Abstract
Chromogranin (Cg) B is an acidic glycoprotein present in neuroendocrine tissue. The sequence shows several dibasic amino acid positions susceptible to proteolytic cleavage. The purpose of this study was to elucidate the expression of CgB epitopes in the human endocrine pancreas. Tissue sections of six human pancreata were immunostained with 16 different region-specific antibodies to the CgB molecule, using double immunofluorescence techniques. The CgB epitope pattern varied in the four major islet cell types. B (insulin)-cells expressed immunoreactivity to all region-specific antibodies. The antibodies to the N-terminal and mid-portions of CgB showed moderate immunoreactivity, the C-terminal antibodies weak. A (glucagon)-cells were reactive only to the N-terminal and mid-portion antibodies but, after microwave pretreatment, to all antibodies, whereas D (somatostatin)-cells expressed only the sequence CgB 244–255 and a subpopulation CgB 580–595. PP (pancreatic polypeptide) cells were immunostained with antibodies between CgB 1–417 and a few with CgB 580–593. The fragment CgB 244–255 was expressed in all four cell types. The cause of these differences may be cell-specific cleavage or masking of the molecule, but varying translation of CgB mRNA is also possible. The extent to which these epitopes reflect fragments having biological functions remains to be evaluated.
C
Most studies on CgB have been focused on biochemical analysis but only a few studies have centered on its cellular localization and function. Post-translational processing of CgB has been reported in various NE tissues from humans and other species. Biochemical characterization of plasma and tissue extracts has shown that CgB processing varies in brain and in adrenal and pituitary glands (Iguchi et al. 1988; Gill et al. 1991; Winkler et al. 1998; Marksteiner et al. 1999). Immunohistochemical (IHC) analyses of CgB have been performed on human endocrine pancreas with antibodies raised to three CgB sequences (Bishop et al. 1989; Schmid et al. 1989; Kimura et al. 2000). Buffa et al. (1989) also applied two monoclonal antibodies with unknown epitopes. The staining results varied with the antibodies used.
We have recently studied the IHC expression and processing of CgA in the different cell types in human pancreatic islets by using 12 region-specific antibodies (Portela–Gomes and Stridsberg 2001). In the present investigation we systematically examined the CgB-immunoreactive patterns of the four major NE cell types of human pancreas, using antibodies raised against 16 different regions of the human CgB molecule. The antibody-binding regions were chosen to display possible processing of the CgB molecule, i.e., the localization of the potential cleavage sites was important for the selection of the antibodies. Furthermore, knowledge about isolated CgB-related peptides, with or without known biological activity, and conserved interspecies amino acid homology was also of importance for the choice of the epitopes for antibody production. Chromogranin-derived peptides have been postulated to take part in the secretory granule generation and function, and CgB fragments may possess biological activity. Therefore, the present study, using CgB region-specific antibodies, can increase the knowledge about the expression and/or possible processing of the molecule in normal islet cells, thus providing additional insight into the functional role of CgB.
Materials and Methods
Tissue specimens from six adult human pancreata were obtained from surgical samples removed at surgery for pancreatic adenocarcinoma. The specimens were taken from both macro- and microscopically normal glandular regions at least 3 cm distant from the neoplasm. Four of the specimens were taken from the body–tail region and two from the head (processus uncinatus).
The pancreatic specimens were fixed for 18–20 hr at room temperature (RT) in 10% buffered neutral formalin, then processed routinely to paraffin. Sections 5 μm thick were cut and attached to poly-
Immunohistochemistry
The streptavidin–biotin complex (ABC) technique (Hsu et al. 1981) with diaminobenzidine as chromogen was applied as a single immunostain to reveal the distribution pattern of endocrine cells in the pancreas and to perform the control stainings specified below. Some sections were pretreated in a microwave oven (Philips Whirlpool Nordic; Stockholm, Sweden) twice for 5 min at 750 W, using a citrate buffer, pH 6.0, as retrieval solution.
