Pharmaceutical Research, Vol. 21, No. 7, July 2004 (© 2004)
`
`Research Paper
`
`Mishandling of the Therapeutic
`Peptide Glucagon Generates
`Cytotoxic Amyloidogenic Fibrils
`
`Satomi Onoue,1,2,6,7 Keiichi Ohshima,3
`Kazuhiro Debari,4 Keitatsu Koh,4 Seiji Shioda,5
`Sumiko Iwasa,1 Kazuhisa Kashimoto,2 and
`Takehiko Yajima1
`
`Received March 8, 2004; accepted March 23, 2004
`
`Purpose. Some therapeutic peptides exhibit amyloidogenic properties
`that cause insolubility and cytotoxicity against neuronal cells in vitro.
`Here, we characterize the conformational change in monomeric
`therapeutic peptide to its fibrillar aggregate in order to prevent amy-
`loidogenic formation during clinical application.
`Methods. Therapeutic peptides including glucagon, porcine secretin,
`and salmon calcitonin were dissolved in acidic solution at concen-
`trations ranging from 1 mg/ml to 80 mg/ml and then aged at 37°C.
`Amyloidogenic properties were assessed by circular dichroism (CD),
`electron microscopy (EM), staining with ␤-sheet-specific dyes, and
`size-exclusion chromatography (SEC). Cytotoxic characteristics were
`determined concomitantly.
`Results. By aging at 2.5 mg/ml or higher for 24 h, monomeric gluca-
`gon was converted to fibrillar aggregates consisting of a ␤-sheet-rich
`structure with multimeric states of glucagon. Although no aggrega-
`tion was observed by aging at the clinical concentration of 1 mg/ml for
`1 day, 30-day aging resulted in the generation of fibrillar aggregates.
`The addition of anti-glucagon serum significantly inhibited fibrillar
`conversion of monomeric glucagon. Glucagon fibrils induced signifi-
`cant cell death and activated an apoptotic enzyme, caspase-3, in PC12
`cells and NIH-3T3 cells. Caspase inhibitors attenuated this toxicity in
`a dose-dependent manner, indicating the involvement of apoptotic
`signaling pathways in the fibrillar formation of glucagon. On the
`contrary to glucagon, salmon calcitonin exhibited aggregation at a
`much higher concentration of 40 mg/ml and secretin showed no ag-
`gregation at the concentration as high as 75 mg/ml.
`Conclusions. These results indicated that glucagon was self-
`associated by its ␤-sheet-rich intermolecular structure during the ag-
`ing process under concentrated conditions to induce fibrillar aggre-
`gates. Glucagon has the same amyloidogenic propensities as patho-
`logically related peptides such as ␤-amyloid (A␤)1–42 and prion
`protein fragment (PrP)106–126 including conformational change to a
`␤-sheet-rich structure and cytotoxic effects by activating caspases.
`These findings suggest that inappropriate preparation and application
`of therapeutic glucagon may cause undesirable insoluble products
`and side effects such as amyloidosis in clinical application.
`KEY WORDS: aggregation; fibril toxicity; glucagon; salmon calcitonin.
`
`INTRODUCTION
`
`Glucagon is a polypeptide hormone that consists of 29
`amino acid residues and plays a central role in the mainte-
`
`1 Department of Analytical Chemistry, Faculty of Pharmaceutical
`Sciences, Toho University, Funabashi, Chiba 274-8510, Japan.
`2 Health Science Division, Itoham Foods Inc., Moriya, Ibaraki 302-
`0104, Japan.
`3 Applied Genome Informatics Division, Shizuoka Cancer Center
`Research Institute, Shizuoka 411-8777, Japan
`
`nance of normal circulating glucose levels (1). The adminis-
`tration of glucagon induces an increase in hepatic glycogen-
`olysis and gluconeogenesis, and attenuates the ability of in-
`sulin to inhibit these processes. Glucagon is widely used for
`peroral endoscopy, clinical diacrisis, and treatment of hypo-
`glycemia. Water-insoluble glucagon is usually solubilized at
`acidic pH. However, there is at least one serious problem
`regarding the physicochemical property of glucagon in solu-
`tion, resulting in glucagon forming gel-like fibrillar aggregates
`in dilute acid (2). Glucagon is largely unfolded with few stable
`intramolecular bonds under clinical usage, while the confor-
`mation in glucagon fibrils is mainly ␤-sheet-rich. Other thera-
`peutic peptides such as insulin (3), GLP-1 analog (4), and
`growth hormone (5) also display conformational changes into
`␤-sheet-rich fibrils. These insoluble products are attributed to
`the formation of partially unfolded intermediates with an ex-
`posed hydrophobic region that drives the aggregation toward
`the pharmaceutically undesirable form (6).
