Islet amyloid, formed by aggregation of islet amyloid polypeptide (IAPP; amylin), is a pathological characteristic of the pancreas in type 2 diabetes and may contribute to the progressive loss of β-cells in this disease. We tested the hypothesis that impaired processing of the IAPP precursor proIAPP contributes to amyloid formation and cell death. GH3 cells lacking the prohormone convertase 1/3 (PC1/3) and IAPP and with very low levels of prohormone convertase 2 (PC2) were transduced with adenovirus (Ad) expressing human or rat (control) proIAPP linked to green fluorescent protein, with or without Ad-PC2 or Ad-PC1/3. Expression of human proIAPP increased the number of transferase-mediated dUTP nick-end labeling (TUNEL)-positive cells 96 h after transduction (+hIAPP 8.7 ± 0.4% vs. control 3.0 ± 0.4%; P < 0.05). COOH-terminal processing of human proIAPP by PC1/3 increased (hIAPP+PC1/3 10.4 ± 0.7%; P < 0.05), whereas NH2-terminal processing of proIAPP by addition of PC2 markedly decreased (hIAPP+PC2 5.5 ± 0.5%; P < 0.05) the number of apoptotic GH3 cells. Islets from mice lacking PC2 and with β-cell expression of human proIAPP (hIAPP+/+/PC2−/−) developed amyloid associated with β-cell death during 2-week culture. Rescue of PC2 expression by ex vivo transduction with Ad-PC2 restored NH2-terminal processing to mature IAPP and decreased both the extent of amyloid formation and the number of TUNEL-positive cells (−PC2 26.5 ± 4.1% vs. +PC2 16.1 ± 4.3%; P < 0.05). These findings suggest that impaired NH2-terminal processing of proIAPP leads to amyloid formation and cell death and that accumulation of the NH2-terminally extended human proIAPP intermediate may be a critical initiating step in amyloid formation.
Type 2 diabetes is characterized by peripheral insulin resistance and a decrease in insulin secretion associated with loss of islet β-cells (1–3). Islet amyloid deposits are a pathological characteristic of the pancreas in type 2 diabetes that likely contribute to the progressive loss of β-cells in this disease (4,5). Islet amyloid is formed by aggregation of islet amyloid polypeptide (IAPP; amylin) (6,7), a 37–amino acid hormone that is colocalized and cosecreted with insulin from β-cells (8,9). IAPP is produced by sequential cleavage of its precursor molecule, proIAPP, at two dibasic (Lys-Arg) sites: first, at its COOH-terminus by prohormone convertase 1/3 (PC1/3), followed by cleavage of the resulting NH2-terminally extended intermediate form at its NH2-terminus by prohormone convertase 2 (PC2) (10–13). The remaining COOH-terminal basic residues following PC1/3 cleavage are removed by carboxypeptidase E during the formation of mature IAPP (14). Proinsulin is similarly processed by sequential cleavage by an endoprotease (followed in each instance by trimming of residual COOH-terminal basic residues by carboxypeptidase E) first by PC1/3 to produce an intermediate form (des-31,32 proinsulin), and then by PC2 to produce mature insulin and C-peptide (15–17).
IAPP-derived amyloid fibrils have been shown to be toxic to β-cells in vitro (18), and studies on both human and nonhuman primates have shown an association between the development of islet amyloid deposits and β-cell loss in type 2 diabetes (19–21). Despite considerable study during the past decade, it is still not understood why islet amyloid forms in type 2 diabetes. The presence of an amyloidogenic sequence in the human IAPP molecule is clearly essential (4,22,23). Elevated IAPP production and secretion from β-cells associated with an increased demand for insulin in type 2 diabetes is likely an important contributor but does not itself appear to be sufficient for islet amyloid formation (24–26). It has been proposed that defective trafficking and/or processing of proIAPP associated with β-cell dysfunction in type 2 diabetes might trigger IAPP aggregation (27–29). A hallmark characteristic of the β-cell defect in type 2 diabetes is impaired processing of proinsulin, leading to disproportionate secretion of proinsulin relative to insulin (30,31). Because proinsulin and proIAPP are processed in parallel by PC1/3 and PC2, it seems probable that like proinsulin, processing of proIAPP will also be impaired in type 2 diabetic patients (23,27,29).