Double immunofluorescence techniques were used to identify the CgB epitopes in the various cell types of the endocrine pancreas. The sections were incubated overnight at RT with a cocktail of two antibodies, either one monoclonal and one polyclonal antibody or two polyclonal antibodies raised in different animal species (anti-rabbit, anti-guinea pig, and/or anti-chicken). Then the sections were incubated for 30 min at RT with biotinylated swine anti-rabbit IgG and then transferred to a mixture of streptavidin–Texas Red and fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse, anti-guinea pig, or anti-chicken IgG. Before application of the respective primary antibodies, the sections were incubated with non-immune serum from the animal species producing the secondary antibodies, at a dilution of 1:10. The secondary antibody in question was pre-incubated overnight at 4C with 10 μl/ml normal serum, both from the animal species recognized by the other secondary antibody and from the species producing that antibody. Between each staining step, the sections were carefully washed with PBS.
The CgB region-specific antibodies used are characterized in Table 1. The other primary antibodies were as follows: mouse monoclonal antibodies against human somatostatin (Novo Nordisk, Bagsvaerd, Denmark; clone Som-018); sheep polyclonal antibodies against human glucagon (The Binding Site, Birmingham, UK; code no. PH519); guinea pig antibodies against human insulin (P. Westermark, Dept. Gen. Pathology, Unit of Pathology, Uppsala, Sweden; code no. Ma 47); and chicken antibodies against human pancreatic polypeptide (A. Larsson, Dept. Med Sci., Clin. Chemistry; Uppsala, Sweden). The working dilutions for immunofluorescence were 1:80, 1:10, 1:80, 1:200, and 1:80, respectively.
Chromogranin B region-specific antibodies used; all antibodies were raised in rabbit
aDilutions used in immunofluorescence staining; for the ABC staining, the dilutions were 10- to 20-fold higher.
The labeled secondary antisera were as follows: biotinylated swine anti-rabbit IgG (DAKO; Glostrup, Denmark); Texas Red-labeled streptavidin, FITC-conjugated goat anti-sheep (Vector Laboratories; Burlingame, CA); FITC-conjugated goat anti-mouse, anti-guinea pig, and anti-chicken IgG (Sigma Chemical; St Louis, MO).
The control stainings entailed (a) omission of the primary antisera, (b) replacement of the first layer of antibody by non-immune serum diluted 1:10 and by the diluent alone, or (c) preincubation (24 hr) of primary antiserum with the relevant antigen (10 nmol per ml diluted antibody solution, respectively) before application to the sections. The secondary antibodies were also tested in relation to the specificity of the species in which the primary antibodies had been raised, by using single staining, the secondary antibody in question being replaced by secondary antibodies from different animal species, e.g., a primary antibody raised in guinea pig was tested with an anti-rabbit-labeled secondary antibody. These control tests were performed with ABC (single staining) and immunofluorescence techniques (co-localization studies). Neutralization tests were performed to exclude crossreactivity among the CgB region-specific antibodies and between these and chromogranin A, insulin, glucagon, somatostatin (amino acid sequence 1–14), and pancreatic polypeptide. The CgB fragment antigens used were the present synthesized peptides (see below). The other antigens used were obtained from Sigma.
For co-localization studies, the sections were examined in a Vanox AHBS3 fluorescence microscope (Olympus; Tokyo, Japan) equipped with filters (Olympus) producing excitation at wavelengths 475–555 nm for Texas Red (filter no. 32821, dichroic mirror BH2-DMG), and 453–488 nm for FITC (no. 32822, BH2-DMIB), respectively, while a double-band filter set (no. 39538, BH2-DFC5) was also used for simultaneous visualization of Texas Red- and FITC-labeled cells (excitation at 550–570 nm and 480–495 nm, respectively). Photographs were taken with Fujicolor 400 film.
Antibody Production
Deduced from the amino acid sequences of human CgB (Benedum et al. 1987), 16 polypeptides were synthesized by a solid-phase system using Fmoc chemistry (Applied Biosystems model 430A; Foster City, CA). The peptides were purified by reversed-phase chromatography and analyzed by plasma desorption mass spectrometry (PDMS, Bioion 20; Bioion Nordic, Uppsala, Sweden). The sequences were selected to be specific for human CgB and the homology was less than 50% to any other known protein sequence in the database SWISS-PROT (release 37, December, 1998), (Shane S. Surrock and John F. Collins 1993; Biocomputing Research Unit, University of Edinburgh, Scotland), except for the respective sequences of CgB from other species. The selected sequences for the respective peptides are shown in Figure 1 and Table 2. To facilitate the coupling to the carrier protein (see below), a cysteine residue was added to either the C-terminal or the N-terminal of the synthesized peptides. To facilitate labeling of the peptides with iodine, a tyrosine residue was added in the position between the cysteine residue and the peptide if the CgB–peptide sequence lacked a tyrosine residue.
