Dominant Expression of Chromogranin B in Pituitary Corticotrophs and Its Putative Role in Interaction With Secretogranin III

Summary

Previous studies have suggested that chromogranin A (CgA) is a partner molecule of secretogranin III (SgIII). In mouse pituitary corticotroph-derived AtT-20 cells, SgIII plays a role in sorting CgA/hormone aggregates into secretory granules (SGs). Although CgA expression is equivocal, CgB is clearly detectable in the rat pituitary corticotrophs. Therefore, we hypothesized that CgB shares a function with CgA in pituitary corticotrophs. In the binding assays, CgB, similar to CgA, showed binding activity to SgIII under weakly acidic conditions and in the presence of Ca²⁺. Considering the differences in animal species, the different abilities of antibodies, and the conditions of tissue fixation and thin sectioning in immunofluorescence histochemistry, we found that CgA was expressed in a small population (approximately 10%), and its expression intensity was weaker than that of CgB (>98%) in rodent pituitary corticotrophs. In addition, similar to CgA, CgB and SgIII were colocalized in adrenocorticotropic hormone (ACTH) granules. The labeling of CgA and CgB was not completely consistent, and CgB colocalized with SgIII in many granules. These results suggest that there are multiple sorting systems for ACTH granules in pituitary corticotrophs and that the SgIII/CgB complex behaves more dominantly than the SgIII/CgA complex, which has somewhat different properties:

Introduction

Granins are found in SGs containing peptide hormones and bioactive amines in endocrine and neural cells, and are known as proteins belong to a family of acidic proteins.^1–6 Chromogranin A (CgA), chromogranin B (CgB), secretogranin II (SgII), and secretogranin III (SgIII) have been extensively studied for their roles in the formation, sorting, and transport of SGs in peptide hormone-producing cells.^7,8 Granins are highly acidic proteins that are sorted into budding granules due to their property of selective aggregation at low pH in the presence of Ca²⁺ in the trans-Golgi network (TGN).^3,8,9–11

SgIII functions in the process essential for secretory granule (SG) biosynthesis owing to its diverse binding properties. The domain identified as amino acid residues (aa) 214–373 of rat SgIII binds strongly to the N-terminal domain of CgA (rat aa 48–111) in the presence of 10 mM Ca²⁺ at pH 5.5.^3,12 In addition, the N-terminal domain of SgIII (rat aa 23–186) specifically binds to cholesterol and localizes to the cholesterol-rich microdomains of TGN and SGs membranes in mouse pituitary corticotroph-derived AtT-20 cells.^3,13 Based on these findings, SgIII is thought to act as a receptor for sorting CgA/prohormone aggregates into SGs by anchoring them to the cholesterol-rich microdomains of the TGN membrane in the regulated secretory pathway.^8,14 The localization of SgIII was observed in the peripheral regions of SGs of rat pituitary mammotrophs and thyrotrophs, pancreatic α-, β-, and δ-cells by electron microscopic immunohistochemistry, suggesting that SgIII plays a role in sorting CgA/prohormone aggregates from TGN to SGs in peptide hormone-producing cells in vivo.^15,16

In the rat pituitary, CgA expression has been demonstrated using immunohistochemistry ^7,17–20 and in situ hybridization.^21,22 CgA mRNA expression was detected in all gland cells of the anterior pituitary lobes. ²² However, the cell type-specific expression of CgA at the protein level in pituitary cells has not been consistently interpreted, with reports that CgA expression was detected in gonadotrophs ^7,17,19 and thyrotrophs, ¹⁷ but not in somatotrophs, ¹⁹ thyrotrophs, ¹⁹ lactotrophs, ¹⁷ and corticotrophs,^17,19 or that CgA expression was not detected in all cells except gonadotrophs. ¹⁵ Thus, CgA has been expressed in AtT-20 cells derived from corticotrophs,^12,23–25 but there is confusion about its expression in pituitary corticotrophs. These discrepancies may be related to basic factors such as the conditions under which the sections were prepared and the reactivity of the antibodies used. Another question is why CgA is expressed in AtT-20 cells but not in pituitary corticotrophs. As AtT-20 cells are derived from mouse pituitary corticotrophs, it has not been determined whether there are differences in CgA expression between mice and rats, or even in rodents. In contrast, CgB expression has been demonstrated in both AtT-20 cells^25,26 and rat pituitary corticotrophs,^15,22 and its expression pattern in glandular cells is consistent with that of SgIII. ¹⁵

It has been reported that CgB shares several features with CgA. CgB shares structural homology with CgA near the N-terminus and at the C-terminus and has a disulfide bond loop structure near the N-terminal domain.^1,27 When bound to Ca²⁺, CgB, like CgA, exhibits greater aggregation properties at pH 5.5 than at pH 7.5.^10,28–30 In addition, the transfection of non-neuroendocrine NIH3T3 and COS-7 cells with CgA or CgB induces the formation of SGs.^31,32 Furthermore, the transfection of CgA or CgB short interfering RNAs into PC12 neuroendocrine cells reduced the number of SGs. ³² These results suggested that CgB, which is functionally similar to CgA, may play a similar role in the absence of CgA expression in rat pituitary corticotrophs. However, a direct analysis has not been adequately performed.

In this study, we first performed light microscopic double immunofluorescence staining of each granin in AtT-20 cells and analyzed the binding activities of CgB and SgIII by comparing their similarities to those of CgA. We analyzed the expression rates of CgA, CgB, and SgIII in mouse and rat pituitary corticotrophs and the colocalization of CgA/SgIII or CgB/SgIII in rat pituitary corticotrophs using triple immunofluorescence labeling and light microscopy. Finally, we analyzed ACTH-containing granules using triple immunogold labeling via electron microscopy. In contrast to AtT-20 cells, a lower percentage of CgA-expressing cells and a higher percentage of CgB-expressing cells were found in the pituitary corticotrophs, suggesting a more active role for CgB at the tissue level. The results obtained here suggest that CgA and CgB play complementary roles in sorting SGs through their interaction with SgIII, despite the morphological differences in their colocalization patterns on hormone granules. We also discuss the similarities and differences in the molecular properties of CgA and CgB via their interactions with SgIII.

Materials and Methods

Cell Culture

The AtT-20 mouse corticotroph-derived cell line was maintained in Dulbecco’s modified Eagle’s medium (DMEM, Gibco 11965-092, Thermo Fisher Scientific Inc., Waltham, MA) supplemented with 10% fetal bovine serum (FBS; 830-257-0005, Equitech-Bio, Inc., Kerrville, TX), 1% penicillin–streptomycin–glutamine (Gibco 11140-050), and 1 mM sodium pyruvate (Gibco 11360-070). AtT-20 cells were seeded in Petri dishes and cultured in an incubator at 37C and 5.0% CO₂ as described previously.^12,13

Construction of FLAG-Tagged SgIII Expression Vector and Transfection

Full-length rat SgIII cDNA fragments fused to a FLAG Tag (pCMV-rSg3-FLAG) were constructed by in-frame insertion of the SgIII gene into the pCMV vector with a FLAG epitope sequence (Stratagene, La Jolla, CA) and transfected into AtT-20 cells using Lipofectamine 2000 Transfection Reagent (1166827, Invitrogen/Thermo Fisher Scientific Inc.), as described previously. ¹³ Forty percent confluence of AtT-20 cells were seeded in 10 cm² Petri dishes and incubated for 24 h. After cell washing with phosphate buffered saline (PBS, 137 mM NaCl, 27 mM KCl, 8.1 mM Na₂HPO₄, and 1.5 mM KH₂PO₄), the medium was changed to Opti-MEM I Reduced Serum Medium (31985062, Invitrogen/Thermo Fisher Scientific Inc.) and incubated with the mixture of pCMV-rSg3-FLAG plasmid (10 μg) and transfection reagent for 24 h. After transfection, the medium was replaced with a regular DMEM-based medium and incubated for another 24 h for subsequent biochemical analysis.