In the present study, we tested the hypothesis that impaired processing of proIAPP may lead to amyloid formation and IAPP-induced cell death and sought to identify the predominant molecular forms of (pro)IAPP that potentiate IAPP fibril formation. We transduced GH3 (rat anterior pituitary) cells, which lack PC1/3 and have very low levels of PC2 expression (13,32), with recombinant adenovirus expressing human proIAPP alone or with PC2 and/or PC1/3 as an in vitro model to investigate the role of proIAPP and its intermediate forms in IAPP-induced cell death. We also cultured islets from mice lacking PC2 (33) and expressing human proIAPP as an in situ model to investigate whether impaired NH2-terminal processing of proIAPP will lead to islet amyloid formation and β-cell death.
RESEARCH DESIGN AND METHODS
Dulbecco’s modified Eagle’s medium, RPMI-1640, Ham’s-F10, trypsin-EDTA, and fetal bovine serum (FBS) were from Invitrogen Canada (Burlington, ON, Canada). Thioflavine S, Hoechst-33342, dithizone, Avertin, collagenase (type XI), DNase, BSA, dextran, 3-isobutyl-1-methylxanthine, phenylmethylsulfonyl fluoride, aprotinin, pepstatin A, and leupeptin were obtained from Sigma-Aldrich (Oakville, ON, Canada). All electrophoresis chemicals were from Bio-Rad Laboratories (Mississauga, ON, Canada). Synthetic human and rat IAPP (1–37) were from Bachem (Torrance, CA).
Transgenic mice expressing human proIAPP were generated in the animal facility of the Child and Family Research Institute, using previously described methods (24). Founders were generated on a C57Bl/6J × CBA/J hybrid background and bred to Bl/6 mice for at least 10 generations. Expression of human IAPP was confirmed in extracts of isolated islets using a human IAPP-specific enzyme-linked immunosorbent assay (Linco Research, St. Charles, MO). Mice lacking active PC2 were described previously (33). Homozygous mice expressing human IAPP and lacking PC2 (hIAPP+/+/PC2−/−) were generated by crossbreeding homozygous human IAPP mice (hIAPP+/+) with heterozygous PC2-null mice (PC2+/−). The animals were cared for in accordance with the principles and guidelines of the Canadian Council on Animal Care. The animals were maintained on 9% fat diet (PMI Feeds, Richmond, IN). Age-matched animals (10–12 weeks old) were used for all studies. Islets were isolated by collagenase digestion as previously described (14). Purity of the islets as assessed by dithizone staining was >95% in all experiments.
Cell cultures.
INS-1 (832/13) β-cells were a gift of Dr. Christopher Newgard (Duke University Medical Center, Durham, NC). GH3 rat anterior pituitary cells were obtained from the American Type Culture Collection (Manassas, VA). INS-1 cells were grown in RPMI-1640 and GH3 cells in Dulbecco’s modified Eagle’s medium containing 11.1 and 25 mmol/l glucose, respectively, supplemented with 10% FBS, β-mercaptoethanol (50 μmol/l; for INS-1 only), penicillin (50 units/ml), and streptomycin (50 μg/ml).
Antisera and recombinant adenoviruses.