Before immunization, the peptides were coupled to a carrier protein. Peptides (1 mg) were dissolved in 100 μl DMSO, after which 100 μl (1 mg) of Imject Æ Maleimide Activated Keyhole Limpet Hemocyanin (aKLH; Pierce, Boule Nordic) was added. This mixture was allowed to react for 2 hr at RT. The coupled peptides were then purified on a PD-10 column (Pharmacia Biotech; Uppsala, Sweden) with PBS as the moving phase. Aliquots of 200 μg (n=5) coupled peptide were frozen and stored at −20C until immunization. The peptide complexes were injected into New Zealand White rabbits to produce polyclonal antibodies, as described earlier (Stridsberg et al. 1995).
Results
The immunostaining pattern with the different CgB region-specific antibodies in the endocrine pancreas varied among the different cell types (see below). Some minor variations in the staining intensity were noted among the various cases. No immunostaining was elicited in the exocrine tissue. Microwave pretreatment affected the intensity of the staining, whereby faint staining without such pretreatment became moderate after such pretreatment, especially the antibodies to CgB 153–167 and 259–270, which stained faintly without microwave pretreatment. The distribution pattern of the endocrine cells was not affected by microwave pretreatment, except for glucagon cells (see below).
Schematic representation of the human chromogranin B molecule, with the amino acid sequences of the 16 peptides used for raising antibodies placed in their appropriate positions (B1–16 to B647–657). The disulfide bridge (S) is shown on the N-terminal. The 15 potential cleavage sites (first amino acid of the dibasic pair) are indicated by vertical lines (the numbers indicate positions 45–633).
Co-localization of chromogranin B related peptides in the human pancreatic endocrine cells using region-specific antibodies a
a-, no immunoreactive cells; +, fewer than 10% positive cells; ++, 10–50%; +++, more than 50% positive cells.
∗Weak, ∗∗moderate, ∗∗∗strong intensity of immunoreactivity; MV, microwave pretreatment; no immunoreactivity was evident without this pretreatment.
In the pancreatic islets, the number of immunoreactive cells with the different CgB region-specific antibodies varied from a few to virtually all cells. The staining intensity also varied among the different antibodies. On the whole, the staining intensity with the antibodies to epitopes in the C-terminal sequences was weaker than that of those in the N-terminal and mid-portion sequences. Because the staining intensity with the antibodies to CgB 16–37 was too weak to allow double staining, these results are not presented. The immunostaining results are summarized in Table 2.
Immunoreactivity to the CgB Region-specific Antibodies in Different Endocrine Cell Types
B (Insulin)-cells. Virtually all B-cells expressed immunoreactivity to all CgB region-specific antibodies (Figures 2–4). The intensity of the immunoreactivity was moderate with most N-terminal and mid-portion antibodies but weak with the C-terminal antibodies.
A (Glucagon)-cells. The frequency of the A-cells displaying immunoreactivity to the different region-specific antibodies varied. Some N-terminal and mid-portion antibodies stained from a few to virtually all (Figures 5 and 6). The intensity of the immunoreactivity was moderate to strong. No immunoreactivity was observed with the region-specific antibodies from CgB 555 to the C-terminus unless microwave pretreated, when moderate to strong staining became evident.
D (Somatostatin)-cells. D-cells were non-immunoreactive to all CgB region-specific antibodies except the 244–255 antibody, which stained virtually all D-cells (Figure 7). A subpopulation of D-cells was also stained by the 580–595 antibody (Figure 8). The intensity of the immunoreactivity with both antibodies was strong.
PP (Pancreatic Polypeptide)-cells. PP cells were stained with antibodies raised to epitopes located in sequences 1 and 417 (Figure 9). A few PP cells were also immunostained by antibodies raised to 506–517 and 568–593 (Figure 10). The staining intensity with the different antibodies varied from weak to strong.
Control Stainings
In double immunofluorescence stainings, our control tests showed that omission of one of the primary antibodies gave a staining pattern corresponding to that of the remaining primary antibody. The other staining controls performed were all negative (Figure 11).
