Specific Localization of Membrane Dipeptidase and Dipeptidyl Peptidase IV in Secretion Granules of Two Different Pancreatic Islet Cells

Abstract

Endocrine cells require several protein convertases to process the precursors of hormonal peptides that they secrete. In addition to the convertases, which have a crucial role in the maturation of prohormones, many other proteases are present in endocrine cells, the roles of which are less well established. Two of these proteases, dipeptidyl peptidase IV (EC 3.4.14.5) and membrane dipeptidase (EC 3.4.13.19), have been immunocytochemically localized in the endocrine pancreas of the pig. Membrane dipeptidase was present exclusively in cells of the islet of Langerhans that were positive for the pancreatic polypeptide, whereas dipeptidyl peptidase IV was restricted to cells positive for glucagon. Both enzymes were observed in the content of secretory granules and therefore would be released into the interstitial space as the granules undergo exocytosis. At this location they could act on secretions of other islet cells. The relative concentration of dipeptidyl peptidase IV was lower in dense glucagon granules, where the immunoreactivity to glucagon was higher, and vice versa for light granules. This suggests that, in A-cells, dipeptidyl peptidase IV could be sent for degradation in the endosomal/lysosomal compartment during the process of granule maturation or could be removed from granules for continuous release into the interstitial space. The intense proteolytic activity that takes place in the endocrine pancreas could produce many potential dipeptide substrates for membrane dipeptidase.

Keywords

endocrine pancreas dipeptidyl peptidase IV membrane dipeptidase α-cells PP cells secretion granule glucagon pancreatic polypeptide

Endocrine cells require an elaborate set of proteolytic convertases for the maturation of their secreted hormones because all of them are synthesized in the form of precursors. Most of the endoproteases directly involved in the maturation of hormone precursors are well characterized (see Halban and Irminger 1994 for review). However, several additional proteases present in endocrine cells have not been clearly identified with a precise role in endogenous hormone maturation. Dipeptidyl peptidase IV (DPP-IV; CD26; EC 3.4.14.5), which is specifically localized in A-cell secretory granules (Poulsen et al. 1993), is an example of one such protease. DPP-IV is a widely expressed glycosylated ectoenzyme that preferentially cleaves dipeptides from the N-terminus of peptides when Pro is the P1 residue (see Trugnan et al. 1997 for review). Pro may be replaced by Ala and, to a lesser extent, by Ser, Gly, Val, and Leu. DPP-IV is a serine protease that has been shown to cleave a variety of biologically active peptides, including substance P, β-casomorphin, members of the vasoactive polypeptide-glucagon family (gastric inhibitory peptide, truncated glucagon-like peptide-1), and the neuropeptide Y-peptideYY family (Frohman et al. 1989; Mentlain et al. 1993; Mentlein et al. 1993; Medeiros and Turner 1996). Another widely distributed ectoenzyme is membrane dipeptidase (MDP; leukotriene D4 hydrolase; EC 3.4.13.19) (see Keynan et al. 1996 for review). MDP cleaves dipeptides with a broad spectrum of specificity and is involved in the in vivo metabolism of glutathione and leukotriene D4. MDP is a zinc metalloenzyme that is attached to the plasma membrane by a glycosyl-phosphatidylinositol (GPI) anchor (Keynan et al. 1996). In previous studies we have reported that MDP is present in the pancreatic acinar cell membranes, with a particularly high level in the secretory granule membrane (Hooper et al. 1997a; LeBel et al. 1998). Although MDP is synthesized in the rough endoplasmic reticulum, where it is attached to a GPI anchor, the protein is released from the membrane in the granule matrix to end up in exocrine secretions (Hooper et al. 1997a). The role of MDP in the exocrine pancreas is not immediately apparent, because secretory products do not require proteolytic maturation before their discharge into the duodenum.

In this study we confirm that DPP-IV is strictly confined to pancreatic A-cells and that MDP is also present in only one type of islet of Langerhans cell in the pig, the pancreatic polypeptide (PP) cells. Like DPP-IV, MDP is also present in the content of secretion granules. Finally, both enzymes were observed in granules undergoing exocytosis for their release within the islet interstitial space, where they could act on secretory products of neighboring cells.