Antisera specific for NH2-terminal and COOH-terminal flanking regions of rat and human proIAPP were generated in rabbits using peptides corresponding to amino acids 2–14 (rat, 8541) or 4–16 (human, 8546) and 52–64 (rat, 8543) or 44–56 (human, 8548) of proIAPP, conjugated to keyhole limpet hemocyanin. Antibody titers were assessed by enzyme-linked immunosorbent assay using the synthetic peptides as standard. Antibodies had no detectable cross-reaction with mature IAPP as determined by Western blot using synthetic human or rat IAPP. Anti-human and anti-rodent IAPP antibody (RGG-7321 and -7323) were obtained from Peninsula Laboratories (Belmont, CA), and rabbit antiserum against PC1/3 was provided by Dr. Iris Lindberg (Louisiana State University, New Orleans, LA). Rabbit antiserum against PC2 and adenoviruses expressing human (Ad-hProIAPP) or rat (Ad-rProIAPP) proIAPP without enhanced green fluorescent protein (EGFP) expression, PC2 (Ad-PC2) or PC1/3 (Ad-PC1/3), were generated as previously described (13,34,35). Adenoviruses expressing human or rat proIAPP-EGFP chimeras (Ad-hProIAPP–EGFP and Ad-rProIAPP–EGFP, respectively) were generated, amplified, and purified as previously described (36), using the bacterial recombination method in BJ5183 Escherichia coli cells.
Transduction with recombinant adenovirus.
GH3 cells at ∼70% confluency were transduced with Ad-rProIAPP–EGFP (multiplicity of infection [MOI] 5) or Ad-hProIAPP–EGFP (5) alone or cotransduced with Ad-hProIAPP–EGFP (5) and either Ad-PC2 (3) or Ad-PC1/3 (18) or both for 2 h. After transduction with each adenovirus, cells were incubated with fresh medium containing 10% FBS for 4 h to allow recovery. Expression of human (pro)IAPP in transduced GH3 cells was assessed by detection of green fluorescent protein (GFP) expression in cells by fluorescence microscopy 36 h after adenoviral transduction.
For experiments on primary β-cells, islets from transgenic hIAPP+/+/PC2−/− mice were transduced with Ad-PC2 (MOI 3) overnight, washed, and incubated with Ham’s-F10 medium containing glucose (10 mmol/l), BSA (0.5%), NaHCO3 (14 mmol/l), nicotinamide (10 mmol/l), l-glutamine (2 mmol/l), CaCl2 (1.6 mmol/l), 3-isobutyl-1-methylxanthine (50 μmol/l), penicillin (50 units/ml), and streptomycin (50 μg/ml) for 2 weeks. Medium was changed every 3 days. Expression of PC2 in transduced islets was detected by Western blot in islet lysates and by double immunostaining of paraffin-embedded islet sections for PC2 and insulin.
Transferase-mediated dUTP nick-end labeling, thioflavine S, and immunostaining.
To detect apoptotic cells, GH3 cells were fixed in 4% paraformaldehyde (20 min), permeabilized with 0.5% Triton X-100 in PBS, and incubated with transferase-mediated dUTP nick-end labeling (TUNEL) reaction mixture (Roche Diagnostics, Laval, QC, Canada) for 1 h at 37°C and then stained with Hoechst-33342 for 10 min. For double insulin and thioflavine S staining, islet sections (5 μm) were blocked in PBS containing 2% normal goat serum (Vector Laboratories, Burlingame, CA), rinsed, and incubated with guinea pig anti-insulin antibody (Dako, Carpinteria, CA) at a 1:100 dilution in PBS/1% BSA at 4°C overnight, followed by incubation with Texas Red–conjugated goat anti–guinea pig antibody (Jackson ImmunoResearch Labs, West Grove, PA) for 1 h (1:100) at room temperature. The slides were rinsed and incubated in 0.5% thioflavine S solution for 5 min. For double insulin and PC2 immunostaining, after incubation with guinea pig anti-insulin antibody (overnight, 4°C), slides were incubated with Alexa 488–conjugated goat anti–guinea pig antibody (Molecular Probes, Eugene, OR) for 1 h (1:100) followed by incubation with rabbit anti-PC2 antibody (1:50, overnight, 4°C) and Texas Red donkey anti-rabbit antibody (1:100, 1 h). For double insulin and TUNEL assay, after insulin staining, islet sections were incubated with TUNEL reaction mixture for 30 min at 37°C. The slides were viewed using a Zeiss Axioplan 2 microscope equipped for epifluorescence with a Sensys high-performance charge-coupled device camera (Photometrics, Tucson, AZ) and Quips Pathvysion imaging software (Applied Imaging, Santa Clara, CA). Images were captured with green, red, or Dapi filters and pseudocolored in Pathvysion to create the final image.