`The generation of insoluble peptide/protein fibrils is
`well-confirmed in amyloidosis, complex disorders character-
`ized by the polymerization and aggregation of normally in-
`nocuous and soluble proteins or peptides followed by extra-
`cellular insoluble fibrils with resistance to peptidases (7,8).
`There are at least 16 proteins forming amyloid fibrils in clini-
`cally diverse conditions, which include ␤-amyloid (A␤) in
`Alzheimer’s disease (9), amylin in type II (non-insulin-
`dependent) diabete mellitus (10), prion protein (PrP) in
`Creutzfeldt-Jakob disease and spongiform encephalopathy
`(11), and polyglutamine in Huntington’s disease (12). These
`pathologically related amyloidogenic protein/peptide fibrils
`share a distinct conformational feature in the richness of the
`␤-sheet structure (13). In addition, there are similar charac-
`teristics of polarity, hydrophobicity, and the size of side-chain
`among certain segments containing 10–15 amino acid residues
`of amyloid-forming peptides such as insulin, A␤, and amylin,
`and these factors were indicative of a consensus sequence as
`a recognition motif of Congo Red, a specific dye for amy-
`loidogenic protein/peptide fibrils with ␤-sheet dependency
`(14,15). Because there are similar neurotoxic effects of fibrils
`from pathologically related peptides and non-pathologically
`related peptides including glucagon (16), it is plausible that
`there is also a common toxic mechanism related to their sec-
`
`4 Laboratory of Electron Microscopy, Showa University School of
`Medicine, Shinagawa, Tokyo 142-8555, Japan.
`5 Department of Anatomy I, Showa University School of Medicine,
`Shinagawa, Tokyo 142-8555, Japan.
`6 To whom correspondence should be addressed. (e-mail onoue@
`fureai.or.jp)
`7 Current address: Pfizer Global Research and Development, Nagoya
`Laboratories, Pfizer Japan Inc., 5-2 Taketoyo, Aichi 470-2393, Ja-
`pan. Tel: +81-569-74-4855; Fax: +81-569-74-4748; E-mail:
`onoue@fureai.or.jp
`ABBREVIATIONS: A␤, amyloid ␤ peptide; PrP, prion protein;
`WST-8, 4-[3-(2-methoxy-4-nitrophenyl)-2-(4-nitrophenyl)-2H-5-tetra-
`zolio]-1, 3-benzene disulfonate sodium salt; Ac-DEVD-CHO, acetyl-
`Asp-Glu-Val-Asp-1-al; Z-VAD-FMK, N-benzyloxycarbonyl-Val-
`Ala-Asp(O-Me) fluoromethyl ketone; TEM, transmission electron
`microscopy; CD, circular dichroism; ThT, thioflavin T; SEC, size-
`exclusion chromatography; DMEM, Dulbecco’s modified Eagle’s
`medium; LDH, lactate dehydrogenase.
`
`0724-8741/04/0700-1274/0 © 2004 Plenum Publishing Corporation
`
`1274
`
`Page 1
`
`NPS EX. 2046
`CFAD v. NPS
`IPR2015-01093
`
`

`
`Amyloidogenic Properties of Glucagon
`
`1275
`
`ondary and macromolecular structures for different amyloid-
`forming proteins/peptides.
`In this investigation, we have characterized the physico-
`chemical and physiological properties of glucagon fibrils using
`biophysical techniques including circular dichroism (CD),
`transmission electron microscopy (TEM), ␤-sheet-imaging
`probes, and size-exclusion chromatography (SEC). As well as
`glucagon, we also investigated the characteristics of other
`therapeutic peptides, which include porcine secretin that ex-
`hibits 52% amino acid sequence homology to glucagon (17)
`and salmon calcitonin, a therapeutic peptide used for the
`treatment of osteoporosis, Paget’s disease, and hypercalcemia
`(18). We have demonstrated that glucagon requires the low-
`est concentration for fibril formation among those three
`therapeutic peptides and the peptide fibrils of glucagon and
`salmon calcitonin possess the same conformational properties
`and cytotoxic apoptotic signaling pathways by activating
`caspases as fibrils derived from pathologically-related pep-
`tides including A␤1–42 and PrP106–126. Here, we provide fur-
`ther insights into the associative behavior of glucagon, show-
`ing that its noncovalent aggregation was dependent on the
`condition for storage.
