Inhibition of glycosaminoglycan synthesis and protein glycosylation with WAS-406 and azaserine result in reduced islet amyloid formation in vitro

Rebecca L. Hull, Sakeneh Zraika, Jayalakshmi Udayasankar, Robert Kisilevsky, Walter A. Szarek, Thomas N. Wight, Steven E. Kahn

Abstract

Deposition of islet amyloid polypeptide (IAPP) as amyloid in the pancreatic islet occurs in ∼90% of individuals with Type 2 diabetes and is associated with decreased islet β-cell mass and function. Human IAPP (hIAPP), but not rodent IAPP, is amyloidogenic and toxic to islet β-cells. In addition to IAPP, islet amyloid deposits contain other components, including heparan sulfate proteoglycans (HSPGs). The small molecule 2-acetamido-1,3,6-tri-O-acetyl-2,4-dideoxy-α-d-xylo-hexopyranose (WAS-406) inhibits HSPG synthesis in hepatocytes and blocks systemic amyloid A deposition in vivo. To determine whether WAS-406 inhibits localized amyloid formation in the islet, we incubated hIAPP transgenic mouse islets for up to 7 days in 16.7 mM glucose (conditions that result in amyloid deposition) plus increasing concentrations of the inhibitor. WAS-406 at doses of 0, 10, 100, and 1,000 μM resulted in a dose-dependent decrease in amyloid deposition (% islet area occupied by amyloid: 0.66 ± 0.14%, 0.10 ± 0.06%, 0.09 ± 0.07%, and 0.004 ± 0.003%, P < 0.001) and an increase in β-cell area in hIAPP transgenic islets (55.0 ± 2.6 vs. 60.6 ± 2.2% islet area for 0 vs. 100 μM inhibitor, P = 0.05). Glycosaminoglycan, including heparan sulfate, synthesis was inhibited in both hIAPP transgenic and nontransgenic islets (the latter is a control that does not develop amyloid), while O-linked protein glycosylation was also decreased, and WAS-406 treatment tended to decrease islet viability in nontransgenic islets. Azaserine, an inhibitor of the rate-limiting step of the hexosamine biosynthesis pathway, replicated the effects of WAS-406, resulting in reduction of O-linked protein glycosylation and glycosaminoglycan synthesis and inhibition of islet amyloid formation. In summary, interventions that decrease both glycosaminoglycan synthesis and O-linked protein glycosylation are effective in reducing islet amyloid formation, but their utility as pharmacological agents may be limited due to adverse effects on the islet.

  • islet amyloid polypeptide
  • heparan sulfate
  • proteoglycan
  • β-cell mass
  • β-cell dysfunction

islet amyloid deposition is a pathogenic hallmark of the islet in Type 2 diabetes, occurring in the vast majority of individuals with the disease (33), and is associated with decreased β-cell mass and function (5, 32). The unique amyloidogenic component of islet amyloid is the β-cell peptide islet amyloid polypeptide (IAPP, amylin) (6, 35). Human IAPP (hIAPP) is amyloidogenic and in vitro studies have shown that early aggregates or oligomers of hIAPP are cytotoxic, leading to β-cell death via apoptosis (12, 22). In contrast, the rodent (rat and mouse) forms of IAPP differ from hIAPP in a number of critical amino acids, rendering rodent IAPP non-amyloidogenic and nontoxic (34). Because of these species-specific differences, several groups have produced transgenic mice expressing hIAPP in their pancreatic islet β-cells to create models of islet amyloid deposition. In our colony of hIAPP transgenic mice, male mice develop islet amyloid deposits in vivo following 1 year of high-fat feeding (29). We have also recently developed a rapid in vitro model of islet amyloid deposition by culturing isolated islets from our hIAPP transgenic mice in high glucose for 7 days (10, 38).

Besides the amyloidogenic peptide IAPP, islet amyloid contains other components that are common to all amyloidoses, including those formed in Alzheimer's disease (Aβ amyloid) and chronic inflammation (AA amyloid). These include apolipoprotein E (4), serum amyloid P component (24), and heparan sulfate (HS) proteoglycans (HSPGs) (36), all of which may contribute to hIAPP amyloidogenesis and its related cytotoxicity.