Alamar Blue reduction assay.
For assessment of cell viability, attached cells or trypsinized cells in suspension (30,000/well, triplicate) were incubated with Alamar Blue dye (Biosource International, Camarilli, CA) diluted 1:10 in medium for 3 h at 37°C. Fluorescence in samples was measured (excitation 530; emission 590 nm) by a Fluoroskan Ascent plate reader (Thermo Labsystems, Helsinki, Finland) and normalized to blank, and data were calculated as percentage of control.
Electrophoresis and immunoblotting.
GH3 cells or islets were lysed in 100 μl (30 μl for islets) NP-40 lysis buffer as described previously (37). Aliquots of protein (10 or 15 μg) from islet or cell lysates were electrophoresed on a polyacrylamide gel (using Tris-tricine buffer for IAPP) followed by immunoblot analysis using appropriate antisera at the following dilutions: NH2-terminal (8541, 8546) and COOH-terminal proIAPP antisera (8543, 8548) at 1:100; PC2 antiserum, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), and anti-rodent IAPP antibodies at 1:1,000; and PC1/3 antiserum at 1:2,000 for 1 h at room temperature. Membranes were then washed and incubated with horseradish peroxidase–conjugated anti-rabbit IgG (or anti-mouse IgG for GAPDH) (Amersham, Baie dUrfe, QC, Canada) diluted 1:5,000 (1:2,000 for PC1/3) for 1 h. Immunodetection was performed using an enhanced chemiluminescence detection kit (Amersham). Protein bands on films were analyzed by densitometry using Quantity One 4.2.1 (Bio-Rad Laboratories, Hercules, CA).
Statistical analysis.
Data are expressed as means ± SE. Statistical analyses were performed using one-way ANOVA followed by a Newman-Keuls test or Student’s t test as appropriate. P < 0.05 was taken as significant.
RESULTS
Enhanced NH2-terminal processing of human proIAPP by adenoviral expression of PC2 reduces (pro)IAPP-induced cell death in GH3 cells.
To investigate whether impaired processing of proIAPP may contribute to (pro)IAPP-induced cell death, GH3 cells were transduced with Ad-hProIAPP–EGFP or cotransduced with Ad-hProIAPP–EGFP plus Ad-PC2. We have previously shown that PC2 is the essential enzyme for cleaving proIAPP at its NH2-terminus in islet β-cells (12,13). Cells transduced at the same MOI with Ad-rProIAPP–EGFP (or Ad-rProIAPP without EGFP) were used as controls to test the potential toxicity associated with GFP expression and adenoviral transduction. GH3 cells expressing human proIAPP had a markedly higher number of apoptotic (TUNEL-positive) cells compared with nontransduced cells (Fig. 1). Strikingly, addition of PC2 was associated with a reduced number of apoptotic cells and enhanced cell viability compared with those expressing human proIAPP alone (Fig. 1). Enhanced NH2-terminal processing of proIAPP in GH3 cells after addition of PC2 was confirmed by Western blot (Fig. 2). As expected (13), addition of PC2 resulted in a significant decrease (60 ± 7%; P < 0.05) in levels of the NH2-terminally unprocessed proIAPP intermediate as well as increased production of mature IAPP. Transduction with PC2 alone did not have any detectable effect on cell survival (Fig. 1B). GH3 cells expressing rat proIAPP with (Fig. 1B) or without GFP (data not shown) had comparable numbers of apoptotic cells compared with nontransduced cells.
Enhanced processing of human proIAPP at its COOH- terminus by adenoviral expression of PC1/3 potentiates (pro)IAPP-induced cell death in GH3 cells.