Antibodies

Details of the primary and secondary antibodies and their combinations are presented in Tables 1 to 3. Rabbit polyclonal anti-CgA (C#101)^12,15 and goat polyclonal anti-CgB (sc-1489) antibodies were used for immunoblotting. Rabbit polyclonal anti-CgA (YII-Y291), goat polyclonal anti-CgB (C-19: sc-1489), and rabbit polyclonal anti-SgIII (H-300: sc-50289 and C-19: sc-1492) antibodies were used for immunocytochemical staining of AtT-20 cells. The Rabbit polyclonal anti-SgIII antibody (C#6), ¹³ the mouse monoclonal anti-CgA antibodies (BM5082, MAB5268, and C-12: sc-393941), the rabbit polyclonal anti-CgA antibody (ab45179), the mouse monoclonal anti-CgB antibody (#32: sc-517541), the rabbit anti-CgB antibody (259 103), the rabbit polyclonal anti-ACTH antibody (AB902), and the guinea pig polyclonal anti-ACTH antibody (452 005) were used for immunofluorescence staining or immunoelectron histochemistry. The anti-CgA monoclonal antibodies, BM5082 and MAB5268, were produced from the same hybridoma clone, LK2H10. The mouse monoclonal anti-DYKDDDDK tag (anti-FLAG tag) antibody, FG4R (MA1-91878) was used for immunoprecipitation. Normal rabbit IgG (sc-2027), normal goat IgG (sc-2028), normal mouse IgG (sc-2025), and normal guinea pig IgG (sc-2711) were purchased from Santa Cruz Biotechnology, Inc. (Dallas, TX), diluted (1:800) and used for negative control staining.

Table 1.

Primary Antibodies Used in the Present Study.

No.	Marker	Species	Dilution (Usage)	Products a (Catalog No. or Reference)	Immunogen	Cross-Reactivity
1	SgIII	Rb	1:100 (IC)	Santa Cruz (H-300: sc-50289)*	Human SgIII aa 1–300	Hu, Rt, Mo
2	SgIII	Rb	1:2000 (IH, IE)	C#6 (Hosaka et al., 2004)*	Rat SgIII aa 373–471	Rt, Mo, Ca, Ch
3	SgIII	Gt	1:100 (IC)	Santa Cruz (C-19: sc-1492)*	Human SgIII C-terminus	Hu, Rt, Mo
4	CgA	Rb	1:1000 (IB)	C#101 (Sakai et al., 2003)	Rat CgA aa 430–442	Hu, Rt, Mo
5	CgA	Rb	1:100 (IC)	Cosmo Bio/Yanaihara (YII-Y291)	Rat CgA aa 94–130	Rt, Mo
6	CgA	Rb	1:250 (IH, IE)	Abcam (ab45179)	Human CgA aa 1–100	Hu, Rt, Mo, Ca
7	CgA	Mo	1:10 (IE)	Acris (BM5082)**	Human CgA aa 250–284	Hu, Rt, Mo, Po
8	CgA	Mo	1:500 (IH)	Merck Sigma-Aldrich (MAB5268)**	Human CgA aa 250–284	Hu, Rt, Mo, Po
9	CgA	Mo	1:50 (IH)	Santa Cruz (C-12: sc-393941)	Human CgA aa 442–457	Hu, Rt, Mo, Eq
10	CgB	Rb	1:1000 (IH, IE)	Synaptic Systems (259 103)*	Mouse CgB aa 407–677	Rt, Mo
11	CgB	Gt	1:1000 (IC, IB)	Santa Cruz (C-19: sc-1489)*	Human CgB C-terminus	Hu, Rt, Mo
12	CgB	Mo	1:50 (IH, IE)	Santa Cruz (#32: sc-517541)*	Rat CgB aa 200–389	Hu, Rt
13	ACTH	Rb	1:10000 (IH)	Chemicon (AB902)	Synthetic ACTH aa 1–24	Hu, Rt, Mo, Ca
14	ACTH	Gp	1:10000 (IH, IE)	Synaptic Systems (452 005)	Mouse POMC aa 134–162	Rt, Mo
15	FG4R	Mo	1:1000 (IP)	Invitrogen (MA1-91878)	Synthetic peptide (DYKDDDDK)

Abbreviations: SgIII, secretogranin III; CgA, chromogranin A; CgB, chromogranin B; ACTH, adrenocorticotropic hormone; Rb, rabbit; Gt, goat; Mo, mouse; Gp, guinea pig; Hu, human; Rt, rat; Ca, canine, Po, porcine; Ch, chicken; Eq, equine; IC, immunofluorescence cytochemistry; IH, immunofluorescence histochemistry; IE, immunoelectron microscopy; IB, immunoblotting; IP, immunoprecipitation.

^aSanta Cruz Biotechnology Inc. (Dallas, TX), Cosmo Bio Co., Ltd. (Tokyo, Japan), Yanaihara Institute Inc. (Shizuoka, Japan), Abcam Ltd. (Cambridge, UK), Acris Antibodies GmbH/Origene Technologies Inc. (Rockville, MD), Merck/Chemicon Sigma-Aldrich, Co. LLC (St. Louis, MO), Synaptic Systems GmbH. (Göttingen, Germany), Invitrogen/Thermo Fisher Scientific Inc. (Waltham, MA).

^*Knockout mouse tissue or knockdown cell validated antibodies. **Prepared from a same hybridoma clone, LK2H10.

Table 3.

Combination of Primary and Secondary Antibodies Used for the Cytochemical and Histochemical Analysis.

Combination of Antibodies	Primary-1 (No.)	Secondary-1 (Alphabet)	Primary-2 (No.)	Secondary-2 (Alphabet)	Primary-3 (No.)	Secondary-3 (Alphabet)	Corresponding Figures
SgIII/CgA	3	A	5	B	–	–	Fig. 1A–D
CgB/SgIII	11	A	1	B	–	–	Fig. 1E–H
CgB/CgA	11	A	5	B	–	–	Fig. 1I–L
CgB/CgA	11	E	5	C	–	–	Fig. 1M–O
SgIII/CgB	2	O	10	N	–	–	Fig. 2
ACTH/SgIII	14	H	2	F	–	–	Fig. 4A–C
ACTH/CgA	14	H	6	F	–	–	Fig. 4D–F
ACTH/CgB	14	H	10	F	–	–	Fig. 4G–I
ACTH/SgIII	14	H	2	F	–	–	Fig. 5A–C
ACTH/CgA	13	C	8	G	–	–	Fig. 5D–F
ACTH/CgA	14	H	6	F	–	–	Fig. 5G–I
ACTH/CgB	13	C	12	G	–	–	Fig. 5J–L
ACTH/CgB	14	H	10	F	–	–	Fig. 5M–O
ACTH/SgIII/CgA	14	H	2	F	9	D	Fig. 6A–D
ACTH/SgIII/CgB	14	H	2	F	12	D	Fig. 6E–H
ACTH/CgA/CgB	14	H	6	F	12	D	Fig. 7A–L
ACTH/CgA/CgB	14	H	9	D	10	F	Fig. 7M–T
ACTH/SgIII	14	L	2	I	–	–	Fig. 8A
ACTH/CgA	14	L	6	I	–	–	Fig. 8B
ACTH/CgB	14	L	10	I	–	–	Fig. 8C
ACTH/SgIII	14	K	2	M	–	–	Fig. 8D
ACTH/CgA	14	K	6	M	–	–	Fig. 8E
ACTH/CgB	14	K	10	M	–	–	Fig. 8F
ACTH/SgIII/CgA	14	L	2	I	7	J	Fig. 9A
ACTH/SgIII/CgB	14	L	2	I	12	J	Fig. 9B

Abbreviations: SgIII, secretogranin III; CgA, chromogranin A; CgB, chromogranin B; ACTH, adrenocorticotropic hormone.

Table 2.

Secondary Antibodies Used in the Present Study.

Alphabet	Antibody Against	Species	Dilution (Usage)	Products a (Catalog No.)
A	Alexa Fluor 488-conjugated anti-goat IgG	Dk	1:100 (IC)	Invitrogen (A-11055)
B	Alexa Fluor 594-conjugated anti-rabbit IgG	Dk	1:100 (IC)	Invitrogen (A-21207)
C	Alexa Fluor 488-conjugated anti-rabbit IgG	Dk	1:2000 (IC, IH)	Invitrogen (A-21206)
D	Alexa Fluor 488-conjugated anti-mouse IgG	Dk	1:2000 (IH)	Invitrogen (A-21202)
E	Alexa Fluor 555-conjugated anti-goat IgG	Dk	1:2000 (IC)	Invitrogen (A-21432)
F	Alexa Fluor 555-conjugated anti-rabbit IgG	Dk	1:2000 (IH)	Invitrogen (A-31572)
G	Alexa Fluor 555-conjugated anti-mouse IgG	Gt	1:2000 (IH)	Invitrogen (A-21424)
H	Alexa Fluor 647-conjugated anti-guinea pig IgG	Gt	1:2000 (IH)	Invitrogen (A-21450)
I	Gold particle (18 nm)-conjugated anti-rabbit IgG	Dk	1:100 (IE)	Jackson (711-215-152)
J	Gold particle (12 nm)-conjugated anti-mouse IgG	Dk	1:100 (IE)	Jackson (715-205-150)
K	Gold particle (12 nm)-conjugated anti-guinea pig IgG	Dk	1:100 (IE)	Jackson (706-205-148)
L	Gold particle (6 nm)-conjugated anti-guinea pig IgG	Dk	1:100 (IE)	Jackson (706-195-148)
M	Gold particle (6 nm)-conjugated anti-rabbit IgG	Dk	1:100 (IE)	Jackson (711-195-152)
N	Gold particle (20 nm)-conjugated anti-rabbit IgG	Gt	1:50 (IE)	BBI (EMGAR20)
O	Gold particle (10 nm)-conjugated anti-rabbit IgG	Gt	1:50 (IE)	BBI (EMGAR10)
P	Peroxidase-conjugated anti-rabbit IgG	Gt	1:5000 (IB)	Jackson (111-035-144)
Q	Peroxidase-conjugated anti-goat IgG	Gt	1:5000 (IB)	Jackson (705-035-147)
R	Peroxidase-conjugated anti-mouse IgG	Gt	1:5000 (IB)	Jackson (111-035-146)

Abbreviations: Dk, donkey; Gt, goat; IC, immunofluorescence cytochemistry; IH, immunofluorescence histochemistry; IE, immunoelectron microscopy; IB, immunoblotting.