`
`MATERIALS AND METHODS
`
`Chemicals
`
`Human glucagon and porcine secretin were synthesized
`by the solid-phase strategy employing optimal side-chain pro-
`tection as reported previously (19). Salmon calcitonin, A␤1–42
`and PrP106–126 were purchased from American Peptide Com-
`pany (Sunnyvale, CA, USA). Congo Red and thioflavin T
`(ThT) were purchased from Wako (Osaka, Japan), and
`WST-8 [2-(2-methoxy-4-nitrophenyl)-3-(4-nitrophenyl)-5-
`(2,4-disulfophenyl)-2H-tetrazolium, monosodium salt] was
`obtained from Dojindo (Kumamoto, Japan). Ac-DEVD-
`CHO and Z-VAD-FMK, caspase inhibitors, were purchased
`from Promega (Madison, WI, USA). Calibration standards
`for size-exclusion column chromatography (SEC) were ob-
`tained from Sigma (St. Louis, MO, USA). Human glucagon
`antibody YP040 was obtained from Yanaihara Institute, Inc.
`(Fujinomiya, Japan).
`
`Aging Treatment of Peptides
`
`Glucagon, salmon calcitonin, and porcine secretin were
`dissolved in 0.01 M HCl at the concentration indicated in the
`text. A␤1–42 and PrP106–126 were prepared in sterile water as
`5 and 10 mg/ml stocks, respectively (20,21). Peptides were
`incubated at 37°C for the periods indicated in the text and
`then diluted to the required concentration.
`
`Transmission Electron Microscopy
`
`An aliquot (2 ␮l) of the peptide gels or solution was
`placed on a carbon-coated Formvar 200 mesh nickel grid. The
`sample was allowed to stand for 15–30 s, and then any excess
`solution was removed by blotting. The samples were nega-
`tively stained with 2% (w/v) uranyl acetate and allowed to
`dry. The samples were then visualized under a Hitachi H-7000
`transmission electron microscope operating at 75 kV. The
`magnification ranged from ×12,000 to ×60,000.
`
`Circular Dichroism Analysis of Amyloidogenic Peptides
`Aged preparations of peptides were dissolved in 20 mM
`Tris-HCl buffer (pH 7.4) or 50% methanol (MeOH)/20 mM
`Tris-HCl buffer, and circular dichroism (CD) spectra (aver-
`age of ten scans) were collected from samples (2 ml) at 0.5 nm
`intervals between wavelengths of 200 and 400 nm using a
`Jasco model J-720 spectropolarimeter (Tokyo, Japan).
`Samples were incubated at room temperature and a baseline
`spectrum was subtracted from the collected data.
`
`Congo Red Binding Assay
`Congo Red-reactive fibrils were measured as described
`previously (22). Aged preparations of peptides were adjusted
`to a concentration of 1 mg/ml, and 40 ␮l of each dilution was
`added to 960 ␮l of 25 ␮M Congo Red in 20 mM potassium
`phosphate buffer (PBS, pH 7.4) containing 150 mM NaCl.
`After a 30-min incubation, absorbance was read at 540 and
`477 nm, and the concentration of bound Congo Red (cb) was
`calculated from the equation, cb ⳱ (A540/25,295–A477/46,306).
`Thioflavin T Binding Assay
`Formation of peptide fibrils was fluorimetrically quanti-
`fied by thioflavin T (ThT) binding (23). An aliquot of 20 ␮l of
`each aged preparation was added to 1980 ␮l of 5 ␮M ThT in
`20 mM PBS (pH 6.0) containing 150 mM NaCl. Fluorescence
`was immediately measured on an RF-5000 spectrofluoropho-
`tometer (Shimadzu, Tokyo, Japan) with excitation and emis-
`sion maxima of 450 and 482 nm, respectively.
`
`Turbidity
`Glucagon was aged at concentrations of 1.0, 2.5, or 5.0
`mg/ml for 24 h in 0.01 M HCl. For turbidity analysis, samples
`aged at the concentrations of 2.5 and 5.0 mg/ml were diluted
`to a final concentration of 1 mg/ml with 0.01 M HCl. An
`aliquot of 200 ␮l of each aged sample as well as non-aged
`glucagon (1.0 mg/ml) was transferred into a 96-well plate and
`turbidity (OD) was measured at 405 nm with a microplate
`reader (Bio-Tek, Winooski, VT, USA).
`
`Size-Exclusion Chromatography
`Aged preparations of peptides were fractionated on a
`Zorbax GF-250 column (Agilent Technologies, Palo Alto,
`CA, USA) at 25°C using a Shimadzu LC-10A HPLC system.
`The column was equilibrated with the mobile phase (20 mM
`citrate buffer containing 130 mM NaCl, pH 3.0), and peptides
`were eluted under constant flow at 1.0 ml/min and monitored
`at 220 nm. The column was calibrated with blue dextran
`(2,000,000 Da) and a series of molecular weight protein stan-
`dards including, sweet potato ␤-amylase (200,000 Da), human
`serum albumin (66,500 Da), chicken albumin (45,000 Da),
`human growth hormone (22,125 Da), porcine insulin (5,777
`Da), and buserelin (1239 Da).