The finding that enhanced processing of proIAPP at its NH2-terminus by PC2 reduced its toxicity raised the possibility that NH2-terminally unprocessed proIAPP might be an important intermediate form mediating the toxic effects of proIAPP. To test this hypothesis, GH3 cells were cotransduced with Ad-hProIAPP–EGFP and Ad-PC1/3 using the same conditions described above. Parallel experiments were performed using nonfibrillogenic Ad-rProIAPP–EGFP as a control. Partial (COOH-terminal) processing of human proIAPP by PC1/3 in GH3 cells was associated with a 42 ± 9% increase in levels of the NH2-terminal proIAPP intermediate (P < 0.05; Fig. 2), a 57 ± 8% decrease in levels of the COOH-terminal proIAPP intermediate (P < 0.05; Fig. 2), and a higher number of TUNEL-positive cells compared with GH3 cells expressing human proIAPP alone (Fig. 3). After transduction with both Ad-PC2 and Ad-PC1/3 in GH3 cells expressing human proIAPP, the number of TUNEL-positive cells was even lower than that seen with Ad-hProIAPP–EGFP plus Ad-PC2, suggesting that complete processing of proIAPP to IAPP is protective. Transduction of GH3 cells with Ad-PC1/3 alone or together with Ad-PC2 did not have any detectable effect on cell death in the absence of Ad-hProIAPP–EGFP (Fig. 3B).
Impaired processing of human proIAPP is associated with cell death in INS-1 cells.
If unprocessed or partially processed form(s) of human proIAPP are associated with (pro)IAPP-mediated cell death, then one would expect β-cells, which are able to process proIAPP to IAPP, to be more resistant than GH3 cells to the toxic effects of human proIAPP. INS-1 cells, a β-cell line that exhibits most of the characteristics of primary islet β-cells (38), including proIAPP processing (13), were transduced with Ad-hProIAPP at MOI 4 and 10, the higher MOI being used to induce a level of expression at which proIAPP processing might be impaired. IAPP-immunoreactive forms and apoptotic cells were detected by Western blot and TUNEL assay, respectively. When human proIAPP was expressed in INS-1 cells at lower levels (MOI 4) that did not outstrip the cells’ ability to process it to mature IAPP (Fig. 4A), no increase in cell death was observed (Fig. 4B). Interestingly, expression of proIAPP at higher levels (MOI 10) resulted in impaired processing of proIAPP (manifest as elevated levels of unprocessed and partially processed form[s] of proIAPP in cell lysates; Fig. 4A) and a marked increase in the number of TUNEL-positive cells in transduced cells compared with nontransduced cells (Fig. 4B). These findings are compatible with the hypothesis that cell death occurs when expression of human proIAPP exceeds the capacity of the β-cell to process it, resulting in accumulation of toxic, incompletely processed forms.
To ensure that the lower number of apoptotic cells detected in INS-1 cells transduced with Ad-hProIAPP compared with GH3 cells was related to their superior ability to handle proIAPP and not a lower susceptibility than GH3 cells to the toxic effects of IAPP fibrils, similar numbers of INS-1 and GH3 cells were cultured in the presence or absence of synthetic human IAPP (40 μmol/l). Although cell viability was, as expected, markedly lower in both GH3 and INS-1 cells exposed to synthetic human IAPP, it was much lower in the latter, suggesting that INS-1 cells are more susceptible than GH3 cells to the toxic effects of human IAPP (Fig. 5). This finding makes it highly unlikely that the greater resistance of INS-1 cells to Ad-hProIAPP–mediated cell death is due to an inherent resistance of the cells to IAPP-induced cell death and rather supports the idea that they are better equipped to handle and process the potentially toxic precursors.
Overexpression of PC1/3 and PC2 in INS-1 β-cells enhances processing of proIAPP at its COOH- and NH2-termini, respectively.