^aInvitrogen/Thermo Fisher Scientific Inc. (Waltham, MA), Jackson ImmunoResearch Laboratories, Inc. (West Grove, PA), BBI Solutions. (Crumlin, UK).

Immunoprecipitation

After transfection with pCMV-rSg3-FLAG, AtT-20 cells were washed with PBS and collected by scraping. Cell extracts were lysed with lysis buffers containing MES (50 mM pH 5.5 or 10 mM pH 7.4), 100 mM NaCl, 1 mM EDTA, 0.4 mM phenylmethanesulfonyl fluoride (06297-02, nacalai tesque, Kyoto, Japan), and 1% IGEPAL CA-630 (18896-100ML, Merck Sigma-Aldrich, Co. LLC, St. Louis, MO) for 1 h on-ice with several vortex mixing and then centrifuged at 13,000 × g for 10 min at 4C to obtain protein sample solution. Untransfected AtT-20 cells were processed using the same procedure and a sample solution was prepared. The protein concentrations of the sample solutions were determined using CBB protein assay reagent (29449-15, nacalai tesque). The protein sample solutions at a concentration of 500 µg/ml were used for immunoprecipitation with 1 µL of anti-FLAG tag antibody in the conditions with and without 10 mM CaCl₂. ¹³ As a negative control, immunoprecipitation was performed in the absence of the anti-FLAG antibody. The precipitated immunocomplexes were captured by adding 20 µL of 50% slurry of Protein A Sepharose 4 Fast Flow (17-6002-35, GE healthcare, Bio-Sciences KK, Tokyo, Japan) and incubated for 16 h at 4C with gentle rotation. The immunoprecipitates were then wash three times with lysis buffers by centrifugation at 2300 × g for 5 min and prepared for SDS-PAGE by adding 100 µL of 1 × SDS sample buffer (50 mM Tris-HCl, pH 6.8, 100 mM dithiothreitol, 2% SDS, 0.1% bromophenol blue, and 10% glycerol).

Immunoblotting

The protein samples (20 µL/lane) were subjected to SDS-PAGE and transferred to an Immobilon-P transfer membrane (Millipore, Billerica, MA). After blocking with 5% skim milk in Tris-buffered saline (pH 7.4) supplemented with 0.05% Tween 20 (TBS-T) for 1 h, the membranes were incubated with primary antibodies overnight at 4C. Peroxidase-conjugated secondary antibodies (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA), anti-rabbit IgG (111-035-144), anti-goat IgG (705-035-147), and anti-mouse IgG (111-035-146) were diluted and incubated at 25C for 2 h. Signals were detected using enhanced chemiluminescence western blotting detection reagents (PRN2106; GE Healthcare Bio-Sciences KK). Immunoreactive signals were detected using a LAS4000 (General Electric, Fairfield, CT), as described previously. ³³

Immunocytochemistry

AtT-20 cells were seeded at 20% confluence in 100 mm Petri dishes and cultured for 48 h. After washing with PBS, cells were fixed with fixative containing 4% paraformaldehyde (4PFA) and 3% sucrose in 0.1 M phosphate buffer (pH 7.2) for 30 min at 20C and then permeabilized with high salt TPBS (10 mM sodium phosphate buffer, 500 mM NaCl, and 0.1% Tween 20, pH 7.2) containing 0.1% Triton X-100 for 30 min. The cells were blocked by incubation with 2% FBS (S1780-500, Japan Bioserum, Hiroshima, Japan) diluted in low salt TPBS (PBS containing 0.05% Tween 20 and 0.1% NaN₃) and then reacted with primary antibodies diluted with low salt TPBS for 18 h at 4C. After incubation with primary antibodies, cells were incubated with a combination of secondary antibodies for 1 h at 25C: Alexa Fluor 488-conjugated donkey anti-goat IgG (A11055; Invitrogen/Thermo Fisher Scientific Inc.) and Alexa Fluor 594-conjugated anti-rabbit IgG (A21207). Cells were counterstained with Hoechst 33342 (04929-82; 0.05 µg/ml concentration, nacalai tesque). Fluorescence images were captured using a BZ-X800 All-in-One Fluorescence Microscope (KEYENCE Co., Osaka, Japan).

For double fluorescence immunocytochemistry using the semithin sections (1-µm thick) of LR-White resin-embedded blocks, AtT-20 cells grown at approximately 100% confluence in 60 mm Petri dishes were fixed with 4PFA in 0.1 M phosphate buffer (pH 7.2) for 20 h at 4C. The cells were scraped from the dishes and centrifuged (2000 × g) to obtain cell pellets. For the preparation of immunoelectron microscopy, AtT-20 cells were fixed first with 4PFA-0.1% glutaraldehyde in 0.1 M phosphate buffer (pH 7.2) (4PFA-0.1GA) for 10 min at 4C, and subsequently with 4PFA containing 3% sucrose for 60 min at 4C. After three rinses in 0.1 M phosphate buffer (pH 7.2) containing 7.5% sucrose at 4C, the cells were scraped from the dishes and immersed and centrifuged in the same buffer at 4C.^12,34 Pellets of fixed AtT-20 cells were subsequently dehydrated with 70% ethanol and infiltrated with LR-White resin (London Resin Company, Ltd., Berkshire, UK) at 4C. ³⁵ The samples were then placed at the bottom of gelatin capsules filled with fresh resin and polymerized using UV irradiation in a TUV-200 Ultraviolet Polymerizer (Dosaka EM, Co. Ltd., Kyoto, Japan), as described previously. ³⁶ The semithin (1-µm-thick) and ultrathin (90-nm-thick) sections were cut using an ultramicrotome (Leica Microsystems; EM UC7 ultramicrotome) and mounted on silane-coated glass slides (New Silane II; Muto Pure Chemicals, Tokyo, Japan) or nickel grids (200 mesh, NISSIN EM Co., Ltd., Tokyo, Japan).

Animals and Tissue Preparation

Wistar rats (7-weeks-old male) and C57BL/6J mice (16-weeks-old male) were purchased from CLEA Japan Inc. (Tokyo, Japan). The animals were deeply anesthetized with isoflurane inhalation and intraperitoneal injection of sodium pentobarbital (0.15 mg/g body weight) and were perfused intracardially with 50 ml of saline followed by 250 ml of 4PFA (n=2) or 250 ml of 4PFA-0.1GA (n=2) at 4C. For mouse tissue fixation, intracardial perfusion with 10 ml saline followed by 50 ml of 4PFA (n=2) was performed. After fixation, the pituitaries were excised, cut into small pieces, and postfixed by immersion in the same fixative for 1 d at 4C. Tissue samples were washed in 0.1 M phosphate buffer, dehydrated with 70% ethanol, and infiltrated into LR-White resin at 4C. The samples were then placed at the bottom of gelatin capsules filled with fresh resin and polymerized using UV irradiation. All animal experiments were performed in compliance with the relevant laws and guidelines of the Care and Use of Laboratory Animals of the Research Council of the Akita Prefectural University (Approval Number: 21-14).