`
`Cell Cultures
`Rat pheochromocytoma (PC12) cells were obtained from
`the RIKEN Cell Bank (Ibaraki, Japan). PC12 cells were cul-
`tured in Dulbecco’s modified Eagle’s medium (DMEM,
`Sigma) supplemented with 5% (v/v) horse serum (HS, Gibco-
`BRL, Grand Island, NY, USA) and 5% (v/v) newborn calf
`serum (CS, Gibco-BRL) as described previously (24). NIH-
`3T3 cells were purchased from American Type Culture Col-
`lection (Manassas, VA, USA) and cultured in DMEM supple-
`
`Page 2
`
`

`
`1276
`
`Onoue et al.
`
`mented with 10% CS. Cells were maintained in 5% CO2/95%
`humidified air at 37°C.
`
`LDH and WST-8 Assay
`
`Cells were seeded at 1 × 104 cells/well in 96-well plates
`coated with type I collagen (Becton Dickinson Labware, Bed-
`ford, MA, USA) at least 24 h before the experiment and
`cultured in serum-free DMEM supplemented with 2 ␮M in-
`sulin. For preparation of peptide fibrils, therapeutic peptides,
`glucagon and salmon calcitonin, were incubated for 24 h at
`5.0 mg/ml and at 60 mg/ml, respectively, and their fibril pro-
`duction was assessed by the Congo Red binding assay. Each
`aged or nontreated peptide was diluted and added to the cell
`culture at the indicated final concentrations. The degree of
`cell death was assessed by measurement of the activity of
`lactate dehydrogenase (LDH) released from the dead cells as
`reported previously (20). LDH activity in the culture medium
`was determined using a commercially available kit (Wako,
`Osaka, Japan) according to the manufacturer’s protocol. In
`addition to LDH measurement in the medium, cell mortality
`was also assayed by WST-8 conversion (25). A volume of 10
`␮l of WST-8 (5 mM WST-8, 0.2 mM 1-methoxy-5-methyl-
`phenazinium methylsulfate, and 150 mM NaCl) was added to
`each well and the reaction was continued for 4 h at 37°C. The
`absorbance of the sample at 450 nm was measured using a
`microplate reader (Bio-Tek) with a reference wavelength of
`720 nm.
`
`Caspase-3-like Activity
`
`Caspase-3-like activity in culture was measured using an
`Apo-ONE Homogeneous Caspase-3/7 Assay Kit (Promega)
`according to the manufacturer’s instructions. Briefly, cells
`(5 × 104 cells/well) in type I collagen-coated 96-well plates
`(Becton Dickinson Labware) were rinsed twice with PBS.
`The cultures were incubated with or without the indicated
`stimulators in DMEM (50 ␮l) at 37°C in an atmosphere of
`95% air and 5% CO2. The cells were lysed in 50 ␮l of Ho-
`mogeneous Caspase-3/7 Buffer containing the caspase-3 sub-
`strate Z-DEVD-Rhodamine 110, and the cell lysates were
`incubated for 12 h at room temperature. After incubation, the
`fluorescence (excitation 480 nm and emission 535 nm) of the
`cell lysates (50 ␮l) was measured with a GEMINIxs spectro-
`fluorophotometer (Molecular Devices, Kobe, Japan).
`
`Statistical Analysis
`
`Statistical evaluation was performed by the Student’s
`t test or one-way analysis of variance (ANOVA) along with
`pairwise comparison by the Fisher’s least significant differ-
`ence procedure. p values less than 0.05 were considered to be
`significant in all analyses.
`
`RESULTS
`
`CD Spectral Analyses on Aged Peptides
`
`nm and a negative extremal band at 218 nm (27), and the
`CD spectrum of ␣-helical structure shows an intense positive
`peak at 192 nm and two negative peaks at 209 nm and 222 nm
`(26). The well-known amyloidogenic peptides, PrP106–126 and
`A␤1–42, gave typical CD spectra for the presence of a ␤-sheet
`structure after aging under both hydrophilic (20 mM Tris-HCl
`buffer, pH 7.4) and hydrophobic (50% MeOH/20 mM Tris-
`HCl buffer, pH 7.4) conditions (Figs. 1D and 1E). When glu-
`cagon was dissolved in the hydrophilic buffer at 10 mg/ml,
`non-aged glucagon gave the characteristic CD spectrum of a
`random coil conformation (Figs. 1A-I). However, when glu-
`cagon was aged for 24 h, a transition from random coil to
`␤-sheet structure was observed for preparations dissolved at
`5 mg/ml or higher. In the hydrophobic condition, the CD
`spectra of non-aged glucagon exhibited a typical ␣-helical
`structure (Fig. 1A-II), the content of which was estimated to
`be 55% according to the calculation established by
`Greenfield&Fasman (26). When aged at 1 mg/ml for 24 h, the
`␣-helical content of glucagon was not changed, whereas aging
`at 2.5 mg/ml or higher resulted in a significant decrease in the
`␣-helical content to ca. 1%.