HSPGs in particular may play a role in islet amyloidogenesis. The HSPG perlecan has been shown to be present in human β-cells from individuals with and without Type 2 diabetes (13), and the β-cell synthesizes several HSPGs that are capable of binding amyloidogenic hIAPP but not non-amyloidogenic rodent IAPP (25). Furthermore, binding of amyloidogenic peptides, including IAPP, to HSPGs via their HS glycosaminoglycan (GAG) chains has been shown to stimulate amyloid fibril formation (2, 3). Thus HSPGs may play a critical role in islet amyloid formation, and evidence that decreasing GAG synthesis reduces islet amyloid formation would provide further evidence to support this hypothesis.

We have generated a series of N-acetylglucosamine analogs that act as small molecule inhibitors of GAG synthesis (17–20). These compounds are effective in reducing amyloid formation in a mouse model of AA amyloidosis (20) and in a transgenic mouse model of CNS Aβ amyloid (16). In the present study we examined the effect of one of these compounds, 2-acetamido-1,3,6-tri-O-acetyl-2,4-dideoxy-α-d-xylo-hexopyranose (WAS-406), on β-cell GAG synthesis and on islet amyloid formation in vitro.

Precursors for GAG synthesis are synthesized via the hexosamine biosynthesis pathway (HBP). The HBP is a nutrient-sensing pathway that has many additional effects in the cell, including regulation of O-linked protein glycosylation. Azaserine, an inhibitor of glutamine:fructose-6-phosphate amidotransferase (GFAT), the rate-limiting enzyme of the HBP, has previously been shown to have no effect on GAG synthesis in arterial smooth muscle cells (28). Therefore, we compared the effects of azaserine and WAS-406 on GAG synthesis and islet amyloid formation in vitro.

RESEARCH DESIGN AND METHODS

Isolation and culture of mouse islets.

Hemizygous hIAPP transgenic and nontransgenic mice on a F1 C57BL/6 × DBA/2 background were bred and utilized under protocols approved by the Institutional Animal Care and Use Committee of the Veterans Affairs Puget Sound Health Care System. Eight to ten-week-old male and female mice were anesthetized with pentobarbital sodium (100 mg/kg ip), and pancreatic islets were isolated by collagenase digestion (Collagenase P, 0.5 mg/ml, Roche Applied Science, Indianapolis, IN) via bile duct cannulation, followed by pancreas excision. Islets were purified on a Histopaque gradient, hand picked, and cultured overnight in RPMI-1640 containing 10% fetal bovine serum, 100 U/ml penicillin, 100 μg/ml streptomycin, and 11.1 mM glucose. Islets were then cultured for up to 7 days in medium containing 16.7 mM glucose alone or together with WAS-406 (10–1,000 μM for dose-response studies, 100 μM thereafter). In a subset of experiments, islets were also cultured in the presence of the HBP inhibitor azaserine (20 μM). Stock solutions of WAS-406 (100 mM) and azaserine (20 mM) were prepared in sterile water and then diluted 1:1,000 in culture medium. Culture medium was changed, and thus new compound was supplied to islets every 48 h.

Histological determination of islet amyloid and β-cell area.

At the end of each experiment, islets were fixed in 4% (wt/vol) phosphate-buffered paraformaldehyde. Islets were embedded in agar, refixed in 4% paraformaldehyde, embedded in paraffin, and processed for histology. Five-micrometer sections were cut throughout the islet pellet, and sections at 100-μm intervals were stained for amyloid with thioflavin S and for insulin as previously described (9). Islet area was determined by circumscribing each islet image with a video cursor; the outline of islets is clearly visible when viewed in the thioflavin S channel. Islet-, thioflavin S-positive-, and insulin-positive areas were determined for each islet cross-section in an average of 25 islets per experimental condition. From these data, the following measures were determined: islet amyloid prevalence (%islets containing thioflavin S-positive staining), islet amyloid severity (∑ thioflavin S area/∑ islet area × 100%), and β-cell area (∑ insulin area/∑ islet area × 100%).

Insulin secretion and content.

Insulin secretion was determined after 7 days of culture. Islets (100–150 per condition) were loaded into a perifusion chamber and were preincubated in Krebs-Ringer bicarbonate buffer containing basal (1.67 mM) glucose for 1 h, followed by measurement of basal insulin secretion at 1.67 mM glucose every 2 min for 8 min. Glucose-stimulated insulin secretion was then assessed by perifusion of islets with Krebs-Ringer bicarbonate buffer containing 16.7 mM glucose for 30 min (fractions collected every 2–5 min). Glucose-stimulated insulin secretion was expressed as the incremental area under the curve. Secretion data from these studies using only the hIAPP transgenic and nontransgenic islets cultured for 7 days in 16.7 mM glucose alone have been previously published (38).