Double and Triple Fluorescence Immunohistochemistry

The semithin sections (1-µm thick) of AtT-20 cells pellets and pituitary tissues embedded in LR-White resin were cut using an ultramicrotome (Leica Microsystems; EM UC7 ultramicrotome) and mounted on silane-coated glass slides (New Silane II; Muto Pure Chemicals, Tokyo, Japan). Sections were subjected to microwave antigen retrieval by boiling in 10 mM sodium citrate buffer (pH 6.0) for 5 min, rinsing sequentially with Milli-Q water and PBS, and incubating with 50 mM NH₄Cl in PBS for 15 min. The sections were then blocked with 5% normal donkey serum or 5% mixture of normal goat and donkey sera in PBS for 30 min and incubated with a mixture of primary antibodies for 24 h at 4C. To analyze the expression of CgA and CgB in AtT-20 cells, the following combination of primary antibodies was used for double fluorescence immunostaining: rabbit antibody against CgA (YII-Y291) and goat antibody against CgB (C-19). After incubation with primary antibodies for 18 h at 4C, cells were incubated with a combination of secondary antibodies for 1 h at 20C: Alexa Fluor 488-conjugated donkey anti-rabbit IgG (A21206; Invitrogen/Thermo Fisher Scientific Inc.) and Alexa Fluor 555-conjugated anti-goat IgG (A21432). As a negative control for double fluorescence labeling, a combination of normal goat IgG and rabbit IgG (Appendix Fig. A1). To analyze the expression of SgIII, CgA, and CgB in mouse pituitary corticotrophs, the following combinations of primary antibodies were used for double fluorescence immunohistochemistry: guinea pig anti-ACTH antibody (452 005) and rabbit antibodies against SgIII (C#6), CgA (ab45179), or CgB (259 103). In addition to the primary antibody combinations used for immunostaining the mouse pituitary, the following combinations were used for the rat pituitary: rabbit anti-ACTH antibody (AB902) and mouse antibodies against CgA (MAB5268) or CgB (#32). After rinsing with PBS, sections were incubated with Alexa Fluor-conjugated secondary antibodies (Invitrogen/Thermo Fisher Scientific Inc.). Secondary antibody combinations of Alexa Fluor 488-conjugated anti-rabbit IgG (A-21206) and Alexa Fluor 555-conjugated anti-mouse IgG (A-21424), or Alexa Fluor 647-conjugated anti-guinea pig IgG (A-21450) and Alexa Fluor 555-conjugated anti-rabbit IgG (A-31572) were used to analyze the coexpression of ACTH and granins. As a negative control for double fluorescence labeling, either a combination of normal guinea pig and rabbit IgG (Appendix Fig. A1 and A2) or normal rabbit and mouse IgG was used (Appendix Fig. A2). In addition, as negative controls for granin immunolabeling in corticotrophs, the sections were incubated with a mixture of guinea pig anti-ACTH antibody and normal rabbit IgG (Appendix Fig. A1 and A2) or rabbit anti-ACTH antibody and normal mouse IgG instead of anti-granin antibodies (Appendix Fig. A2). For triple fluorescence immunohistochemistry to observe the localization of SgIII and Cgs (CgA and CgB) in rat pituitary corticotrophs, the following combinations of primary antibodies were used: guinea pig anti-ACTH, rabbit anti-SgIII, and mouse anti-Cg. Furthermore, to analyze the localization of CgA and CgB in the ACTH granules, a mixture of guinea pig anti-ACTH, rabbit anti-CgA, and mouse anti-CgB antibodies was reacted with the sections. A mixture of secondary antibodies of Alexa Fluor 647-conjugated anti-guinea pig IgG (A-21450), Alexa Fluor 555-conjugated anti-rabbit IgG (A-31572), and Alexa Fluor 488-conjugated anti-mouse IgG (A-21202) was used for the triple fluorescence labeling. A mixture of normal guinea pig, rabbit, and mouse IgG instead of primary antibodies or a mixture of guinea pig anti-ACTH antibody (452 005), normal rabbit, and mouse IgG was used as the negative control for triple fluorescence labeling (Appendix Fig. A3). Fluorescence images were acquired using a BX51 microscope equipped with a DP73 CCD camera and digitized using the cellSens Standard 1.8.1 software (Olympus, Tokyo, Japan). A 60× objective lens (UPlanSApo/1.35 Oil) in a 4140 × 3096 pixel-format was used, and the acquired images were processed by linear adjustments in brightness and contrast. In double staining, ACTH was represented by magenta and granin by green. In the color representation, ACTH is shown as magenta and granins as green in double staining, and ACTH is shown as white, CgA as red, and CgB as green in triple staining, using Adobe Photoshop 21.1.3. To evaluate the expression rate of each granin in corticotrophs, non-overlapping areas were randomly photographed at 600× magnification (five micrographs for each granin) after double immunostaining for granins (SgIII, CgA, or CgB) and ACTH, and the number of cells reactive to granins or ACTH was counted. Images were analyzed using the open-source image processing software Fiji/ImageJ (https://imagej.net/software/fiji/) Cell Counter, a tool that records the number of clicks/marks made on the images as the user identifies and determines the cells.

Immunoelectron Microscopy

Ultrathin sections (90-nm-thick) cut from LR-White resin-embedded tissue blocks were mounted on nickel grids (200 mesh, NISSIN EM Co., Ltd.). The sections were immersed in 50 µL of 10 mM sodium citrate buffer (pH 6.0) and processed for antigen retrieval at 95C for 5 min using a GeneAmp PCR System 9700 Fast Thermal Cycler (Applied Biosystems Inc., Waltham, MA). For double immunolabeling of AtT-20 cells, the two-face technique ³⁷ was used. Briefly, sections were thoroughly washed with Milli-Q water and dried. One face was transferred to a drop of 5% normal goat serum in 20 mM Tris-HCl buffer (TB) (pH 7.5) containing 150 mM NaCl (TBS), and incubated at 20C for 30 min. This side of the grid was then incubated with the first antibodies diluted in TBS containing 1% BSA (A9647, Merck Sigma-Aldrich, Inc.) for 60 min at 20C. To immunolabel granins, the sections were incubated with the secondary antibody for 60 min at 20C. Between each step, the sections were washed by successively floating on four drops of TB (pH 7.5) containing 500 mM NaCl and 0.1% BSA. After washing, sections were thoroughly rinsed with Milli-Q water and dried. Immunostaining was performed on the opposite side of each section using a similar procedure. The following pairs of primary antibodies were used for double immunogold labeling: rabbit antibodies against CgB (259 103) and SgIII (C#6). The intracellular localization of each granin was distinguished by labeling with colloidal gold particles of different sizes conjugated to goat anti-rabbit IgG (EMGAR20, 20 nm or EMGAR10, 10 nm; BBI solutions, Crumlin, UK). For the immunolabeling of CgA/CgB, the secondary antibodies were switched of between the two sides of a grid to control the reactivity related to the size of the colloidal gold particles. As a negative control, normal rabbit IgG was used instead of rabbit anti-granin antibody (Appendix Fig. A4).

For rat pituitary staining, sections were incubated with mixtures of primary or secondary antibody solutions without using the two-face technique. After antigen retrieval processing, the sections were washed with Milli-Q water, incubated with 5% normal donkey serum in TBS for blocking at 20C for 30 min, and then incubated with a mixture of primary antibodies diluted in TBS containing 1% BSA for 16 h at 4C. For double immunolabeling of the rat pituitary glands fixed with 4PFA-0.1GA, a guinea pig anti-ACTH antibody (452 005) was combined with either rabbit antibodies against SgIII (C#6), CgA (ab45179), or CgB (259 103). A combination of normal rabbit IgG and guinea pig anti-ACTH antibodies was used as a negative control for the granin immunolabeling of ACTH granules (Appendix Fig. A4). For triple immunolabeling to analyze the colocalization of SgIII/CgA and SgIII/CgB in ACTH granules in the rat pituitary glands, two sets of primary antibody mixtures were used: (1) anti-ACTH (452 005)/anti-SgIII (C#6)/anti-CgA (BM5082) antibodies, and (2) anti-ACTH (452 005)/anti-SgIII (C#6)/anti-CgB (#32) antibodies were used. A mixture of normal rabbit and mouse IgG and guinea pig anti-ACTH antibodies was used as a negative control for the granin immunolabeling of ACTH granules (Appendix Fig. A4). To evaluate the intracellular localization of ACTH and granins in corticotroph, donkey secondary antibodies conjugated to colloidal gold particles of different diameters (18 nm, anti-rabbit IgG; 12 nm, anti-mouse and anti-guinea pig IgG; 6 nm, anti-rabbit and anti-guinea pig IgG, Jackson ImmunoResearch Laboratories, Inc.) were used. After incubation with primary and secondary antibodies, the grids were washed with TB (pH 7.5) containing 250 mM NaCl and 0.1% BSA. The sections were postfixed in 1% glutaraldehyde/TBS for 15 min and then contrasted with 10% gadolinium acetate tetrahydrate dissolved in Milli-Q water.

Immunogold-labeled granule images were acquired using a JEM-1011 transmission electron microscope (JEOL, Tokyo, Japan) equipped with a Mega-ViewG3 side-mount TEM CCD camera (1936 × 1456 pixels) and RADIUS camera control/image analysis software (EMSIS GmbH, Münster, Germany). Images were acquired at 100,000× magnification and processed by linear brightness and contrast adjustment using the Adobe Photoshop software.

