Degradation of insulin amyloid by antibiotic minocycline and formation of toxic intermediates

Insulin balls, localized insulin amyloids formed at subcutaneous insulin-injection sites in patients with diabetes, cause poor glycemic control owing to impairments in insulin absorption. Our previous study has shown that some insulin balls are cytotoxic, but others are not, implying amyloid polymorphism. Interestingly, the patient with toxic insulin balls had been treated with antibiotic minocycline, suggesting a possible relationship between toxicity of insulin balls and minocycline. However, the direct effect of minocycline on the structure and cytotoxicity of the insulin amyloid is still unclear. Herein, we demonstrated that that minocycline at physiological concentrations induced degradation of insulin amyloids formed from human insulin and insulin drug preparations used for diabetes patients. Interestingly, the process involved the initial appearance of the toxic species, which subsequently changed into less-toxic species. It is also shown that the structure of the toxic species was similar to that of sonicated fragments of human insulin amyloids. Our study shed new light on the clarification of the revelation of insulin balls and the development of the insulin analogs for diabetes therapy.

pFTAA and BTD21 assay for the insulin amyloid samples incubated with minocycline. pFTAA and BTD21 assay of (i)-amyloid (upper) and (r)-amyloid (lower) incubated in the presence of 0-100 μM minocycline. The left images are the spectrum of pFTAA or BTD21 of insulin amyloids incubated for 72 h: probe (thin solid line), amyloid (thick solid line), and 25, 50, and 100 μM minocycline (thick-dashed, thick-dotted, and solid-dashed line, respectively). The right pictures are plots of the intensities normalized for insulin amyloid as 100%: amyloid (white squares) and 25, 50, and 100 μM minocycline (black, gray and white circles, respectively). and (r)-amyloid (upper right) were incubated with minocycline for 1 week at the indicated concentrations. The solutions were centrifuged at 15,000 rpm for 15 min. The protein concentration in the supernatant was adjusted to the same protein concentration (17.2 µM (0.1 mg/mL)) as determined by a BCA assay. The size was analyzed using Zetasizer Nano-ZS (Malvern, Worcestershire, UK). The average values of three repeated measurements are shown. (B) (i)-amyloid (upper left) and (r)-amyloid (upper right) incubated with a fixed minocycline concentration (50 µM) for the indicated periods.

Figure S3 Cytotoxicity of minocycline against HeLa cells using MTT assay
Cytotoxicity of minocycline against HeLa cells using the MTT assay. The cell viability with PBS was normalized to 100%.  (upper) or with a fixed minocycline concentration (50 µM) for the indicated periods (lower). Samples incubated without minocycline for 7 days were shown as controls (0 µM and 0 day). All samples were quantified by BCA assay and were diluted to the same protein concentration: 0.1, 0.5, and 1 μM (gray, white, and black, respectively). The absorbance was normalized to 100% for PBS. (*P <0.05 , **P <0.01 , ***P <0.005)

Figure S5 Cytotoxicity of the uncentrifuged insulin amyloid samples incubated with minocycline (uncentrifuged) using MTT assay against HeLa cells (A) and PC12 cells (B).
Cytotoxicity of the uncentrifuged samples against HeLa cells (A) and PC12 cells (B) using MTT assay. The (i)-amyloid (left) and (r)-amyloid (right) samples were incubated with minocycline for 1 week at the indicated concentrations and centrifuged (upper) or with a fixed minocycline concentration (50 µM) for the indicated periods (lower). Samples incubated without minocycline for 7 days were shown as controls (0 µM and 0 day). All samples were quantified by BCA assay and were diluted to the same protein concentration: 0.1, 0.5, and 1 μM (gray, white, and black, respectively). The absorbance was normalized to100% for PBS. (*P <0.05 , **P <0.01 , ***P <0.005) The (i)-amyloid (upper) and (r)-amyloid (lower) samples were incubated with a fixed minocycline concentration (50 µM) for the indicated periods, and the precipitate and supernatant samples were evaluated by ANS (A) and hFTAA (B). Left panels are the probe spectrum with: native insulin (thindotted line), probe (thin solid line), and samples at 0, 1, 3, and 7 days (thick solid, thick-dashed, thick dotted, and solid-dashed lines, respectively. The right panels are the peak fluorescence values at 480 nm (ANS) and at 560 nm (hFTAA): precipitate (gray), supernatant (white) and native insulin (black). The protein concentrations were adjusted to 5 µM. (C) Dot blot assay of the degradable amyloid samples in the precipitate by anti-insulin A chain and B chain antibody. The (i)-amyloid and (r)amyloid samples were incubated with a fixed minocycline concentration (50 µM) for the indicated periods, and 5 μM precipitate samples after centrifugation were used. Graphs show the intensity analysis using the anti-A-chain antibody (left) and anti-B-chain antibody (right) by ImageJ: (i)amyloid (gray) and (r)-amyloid (white). Intensities were normalized to 100 % for each untreated amyloid. (*P <0.05 , **P <0.01 , ***P <0.005)

Figure S7 Cytotoxicity of the sonicated insulin amyloids using MTT assay against PC12 cells.
Cytotoxicity of (i)-amyloid (left) and (r)-amyloid (right) sonicated for various times against PC12 cells, as evaluated by MTT assay. All samples were quantified by BCA assay and were diluted to the same protein concentration: 0.1, 0.5, and 1 μM (white, gray, and black, respectively). The absorbance was normalized to 100% for PBS.

Figure S13 Degradation of insulin amyloids by minocycline at low concentrations
ThT assay of (i)-amyloid (left) and (r)-amyloid (right) incubated for an extended period of time in the presence of 0-10 μM minocycline. The intensities without incubation were normalized as 100%.