`We also tested the effect of aging and concentration on
`the conformational change for other therapeutic peptides,
`porcine secretin and salmon calcitonin. Porcine secretin has
`high sequential homology to glucagon and the CD spectra of
`both non-aged peptides exhibited high similarity (Figs. 1A
`and 1B). On the contrary to glucagon, the CD spectrum of
`aged secretin at the concentration of 75 mg/ml was almost
`identical to that of non-aged secretin (Figs. 1B-I and 1B-II),
`suggesting no conformational changes to peptide aggregate of
`␤-sheet-rich structures. Salmon calcitonin, without aging at
`concentrations as high as 75 mg/ml or with aging at a low
`concentration of 1 mg/ml, showed the presence of a random
`coil structure in the hydrophilic environment (Fig. 1C-I) and
`a typical ␣-helical structure in the hydrophobic environment
`(Fig. 1C-II). However, as in the case of glucagon, a confor-
`mational transition from an ␣-helical structure to ␤-sheet
`structure was observed after aging at high concentrations of
`50 and 75 mg/ml (Fig. 1C). It was thus shown that salmon
`calcitonin requires about a 10-fold higher concentration than
`glucagon for structural changes to occur, suggesting that glu-
`cagon has a high propensity to yield ␤-sheet structure.
`
`Electron Microscopic Studies on Peptide Aggregates
`
`Initially, human glucagon was incubated at a concentra-
`tion of 10 mg/ml in 0.01 M HCl for 24 h at 37°C and diluted
`to a final concentration of 1.0 mg/ml prior to applying on
`transmission electron microscopy (TEM). TEM showed well-
`defined fibrils (Fig. 2A), which morphologically resembled
`the classic amyloid fibrils such as A␤1–42 and PrP106–126
`(28,29). The fibril morphologies included largely disordered,
`rigid, and branching fibrils stacked together edge to edge with
`a width of 10–50 nm and various lengths. Similar fibril struc-
`tures were observed for salmon calcitonin aged at the con-
`centration of 60 mg/ml for 24 h (Fig. 2B).
`
`CD is displayed when an optically active substance pref-
`erentially absorbs left or right-handed circularly polarized
`light and it provides useful information when conformational
`alterations occur in peptides/proteins (26). The CD spectrum
`of a ␤-sheet structure shows an intense positive band at 198
`
`Physicochemical Properties of Peptide Aggregates
`
`TEM analysis and CD spectral analyses on aged and con-
`centrated glucagon and salmon calcitonin revealed the pres-
`ence of amyloidogenic fibrils and conformational alterations
`
`Page 3
`
`

`
`Amyloidogenic Properties of Glucagon
`
`1277
`
`Fig. 1. CD spectra representative of glucagon, salmon calcitonin, PrP106–126,
`and A␤1–42. (A) CD spectra of non-aged or aged glucagon in 20 mM Tris-HCl
`buffer, pH 7.4 (A-I) or 50% MeOH/20 mM Tris-HCl buffer, pH 7.4 (A-II).
`Solid line, non-aged glucagon; dotted line, glucagon aged at 1 mg/ml; dashed
`line, aged at 2.5 mg/ml; broken line, aged at 5 mg/ml; and chain line, aged at
`10 mg/ml. (B) CD spectra of non-aged or aged secretin in 20mM Tris-HCl
`buffer, pH 7.4 (B-I) or 50% MeOH/20 mM Tris-HCl buffer, pH 7.4 (B-II).
`Solid line, non-aged porcine secretin; and chain line, aged at 75 mg/ml. (C) CD
`spectra of non-aged or aged salmon calcitoin in 20mM Tris-HCl buffer, pH 7.4
`(C-I) or 50% MeOH/20 mM Tris-HCl buffer, pH 7.4 (C-II). Solid line, non-
`aged salmon calcitonin; dotted line, salmon calcitoin aged at 1 mg/ml; dashed
`line, aged at 50 mg/ml; and chain line, aged at 75 mg/ml. (D) CD spectra of
`non-aged or aged PrP106–126 in 20 mM Tris-HCl buffer, pH 7.4 (solid line) or
`50% MeOH/20 mM Tris-HCl buffer, pH 7.4 (dashed line). (E) CD spectra of
`normal or aged A␤1–42 in 20 mM Tris-HCl buffer, pH 7.4 (solid line) or 50%
`MeOH/20 mM Tris-HCl buffer, pH 7.4 (dashed line).