For determination of insulin content islets were solubilized in acid ethanol [0.2 M HCl, 48% (vol/vol) ethanol]. Total protein was determined (Coomassie Plus Protein Assay; Pierce Biotechnology, Rockford, IL), and insulin content was measured by radioimmunoassay as described previously (10).

Islet cell viability.

Islet cell viability after 7 days of culture was assessed using the Cell Proliferation Kit I (Roche) according to the manufacturer's instructions. Twenty islets per well (triplicates for each condition) were incubated for 4 h in culture medium containing 0.5 mg/ml 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT). Islets were solubilized by overnight incubation in solubilization buffer, and solubilized formazan was quantified spectrophotometrically (590 nm).

Assessment of GAG synthesis in islets and β-TC3 cells.

Islets were isolated from hIAPP transgenic and nontransgenic mice and recovered overnight as described above. The immortalized islet β-cell line β-TC3 was plated at 1.2 × 106 cells/ml and cultured for 5 days before being studied in DMEM containing 10% fetal bovine serum, 100 U/ml penicillin, 100 μg/ml streptomycin, and 5.5 mM glucose. Islets (450 per plate) and β-TC3 cells were then metabolically labeled for 48 h with [35S]Na2SO4 (100 μCi/ml; MP Biomedicals, Costa Mesa, CA) in medium containing 16.7 mM glucose alone or together with WAS-406 (100 μM for islets and 10–1,000 μM for β-TC3 cells) or azaserine (20 μM; in a subset of experiments). This approach labels newly synthesized GAG chains on proteoglycans.

Incorporation of [35S]Na2SO4 into GAGs was assessed in combined medium plus cell preparations by cetylpyridinium chloride precipitation (31). Labeled media were collected in the presence of protease inhibitors (5 mM benzamidine, 100 mM 6-aminohexanoic acid, and 0.1 mM phenylmethylsulfonyl fluoride), and labeled islets/cells were solubilized in 8 M urea buffer (8 M urea, 2 mM EDTA, 0.25 M NaCl, 50 mM Tris·HCl, and 0.5% Triton-× 100 detergent, pH 7.4) containing protease inhibitors.

To determine the effects of WAS-406 to reduce GAG synthesis and its specificity for HS versus chondroitin-dermatan sulfate (CS/DS), labeled GAGs were purified from β-TC3 cells. [35S]Na2SO4 labeling was performed as described above, and cells were solubilized in phosphate-buffered saline (PBS, pH 7.2) containing 1% (vol/vol) Triton X-100, pooled together with medium and incubated with pronase (Streptomyces griseus 100 μg/ml, Roche) overnight at 37°C. GAGs were isolated by application to DEAE Sephacel equilibrated in PBS and eluted over a 0.15–0.8 M sodium chloride gradient. Residual core protein fragments were removed by alkaline elimination and borohydride reduction, and the resulting GAG preparations were digested with no enzyme (control), heparinase I, II plus III (0.8, 0.4, and 0.8 U per digestion, respectively, Sigma, St. Louis, MO), chondroitinase ABC (0.03 U per digestion, Seikagaku, Cape Cod, MA), or a combination of heparinase I, II, III + chondroitinase ABC for 3 h at 37°C. Reaction products were analyzed by molecular sieve chromatography (Sepharose CL-6B, equilibrated in 0.2 M Tris, 0.2 M NaCl, pH 7.0).

Western blot analysis.

Islet lysates (75 islets per culture condition) were prepared by sonication on ice in 20 mM Tris·HCl, 150 mM NaCl, 1% (vol/vol) Nonidet P-40 (pH 7.5) followed by centrifugation for 20 min at 12,000 g. Total protein concentration was determined using the Coomassie Plus Protein Assay (Pierce Biotechnology), and from this, equal protein (15 μg) was loaded for each sample. Samples were separated by SDS-PAGE and transferred to polyvinylidene fluoride membrane, and nonspecific binding was blocked by incubation in 50 mM Tris·HCl, 150 mM NaCl, 0.1% (vol/vol) Tween-20, and 5% (wt/vol) nonfat dry milk, pH 7.5. Membranes were probed with antisera directed against O-linked N-acetylglucosamine (1:100 dilution, RL2, Affinity BioReagents, Golden, CO). Primary antibody binding was detected using peroxidase-conjugated anti-mouse IgGs followed by enhanced chemiluminescence (PerkinElmer, Wellesley, MA).