These findings suggest that the amount of PC2 relative to proIAPP might be an important determinant of the ability of a cell to process proIAPP and resist amyloid-induced cell toxicity. We therefore investigated whether increasing prohormone convertase expression in INS-1 cells would enhance processing of endogenous proIAPP. INS-1 cells were transduced with either Ad-PC1/3 or Ad-PC2, and IAPP-immunoreactive forms were detected by Western blot in cell lysates. Overexpression of PC1/3 enhanced processing of proIAPP at its COOH-terminus manifest as a 32 ± 7% increase in levels of NH2-terminally unprocessed proIAPP (P < 0.05), whereas PC2 overexpression enhanced proIAPP processing at its NH2-terminus manifest as a 36 ± 5% decrease in levels of the NH2-terminally intermediate form (P < 0.05) compared with nontransduced INS-1 cells (Fig. 6). Furthermore, the lower levels of intermediate proIAPP forms observed in INS-1 cells transduced with Ad-PC2 (Fig. 6A) confirmed our previous finding that PC2 is capable of processing proIAPP at both its NH2- and COOH-termini to form mature IAPP (13).
Expression of PC2 in hIAPP+/+/PC2−/− mouse islets restores NH2-terminal proIAPP processing, decreases amyloid formation, and enhances cell survival.
PC2 knockout (PC2−/−) mouse islets express only an inactive form of proPC2 and have blocked NH2-terminal proIAPP processing, resulting in accumulation of the NH2-terminally extended proIAPP intermediate in islets (12). We crossbred mice with β-cell expression of human proIAPP (mIAPP+/+/hIAPP+/+) onto the PC2−/− background to test whether impaired NH2-terminal processing of human proIAPP might lead to islet amyloid formation and β-cell death in situ. To determine whether restoring expression of PC2 in hIAPP+/+/PC2−/− islets corrects proIAPP processing at its NH2-terminus, islets were transduced with Ad-PC2 (MOI 3) overnight and cultured for 2 weeks. Wild-type, hIAPP−/−/PC2−/−, and hIAPP+/+/PC2+/+ islets (all expressing endogenous mouse IAPP) were used as controls. As expected (33), no mature PC2 was detected by Western blot in hIAPP+/+/PC2−/− islets, but after transduction with Ad-PC2, levels of mature PC2 were comparable with those in wild-type islets (Fig. 7A). Immunostaining confirmed the lack of PC2 expression in hIAPP+/+/PC2−/− islets and its restoration after Ad-PC2 transduction in a high proportion of both β- and non–β-islet cells (Fig. 7B). The adenoviral-mediated rescue of PC2 expression in hIAPP+/+/PC2−/− islets restored normal processing of proIAPP at its NH2-terminus, as demonstrated by a striking increase in levels of the ∼4-kDa mature form of the peptide (Fig. 7C).
When isolated islets from hIAPP+/+/PC2−/− mice were cultured for 2 weeks, they developed amyloid deposits (Fig. 8A) and increased cell death (Fig. 8B) compared with mice expressing both human IAPP and PC2. Remarkably, rescue of PC2 expression and proIAPP processing was associated with a marked decrease in amyloid deposits (Fig. 8A) and in the number of TUNEL-positive cells (Fig. 8B and C), to a level that was comparable with PC2 knockout and wild-type islets. Adenoviral expression of PC2 in islets from hIAPP−/−/PC2−/− mice did not have any significant effect on cell survival, suggesting that the enhanced survival of cultured hIAPP+/+/PC2−/− islets transduced with Ad-PC2 was due to the restoration of proIAPP processing and inhibition of amyloid formation and not due to any other effects of PC2 expression (Fig. 8).