`
`Page 4
`
`

`
`1278
`
`Onoue et al.
`
`Fig. 2. Electron micrographs of negatively stained fibrils formed by aged (A) glucagon and (B) salmon
`calcitonin. Fibrils of glucagon and salmon calcitonin were prepared by aging at the concentration of 10
`and 60 mg/ml, respectively, in 0.01 M HCl for 24 h. Prior to TEM analysis, the samples were diluted
`to the final concentration of 1 mg/ml. The bars represent 0.2 ␮m.
`
`to the ␤-sheet structures, respectively. Here, in order to put
`these observations together, we performed physicochemical
`analyses including the binding assay using Congo Red and
`ThT, and turbidity assays. Congo Red selectively binds to
`amyloid-like aggregates having a ␤-sheet rich conformation,
`and is used to detect amyloid fibrils in pathological specimens
`(30). When glucagon was incubated at various concentrations
`ranging from 1 to 10 mg/ml at 37°C, bathochromic effect was
`observed upon staining with Congo Red, indicating the for-
`mation of aggregate at 5 mg/ml or higher (Fig. 3A). The rate
`of aggregation was in a time-dependent manner; it reached
`the maximal level approximately 6 min after aging at 7.5 and
`10 mg/ml and 25 min at 5 mg/ml. Glucagon at 1 and 2.5 mg/ml
`did not show any signs of fibrillar formation within 24 h aging.
`In addition to Congo Red, ThT has also been used com-
`monly as an amyloid-binding dye for in vitro studies (23), and
`this fluorochrome is considered to be a potential pharmaco-
`phore for further design of amyloid-imaging agents. ThT
`binding activities were observed in the glucagon preparations
`aged at 2.5 and 5 mg/ml for 24 h, and their fluorescence in-
`tensities at each concentration were 5-fold and 20-fold higher
`than those of glucagon with or without aging at 1 mg/ml,
`respectively (Fig. 3B). There was also a significant increase in
`the turbidity 24 h in the glucagon preparation aged at 2.5 and
`5 mg/ml for 24 h, whereas the turbidity of the preparation
`aged at 1 mg/ml was almost equal to that of the non-aged
`preparation at 1 mg/ml (Fig. 3C). Hence, these physicochem-
`ical analyses indicated that aging at higher concentrations
`(>5 mg/ml) accelerated the misfolding of glucagon, producing
`amyloidogenic fibrils readily.
`Surprisingly, when glucagon was aged at 1 mg/ml for 30
`days at room temperature, the presence of a ␤-sheet structure
`was detected by both Congo Red and ThT binding analyses
`(data not shown). It should also be pointed out that a con-
`comitant formation of a fibrillar aggregate was observed in
`TEM analysis (data not shown).
`The physicochemical properties on salmon calcitonin are
`summarized in Table I. Peptide aggregation was indicated by
`the increases in turbidity, Congo Red and ThT bindings after
`24 h aging. The peptide concentration requisite for aggrega-
`tion was 40 mg/ml in turbidity and ThT binding analyses and
`60 mg/ml in Congo Red binding analysis, which was almost
`consistent with the result of the CD study.
`
`Size-Exclusion Chromatography of Glucagon Aggregate
`
`We performed size exclusion chromatography (SEC) to
`check the size and molecular pattern of the glucagon aggre-
`gate. Each glucagon preparation aged for 0, 1, 3, and 6 h was
`applied to a SEC column (ZORBAX Bio Series GF-250) at
`7.5 mg/ml. The SEC analysis performed with 20 mM sodium
`phosphate buffer containing 130 mM NaCl (pH 7.4) revealed
`that there was only one peak at 3.5 kDa corresponding to a
`molecular weight of monomeric glucagon even with the
`preparations of glucagon aggregate or non-aged glucagon,
`suggesting that the aggregated glucagon was immediately dis-
`aggregated on SEC. On the other hand, when 20 mM citrate-
`HCl buffer containing 130 mM NaCl (pH 3.0) was used as the
`mobile phase, a main peak appeared at 2000 kDa for the
`glucagon aggregate (Fig. 4A), suggesting that glucagon fibril
`was much stable in the acidic condition as compared to the
`neutral condition. During the time course from the non-aging
`state to 6 h aging, the peak at 3.5 kDa of monomeric glucagon
`was shifted to generate a peak at 2000 kDa, suggesting that
`intermolecular association is likely to be involved in the gen-
`eration of the glucagon aggregate.
`Under the acidic condition, the retention times for mo-
`lecular weight standards and non-aged glucagon were well
`associated with each molecular weight (MW) in the range
`from 1.2 kDa to 2000 kDa. The correlation coefficient be-
`tween retention time and Mw was estimated to be r2 ⳱ 0.95
`(p < 0.01) according to the log(MW) vs. elution time standard
`curve (Fig. 4B).