Results

Localization of SgIII, CgA and CgB in AtT-20 Cells

We examined the subcellular localizations of SgIII, CgA, and CgB in AtT-20 cells using double fluorescence immunocytochemistry. SgIII and CgA were localized around the Golgi apparatus and at the cell tip, where SGs accumulated (Fig. 1A and B). Most CgA colocalized with SgIII in the granules, but only a few SgIII-only signals were observed (Fig. 1C and D). Similarly, CgB and SgIII were localized around the Golgi apparatus and at the cell tip (Fig. 1E and F), and both signals colocalized in some granules, similar to SgIII alone (Fig. 1G and H). CgB and CgA appeared to be colocalized in some granules, but many granules were independently labeled with CgB or CgA (Fig. 1I–L). For a more detailed analysis, AtT-20 cell pellets were embedded in LR-White resin and immunolabeled for CgB and CgA using semithin sections. Consistent with the results of the immunocytochemical analysis (Fig. 1K and L), a small number of granules colocalized with CgB and CgA, but many granules were localized independently (Fig. 1M–O). The percentage of granules colocalizing with CgB and CgA was 6.6% (Fig. 1P). Therefore, the localization of CgB and CgA in the granules did not appear to be completely consistent in AtT-20 cells despite their colocalization with SgIII.

Figure 1.

Localization of SgIII, CgA, and CgB in AtT-20 cells. Localization of SgIII, CgA, and CgB in AtT-20 cells is observed around the Golgi apparatus (asterisks) and cell tips (arrows). (A–D) Most of CgA colocalized with SgIII in the SGs (black arrowheads), although there were a few SgIII-only labeled granules (white arrowheads). (E–H) CgB localized similarly to CgA around the Golgi apparatus and at the cell tips, and SGs colocalized with SgIII (black arrowheads) and labeled only with CgB (white arrowheads). (I–L) CgB (green arrowheads) and CgA (red arrowheads) were not fully colocalized and were only partially colocalized (yellow arrowheads). Enlarged images (D, H, and L) show the rectangular areas in the merged images (C, G, and K). The blue fluorescence in the merged images indicates nuclear counterstaining with Hoechst 33342. (M–O) Immunolabeling of CgB and CgA using the semithin sections of LR-White resin-embedded AtT-20 cell pellets. SGs colocalized with CgB (red arrowheads) and CgA (green arrowheads) were restricted (yellow arrowheads) and independently localized. (P) Percentages of CgB-, CgA-, and CgB/CgA-labeled granules calculated from 20 AtT-20 cells. Scales, 20 µm and 4 µm (magnified images and panel O).

The colocalization of SgIII and CgA in AtT-20 cells has been reported using immunoelectron microscopy using the two-face labeling technique.^12,23 However, the colocalization of SgIII and CgB has not yet been investigated. Therefore, we analyzed the colocalization of SgIII and CgB using a similar method. The SGs labeled with both CgB and SgIII and those labeled with CgB only were observed in AtT-20 cells (Fig. 2). This labeling pattern was generally consistent with the results of immunofluorescence staining (Fig. 1E–H).

Figure 2.

Colocalization of CgB and SgIII using immunogold labeling in AtT-20 cells. A two-face immunogold labeling technique was used. Sections were incubated with rabbit antibodies against CgB and SgIII. Two sizes (20 and 10 nm in diameter) of colloidal gold particle-conjugated secondary antibodies were labeled, indicating that CgB (20 nm gold) and SgIII (10 nm gold), colocalized in SGs (closed arrows), whereas only a few granules labeled with CgB alone were observed (open arrow). The inset shows a magnified view of the CgB/SgIII-colocalized granules. Scales, 200 and 100 nm (inset).

CgB, like CgA, Forms a Complex With SgIII

To investigate the binding activity of CgB and SgIII, cell extracts were prepared from mouse pituitary corticotroph-derived AtT-20 cells transfected with pCMV-rSg3 (Full)-FLAG and immunoprecipitation was performed. Immunoprecipitates captured using the anti-FLAG-tag antibody were subjected to immunoblotting using anti-granin antibodies. CgB coprecipitated with SgIII and showed stronger binding under conditions of weak acid (pH 5.5) and high Ca²⁺ concentration (10 mM), but not under neutral conditions (pH 7.4) (Fig. 3).

Figure 3.

CgB, like CgA, binds to SgIII under conditions of weak acid and high Ca²⁺ concentration. Immunoprecipitation (IP) was performed on extracts from AtT-20 cells transfected with pCMV-rSg3 (Full)-FLAG. The cell extracts were incubated with an anti-FLAG-tag antibody (lanes 1–4). A negative control for IP was performed without the antibody (lanes 5–8). The immunoprecipitates were analyzed by immunoblotting to detect granins using anti-CgB (C-19) and anti-CgA (C#101) antibodies. Under weakly acidic conditions (pH 5.5) and high Ca²⁺ concentrations (10 mM), CgB coprecipitated with SgIII as CgA (lanes 3 and 4). Cell extracts (C.E.) of AtT-20 were loaded as a positive control for immunoblotting (lane 9). kDa, kilodaltons.

Expression of SgIII, CgA, and CgB in Mouse Pituitary Corticotrophs

The expression of SgIII, CgA, and CgB in mouse pituitary corticotrophs, the origin of AtT-20 cells, was analyzed by double immunofluorescence staining using a combination of guinea pig antibodies against ACTH and rabbit antibodies against these granins. In this approach, we used LR-White resin-embedded semithin sections (1-µm thick) and were able to detect the punctate immunoreactive signals of SGs with higher resolution compared to conventional paraffin-embedded sections (5-µm thick). Coexpression of ACTH and SgIII was observed in most corticotrophs (Fig. 4A–C). CgA was expressed in small subpopulations, but was barely detectable in many corticotrophs. The CgA expression pattern was divided into three subpopulations (moderately positive, weakly positive, and barely detectable) (Fig. 4D–F). In contrast, coexpression of ACTH and CgB was detected in many corticotrophs (Fig. 4G–I). The expression intensity of these two chromogranins in corticotrophs was weaker than that in other chromogranin-expressing cells and appeared to be related to the lower number and density of hormone granules contained in corticotrophs. After double immunofluorescence staining for ACTH and each granin, the number of cells with different expression intensities was counted and the percentages of SgIII, CgA, and CgB expression were calculated (Table 4). SgIII was expressed in 99.5% of corticotrophs. CgA was expressed in 15.6% of corticotrophs but not in 84.4% of corticotrophs. However, CgB was expressed in 98.1% of corticotrophs but was not expressed in 1.9% of corticotrophs.

Figure 4.

Double immunofluorescence staining for ACTH and granins in mouse pituitary gland. Sections were incubated with primary antibodies against ACTH (452 005, raised in guinea pigs) in combination with rabbit polyclonal antibodies against SgIII (C#6) (A–C), CgA (ab45179) (D–F), or CgB (259 103) (G–I). Variations in the intensity of granin expression were observed in corticotrophs. Corticotrophs with moderate granin expression are indicated by arrows, those with weak expression are indicated by white arrowheads, and those with barely detectable expression are indicated by black arrowheads. Scale, 10 µm.

Table 4.

The Percentage of Granin Expression in Mouse Pituitary Corticotrophs.

Granin (Antibody)	Expression Intensity			Total Number of Cells
	–	+	++
SgIII (C#6)
Cell count (%)	0 (0)	6 (2.3)	257 (97.2)	263
CgA (ab45179)
Cell count (%)	211 (84.4)	24 (9.6)	15 (6.0)	250
CgB (259 103)
Cell count (%)	5 (1.9)	28 (10.8)	226 (87.3)	259

Percentages were calculated as the number of granin-expressing cells per total number of corticotrophs from double immunofluorescence staining for ACTH and granins, as shown in Fig. 4. The intensity of granin immunoreaction was categorized as barely detectable (−), weakly positive (+), and moderately positive (++). Data were obtained from two individual mouse pituitary glands and are presented as average values. Abbreviations: Sg III, secretogranin III; Cg A, chromogranin A; Cg B, chromogranin B.