DISCUSSION
In this study, we show that impaired processing of proIAPP at its NH2-terminal cleavage site is an important factor initiating islet amyloid formation and amyloid-induced β-cell death. Consistent with a previous finding in COS-1 cells transfected to express human proIAPP (39), adenoviral-mediated expression of human (but not rat) proIAPP in GH3 cells markedly increased cell death. Coexpression of (pro)IAPP and PC2, the essential enzyme for processing of proIAPP at its NH2-terminus (12), significantly reduced cell death compared with cells expressing human (pro)IAPP alone, whereas coexpression of human proIAPP and PC1/3, which exclusively cleaves proIAPP at its COOH-terminus (13), resulting in the accumulation of NH2-terminally unprocessed proIAPP, potentiated (pro)IAPP-induced cell death. Perhaps even more importantly, complete restoration of proIAPP processing by adenoviral expression of both PC1/3 and PC2 almost totally prevented (pro)IAPP-induced cell death. In support of this, other studies have found that expression of human proIAPP in transformed cell lines lacking the prohormone convertases results in the formation of amyloid fibrils (39,40). Expression of PC2 and PC1/3 alone or together in GH3 cells did not have any detectable effect on cell survival, ruling out any significant effect of these enzymes on processing or production of a biologically active form of a pro-survival molecule.
Islets from hIAPP+/+/PC2−/− mice provide a valuable model to study the role of NH2-terminally unprocessed proIAPP molecular forms in amyloid formation and β-cell death. Although hIAPP+/+/PC2−/− mouse islets form amyloid rapidly during culture with high glucose, these animals in our hands do not develop significant levels of islet amyloid in vivo (L.M., C.B.V., unpublished observations), likely because of their lower blood glucose levels and resulting decreased IAPP synthesis associated with their impaired glucagon production in the absence of PC2 (41). Islets from hIAPP+/+/PC2−/− mice have a much higher rate of cell death in culture compared with hIAPP+/+/PC2+/+ mouse islets, which are able to completely process proIAPP to mature IAPP. Strikingly, rescue of proIAPP processing in hIAPP+/+/PC2−/− islets by adenoviral expression of PC2 was associated with the presence of very low or undetectable levels of amyloid deposits and a marked decrease in the number of TUNEL-positive cells, suggesting a direct correlation between impaired processing of proIAPP at its NH2-terminus and β-cell death associated with amyloid formation. Importantly, these observations were made in intact islets and did not require massive overexpression of human proIAPP, showing for the first time in primary cells that impaired NH2-terminal processing of human proIAPP in situ leads to amyloid formation and cell death.
We previously have shown that GH3 cells transduced to express proIAPP have inefficient proIAPP processing manifest as an accumulation of unprocessed and partially processed form(s) of proIAPP (13). Paulsson et al. (40), using an immunostaining approach, recently reported that neither incompletely processed proIAPP forms nor amyloid were detectable in GH3 cells transfected to express human proIAPP. This discrepancy may be either due to the low sensitivity of immunostaining as a means of assessing prohormone processing or the higher level of PC2 expression in the GH3 cells used by those authors compared with those used in our study. In any case, our findings are in general agreement with the conclusion of Paulsson et al. (40) that aberrant proIAPP processing leads to amyloid formation and add significantly to that study by demonstrating clearly that impaired human proIAPP processing causes cell death as well as amyloid formation, that the critical processing site in proIAPP for preventing amyloidogenesis and cytotoxicity is the NH2-terminal cleavage site, and finally that these findings hold true in primary islets as in cell lines.
Considering the very low (∼5%) transfection efficiency with human proIAPP cDNA in GH3 cells in the study of Paulsson et al. (40) compared with the high (∼70%) transduction efficiency with Ad-hProIAPP in our study, our findings also raise the possibility that the ratio of proIAPP to its converting enzymes might be an important factor in proIAPP processing and amyloid-induced cell death. This idea is further supported by the finding that adenoviral-mediated expression of human proIAPP in INS-1 cells did not cause cytotoxicity unless proIAPP expression was high enough to cause impaired proIAPP processing. Furthermore, overexpression of PC1/3 and PC2 in INS-1 cells accelerates conversion of both proIAPP (Fig. 6) and proinsulin (35). Taken together, these data suggest that the ratio of proIAPP to active PC2 might be very important for efficient processing of proIAPP in β-cell granules. Prolonged exposure to high glucose levels has been shown to result in a decrease in the islet content of PC2 and PC1/3 and in impaired processing of both (pro)insulin (30) and (pro)IAPP (29).