`
`Inhibition of Glucagon Antiserum on the Formation of
`Glucagon Aggregate
`
`The antisera raised against A␤1–42, PrP106–126, and poly-
`glutamine have the ability to suppress their respective aggre-
`gations in vitro and in vivo (29,31,32). Thus, we examined the
`effect of antiserum against glucagon on the formation of glu-
`cagon fibrils. According to X-ray analysis, glucagon adopts a
`mainly ␣-helical conformation in the amino acid sequence
`between positions 10 and 25, which is stabilized by the hy-
`drophobic interactions between glucagon molecules related
`by 3-fold symmetry (33). The glucagon antiserum used rec-
`ognized the entire glucagon molecule but not a fragment of
`glucagon1–12 (data not shown), indicating that the ␣-helical
`
`Page 5
`
`

`
`Amyloidogenic Properties of Glucagon
`
`1279
`
`Table I. Physicochemical Analyses of Aged Salmon Calcitonin
`
`Misfolding properties
`
`Aging
`concentration
`(mg/ml)
`
`Turbidity
`at 405 nm
`(% of control)
`
`Bound
`Congo Red
`(⌬ ␮M)
`
`Thioflavin T
`fluoresecence
`(% of control)
`
`20.0
`40.0
`60.0
`80.0
`
`100 ± 09
`154 ± 16
`355 ± 46
`464 ± 42
`
`0
`0.3
`9.9 ± 1.2
`11.5 ± 1.0
`
`100 ± 04
`351 ± 13
`665 ± 34
`713 ± 56
`
`Salmon calcitonin was dissolved in 20 mM PBS (pH 7.4) containing
`150 mM NaCl at the concentrations of 20, 40, 60, and 80 mg/ml and
`aged for 24 h at 37°C. Misfolding properties of salmon calcitonin were
`assessed by turbidity, Congo Red staining, and ThT fluorometric as-
`says. Data represent the mean ± SD of four experiments.
`
`region in the C-terminal moiety of glucagon represents the
`epitope of this antiserum. When the glucagon antiserum was
`added at the dilution of 1:20 or 1:50 to glucagon dissolved at
`5 mg/ml and aged at 37°C for 24 h, it decreased the fibrillar
`formation of glucagon by 70% and 50%, respectively, as mea-
`sured by Congo Red binding assay (Fig. 5) as well as ThT
`binding fluorescence assay (data not shown). The addition of
`unrelated control serum had no effect on fibril formation.
`
`Cytotoxicity of Peptide Fibrils
`
`To assess the cytotoxicity of the glucagon fibrils, PC12
`cells and NIH-3T3 cells were exposed to 0.1–100 ␮M peptide
`aggregate for 72 h followed by the measurement of cell
`viability determined by WST-8 assay and released LDH.
`A significant decrease in cell viability was observed in cul-
`tures exposed to 10–100 ␮M aged glucagon (p < 0.01) but not
`in cultures treated with 100 ␮M non-aged glucagon (Fig. 6A,
`hatched bars). To determine whether the loss of cell viability
`was equivalent to cell death, the release of LDH was mea-
`sured. Treatment with 10 ␮M aged glucagon induced a sig-
`nificant increase in LDH release compared to control and no
`significant increase in LDH release was observed in cultures
`treated with non-aged glucagon at 100 ␮M and aged glucagon
`at 1 ␮M or lower (data not shown). Thus, glucagon fibrils
`were found to be highly toxic to PC12 cells, which was similar
`to the case of aged PrP106–126 and A␤1–42 (>10 ␮M) (20,21).
`Aged salmon calcitonin also displayed significant cytotoxicity
`in PC12 cells (Fig. 6A, filled bars), whereas non-aged salmon
`calcitonin did not induce significant cell death. As shown in
`Fig. 6B, the incubation of NIH-3T3 cells with these peptide
`fibrils at 25 ␮M or higher for 48 h resulted in a significant
`decrease in cell viability. On the contrary, non-aged peptides
`were not toxic to NIH-3T3 cells. These findings suggested that
`the presence of amyloid fibrils was responsible for the cyto-
`toxicity of aged glucagon or aged salmon calcitonin toward
`both PC12 cells and NIH-3T3 cells. These results were con-
`sistent with a previous report showing that peptide fibrils in-
`duced neurotoxicity in cultured hippocampal neurons while
`native peptide did not affect the cell survival as measured by
`MTT assay (34).