Discussion
CgB and CgA are two members of the same family. They are structurally closely related proteins present in most NE cell types. There is often co-expression of these two glycoproteins, although CgA frequently predominates in NE cells. In the present study, all region-specific antibodies to CgB revealed immunoreactive cells in the human pancreatic islets but the distribution patterns varied demonstrably. The antibodies raised against the N-terminal and mid-parts of the CgB molecule stained more cells and with a generally stronger immunoreactivity than did the C-terminal antibodies. B-cells reacted with all our region-specific antibodies, although the immunoreactivity was usually weak. A-cells showed strong immunoreactivity to the antibodies directed towards the N-terminal and mid-portion of the molecule. Without microwave pretreatment, no immunoreactivity to the C-terminal antibodies was seen, but after such treatment intense immunoreactivity was apparent. The presence of CgB immunoreactivity in human D-cells is reported for the first time. However, these cells reacted only with the mid-portion antibody (CgB 244–255), although a subpopulation of the cells was also immunostained with the C-terminal antibody (580–593). The N-terminal and mid-portion antibodies, as well as some C-terminal antibodies, immunostained the PP cells. Our results show similarities and differences in expression of both CgB and CgA (Portela–Gomes and Stridsberg 2001) in the pancreatic islet cells. Generally speaking, B- and A-cells expressed all CgA and CgB region-specific antibodies. Our previous study showed that D-cells reacted with only one of the 12 region-specific CgA antibodies, raised against an epitope in the C-terminus. In the present study, two epitopes of CgB were expressed in this cell type. Interestingly, one of our CgB antibodies (580–595) immunostained only a subfraction of the D-cells, indicating some heterogeneity of these cells. N-terminal and mid-portion CgB antibodies, as well as some C-terminal antibodies, were expressed in PP cells, whereas this cell type was reactive only to mid-portion CgA antibodies. It is interesting to note the presence of CgB 244–255 epitopes in all pancreatic cell types. In our previous CgA study, none of the epitopes was expressed in all cell types. This constant presence suggests that the CgB 244–255 part of the molecule may be of physiological importance.
The observed differences in staining pattern reflect the number of existing or available CgB epitopes. This indicates that, in some cells, the CgB molecule is cleaved in such a way that some of our antibodies do not bind to the molecule. This would be consistent with biochemical data of variable CgB processing in different NE cells, being more marked in brain than in adrenal glands but less marked in the pituitary glands (Iguchi et al. 1988; Gill et al. 1991; Winkler et al. 1998; Marksteiner et al. 1999). The nonexistent or weak reaction may also indicate that the epitopes can be masked by other granule-related proteins or by the histological processing. Microwave pretreatment influenced the results of the C-terminal antibodies in A-cells, indicating that the molecular C-terminus is demasked, because the staining appeared positive after this pretreatment. This microwave treatment increased the staining intensity but not the frequency of the other cell types. The lack of reaction may also be due to fragments present at too low a concentration for detection by the techniques used, or else are nonexistent. Prohormone convertases, predominantly PC2 and PC3 (PC3/PC1), are important for the processing of chromogranins (Laslop et al. 1998,2000). It is possible that variable expression of the two proconvertases may explain some of the differences in the expression of CgB fragments in the different islet cell types. Another thrilling explanation could be that only some molecular regions are translated from the mRNA. Knowledge about the gene structure for CgB is limited, but the CgA gene contains eight exons and seven introns (Mouland et al. 1994). The genes for human CgB display similar structure and the genes are thought to stem from a common ancestor gene (Wu et al. 1991). Therefore, there is a possibility that mRNA splicing and differential expression of individual exons may occur, which to some extent could explain the different immunostaining patterns in the different endocrine cell types.
Previous IHC studies using region-specific antibodies to CgB are sparse. GAWK (amino acid sequence CgB 420–493) immunoreactivity was detected by electron microscopy in A-cells by Bishop et al. (1989), in which more cells were stained with antibodies to the N-terminal than to the C-terminal amino acid sequences of the molecule. Our antibodies directed to the the N- and C-terminal GAWK (CgB 420–432 and 439–451) showed a similar immunostaining pattern in A-cells but, unlike Bishop et al. (1989), we also found weak GAWK immunoreactivity in some B-cells. This difference in immunostaining is probably related to the fixation/embedding procedure, which can impair antigenicity, or the possibility that fewer cells can be studied by electron microscopy. In addition, the antibodies used in the mentioned study were raised against rat CgB whereas ours were against human CgB. Although rat and human peptides both have a high degree of homology, this could account for some differences in the results. Another investigation reported CgB immunoreactivity in human islet cells (Buffa et al. 1989). These authors found that both PP- cells and A-cells were stained with the two CgB antibodies, but the staining intensity varied. However, it was not established to which part of the molecule these antibodies were bound. Compared with our results, those antibodies probably lay in the N-terminal or mid-portions. The other two IHC studies in human pancreas demonstrated the presence of human CgB 306–326 in B- and A-cells (Schmid et al. 1989) and of human CgB 548–564 in a few unspecified islet cells (Kimura et al. 2000). The amino acid sequences to which these antibodies were raised did not match our antibodies closely and our results were somewhat different.