The location of MDP in the endocrine pancreas is better justified than in the exocrine pancreas, because intense proteolytic processing takes place in the islets, thus producing numbers of potential substrates.

Materials and Methods

Antibodies

Antibodies were generated against affinity-purified pig kidney MDP and against pig kidney DPP-IV as previously described (Littlewood et al. 1989; Howell et al. 1993). Antibodies specific to porcine glucagon (NCL-GLUCp) were from Novo Castra Labs (Newcastle, UK). Antibodies to somatostatin (S309-JA3-81), a kind gift of Dr. R. Benoit (McGill University, Montreal, Canada), were previously characterized as specific for somatostatin 28 (Ravazzola et al. 1983). Antibodies to serotonin (B56-1) and PP (B32-1) were from Euro-Diagnostica (Malmö, Sweden). Anti-insulin antibodies (A564) were from Dako (Carpinteria, CA). All the primary antibodies were produced in rabbits.

Tissue Fixation and Embedding

Immediately after sacrifice, pieces of tissue (1 mm³) were taken from different regions of the pancreas of a 47-day-old pig (20 kg) and were fixed at room temperature (RT) for 120 min in 2% formaldehyde, 0.25% glutaraldehyde buffered with 100 mM PIPES, pH 7.4. After washing in the same buffer, samples were progressively dehydrated in ethanol at −35C and embedded at the same temperature in Unicryl for 24 hr. Polymerization was brought about by diffuse UV irradiation at 360 nm for 24 hr at −35C, 24 hr at −20C, 24 hr at −10C, 24 hr at 4C, and finally 24 hr at RT. Experimental procedures on animals in this study were performed in compliance with Canadian Council on Animal Care guidelines.

Immunocytochemistry

Thin sections were immunocytochemically stained with protein A-gold using 10-nm gold particles at a dilution of OD_{520 nm} = 0.1 in PBS for 25 min at RT as previously reported (LeBel et al. 1998) and were counterstained with uranyl acetate and lead citrate for observation by electron microscopy. Three control incubations were performed: without primary antibodies, with preadsorbed antibodies, and with preimmune serum. The antigens used for preadsorption were the corresponding porcine hormones, or kidney microvilli in the case of MDP and DPP-IV antibodies. Under these control conditions, only a few randomly distributed gold grains were observed. Quantitative evaluation of labeling concentration over secretory granules was carried out by counting gold particles over 15 granules. Differences between groups were tested with Student's t-test and only significant differences (p>0.005) are mentioned. For double staining, sections labeled with one primary antibody (anti-MDP or anti-DPP-IV) and protein A-gold particles of 10 nm were submitted to controlled silver enhancement (AURION R-Gent Silver Enhancement Kit; Cedarlane Labs, Hornby, ON, Canada) for 3 min. Sections were rinsed and submitted to immunocytochemical staining with the second antibody (anti-PP or anti-glucagon) using 10-nm colloidal gold particles.

Results

Immunocytochemistry of MDP in the exocrine pancreas gave a clear localization of the protein in the membranes of the acinar cell secretory pathway, and particularly in the granule membrane and content (LeBel et al. 1998). In the endocrine portion of the pancreas, MDP was restricted to only one cell type, a type that is not a major constituent of the islet of Langerhans. As shown in Figure 1, the identity of the cell could not be determined from the morphology of its granules, one of the reasons being that the morphology of the pig endocrine pancreas differs markedly from that of rodents. For example, the typical morphology observed in rodent B-cell β-granules, with a dense core surrounded by a clear halo, was not observed in the pig endocrine tissue (Figure 2a). The morphology of the pig β-granule is very similar to that of the human (Like and Orci 1972). To determine the identity of the MDP-positive cell, immunocytochemical localization of five endocrine secretory products was carried out in the pig tissue. Figures 2b and 2c show that only A-cells have granules with a prevailing round morphology. Granules of the other cell types have a less homogeneous morphology, like those of D-cells (Figure 2d), PP cells (Figure 2e), and enterochromaffin cells (EC cells) (Figure 2f). As previously shown by Poulsen et al. (1993), we were able to confirm that the only endocrine pancreatic cell expressing DPP-IV was the A-cell (Figure 2c). DPP-IV is also totally absent from the exocrine pancreas (not shown).