Previous studies have suggested that the increased proinsulin-to-insulin ratio characteristic of type 2 diabetes may be due to an increased demand for insulin in this disease, resulting in the release of proinsulin from immature β-cell granules before its processing is complete (30). If true, because proIAPP and proinsulin are processed and secreted from β-cells in parallel, it seems likely that in type 2 diabetes, secretion of proIAPP and particularly the NH2-terminally extended proIAPP intermediate from β-cells will also be disproportionately increased. The first cleavage of proIAPP by PC1/3 occurs early in the secretory pathway, likely in the trans-Golgi network and immature granules, whereas the second cleavage by PC2 occurs later in the pathway, mainly in mature granules (37), suggesting that the latter step may be more susceptible to impairment in conditions of increased secretory demand. In accordance with this hypothesis, it has been shown that chronic exposure to high glucose concentrations results particularly in an increase in the proportion of the NH2-terminally unprocessed proIAPP intermediate in cultured human islet β-cells (29). It has also been suggested that prolonged exposure to high glucose concentrations results in an increase in the synthesis of (pro)IAPP and favors the selective secretion of (pro)IAPP-immunoreactive forms through the constitutive secretory pathway (9,42). Because NH2-terminal processing of proIAPP by PC2 occurs in granules, constitutive secretion bypasses this step and is an alternative route by which increased secretion of NH2-terminally extended proIAPP may occur (37). We therefore propose that increased secretory demand in type 2 diabetes could lead to release of elevated levels of proIAPP and its NH2-terminal intermediate form either via the constitutive pathway or by release of immature granules. After its secretion, NH2-terminally unprocessed proIAPP may bind to heparan sulfate proteoglycans, a component of both cellular basement membranes and islet amyloid deposits, creating a nidus for extracellular amyloid formation (43). Immunoreactivity for the NH2-terminal (but not COOH-terminal) region of proIAPP has been found to be present in human islet amyloid deposits (28), consistent with the idea that the NH2-terminally unprocessed form of proIAPP may contribute to islet amyloid formation.
Alternatively, elevated levels of proIAPP and its NH2-terminally unprocessed intermediate may initiate formation of intracellular protofibrils/fibrils, which may then act as a nidus for further extracellular fibril formation. Intracellular amyloid has been reported in human insulinomas (44) and human islets transplanted into nude mice (45). Recent studies suggest that protofibrils are more toxic than large IAPP fibrils and might be the major mediators of β-cell death (46–48).
In summary, our data suggest that impaired NH2-terminal processing of proIAPP is an important factor initiating amyloid formation and β-cell death. Defects in processing, sorting, and/or secretion of (pro)IAPP associated with β-cell dysfunction in type 2 diabetes and insulinomas may result in production and secretion of elevated levels of proIAPP and its NH2-terminally unprocessed form, leading to intracellular and/or extracellular fibril formation and β-cell apoptosis. Restoration of intact proIAPP processing may therefore be a potential therapeutic approach to prevent or slow (pro)IAPP-induced β-cell apoptosis and maintain β-cell mass in type 2 diabetes.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Article Information
L.M. has received a postdoctoral fellowship award from the Canadian Diabetes Association; L.H. has received Larry L. Hillblom Foundation Grant 2003/1Mc; P.A.H. has received Swiss National Research Foundation Grant 3200B0-101902; and C.B.V. has received Canadian Institutes of Health Research Grant MT-14682. This work was supported by core support for islet isolation from the Childhood Diabetes Research Unit funded by the Michael Smith Foundation for Health Research.
We thank Dr. Steven Kahn and Dr. Rebecca Hull (University of Washington) for their expert advice for thioflavine S staining. Technical support of Galina Soukhatcheva for islet isolation and Jill McCuaig for growing Ad-hProIAPP and Ad-rProIAPP is gratefully acknowledged.