Data analysis.

Data are expressed as means ± SE. For MTT islet viability, data are presented as percentage of nontransgenic islets cultured in 16.7 mM glucose due to variability among experiments resulting from the use of different lots of MTT kits. For metabolic-labeling experiments, n = 2–3 proteoglycan preparations; for all other experiments, n = 3–8 islet preparations. Dose-response data were compared by analysis of variance with Bonferroni post hoc analysis. Two group comparisons were made by t-test or Mann Whitney U nonparametric test if the data were not normally distributed. A P ≤ 0.05 was considered statistically significant.

RESULTS

Dose-dependent inhibition of islet amyloid formation by WAS-406.

Incubation of isolated hIAPP transgenic islets for 7 days in 16.7 mM glucose and increasing concentrations of WAS-406 was associated with a marked, dose-dependent decrease in both the prevalence (Fig. 1A) and severity (Fig. 1B) of islet amyloid formation.

Fig. 1.

Dose-dependent decrease in islet amyloid prevalence (%islets with amyloid; A) and severity (%islet area occupied by amyloid; B) in human islet amyloid polypeptide (hIAPP) transgenic islets cultured for 7 days in the presence of increasing concentrations of 2-acetamido-1,3,6-tri-O-acetyl-2,4-dideoxy-α-d-xylo-hexopyranose (WAS-406). n = 4–8 per condition. *P < 0.05 or #P < 0.005 vs. untreated islets.

Effect of WAS-406 on β-cell area, insulin secretion, insulin content, and islet viability.

Based on the dose-response study, a dose of 100 μM of WAS-406 was selected for further assessments of islet morphology. The reduction in amyloid formation observed in hIAPP transgenic islets in the presence of WAS-406 was associated with an increase in β-cell area as a proportion of islet area (Fig. 2A). In contrast, the proportion of islet area composed of β-cells was not significantly altered in nontransgenic islets cultured in the presence or absence of 100 μM WAS-406 (Fig. 2A). Islet area did not differ among treatment conditions or between hIAPP transgenic and nontransgenic islets (data not shown). Islet insulin content was not different in hIAPP transgenic islets compared with that of nontransgenic islets cultured in 16.7 mM glucose alone (Fig. 2B). However, WAS-406 treatment resulted in a significant increase in insulin content in both hIAPP transgenic and nontransgenic islets (Fig. 2B). Glucose-stimulated insulin secretion (n = 4 per condition) was not significantly different with WAS-406 treatment in hIAPP transgenic islets (11.1 ± 2.8 vs. 18.0 ± 6.9 μU/islet × 30 min for untreated vs. WAS-406) or nontransgenic islets (15.2 ± 1.3 vs. 22.2 ± 3.2 μU/islet per 30 min for untreated vs. WAS-406, P = 0.1). Islet viability was significantly lower in hIAPP transgenic islets compared with nontransgenic islets cultured in 16.7 mM glucose alone (Fig. 2C) as we have previously shown (38), and WAS-406 treatment was associated with a tendency toward a reduction in viability in nontransgenic islets (P = 0.065) but no change in viability in hIAPP transgenic islets.

Fig. 2.

Incubation of hIAPP transgenic islets for 7 days in the presence of 100 μM WAS-406 resulted in a significant increase in β-cell area as a proportion of islet area compared with hIAPP transgenic islets cultured in the absence of WAS-406 (A). β-Cell area in nontransgenic (NT) islets was not significantly affected by WAS-406 (A). Islet immunoreactive insulin (IRI) content was significantly increased in both hIAPP transgenic and nontransgenic islets cultured for 7 days in the presence of WAS-406 (B). Islet viability was determined by MTT assay read at an absorbance (Abs) of 590 nm and significantly lower in hIAPP transgenic islets than in nontransgenic islets cultured for 7 days in the absence of WAS-406 (C). WAS-406 tended to decrease islet viability in nontransgenic islets and had no effect on islet viability in hIAPP transgenic islets. n = 5–8 per condition. *P ≤ 0.05 vs. untreated hIAPP transgenic islets; #P ≤ 0.01 vs. untreated nontransgenic islets.