Double labeling immunocytochemistry of MDP separately with each one of the five endocrine pancreatic secretory products led to the exclusive co-localization of MDP and PP (Figure 3). In PP cells, MDP could be observed in structures that may represent poorly preserved Golgi cisternæ (Figure 3b), in the granule contents, and in granules very close to the membranes, and could therefore be undergoing exocytosis (Figure 3c). In all these locations, MDP was mostly present in the lumen of these organelles (Figures 3a and 3c) and was consequently seen secreted along with PP (Figure 3c). The presence in the lumen of these organelles of a presumably soluble form of a GPI-anchored membrane protein is somewhat unexpected but has previously been observed in the exocrine pancreas with GP-2 (Paquette et al. 1986) and MDP (Hooper et al. 1997a; LeBel et al. 1998). Therefore, it was interesting to examine the localization of a differently anchored membrane protein.

Figure 1

Localization of MDP in the pig endocrine pancreas. Under low magnification (a) MDP immunoreactivity is confined to only one cell type (arrow) distinct in morphology from adjacent A-cells, B-cells, and from another unidentified cell (^∗) (b) An enlarged area of the positive cell in a shows MDP immunoreactivity associated with the secretory granule matrix. Labeling is uniform from one granule to another. Note the low background of labeling in the adjacent cell (^∗). Nu, nucleus.

Localization of DPP-IV, which is a Type II membrane protein anchored by an uncleaved signal pep-tide, was performed. Figure 4 shows that, like MDP in PP cells, DPP-IV is present in the lumen of A-cell granules. The concentration of gold particles per μm² for DPP-IV was separately determined for granules that were electron-lucent and those that were electron-dense. The concentration was 195.24 ± 15.60/μ² in lucent granules, and 145.64 ± 15.25/μm² in dense granules, showing a significant decrease of DPP-IV immunoreactivity in dense granules. This decrease in reactivity could be due to masking of the antigen in denser granules but, on the contrary, glucagon immunoreactivity in the same granules increased very significantly. The concentration of gold particles for glucagon was 42.71 ± 7.47/μm² in electron-lucent granules, and 485.72 ± 29.28/μm² in electron-dense granules. Such a positive concentration of glucagon immunolabeling over dense granules suggests that maturation of the granules is taking place, as previously observed in the exocrine pancreatic secretory pathway (Bendayan 1984). Therefore, in contrast to glucagon, DPP-IV did not undergo concentration in the A-cell dense granules.

Figure 2

Characterization of five pig pancreatic endocrine cell types. To clearly identify the MDP-positive cell, antibodies to five secretory products of pancreatic islet cells were used to distinguish the morphology associated with the secreted hormone. Anti-insulin strongly labeled B-cell granules (a) anti-glucagon A-cell granules (b) anti-somatostatin-28 D-cell granules (d) anti-pancreatic polypeptide PP-cell granules (e) and anti-serotonin EC cell granules (f). (c) The specific immunolocalization of DPP-IV in A-cells (Poulsen et al. 1993) was confirmed, showing that the labeling is low in some granules (arrowhead, winding arrow) and high in others (arrow). In D-cells, round dense granules (large arrow) and irregular light granules (small arrow) were similarly labeled for somatostatin-28.

Figure 3

Co-localization of MDP and PP in pig pancreatic islet cells. (a) Double immunolabeling of PP cells with MDP antibodies (large particles) and PP antibodies (small particles) was performed. Most granules were positive for MDP (arrows). (b) Arrowheads show MDP in structures that could be Golgi cisternae, most of the signal being in the lumen of the cisternæ and little being associated with the membrane. (c) Granules in the close vicinity of the plasma membrane that may be in the process of exocytosis (arrows) are positive for PP and MDP. Nu, nucleus.

Figure 4

Association of DPP-IV with A-cell electron-lucent granules. (a) Glucagon immunodetection in A-cells was much more intense in electron-dense granules (arrows) compared to electron-lucent granules (winding arrows). Conversely, immunocytochemical detection of DPP-IV in A-cells (b) was much stronger and uniform in lucent granules (winding arrows) than in dense granules (arrows), where it was very heterogeneous. (c) Co-localization of DPP-IV (large particles) and glucagon (small particles) in A-cells confirmed that light granules have a much higher signal for DPP-IV than glucagon (bent arrow). Many granules positive for glucagon and DPP-IV are very close to the membrane and could be in the process of exocytosis (arrows). A B-cell with a granule located close to the plasma membrane (winding arrow) shows the specificity of detection for DPP-IV and glucagon, as well as the low level of background.

