 (MUC 7). CF is associated with a shift from a 2-->6 linkage to a 2-->3 linkage on MG1 with some patients showing almost no 2-->6 linkage; 2-->3 linkage on MG2 is similarly increased in disease in some individuals. The expression of fucose on these mucins is also raised in CF patients. These shift to a 2-->3 linkage of sialic acid, and with increased fucosylation this promotes the formation of sialyl-Lewis X antigen detected on CF mucins in our study. These changes will be tested for their correlation with the severity of lung disease. We gratefully acknowledge support from the European Union Biomed-II Programme.

Constitutive secretion of kallikreins in vivo from rat submandibular glands.

In parasympathetic saliva from rat submandibular glands the relative proportions of the various tissue kallikreins differ from those in sympathetic saliva. Kallikreins in sympathetic saliva arise from exocytosis of prepackaged granules in granular tubules, so the kallikreins in parasympathetic saliva must come from a non-granular pool, and are likely to be secreted through a constitutive vesicular route. During periods devoid of stimulation in anaesthetised rats, the kallikreins have been found to accumulate progressively in glandular lumina in parasympathetic-like proportions. As this transport of kallikrein into lumina occurs continuously in vivo, independently of any stimulation or any secretion of fluid, it must arise by constitutive vesicular secretion. During parasympathetic stimulation, the kallikreins are secreted into the saliva at a greater rate than in the resting state but their proportions remain the same and the means by which this increase occurs is open to debate. Constitutively secreted true tissue kallikrein (rK1) has been found to have a different molecular form from that in secretory granules. The submandibular glands also contribute to the kallikreins normally circulating in the blood. Serum levels of kallikrein increased equally during either parasympathetic or sympathetic stimulation and were independent of the amounts secreted into the saliva, so are likely to have arisen from constitutive secretion via the basal sides of the cells, morphological evidence for which has been found in the mouse (Penschow & Coghlan, 1993).

Enzyme histochemistry of rat mast cell tryptase.

Fixation and staining conditions for rat mast cell tryptase and its histochemical distribution in different rat tissues were investigated. Prostate, skin, lung, gut, stomach and salivary glands were fixed in either aldehyde or Carnoy fixatives and then frozen or embedded in paraffin wax. Preservation of tryptase enzymic activity against peptide substrates required aldehyde fixation and frozen sectioning. Of the peptide substrates examined, z-Ala-Ala-Lys-4-methoxy-2-naphthylamide and z-Gly-Pro-Arg-4-methoxy-2-naphthylamide proved the most effective for the demonstration of tryptase. Double staining by enzyme cytochemistry followed by immunological detection of tryptase showed that, in all tryptase-containing mast cells, the enzyme is at least in part active. Conventional dye-binding histochemistry was used to confirm the identity of mast cells. Aldehyde-fixed mucosal mast cells required a much shorter staining time with Toluidine Blue if tissue sections were washed directly in t-butyl alcohol. Double staining by enzyme cytochemistry and dye binding showed that tryptase is absent from mucosal and subepidermal mast cells, which are also smaller in size and appear to contain fewer granules than connective tissue mast cells. This study demonstrates that rat mast cell tryptase, unlike tryptases in other species, is a soluble enzyme. It is stored in an active form and is absent from some mast cell subpopulations in mucosa, skin and lung.

The enzymic potential of tissue kallikrein (rK1) in rat submandibular saliva depends on whether it was secreted via constitutive or regulated pathways.

The enzymic activity and immunoreactivity of rat tissue kallikrein (rK1) secreted at rest by granular duct cells of unstimulated submandibular glands has been compared with that secreted on autonomic nerve stimulation. Although a direct vesicular, constitutive secretory pathway operates for rK1 secretion from granular duct cells of unstimulated and parasympathetically stimulated glands the rK1 was not present in a pro-form and actually showed a greater enzymic activity per unit immunoreactive protein than the granule-derived rK1 in sympathetically evoked saliva. Constitutively secreted rK1 was found to be in a single chain molecular form by reducing SDS gel electrophoresis. In contrast rK1 secreted from the storage granule pool of granular duct cells on sympathetic nerve stimulation was present in much higher amounts and occurred in both one-chain and two-chain forms as revealed by SDS gel electrophoresis under reducing conditions. The lower enzymic potential of rK1 in sympathetically evoked saliva might be accounted for by its conversion to a two-chain form.

Glycosylation of secretory proteins in salivary glands and saliva studied by lectin-probes.

Lectin-probes were used histochemically to study glycosylation patterns in prepackaged secretory material in cat and rat submandibular glands and to assess the secretory changes induced by nerve stimulations. The same probes were also used for correlative biochemical assessments of the constituents in glandular extracts and in saliva from the same experiments, after electrophoretic separations and immobilization on nitrocellulose membranes. These studies provided a broader understanding than the use of either procedure alone. Lectin binding patterns in saliva were more complex than anticipated from the histochemistry and this probably relates to the presence of different constituents in secretory granules and a greater availability of binding sites biochemically. In cats, parasympathetic stimulation caused depletion of lectin staining in acini and to some extent in demilunes, whereas sympathetic stimulation caused depletion of lectin binding in striated ducts and reduction in demilunar size. Biochemically after SDS-PAGE of saliva similar secretory effects were observed but each nerve also evoked some secretion from the cells not showing histochemical change. In rats, sympathetic stimulation caused depletion of lectin positive granules from both acini and granular tubules. After SDS-PAGE of the saliva one zone of lectin binding 94-150 kD (Mr) was identified as acinar mucin. Granular tubule proteins consisting mainly of kallikreins occurred in a zone 25-35 kD (Mr) and lectin bindings suggested that different kallikreins may be differently glycosylated. This unexpected possibility was confirmed on slot-blot preparations of separated kallikreins, which also revealed the novel finding that some kallikreins are O- as well as N-glycosylated.