WAS-406 decreases GAG synthesis in β-TC3 cells and islets.

WAS-406 decreased the incorporation of [35S]sulfate into newly synthesized GAGs in a dose-dependent manner in immortalized β-TC3 cells (Fig. 3A). GAG synthesis was reduced by 62% in β-TC3 cells incubated with 100 μM WAS-406 (P < 0.001 vs. untreated cells). GAG synthesis during a 48-h label was similarly reduced with WAS-406 (100 μM) by 66% in both hIAPP transgenic and nontransgenic islets (Fig. 3B). Ion exchange chromatography showed that β-TC3 cell GAGs synthesized in the presence and absence of WAS-406 eluted from DEAE Sephacel at a similar salt concentration (0.45 M NaCl; Fig. 3C), although, as expected, the abundance of GAGs synthesized in the presence of WAS-406 was reduced compared with untreated cells.

Fig. 3.

Analysis of glycosaminoglycan (GAG) synthesis during a 48-h incubation with WAS-406. Incorporation of [35S]sulfate into newly synthesized GAGs was assessed by cetylpyridinium chloride precipitation (A and B). WAS-406 treatment resulted in a dose-dependent decrease in GAG synthesis in β-TC3 cells (A) and a decrease in GAG synthesis in both hIAPP transgenic and nontransgenic (NT) islets (B). n = 2–3 preparations per condition. C: 35S-labeled GAGs synthesized over a 48-h period by β-TC3 cells in the absence (open circles) and presence (closed triangles) of 100 μM WAS-406 and eluted from DEAE Sephacel over a 0.15–0.8 M NaCl gradient. The abundance of GAGs was decreased in the presence of WAS-406 but the NaCl concentration at which these GAGs eluted from DEAE Sephacel was not different between GAGs synthesized in the presence or absence of WAS-406. *P < 0.05 vs. untreated cells/islets.

WAS-406 inhibits heparan and chondroitin-dermatan sulfate synthesis in β-TC3 Cells.

To determine whether WAS-406 was specific for inhibition of HS synthesis, equal numbers of counts of 35S-labeled GAGs synthesized by β-TC3 cells in the presence or absence of WAS-406 were digested with no enzyme (intact GAG control), heparinases I, II, and III (to digest HS), chondroitinase ABC (to digest CS/DS), or the combination of heparinases I, II, III + chondroitinase ABC. If WAS-406 was specific for inhibition of HS synthesis, the residual GAGs synthesized in the presence of this compound would be expected to be insensitive to heparinase digestion but sensitive to chondroitinase ABC treatment.

Analysis of reaction products by CL-6B Sepharose size exclusion chromatography showed that intact GAGs were of a similar size in control (Fig. 4A) and WAS-406 (Fig. 4E)-treated β-TC3 cells. As we showed previously, β-TC3 cells synthesize predominantly HS (25); thus a greater proportion of GAGs synthesized in control cells (without WAS-406) was sensitive to heparinase I, II, and III digestion (Fig. 4B; shaded area denotes digested material, in this case heparinase-sensitive material) than to chondroitinase ABC digestion (Fig. 4C). As expected, no intact GAGs were present following digestion with the combination of heparinase I, II, and III and chondroitinase ABC (Fig. 4D). β-TC3 GAGs synthesized in the presence of WAS-406 contained a similar proportion of heparinase I-, II-, and III-sensitive material (Fig. 4F) and chondroitinase ABC-sensitive material (Fig. 4G) compared with control cells, and all material was sensitive to the combination of heparinase I, II, and III and chondroitinase ABC treatment (Fig. 4H).

Fig. 4.

Sepharose CL-6B size-exclusion analysis of [35S]sulfate-labeled β-TC3 GAGs synthesized in 16.7 mM glucose alone (AD) or together with WAS-406 (100 μM, EH). GAGs were treated with no enzyme (A and E); heparinase I, II, and III (B and F); chondroitinase ABC (C and G); or heparinase I, II, and III + chondroitinase ABC (D and H). A right shift in the size exclusion profile is indicative of sensitivity to a given enzyme combination (shaded areas correspond to digested material). A greater proportion of β-TC3 GAGs were sensitive to heparinase (B) than chondroitinase (C), with all material being sensitive to digestion with a combination of heparinase + chondroitinase (D). GAGs synthesized in the presence of WAS-406 showed similar sensitivity to heparinase (F), chondroitinase (G), and the combination (H), demonstrating that WAS-406 does not disproportionately alter HS or CS/DS synthesis.