In divergence with a previous report in the mouse (Rombout et al. 1987), no PP immunoreactivity could be detected in the pig A-cell granules (not shown).

Discussion

Maturation and processing of hormone precursors in endocrine cells are performed by a number of exoand endoproteolytic proteases (Halban and Irminger 1994). In this respect, the endocrine pancreas is a particularly rich gland, secreting five different secretory products that are synthesized by five specialized cells. In addition to being endowed with the well-described protein convertase family of endoproteases, carboxypeptidases and the ubiquitous furin (Hook et al. 1994), the A-cell specifically contains DPP-IV in its secretory granules (Poulsen et al. 1993). Here we show by immunocytochemistry that DPP-IV is mostly soluble in granules of A-cells and consequently is secreted in the interstitial space of the islets of Langerhans. DPP-IV releases dipeptides from the N-terminus of proteins. Its cleavage specificity requires a Pro residue (or exceptionally Ala) to be at the C-terminus of the released dipeptide (see Trugnan et al. 1997 for review). According to its specificity, primary substrates for the peptidase in the α-granule would be glucagon-like peptide-2 and the truncated glucagon-like peptide-1, two C-terminal byproducts of proglucagon processing. Although both peptides are good substrates for DPP-IV in vivo (Kieffer et al. 1995; Drucker et al. 1997), the acidic pH of the granule would make their hydrolysis by DPP-IV less likely in this location. The secretory nature of DPP-IV that we report, and the optimal pH of its activity, suggest that the protease could act on the secretory products released into the interstitial space by any of the five types of endocrine cells. In addition to the previously mentioned glucagon maturation byproducts, two other secretory products could be substrates for DPP-IV. These are the propeptides of somatostatin and PP that are specifically secreted by D-cells and PP-cells, respectively. In support for such a role of the secreted DPP-IV on PP are reports that the major form of neuropeptide Y immunoreactivity in the circulation is a 34-residue peptide constituted of residues 3–36 (Medeiros and Turner 1996). Neuropeptide Y is a peptide very similar in sequence and function to PP, which possesses the consensus Ala-Pro at the N-terminus for recognition by DPP-IV. For the somatostatin propeptide, there is only one report for such trimming of its N-terminus (Baldissera 1994). Finally, the release of DPP-IV in the interstitial space also suggests a role for the protease in the inactivation of cytokines and chemokines, many of which have the N-terminal consensus sequence (X-Pro) (Vanhoof et al. 1995). Evidence for an involvement of DPP-IV in anti-inflammatory and antiviral responses has accumulated (Proost et al. 1998), supporting such a role for DPP-IV in the endocrine pancreas.

The release of DPP-IV from the membrane is not a new feature for this protease. It has long been known that acidic conditions favor autolysis of the protein (Macnair and Kenny 1979) and that a soluble form of the protein is present in the circulation (Trugnan et al. 1997). The simple mode of membrane anchorage of DPP-IV by an uncleaved signal peptide reduces the task of releasing the peptidase to the hydrolysis of a single peptide bond. A number of proteins anchored by a single membrane-spanning polypeptide are also known to exist in soluble forms as a result of proteolytic cleavage by a family of proteases termed secretases (Hooper et al. 1997b). It is very likely that the conditions of pH found in vivo in mature [pH 5.5 according to Dittié et al. (1997)] and most particularly in immature granules, where it is lower (Orci et al. 1987), are close to those used in vitro (pH 4.5–5.0) to generate a soluble form of DPP-IV by autolysis (Macnair and Kenny 1979). The higher amount of DPP-IV in electron-lucent granules compared to electron-dense ones is a very interesting finding regarding the biogenesis of secretory granules. If the lower level of DPP-IV immunodetection in dense granules was not due to antigen masking, this suggests that the protein is sorted out of immature granules during maturation, a phenomenon that has been observed with proteins that are soluble under the acidic conditions of granules and that have weak or no affinity for the aggregated cargo. This pathway is called the constitutive-like secretory pathway. It was elegantly demonstrated with the insulin C-peptide (Arvan et al. 1991). The pathway operates when granules are in the process of maturation. According to this scenario, DPP-IV could then be released almost continuously from A-cells. An alternative to the former pathway to explain the decrease of DPP-IV in dense granules could be a degradation of DPP-IV in the endosomal/lysosomal compartment.