The physiological functions of CgB and its fragments are still largely unknown. CgB has been reported to bind calcium (Gorr et al. 1989), to exert a trophic action in neurons (Chen et al. 1992), and to modulate sorting of peptide hormones to the secretory granules (Natori and Huttner 1996). Antibacterial functions have also been shown for two different fragments of CgB, secretolytin (Strub et al. 1995) and chrombacin (Metz–Boutigue et al. 1998). Recently, it was shown that antibodies to a central part of the CgB molecule, 312–331, could inhibit glucose-stimulated insulin release from isolated mouse pancreatic islets (Karlsson et al. 2000). This was the first time that CgB had been shown to have an inhibitory effect on hormone release. Based on the locations of dibasic amino acids, i.e., potential cleavage sites, it appears probable that the peptide responsible for inhibiting insulin release covers the CgB sequence 306–365. Interestingly, our CgB 312–331 antibodies were mostly expressed in B-cells, which may be related to the hormonal effect of this CgB fragment, whereas its expression occurred in only a few A- and PP-cells.
In conclusion, this study has shown a variable IHC expression pattern of CgB epitopes available in the endocrine pancreatic cell types, with different CgB epitopes available in the various cell types. This may be the result of different post-translational processing of the CgB molecule or different expression at the mRNA level, indicating that the fragments probably have different roles in the various cell types. Studies of the expression of CgB region-specific antibodies in different NE tumors and diabetic islets may provide some insight into the functional importance of these CgB fragments.
Human pancreatic islet double immunostained for (
Human pancreatic islet double immunostained for insulin (FITC) and CgB 542–552 (Texas Red). The figure demonstrates that this CgB fragment is present in most A-cells (yellow to yellow-green). Bar = 53 μm.
Human pancreatic islet double immunostained for glucagon (FITC) and CgB 244–255 (Texas Red), analyzed with the double-band filter set. The yellow-stained cells contain both the hormone and the CgB fragment. The absence of green-stained cells indicates that all A-cells contain this CgB fragment. Bar = 53 μm.
Human pancreatic islet double immunostained for glucagon (FITC) and CgB 439–451 (Texas Red). The yellow cells indicate co-localization of the two proteins. Some A-cells (green) do not display the CgB fragment (arrows). Bar = 53 μm.
Human pancreatic islet double immunostained for somatostatin (FITC) and CgB 244–255 (Texas Red). CgB immunoreactivity is evident in all D-cells (yellow). Bar = 53 μm.
Human pancreatic islet double immunostained for somatostatin (FITC) and CgB 647–657 (Texas Red). Immunoreactivity to this CgB fragment was only demonstrated in a fraction of D-cells (yellow), whereas other D-cells (green) were nonreactive (arrows). Bar = 80 μm.
Human pancreatic islet double immunostained for pancreatic polypeptide (FITC) and CgB 244–255 (Texas Red). All PP-cells show CgB immunoreactivity, illustrated by the yellow color. The central dark area does not show any PP-cells. A small amount of very faint CgB fragment staining was visualized with the single filter but not with the double filter. Bar = 80 μm.
Human pancreatic islet double immunostained for pancreatic polypeptide (FITC) and CgB 406–417 (Texas Red) showing co-localization of these two substances in only a fraction of PP-cells (yellow-green). Bar = 53 μm.
Human pancreatic islet immunostained with normal rabbit serum, showing no immunoreactivity. Bar = 53 μm.
Footnotes
Acknowledgements
Supported by a grant from the Swedish Cancer Foundation and by the Ihres Fund.
We thank Prof Lars Grimelius for fruitful discussions and laboratory facilities and Prof Henry Johansson for comments on the manuscript. We are also grateful for the PP antibodies raised in chicken, kindly provided by Assoc Prof Anders Larsson.