WAS-406 decreases O-linked protein glycosylation.

Since WAS-406 is an analog of N-acetylglucosamine, an intermediate of the HBP, we analyzed flux through the HBP by measuring O-linked protein glycosylation in islets treated for 7 days with 16.7 mM glucose alone or with WAS-406 (100 μM) or the HBP inhibitor azaserine (20 μM) as a positive control. WAS-406 and azaserine treatment in both nontransgenic islets (Fig. 5) and hIAPP transgenic islets (not shown) resulted in a decrease in O-linked protein glycosylation as demonstrated by a reduced number of glycosylated proteins detected using an anti-O-linked N-acetylglucosamine antibody. The magnitude of the effect to decrease O-linked protein glycosylation was greater with WAS-406 than with azaserine.

Fig. 5.

Western blot showing O-linked protein glycosylation in nontransgenic islets cultured for 7 days in the presence or absence of 100 μM WAS-406 or 20 μM azaserine. Several O-linked N-acetylglucosamine-modified proteins are detected in islets following 7 days of culture in 16.7 mM glucose. In contrast, a decrease in O-linked N-acetylglucosamine-modified proteins was seen following WAS-406 or azaserine treatment, consistent with a downregulation of flux through the hexosamine biosynthesis pathway.

Effect of the HBP inhibitor azaserine on proteoglycan synthesis and islet amyloid deposition.

Given the similar effects of WAS-406 and azaserine to decrease O-linked protein glycosylation, but the reported lack of an effect of azaserine to reduce GAG synthesis in arterial smooth muscle cells (28), we sought to determine the effect of azaserine on β-cell GAG synthesis and on islet amyloid formation. Unexpectedly, azaserine treatment in β-TC3 cells was associated with a significant reduction in GAG synthesis as determined by cetylpyridinium chloride precipitation (Fig. 6A) and elution from DEAE Sephacel over a 0.15–0.8 M NaCl gradient (Fig. 6B). Furthermore, incubation of hIAPP transgenic islets for 7 days in 16.7 mM glucose and in the presence of azaserine (20 μM) was associated with a marked reduction in islet amyloid prevalence (78 ± 5 vs. 10 ± 6% in untreated vs. treated islets, P < 0.01) and islet amyloid severity (Fig. 6C) and with an increase in the proportion of β-cell area to islet area (Fig. 6D).

Fig. 6.

Analysis of GAG synthesis in β-TC3 cells during a 48-h incubation with or without azaserine (20 μM). Azaserine reduced incorporation of [35S]sulfate into newly synthesized proteoglycans from β-TC3 cells, assessed by cetylpyridinium chloride precipitation (A), and was associated with a reduction in abundance of GAGs assessed by elution from DEAE Sephacel over a 0.15–0.8 M NaCl gradient (B; untreated cells shown in open circles, 20 μM azaserine shown in closed squares). n = 3 per condition *P < 0.05 vs. untreated cells. Incubation of hIAPP transgenic islets for 7 days in 16.7 mM glucose and 20 μM azaserine was associated with a significant reduction in islet amyloid severity (C) and a significant increase in β-cell area (D). n = 4 per condition. #P ≤ 0.01 vs. untreated islets.

DISCUSSION

We have shown for the first time that interventions that decrease GAG synthesis and protein glycosylation in islets are capable of inhibiting islet amyloid deposition and preventing the amyloid-induced loss of β-cell area in vitro. In the present study, WAS-406 and azaserine reduced both GAG synthesis and flux through the HBP, measured as O-linked protein glycosylation, resulting in a marked decrease in islet amyloid deposition.