`
`Signaling Pathways in the Cytotoxicity of Peptide Fibrils
`
`Caspases, a family of cysteine proteases, are key media-
`tors of apoptosis. In particular, the activation of caspase-3 is
`
`Fig. 3. Physicochemical analyses of glucagon on the formation of
`amyloid fibrils. (A) Time course of glucagon aggregation as moni-
`tored by Congo Red binding assay. Glucagon was incubated at vari-
`ous concentrations (〫, 1.0 mg/ml; 䊐, 2.5 mg/ml; 䉮, 5.0 mg/ml; 䉭, 7.5
`mg/ml; 䊊, 10 mg/ml) in 20 mM NaPB (pH 7.4) at 37°C for various
`time periods up to 24 h. Samples aged at the concentrations higher
`than 1 mg/ml were adjusted to 1 mg/ml with 0.01 M HCl and applied
`to Congo Red binding assay as described in “Materials and Meth-
`ods.” (B) ThT binding assay. Glucagon was aged at the concentration
`of 1.0, 2.5, or 5.0 mg/ml for 24 h in 0.01 M HCl. Samples aged at the
`concentration of 2.5 and 5.0 mg/ml were diluted to the final concen-
`tration of 1 mg/ml with 0.01 M HCl. Binding abilities of ThT against
`non-aged (1.0 mg/ml) and aged glucagon were measured as described
`in “Materials and Methods.” (C) Turbidity assay: Glucagon was aged
`at the concentration of 1.0, 2.5, or 5.0 mg/ml for 24 h in 0.01 M HCl.
`Samples aged at the concentration of 2.5 and 5.0 mg/ml were diluted
`to the final concentration of 1 mg/ml with 0.01 M HCl. Turbidities of
`non-aged (1.0 mg/ml) and aged glucagon were measured at OD 405
`nm as described in “Materials and Methods.” Each point represents
`the mean ± SD of four determinations.
`
`Page 6
`
`

`
`1280
`
`Onoue et al.
`
`Fig. 5. Effect of glucagon antiserum on the fibril formation of gluca-
`gon. Glucagon was dissolved in 0.01 M HCl, and then glucagon an-
`tiserum or unrelated antiserum was immediately added. The mixture
`was incubated for 24 h at 37°C. In the mixture, the final concentration
`of glucagon was 5.0 mg/ml. The molar ratio between serum (anti-
`glucagon antiserum, hatched bars; unrelated serum, filled bars) and
`the peptide was 1: 20 and 1: 50. Fibril formation was assessed by the
`Congo Red-binding assay. All data are scaled with glucagon alone
`representing 100% aggregation, and data represent the mean ± SD of
`four determinations. ##p < 0.01 with respect to the control group
`(glucagon alone). **p < 0.01 between indicated groups.
`
`Fig. 4. Size-exclusion chromatography of aged glucagon. (A) The
`elution profile of aged glucagon on a Zorbax GF-250 size exclusion
`column. Glucagon was incubated at the concentration of 7.5 mg/ml at
`37°C. Each sample aged at the periods of 0, 1, 3, and 6 h was applied
`on SEC. Arrows (I–VII) represent elutions of a series of the following
`molecular weight protein standards with their retention times: I, blue
`dextran (2000 kDa); II, sweet potato ␤-amylase (200 kDa); III, hu-
`man serum albumin (66.5 kDa): IV, chicken albumin (45 kDa); V,
`human growth hormone (22.1 kDa); VI, porcine insulin (5.8 kDa);
`VII, buserelin (1.2 kDa); and *, glucagon (3.5 kDa). Samples were
`loaded on a SEC column in 20 mM citrate buffer (pH 3.0), and the
`elution profiles were determined by UV detection at 280 nm. (B)
`Standard curve of the log (Mw) vs. elution times.
`
`versible inhibitor of several members of the caspase family
`(37), were used to investigate whether apoptosis was involved
`in the cytotoxicity by aged glucagon. The addition of Z-VAD-
`FMK (100 ␮M) rescued glucagon-induced cell death close to
`the control level (vehicle only) and this protective effect was
`reduced to half at 50 ␮M (Fig. 7B). On the other hand, Ac-
`DEVD-CHO attenuated the fibril toxicity by 40% only at
`high concentration (100 ␮M). These data indicate that not
`only caspase-3 but also other caspases may play an important
`role in the final execution of the cell death program stimu-
`lated by aged glucagon.
`
`required for the early stages of apoptosis that include DNA
`fragmentation and morphological changes (35). To determine
`whether aged glucagon induces caspase-3 activation in PC12
`cells, we exposed PC12 cells to the aged peptide (50 ␮M) and
`measured caspase-3-like activity in cell lysates obtained by
`the cleavage of a fluorometric caspase-3 substrate, Z-DEVD-
`Rhodamine 110 (34). The caspase-3 activity increased prior to
`the loss of me