Expression of SgIII, CgA, and CgB in Rat Pituitary Corticotrophs

To analyze whether there were species differences in the expression of SgIII, CgA, and CgB in the pituitary corticotrophs between mice and rats, immunofluorescence staining was performed using a combination of guinea pig anti-ACTH antibody and rabbit anti-granin antibody, and rabbit anti-ACTH antibody and mouse anti-CgA and anti-CgB antibodies. Similar to the mouse corticotrophs, the expression intensities of ACTH and granin varied in rat corticotrophs (Fig. 5). SgIII expression was observed in most corticotrophs (Fig. 5A–C). Although the labeling intensity of the anti-CgA monoclonal antibody (MAB5268) was weaker than that of the anti-CgA polyclonal antibody (ab45179), the expression pattern of CgA in rat corticotrophs was similar to that in mice and was confirmed to be positive in a small number of corticotrophs, but undetectable in many (Fig. 5D–I). Conversely, CgB was expressed in many corticotrophs (Fig. 5J–O). We counted the number of corticotrophs labeled with anti-granin antibodies and calculated the percentages of SgIII, CgA, and CgB expression (Table 5). SgIII was expressed in 98.3% of corticotrophs. CgA was expressed in 6.8% to 11.1%, but was not expressed in 88.9% to 93.2%. CgB was expressed in more than 98.5% of the corticotrophs but was not expressed in less than 1.5% of corticotrophs.

Figure 5.

Coexpression of ACTH and granins in rat pituitary gland. Primary guinea pig (452 005) and rabbit (AB902) anti-ACTH antibodies were used in combination with rabbit anti-SgIII polyclonal antibody (C#6) (A–C), anti-CgA mouse monoclonal antibody (MAB5268) (D–F), anti-CgA rabbit polyclonal antibody (ab45179) (G–I), anti-CgB mouse monoclonal antibody (#32) (J–L), and anti-CgB rabbit polyclonal antibody (259 103) (M–O). Variations in the intensity of granin expression were observed in corticotrophs. The pattern of coexpression of ACTH with CgA or CgB was similar for each of the two antibody combinations (D–F vs. G–I for ACTH and CgA and J–L vs. M–O for ACTH and CgB). Corticotrophs with moderate granin expression are indicated by arrows, those with weak expression are indicated by white arrowheads, and those with barely detectable expression are indicated by black arrowheads. Scale, 10 µm.

Table 5.

The Percentage of Granin Expression in Rat Pituitary Corticotrophs.

Granin (Antibody)	Expression Intensity			Total Number of Cells
	–	+	++
SgIII (C#6)
Cell count (%)	6 (1.7)	55 (15.1)	302 (83.2)	363
CgA (MAB5268)
Cell count (%)	337 (88.9)	34 (9.0)	8 (2.1)	379
CgA (ab45179)
Cell count (%)	304 (93.2)	9 (2.8)	13 (4.0)	326
CgB (#32)
Cell count (%)	0 (0)	17 (4.2)	384 (95.8)	401
CgB (259 103)
Cell count (%)	5 (1.5)	40 (11.7)	279 (86.8)	342

Percentages were calculated as the number of granin-expressing cells per total number of corticotrophs from double immunofluorescence staining for ACTH and granins, as shown in Fig. 5. The intensity of granin immunoreaction was categorized as barely detectable (−), weakly positive (+), and moderately positive (++). Data were obtained from two individual rat pituitary glands and are presented as mean values. Abbreviations: Sg III, secretogranin III; Cg A, chromogranin A; Cg B, chromogranin B.

Localization of SgIII, CgA, and CgB in Rat Corticotrophs

To analyze the colocalization of SgIII/CgA and SgIII/CgB in the SGs of rat corticotrophs, we performed triple immunofluorescence staining using a combination of primary antibodies: guinea pig anti-ACTH, rabbit anti-SgIII, and mouse anti-CgA or CgB antibodies. A few granules colocalized with SgIII and CgA, and many granules were independently labeled with SgIII or CgA (Fig. 6A–D). In contrast, many granules were labeled with both SgIII and CgB, and a few granules were labeled with SgIII or CgB alone (Fig. 6E–H). Thus, the colocalization pattern of CgA or CgB with SgIII in rat pituitary corticotrophs was slightly different from that in AtT-20 cells (Fig. 1A–H).

Figure 6.

Localization of SgIII, CgA, and CgB in ACTH granules of rat pituitary corticotrophs. Triple immunofluorescence staining was performed using antibodies against ACTH (452 005), SgIII (C#6), CgA (C-12), and CgB (#32) ab45179) was performed. Corticotrophs expressing CgA are limited in number and not detected in most corticotrophs (asterisks). In a few corticotrophs, CgA-labeled granules colocalized with SgIII (arrows), but signals of CgA (black arrowheads) or SgIII (white arrowheads) alone were also observed (A–D). In contrast, colocalization of SgIII and CgB was observed in many granules (arrows), with few signals from CgB (black arrowheads) or SgIII (white arrowheads) alone (E–H). Scale, 5 μm.

We then analyzed the localization of CgA and CgB in rat corticotrophs by triple immunofluorescence staining using two combinations of primary antibodies: guinea pig anti-ACTH (452 005), rabbit anti-CgA (ab45179), and mouse anti-CgB (#32); and anti-ACTH (452 005), mouse anti-CgA (C-12), and rabbit anti-CgB (259 103). In ACTH-labeled cells, a few overlapping immunoreactive signals of CgA and CgB were detected; however, many signals of these two chromogranins were localized independently (Fig. 7A–H, M–P). In contrast, a relatively strong colocalization of CgA and CgB was detected in ACTH-unlabeled cells, although a few single signals of CgA or CgB were also present (Fig. 7I–L, Q–T).

Figure 7.

Localization of CgA and CgB in ACTH granules of rat pituitary corticotrophs. Triple immunofluorescence staining was performed using antibodies against ACTH (452 005) in combination with either CgA (ab45179) and CgB (#32) (A–L), or with CgA (C-12) and CgB (259 103) (M–T) was performed. Many corticotrophs expressed CgB, but few expressed CgA. Asterisks indicate corticotrophs that do not express CgA. Corticotrophs expressing CgA were limited in number, and a few granules were labeled with both CgA- and CgB-labeled granules in corticotrophs (arrows); however, many granules labeled with CgA alone (white arrowheads) or CgB alone (black arrowheads) were observed (E–L, M–P). In ACTH-unlabeled cells, CgA expression was strong and colocalization with CgB was observed in many granules, although there were granules in which both were labeled alone (double asterisks) (I–L, Q–T). Scale, 5 μm.

Immunoelectron Microscopy Analysis of Granin Localization in ACTH Granules

In rat corticotrophs, we first performed double immunogold labeling of ACTH granules with anti-granin antibodies. Gold particles with SgIII localization were observed at the periphery of the granules, whereas CgB localization was observed at sites farther from the periphery. There was no difference in the distance between small gold particles (6 nm) showing ACTH localization and large gold particles (18 nm) showing CgA or CgB localization (Fig. 8A–C). Considering the labeling efficiency based on the size of the gold colloids, we detected ACTH and granin using secondary antibodies labeled with gold particles of different sizes (ACTH: 6 nm vs. granin: 18 nm; ACTH: 12 nm vs. granin: 6 nm). There was no significant difference in labeling efficiency depending on the size of the gold particles (Fig. 8D–F).

Figure 8.

Double immunogold labeling for ACTH and granins in rat pituitary corticotrophs. Sections were incubated with primary antibody mixtures of anti-ACTH (452 005), (A and D) anti-SgIII (C#6), (B and E) anti-CgA (ab45179), and (C and F) anti-CgB (259 103). Colloidal gold-labeled secondary antibodies were used to detect guinea pig anti-ACTH antibody (small arrows) and rabbit anti-granin antibodies (large arrows). The labeled gold colloids of the secondary antibodies detecting ACTH and granins reacted at different sizes (A–C, ACTH: 6 nm vs. granins: 18 nm; D–F, ACTH: 12 nm vs. granins: 6 nm). Scale, 100 nm.

Triple immunogold labeling was performed to determine whether there were differences in the localization patterns of SgIII/CgA or SgIII/CgB in ACTH granules. SgIII, CgA, and CgB in ACTH granules were labeled by three different size (6, 12, and 18 nm) colloidal gold-labeled secondary antibodies to probe guinea pig anti-ACTH, rabbit anti-SgIII, and mouse anti-CgA or anti-CgB antibodies. At the electron microscopic level, similar to the results obtained by immunofluorescence staining, a high number of ACTH granules labeled with SgIII alone, CgA or CgB alone were observed (Fig. 9A and B). Colocalization of SgIII and CgA was detected in some ACTH granules, but its frequency was limited compared to the colocalization of SgIII and CgB (Fig. 9, insets).

Figure 9.

Triple immunogold labeling for ACTH and granins in rat pituitary corticotrophs. Sections were incubated with primary antibody mixtures of anti-ACTH (452 005), anti-SgIII (C#6), and (A) anti-CgA (MAB5268), or (B) anti-CgB (#32). Colloidal gold-labeled secondary antibodies were used to detect guinea pig anti-ACTH antibody (6 nm gold particles, small arrows), mouse anti-CgA and anti-CgB antibodies (12 nm gold particles, arrowheads), and rabbit anti-SgIII antibody (18 nm gold particles, large arrows). The insets show the ACTH granules doubly labeled with SgIII/CgA and SgIII/CgB. Scale, 100 nm.