Whereas WAS-406 has been shown to be effective in reducing other amyloidoses (16, 20), it was not clear that the same would hold true for islet amyloid formation. In the present study, we found that incubation of hIAPP transgenic islets with this compound resulted in a marked, dose-dependent decrease in both the prevalence and severity of islet amyloid; this was associated with an increase in β-cell area. We also found that WAS-406 was effective in reducing [35S]sulfate incorporation into GAGs synthesized by both intact primary islets and by immortalized β-TC3 cells, in line with previous data that this compound also decreased GAG synthesis in hepatocytes; reducing both [35S]sulfate and [3H]glucosamine incorporation into GAGs (20). In β-TC3 cells, residual GAGs synthesized in the presence of WAS-406 were sensitive to both heparinase and chondroitinase treatment, with the proportions of HS and CS/DS being similar to control cells. These data suggest that WAS-406 is not specific for HS synthesis, which is not unexpected, given that the CS/DS disaccharide constituent N-acetylgalactosamine is formed by the epimerization of N-acetyglucosamine (a component of the repeating disaccharide in HS) and that levels of these two compounds exist in equilibrium, such that any intervention that perturbs the levels of one compound would also be expected to affect levels of the other. Since HS and CS/DS are important components of the cell and extracellular matrix, it is possible that the effect of decreased GAG synthesis with WAS-406 may have confounded the results of the present study. The inclusion of nontransgenic islets treated with WAS-406 allowed us to distinguish the effects of these compounds to decrease amyloid formation from other effects due to decreased GAG synthesis, but the potential adverse effects of long-term use of WAS-406 may limit its utility as a pharmacological agent, as described in more detail below.

Previously, we have shown that a radiolabeled analog of N-acetylglucosamine, similar to WAS-406, can be incorporated into growing GAG chains, blocking their extension (20). It is possible that WAS-406 acts via this mechanism in β-cells, although in the present study we did not evaluate whether it was directly incorporated into GAGs. Alternatively, WAS-406 may reduce islet GAG synthesis via downregulation of the HBP, the biosynthetic pathway responsible for N-acetyglucosamine and N-acetylgalactosamine synthesis. The observed marked decrease in O-linked protein glycosylation following incubation of islets for 7 days with WAS-406 is consistent with downregulation of the HBP.

Unexpectedly, incubation of islets with azaserine also resulted in a reduction of GAG synthesis in β-TC3 cells. This differs from a previous study showing no effect of azaserine in the presence of high glucose to alter GAG synthesis in arterial smooth muscle cells (28). This may be due to the fact that, in that cell type, substrate supply through the HBP was not rate limiting for GAG synthesis at high glucose. However, increasing glucose concentrations have been associated with increased HBP activity in other cell types (7), and this is also likely the case in β-cells, since β-cell glucose metabolism is regulated by the high Km enzyme glucokinase. Thus, as our data suggest, azaserine would be predicted to be more effective at inhibiting GAG synthesis at high glucose in β-cells than in smooth muscle cells. The decreases in GAG synthesis and O-linked protein glycosylation with azaserine were smaller in magnitude than those seen with WAS-406, which is likely a result of the lower dose of azaserine (20 μM vs. 100 μM for WAS-406). This lower dose of azaserine was chosen based on previous studies in mouse islets (37), with a higher dose being avoided due to concerns of toxicity. However, this lower dose of azaserine was extremely effective in inhibiting islet amyloid formation, suggesting that modulation of flux through the HBP is important for islet amyloid formation.

Whereas it is possible that the effects of WAS-406 and azaserine to downregulate HBP flux in general, rather than a specific effect on GAG synthesis, may explain their ability to reduce islet amyloid formation, other lines of evidence support our hypothesis that decreased GAG, and particularly HS, synthesis is critical for the deposition of amyloid. First, several in vitro studies have shown that the HSPG perlecan HS and the related GAG heparin can bind amyloidogenic peptides, including hIAPP (2, 3, 25), and that this interaction leads to increased amyloid fibril formation (2, 3). HS is more effective in increasing hIAPP fibril formation than CS/DS (2), suggesting a specific effect for HS in fibril formation. More direct evidence comes from studies using a murine model of AA amyloidosis together with overexpression of the enzyme heparanase that results in in vivo fragmentation of HS but does not target CS/DS (21). Transgenic mice overexpressing heparanase showed marked resistance to hepatic and renal amyloid deposition but developed amyloid deposition in the spleen in a model of AA amyloidosis. The differential organ sensitivity in this transgenic mouse was inversely correlated with the organ expression of the transgene. That is, the heparanase transgene and its protein were expressed in the kidney and liver but not the spleen. Thus, in the same animal, fragmentation of HS was observed in the kidney and liver but not the spleen rendering only the spleen sensitive to AA amyloid deposition. This argues for a clear role of HS in amyloidogenesis.