Apical secretion of rat submandibular tissue kallikrein continues in the absence of external stimulation: evidence for a constitutive secretory pathway.

Sequential samples of saliva evoked by low frequency parasympathetic nerve stimulation were collected from cannulated submandibular ducts of anaesthetized rats following periods of varying length without stimulation. After such rest periods the first samples of saliva were found to contain much higher levels of tissue kallikrein (rK1) activity and protein than the ensuing samples from the same stimulation period and the latter contained levels similar to those found previously in parasympathetically evoked saliva rK1 activity in first samples increased with the length of the preceding interval, indicating that a continuous secretion of rK1 occurs from ductal cells into glandular lumina and accumulates there in the absence of stimulation or fluid secretion. The protein secretory pattern following rest pauses was unaffected by alpha-adrenergic receptor blockade and a high percentage of the proteinase activity was resistant to soya bean trypsin inhibitor, showing that the secretion was not caused by exocytosis of storage granules from ductal cells and therefore was likely to be a result of a constitutive secretion of newly synthesized enzyme. The pattern of continuous secretion into lumina detected for total protein in the absence of stimulation, suggests that other secretory proteins may also be secreted similarly but at different rates.

Use of lectin-probes for correlative histochemical and biochemical assessments of the glycosylation patterns of secretory proteins, including kallikreins, in salivary glands and saliva.

Labelled lectins were used as probes to study the glycosylation and secretion of submandibular glycoproteins not only in sections of fixed glands but also in glandular extracts and in nerve-induced saliva, after electrophoretic separations and immobilization in nitrocellulose membranes. In cats the glycoproteins in sympathetic saliva differed considerably from those in parasympathetic saliva. In sympathetic saliva they were found to originate mainly from striated ducts, to some extent from demilunar cells and to a small extent from acinar cells, whereas in parasympathetic saliva they arose mainly from acinar cells and demilunes and only to a small extent from striated ducts. In rat submandibular glands sympathetic stimulation caused extensive depletion of lectin stainable granules from granular tubules. Corresponding strong binding occurred with the same lectins to constituents in saliva that ran between 25 and 35 kD on SDS gel electrophoresis and were shown to contain tissue kallikreins. Their binding patterns suggested that individual kallikreins from the same gland may be glycosylated in different ways. This possibility was tested on five different kallikreins after separation from submandibular extracts by isoelectric focussing. Lectin bindings on slot blot preparations of these kallikreins were tested before and after N-glycosidase F, sialidase or endo-alpha-N-acetylgalactosaminidase digestions. Results showed that, despite their close genetic and structural similarities, the kallikreins are in fact differently sialylated and fucosylated and the novel finding that some contain O-glycosidically linked side chains as well as the anticipated N-glycosidically linked side chains was revealed. Thus, correlative histochemical and biochemical assessments of bindings with lectin probes has provided important new information about differences in the glycosylation patterns of individual glycoproteins stored within the same secretory granules.

Apical secretion of rat submandibular tissue kallikrein continues in the absence of external stimulation: evidence for a constitutive secretory pathway.

Sequential samples of saliva evoked by low frequency parasympathetic nerve stimulation were collected from cannulated submandibular ducts of anaesthetized rats following periods of varying length without stimulation. After such rest periods the first samples of saliva were found to contain much higher levels of tissue kallikrein (rK1) activity and protein than the ensuing samples from the same stimulation period and the latter contained levels similar to those found previously in parasympathetically-evoked saliva. rK1 activity in first samples increased with the length of the preceding interval, indicating that a continuous secretion of rK1 occurs from ductal cells into glandular lumina and accumulates there in the absence of stimulation or fluid secretion. The protein secretory pattern following rest pauses was unaffected by alpha-adrenergic receptor blockade and a high percentage of the proteinase activity was resistant to soya bean trypsin inhibitor, showing that the secretion was not caused by exocytosis of storage granules from ductal cells and therefore was likely to be a result of a constitutive secretion of newly synthesized enzyme. The pattern of continuous secretion into lumina detected for total protein in the absence of stimulation, suggests that other secretory proteins may also be secreted similarly but at different rates.

Differences in the glycosylation of rat submandibular kallikreins.

The glycosylations of five different rat submandibular kallikreins, rK1, rK2, rK7, rK9 and rK10, vacuum-blotted onto nitrocellulose membranes, have been studied by means of labelled lectins using enhanced chemiluminescence detection. The results demonstrated the individual submandibular kallikreins are not heavily glycosylated in rats, but consistently show different patterns of glycosylation. Following digestion of slot-blotted enzymes with peptide-N-glycosidase F (PNGase): binding by lectin from Lens culinaris (alpha Man-directed) was abolished, whilst that of lectin from Maclura pomifera (Gal beta 1,3GalNAc-directed) persisted (but could be abolished by periodate oxidation and endo-alpha-N-acetylgalactosaminidase digestion), revealing that there are O- as well as N-linked sugar chains on the kallikreins; a novel observation for this family of enzymes. The presence of GalNAc in addition to GlcNAc, Fuc, Gal, and Man, in sugar chains of rK1 was confirmed by high pH anion exchange chromatography following acid hydrolysis. Different intensities of binding by lectin from Limax flavus (NeuNAc-directed) suggest that sialylation of individual kallikreins differs, whilst sialidase and PNGase digestions suggest that sialic acid is the terminal residue of some N-linked but not O-linked structures.