Discussion

CgA expression has been observed in AtT-20 cells derived from mouse pituitary corticotrophs.^12,23–25 However, CgA expression in rat pituitary corticotrophs has not been confirmed immunohistochemically.^15,17,19 A combined analysis using in situ hybridization and immunohistochemistry was inconsistent, with CgA mRNA being expressed in 32% of corticotrophs. ²² Although CgA mRNA is expressed in pituitary cells, its expression at the protein level may be determined by the type of secretory cells. In addition, CgA expression was confirmed immunohistochemically in a subset of corticotrophs in sheep pituitary. ³⁸ In previous studies, double immunostaining of CgA and ACTH was not performed on the same sections, and a more accurate assessment of CgA-expressing cells may have been insufficient. In addition, the reactivity of anti-CgA antibodies, tissue fixatives, embedding methods, section thickness, and antigen retrieval must be considered factors that may lead to inconsistent results regarding CgA expression in pituitary corticotrophs.

We performed double/triple immunofluorescence staining with multiple CgA and ACTH antibodies to analyze CgA expression in mouse and rat corticotrophs using LR-White resin-embedded antigen retrieval semithin sections of 4PFA-fixed pituitary glands. Among these rodents, 15.6% of corticotrophs expressed CgA in mice and 6.8–11.1% in rats. Although these percentages were not high, they clearly indicated that approximately 10% of corticotrophs expressed CgA. Our group has previously reported the specific binding of CgA and SgIII in AtT-20 cells, which are involved in the sorting and transport of hormone peptides during SG formation.^8,12,13 In this study, we demonstrated the specific binding of CgB and SgIII using immunoprecipitation and immunocytochemical colocalization of both granins in ACTH granules of AtT-20 cells. In contrast, CgB is expressed in rat pituitary corticotrophs,^15,22 and its distribution pattern in the pituitary is consistent with that of SgIII. ¹⁵ Although the expression intensity varied, SgIII and CgB were expressed in most rat corticotrophs (both >98%) in the present analysis. This result differed markedly from that of CgA, which was expressed in only a small number of corticotrophs (~10%). No species differences were observed between rats and mice. These findings also suggest that the individual cellular activities of corticotrophs may influence CgA or CgB expression.

When analyzing the active peptides and the conversion enzymes that process them and their regulatory proteins (i.e., proteins that express their functions after translation and processing), it is possible that using only one type of antibody for analysis may not provide sufficient information about expression specificity. Particularly, even if the immunogen is a full-size or a relatively long region of an antigen, it may not be possible to sufficiently detect the fragments after processing; even if it is the same molecule, the mode of processing may differ depending on the cell in which it is expressed, resulting in differences in antibody reactivity. Like peptide hormones, CgA and CgB are processed by prohormone convertases. ¹

CgA is cleaved at multiple dibasic sites in the intracellular or extracellular space, resulting in the generation of many active peptide fragments, especially in both the N- and C-terminal regions,^39–43 and there are only a limited number of studies that have analyzed in detail the coexistence of individual peptide fragments at the tissue level.^44,45 Furthermore, in contrast to normal endocrine tissues, immunohistochemical expression of different epitopes has been detected in neuroendocrine tumors. ⁴⁵ In this study, we used a variety of available anti-granin antibodies and obtained consistent results that explained the immunohistochemical expression of CgA in pituitary corticotrophs. The peptide immunogen for the anti-CgA ab45179 antibody is human CgA 1–100 N-terminus, and the ab45179 antibody cross-reacts with human, mouse, rat, and dog cells.^6,46–48 It can detect active peptides such as 21kDa β-granin (rat CgA 1–128), vasostatin-1 (CgA 1–76), and vasostatin-2 (CgA 1–113). Similarly, the YII-Y-29-EX antibody uses the N-terminal region of human CgA 94–130 as an immunogen to detect these active peptides. The YII-Y291 antibody is immunogenic against rat CgA but cross-reacts with mouse CgA.^49,50 Detection of β-granin in the mouse pituitary and increased expression (or processing) due to restraint stress was observed using immunoblot with the YII-Y291 antibody, ³³ but the significance of β-granin production in corticotropin is not well understood. In addition, two mouse monoclonal antibodies against CgA (BM5082 and MAB5268) were derived from the same hybridoma clone, LK2H10, using human pheochromocytoma as the immunogen ⁵¹ and recognize the same epitope. The epitope is located between the amino acids 250 and 283 of human CgA in the N-terminal region of pancreastatin. ⁵² Another mouse monoclonal antibody against CgA (C-12: sc-393941) was prepared using human CgA 442–457 (the 16 amino acids at the C-terminus; AELEKVAHQLQALRRG) as the immunogen; the amino acid sequences of humans, mice, and rats were identical. The C-12 monoclonal antibody recognizes full-length (75 kDa) unprocessed CgA and two types of processed C-terminal peptides, ER-37 and GR-44. ⁴¹ In addition, the immunogenic peptide sequence of the C-12 antibody is almost the same as that of the C#101 antibody,^12,15 and it is very likely that the C-12 antibody recognizes full-size unprocessed CgA.

Like CgA, CgB is widely distributed in neuroendocrine cells but shows different expression in terms of species and tissue specificity compared with CgA and the coexisting hormones.^18,53 In bovine CgB, 18 cleavage sites have been identified. The peptide cleaved at amino acid residues 42–43 ⁵⁴ yields the active peptide CgB 1–41, ⁵⁵ and cleavage at the C-terminus yields the antimicrobial peptide secretolytin (CgB 614–626). ⁵⁴ In addition, several peptides with unknown biological activities, such as GAWK (CgB 420–493), CCB (CgB 597–653), and BAM-1745 (CgB 547–560), have been reported in the pituitary and adrenal glands.^56–63 The anti-CgB 259 103 antibody used in this study is based on a relatively broad region of the C-terminal side of mouse CgB 401–667 and shows cross-reactivity with rats, ⁶⁴ and the specificity of the reactivity has been confirmed in tissues from CgA/CgB double KO mice. ⁶⁵ The goat anti-CgB antibody (C-19: sc-1489) uses the C-terminal end of human CgB as an antigen and has been shown to cross-react in humans, mice, and rats. However, it has recently been discontinued, and the mouse anti-CgB monoclonal antibody (#32: sc-517541) has been replaced. The C-19 antibody was prepared through immunization with rat CgB 200-389 and cross-reacts between humans and mice. Both C-19 and #32 are control antibodies that are effective in detecting CgB expression in rat islet cell lines (INS-1 cells) and isolated islets.^66,67

In conjunction with the structural similarities between CgA and CgB, it has been reported that both proteins share similar functional properties, including a disulfide loop structure,^1,27 pH- and Ca²⁺-dependent aggregation,^10,28–30 and induction of SG formation by expression in transfected non-neuroendocrine cells.^31,32 Therefore, we predicted that the specific binding of CgB and SgIII would have a similar effect on the selective transport of hormone peptides to ACTH granules in pituitary corticotrophs. In an immunoprecipitation binding assay using FLAG-tagged SgIII expression in AtT-20 cells, CgB showed binding activity to SgIII under weakly acidic conditions and in the presence of Ca²⁺, where SgIII has been shown to bind to CgA,^12,14 suggesting that the CgB/SgIII complex has a function similar to that of the CgA/SgIII complex. Detailed analysis of the localization of granins in ACTH granules revealed that CgA and CgB colocalized with SgIII in a relatively large number of granules, but not in some granules in AtT-20 cells. A few granules colocalized with CgA and CgB, and many separately labeled granules were observed. In contrast, SgIII colocalized with CgA and CgB, and dominant colocalization of SgIII and CgB was observed in rat pituitary corticotrophs. Therefore, it was speculated that CgB plays a role in aggregating ACTH by binding to SgIII, and its function might be dominant compared to that of CgA in corticotrophs. Assuming that there are two sorting and transport pathways in pituitary corticotrophs, through which CgA and CgB act by binding to SgIII, it is important to determine their functional differences.