Like DPP-IV, MDP is also secreted in the interstitial space. To find a substrate for MDP is easier because the only identifiable one in the endocrine pancreas would be the dipeptide made of the two basic residues that specify the cleavage site of convertases. However, the release of these two basic residues from the C-terminal end of these proteolytic fragments has thus far been attributed to sequential hydrolysis by carboxypeptidases (Loh et al. 1993). Because exocrine pancreatic cells are capable of a full range of prohormone processing reactions without any need for it (Dickinson et al. 1993), it would not be surprising that in some endocrine cells MDP could be an example of such a remaining activity without real utility in processing endogenous secretory products. As for the release of MDP into the endocrine granule lumen, it is quite conceivable that a phospholipase could be involved in cleaving its GPI anchor and releasing the protein from the membrane. A phospholipase A activity has indeed been identified in the exocrine pancreatic acinar cell granule membrane that promotes such a release of MDP (Hooper et al. 1997a; LeBel et al. 1998), but its purification and characterization have not yet been achieved.

Footnotes

Acknowledgements

Supported by the Canadian Cystic Fibrosis Foundation, NSERC (Canada), FCAR (Québec) to DL, by the Medical Research Council of Great Britain to NMH, and by a NATO Collaborative Research Grant.

References

Arvan

Kuliawat

Prabakaran

Zavacki

Elahi

Wang

Pilkey

(1991) Protein discharge from immature secretory granules displays both regulated and constitutive characteristics. J Biol Chem 266:14171–14174.

Baldissera

(1994) Differential processing of regulatory peptide precursors in pancreas and gut: studies of proglucagon and prosomatostatin processing. Danish Med Bull 41:79–92.

Bendayan

(1984) Concentration of amylase along its secretory pathway in the pancreatic acinar cell as revealed by high resolution immunocytochemistry. Histochem J 16:85–108.

Dickinson

Takeuchi

Guo

Y-J

Stadler

Yamada

(1993) Expression and processing of prohormones in nonendocrine cells. Am J Physiol 264:G553–560.

Dittié

Thomas

Tooze

(1997) Interaction of furin in immature secretory granules from neuroendocrine cells with the AP-1 adaptor complex is modulated by casein kinase II phosphorylation. EMBO J 16:4859–4870.

Drucker

Shi

Crivici

Sumner-Smith

Tavares

Hill

DeForest

Cooper

Brubaker

(1997) Regulation of the biological activity of glucagon-like peptide 2 in vivo by dipeptidyl peptidase IV. Nature Biotechnol 15:673–677.

Frohman

Downs

Heimer

Felix

(1989) Dipeptidyl peptidase-IV and trypsin-like enzymatic degradation of human growth hormone-releasing hormone in plasma. J Clin Invest 83:1533–1540.

Halban

Irminger

(1994) Sorting and processing of secretory proteins. Biochem J 299:1–18.

Hook

VYH

Azaryan

Hwang

S-R

Tezapsidis

(1994) Proteases and the emerging role of protease inhibitors in prohormone processing. FASEB J 8:1269–1278.

10.

Hooper

Cook

Lainé

LeBel

(1997a) Identification of membrane dipeptidase as a major glycosyl-phosphatidylinositol-anchored protein of the pancreatic zymogen granule membrane, and evidence for its release by phospholipase A. Biochem J 324:151–157.

11.

Hooper

Karran

Turner

(1997b) Membrane protein secretases. Biochem J 321:265–279.

12.

Howell

Brewis

Hooper

Kenny

Turner

(1993) Mosaic expression of membrane peptidases by confluent cultures of Caco-2 cells. FEBS Lett 317:109–112.

13.