Our observation that decreasing GAG synthesis with an N-acetylglucosamine analog results in reduced islet amyloid formation is in keeping with findings that these compounds reduce amyloid formation in mouse models of AA and CNS amyloidosis (16, 20). However, whereas these effects were similar and amyloid deposits derived from different amyloidogenic peptides contain several invariant components, it is important to recognize that the role of these components in amyloid formation is not equivalent among different amyloidoses. For example, apolipoprotein E is important in the development of Aβ amyloid, as shown by delayed and reduced amyloid deposition in a mouse model of Aβ deposition lacking one or both apolipoprotein E alleles (1). In contrast, we found that whereas apolipoprotein E is a component of islet amyloid, apolipoprotein E deficiency had no effect on islet amyloid formation in human IAPP transgenic mice (30). Furthermore, we found that apolipoprotein E is not synthesized by islet β-cells (30), suggesting that it may be trapped during the deposition of islet amyloid, perhaps by binding to proteoglycans, but that it appears to have no causative role in islet amyloidogenesis.

Whereas we have observed that WAS-406 and azaserine are capable of decreasing amyloid formation, an important aspect of our findings is that they highlight the need for specificity in the development of amyloid inhibitors. Whereas increased flux through the HBP has been shown to be detrimental to the β-cell (14, 27), protein glycosylation is critical for normal cellular function, such that decreased protein glycosylation is associated with impaired insulin secretion (37). Thus the effects of chronic downregulation of the HBP by WAS-406 or azaserine may have effects on the cell that are unrelated to their ability to reduce islet amyloid formation and may be detrimental to the β-cell. Islet viability tended to be decreased in nontransgenic islets treated with WAS-406, and islet viability did not improve when islet amyloid was inhibited with WAS-406 in contrast to our findings with other amyloid inhibitors (Zraika S et al., unpublished observation), consistent with a detrimental effect of WAS-406 on islet viability. In contrast, insulin content was significantly increased in both hIAPP transgenic and nontransgenic islets with WAS-406 treatment, a finding seemingly at odds with the viability data. However, glucose-stimulated insulin secretion was not significantly altered (in islets from either genotype) in the presence of WAS-406, despite this significant increase in insulin content. Thus we speculate that the increase in insulin content may occur, at least in part, due to a failure to adequately increase insulin secretion following WAS-406 treatment, consistent with a role of WAS-406 to decrease β-cell secretory function and impair islet viability.

Even in the absence of downregulation of the HBP, systemic downregulation of GAG synthesis would also be expected to have detrimental effects on the cell, because GAGs at the cell surface and in the extracellular matrix are important for sequestration and signaling of growth factors and chemokines (11). Thus alternative approaches to reduce islet amyloid deposition have to be pursued. To date these have mainly targeted the amyloidogenic peptide IAPP. Several groups have described peptide-based inhibitors that are effective at reducing fibril formation and/or cytotoxicity of synthetic hIAPP (15, 26). However, the bioavailability and thus applicability of these compounds have not been studied in models of de novo islet amyloid formation. Similarly, small molecules such as Congo red and rifampicin have also been shown to reduce hIAPP fibril formation (8, 23), but further work is required to investigate their toxicity and/or efficacy in long-term studies.

In summary, our data demonstrate that small molecules such as WAS-406 and azaserine are highly effective inhibitors of islet amyloid formation. However, further studies with more specific interventions will be required to definitively prove that HS is required for islet amyloid deposition. Understanding of the actions of compounds that are capable of inhibiting islet amyloid will be useful in developing therapeutic interventions to reduce amyloid formation that occurs as part of the islet lesion in Type 2 diabetes.

GRANTS

This work was supported by the Department of Veterans Affairs (to S. E. Kahn); National Institutes of Health Grants DK-17047 (to R. L. Hull), DK-74404 (to R. L. Hull), and RR-16066 (to S. E. Kahn); and the American Diabetes Association (to S. E. Kahn). Preparation and development of WAS-406 was funded by the Canadian Institutes for Health Research Grant MOP-3153 (to R. Kisilevsky), the Natural Sciences and Engineering Research Council of Canada (to W. A. Szarek) and the Institute for the Study of Aging (to R. Kisilivsky and W. A. Szarek).

Acknowledgments

We thank Jeanette Teague, Rebekah Koltz, Rahat Bhatti, Robin Vogel, Shani Wilbur, and Michael Peters for excellent technical support. We are grateful to Michael Kinsella for helpful discussion in the preparation of this manuscript.

Footnotes

  • 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.

REFERENCES

View Abstract