In addition to the similarities between CgA and CgB, there are also differences in their properties of CgA and CgB. The N-terminal disulfide loops of CgA and CgB are thought to be essential for sorting to regulated SGs,^68,69 whereas in other reports, the removal of the N-terminal structure of the disulfide-linked loop in CgA had no effect on targeting from the TGN to the SGs.^12,70,71 In contrast, the loop structure of CgB has been reported to serve as a signal for sorting SGs^3,72–75 Thus, the disulfide loop potential was essential for sorting and transport to SGs in CgB, but not in CgA. Furthermore, to induce the same level of aggregation, CgA requires approximately 165-fold higher concentrations of Ca²⁺ than the same amount of CgB, which is more sensitive to Ca²⁺ than CgA, suggesting that CgB initiates the aggregation reaction before CgA during sorting from TGNs to SGs.^29,30 In addition, CgB-transfected NIH3T3 and COS-7 cells, neither of which are derived from neuroendocrine cells, form more SGs per cell than CgA-transfected cells, indicating that CgB is more effective than CgA in inducing SG formation. ³² Recent individual-level findings in pancreatic β cells have reported that CgB has binding activity with another Sg/Cg family member, VGF (synonym: SgVII), and is involved in SG budding. ⁶⁷ Such differences between CgA and CgB may be related to functional differences in pituitary cells, including corticotrophs. In this study, we observed the granules in which SgIII and CgA or SgIII and CgB were colocalized in ACTH granules in both immunofluorescence and immunogold staining, with more granules colocalizing with CgB than with CgA. We also found that fewer ACTH granules colocalized with CgA and CgB than with granules in non-corticotrophs. Differences in the localization patterns of CgA and CgB in different endocrine cells may be related to granule size, maturation, and chemical properties of the secreted hormones; therefore, it would be invaluable to analyze the subcellular localization of CgA and CgB in peptide hormone-producing cells other than corticotrophs, such as pituitary gonadotrophs and pancreatic islet cells.^7,15,16 Interestingly, the SGs of rat gonadotrophs are classified according to granule size, hormone type, and immunoreactivity to anti-granin antibodies, with follicle-stimulating hormone (FSH), luteinizing hormone (LH), and CgA localized in large granules and LH and SgII in small granules. ⁷ SgIII is localized on large granules where CgA colocalizes in rat gonadotrophs. ¹² CgB was also found to localize to SGs in rat pituitary thyrotrophs and mammotrophs, as well as to SgII and SgIII, and was unevenly distributed in the peripheral region of large mature granules (recognized as the granule membrane) compared to small immature granules in prolactin granules. ¹⁵ These observations are important because they suggest that the localization of CgB in SGs is related to granule maturation and plays a role different from that of CgA through its interaction with SgIII.

In the adrenal medulla of CgA-knockout (KO) mice, vesicle size and number are normal, but the mRNA levels of other granins (CgB, SgII, SgIII, SgV, and SgVI) are higher than those in wild-type mice.^3,76 Furthermore, CgB-KO mice show unchanged granule numbers and morphology in islet β-cells and pituitary corticotrophs.^3,77 Therefore, granins may play a compensatory role in SG biosynthesis. In addition, CgB-KO mice show some compensatory biosynthesis of SgII and CgA in tissues compared to wild-type mice, upregulation of SgII and CgA in the adrenal gland, increased levels of SgII but not CgA in the pituitary gland, and similar levels of SgII and CgA in pancreatic islet cells. ⁷⁷ Because granins complement each other in SG biosynthesis, the ablation of some granins in mouse models does not significantly affect granule biogenesis and morphology.^33,76,77 Since the pituitary tissue is a heterogeneous collection of cells, it would be interesting to determine what compensatory changes in the expression levels of other granins occur with CgB ablation that are restricted to corticotrophs.

In conclusion, considering the differences in animal species, the different types of antibodies used, and conditions of tissue fixation, thin sectioning, and antigen retrieval, we found that CgA was expressed in a small number of rodent pituitary corticotrophs, and its expression intensity was weaker than that of CgB. Furthermore, in mouse corticotroph-derived AtT-20 cells, CgB showed binding activity with SgIII as CgA, suggesting the existence of multiple pathways for sorting and transport of ACTH granules and that the CgB/SgIII interaction may have more predominant effects. The localization of CgB and CgA in the SGs did not coincide despite their respective colocalization with SgIII in the SGs, suggesting that the SgIII/CgA and SgIII/CgB complexes have different properties. Further studies are needed to understand the shared roles of CgA and CgB in the peptide hormone sorting and transport system. The presence of multiple systems for peptide hormone sorting and transport in pituitary corticotrophs facilitates the formation of fresh SGs and may be one of the mechanisms underlying the elaborate control of hormone secretion in the regulatory secretory pathway.

Appendix

Figure A1.

Negative control for double immunofluorescence staining of AtT-20 cells and mouse pituitary gland. Normal rabbit IgG (Rb-IgG) (A) and normal goat IgG (Gt-IgG) (B) were used instead of primary antibodies and labeled by incubation with Alexa Fluor 488-conjugated anti-Rb-IgG and Alexa Fluor 555-conjugated anti-Gt-IgG, respectively, as secondary antibodies. Normal guinea pig IgG (Gp-IgG) (C) and normal Rb-IgG (D) were used instead of primary antibodies and labeled by incubation with Alexa Fluor 647-conjugated anti-Gp-IgG and Alexa Fluor 555-conjugated anti-Rb-IgG, respectively, as secondary antibodies. In addition, a mixture of guinea pig anti-ACTH antibody (452 005) (E) and normal Rb-IgG (F) was used as a control for nonspecific binding of normal Rb-IgG in mouse corticotrophs. Non-specific reactions of normal Rb-IgG were barely detectable in mouse corticotrophs (white arrowheads). Scale, 10 µm.

Figure A2.

Negative control for double immunofluorescence staining of rat pituitary gland. Normal Rb-IgG (A) and normal mouse IgG (Mo-IgG) (B) were reacted instead of primary antibodies and labeled by incubation with Alexa Fluor 488-conjugated anti-Rb-IgG and Alexa Fluor 555-conjugated anti-Mo-IgG as secondary antibodies. In addition, a mixture of anti-ACTH antibody (AB902) (C) and normal Mo-IgG (D) was used as a control for nonspecific binding of normal Mo-IgG in rat corticotrophs. Normal Gp-IgG (E) and normal Rb-IgG (F) were used instead of primary antibodies and labeled by incubation with Alexa Fluor 647-conjugated anti-Gp-IgG and Alexa Fluor 555-conjugated anti-Rb-IgG, respectively, as secondary antibodies. In addition, a mixture of guinea pig anti-ACTH antibody (452 005) (G) and normal Rb-IgG (H) was used to observe the nonspecific binding of normal Rb-IgG in mouse corticotrophs. Nonspecific reactions of normal Rb-IgG, Mo-IgG, and Gp-IgG were barely detectable in rat corticotrophs (white arrowheads). Scale, 10 µm.

Figure A3.

Negative control of triple immunofluorescence staining of rat pituitary gland. A mixture of normal Gp-IgG (A), normal Rb-IgG (B), and normal Mo-IgG (C) was used instead of the primary antibodies, followed by incubation with Alexa Fluor 647-conjugated anti-Gp-IgG, Alexa Fluor 555-conjugated anti-Rb-IgG, and Alexa Fluor 488-conjugated anti-Mo-IgG secondary antibodies. In addition, a mixture of guinea pig anti-ACTH antibody (452 005) (D), normal Rb-IgG (E), and normal Mo-IgG (F) was used to observe nonspecific cross-labeling. Nonspecific reactions of normal Mo-IgG or Rb-IgG were barely detectable in rat corticotrophs (white arrowheads). Scale, 5 µm.

Figure A4.

Negative control for immunoelectron microscopy. (A) To obtain a negative control for the two-face technique of AtT-20 cells, sections were reacted with normal Rb-IgG labeled with gold particle-conjugated secondary antibodies (20 and 10 nm diameter). (B, C) To observe the nonspecific reaction of normal Rb-IgG in rat pituitary ACTH granules fixed with 4% PFA/0.1% GA, sections were incubated with a mixture of guinea pig anti-ACTH antibody (452 005) and normal Rb-IgG. Gold particle-conjugated secondary antibodies were used to detect ACTH granules (6 nm or 12 nm gold particles) and normal Rb-IgG (18 nm gold particles, unlabeled). (D) To observe nonspecific reaction of normal Rb-IgG and normal Mo-IgG (Mo-IgG) in rat pituitary ACTH granules fixed with 4% PFA, sections were incubated with a mixture of guinea pig anti-ACTH antibody (452 005), normal Rb-IgG, and normal Mo-IgG. Gold particle-conjugated secondary antibodies were used to detect ACTH granules (6 nm gold particles, indicated by arrows) and normal Rb-IgG (18 nm gold particles, unlabeled) and Mo-IgG (12 nm gold particles, unlabeled). The presence of ACTH-unlabeled granules (asterisks) indicates specific reaction of guinea pig anti-ACTH antibody (B–D). Scale, 100 nm.