Keynan

Hooper

Turner

(1996) Molecular and functional aspects of membrane dipeptidase. In Hooper

, ed. Zinc Metalloproteases in Health and Disease. London, Taylor & Francis, 285–309.

14.

Kieffer

Mcintosh

CHS

Pederson

(1995) Degradation of glucose-dependent insulinotropic polypeptide and truncated glucagon-like peptide 1 in vitro and in vivo by dipeptidyl peptidase IV. Endocrinology 136:3585–3596.

15.

LeBel

Grondin

Cook

Hooper

(1998) Membrane dipeptidase in the pig exocrine pancreas: ultrastructural localization and secretion. J Histochem Cytochem 46:841–846.

16.

Orci

(1972) Embryogenesis of the human pancreatic islets: a light and electron microscopic study. Diabetes 21:511–534.

17.

Littlewood

Hooper

Turner

(1989) Ectoenzymes of the kidney microvillar membrane. Affinity purification, characterization and localization of the phospholipase C-solubilized form of renal dipeptidase. Biochem J 257:361–367.

18.

Loh

Beinfeld

Birch

(1993) Proteolytic processing of prohormones and proneuropeptides. In Loh

, ed. Mechanisms of Intracellular Trafficking and Processing of Proproteins. Boca Raton, FL, CRC Press, 179–224.

19.

Macnair

Kenny

(1979) Proteins of the kidney microvillar membrane. The amphipathic form of dipeptidyl peptidase IV. Biochem J 179:379–395.

20.

Medeiros

Turner

(1996) Metabolism and functions of neuropeptide Y. Neurochem Res 21:1125–1132.

21.

Mentlain

Gallwitz

Schmidt

(1993) Dipeptidyl peptidase-IV hydrolyses gastric inhibitory peptide (GIP), glucagon-like peptide-1(7-36) amide, peptide histidine methionine (PHM) and is responsible for their degradation in human serum. Eur J Biochem 214:829–835.

22.

Mentlein

Dahms

Grandt

Kruger

(1993) Proteolytic processing of neuropeptide Y and peptide YY by dipeptidyl peptidase IV. Regul Pept 49:133–134.

23.

Orci

Ravazzola

Anderson

RGW

(1987) The condensing vacuole of exocrine cells is more acidic than the mature secretory vesicle. Nature 326:77–79.

24.

Paquette

Leblond

Beattie

LeBel

(1986) Reducing conditions induce a total degradation of the major zymogen granule membrane protein in both its membranous and its soluble form. Immunochemical quantitation of the two forms. Biochem Cell Biol 64:456–462.

25.

Poulsen

Hansen

Dabelsteen

Hoyer

Norén

Sjöström

(1993) Dipeptidyl peptidase IV is sorted to the secretory granules in pancreatic islet A-cells. J Histochem Cytochem 41:81–88.

26.

Proost

De Meester

Schols

Struyf

Lambeir

Wuyts

Opdenakker

De Clercq

Scharpe

Van Damme

(1998) Amino-terminal truncation of chemokines by CD26/dipeptidylpeptidase IV−conversion of RANTES into a potent inhibitor of monocyte chemotaxis and HIV-1-infection. J Biol Chem 273:7222–7227.

27.

Ravazzola

Benoit

Ling

Guillemain

Orci

(1983) Immunocytochemical localization of prosomatostatin fragments in maturing and mature secretory granules of pancreatic and gastrointestinal D cells. Proc Natl Acad Sci USA 80:215–218.

28.

Rombout

Abad

Binkhorst

Taverne-Thiele

(1987) Coexistence of pancreatic polypeptide (PP)- and glucagon-immunoreactivity in pancreatic endocrine cells of mouse. Histochemistry 87:1–6.

29.

Trugnan

Aït-Slimane

David

Baricault

Berbar

Lenoir

Sapin

(1997) Dipeptidyl peptidase IV (DPP-IV/CD26): biochemistry and control of cell-surface expression. In Kenny

Boustead

, eds. Cell-surface Peptidases in Health and Disease. Oxford, BIOS Scientific Publishers, 203–217.

30.

Vanhoof

Goossens

De Meester

Hendriks

Scharpe

(1995) Proline motifs in peptides and their biological processing. FASEB J 9:736–744.