Prion protein modulates glucose homeostasis by altering intracellular iron

Abstract

The prion protein (PrP^C), a mainly neuronal protein, is known to modulate glucose homeostasis in mouse models. We explored the underlying mechanism in mouse models and the human pancreatic β-cell line 1.1B4. We report expression of PrP^C on mouse pancreatic β-cells, where it promoted uptake of iron through divalent-metal-transporters. Accordingly, pancreatic iron stores in PrP knockout mice (PrP^-/-) were significantly lower than wild type (PrP^+/+) controls. Silencing of PrP^C in 1.1B4 cells resulted in significant depletion of intracellular (IC) iron, and remarkably, upregulation of glucose transporter GLUT2 and insulin. Iron overloading, on the other hand, resulted in downregulation of GLUT2 and insulin in a PrP^C-dependent manner. Similar observations were noted in the brain, liver, and neuroretina of iron overloaded PrP^+/+ but not PrP^-/- mice, indicating PrP^C-mediated modulation of insulin and glucose homeostasis through iron. Peripheral challenge with glucose and insulin revealed blunting of the response in iron-overloaded PrP^+/+ relative to PrP^-/- mice, suggesting that PrP^C-mediated modulation of IC iron influences both secretion and sensitivity of peripheral organs to insulin. These observations have implications for Alzheimer's disease and diabetic retinopathy, known complications of type-2-diabetes associated with brain and ocular iron-dyshomeostasis.

Figure 1

PrP is expressed on insulin positive β-cells in mouse pancreas, primary cells from mouse pancreas, and human insulin producing β-cell line 1.1B4: (a) Schematic representation of full length (FL), α-cleaved (C1), and β-cleaved (C2) forms of PrP and antibody reactivity. (b) Probing of pancreatic lysates for PrP shows the expected glycoforms in C6 PrP^+/+ samples and no reactivity in C6 PrP^−/− samples (lanes 1 & 2). β-actin provides a loading control. (c) Probing of lysates from 1.1B4 cells for PrP shows the expected glycoforms that are down-regulated in cells transfected with PrP-specific siRNA. α-tubulin provides a loading control. The full-length blots are provided in Supplementary File (Raw data). (d) Immunoreaction of 1.1B4 cells for PrP shows plasma membrane and endosomal reaction (panels 1 & 2). Scale bar: 20 μm. (e) Immunoreaction of pancreatic sections from Tg40 PrP mice with 3F4 shows a positive reaction for PrP that co-localizes with insulin staining on β-cells (panels 1 & 2). No reaction for PrP is detected in PrP^−/− samples that show a robust reaction for insulin (panels 3 & 4). (f) Primary cultures from Tg40 PrP pancreas show a positive reaction for PrP in insulin positive β-cells (panels 1 & 2). No reaction for PrP is detected in cells from PrP^−/− pancreas though insulin positive β-cells are prominent (panels 3 & 4). Scale bar: 20 μm.

Figure 2

Iron uptake is coupled with β-cleavage of PrP^C: (a) Probing of deglycosylated pancreatic lysates from C6 PrP^+/+ mice with 8H4 shows β-cleavage of majority of PrP. No reactivity is detected in C6 PrP^−/− samples as expected. Most of the PrP from human brain homogenate is cleaved at the α-site (lanes 3). (All samples were fractionated and processed on the same gel. Complete gel is included in supplementary file-raw data). *Represents a non-specific band. (b) Probing of deglycosylated pancreatic lysates from Tg40 PrP mice with 3F4 shows β-cleavage of majority of PrP (lanes 1–6). Systemic iron overload increases FL and the β-cleaved form of PrP (lanes 2, 4 & 6 vs. 1, 3 & 5). Most of the PrP in human brain homogenate is FL (lanes 8 & 9). Pancreatic lysates from PrP^−/− mice show no reactivity for 3F4 as expected. (c) The ratio of C2 vs. full-length PrP is significantly higher in pancreatic lysates relative to the brain. Values are mean + SEM of the indicated n. ***p < 0.001. (d) Exposure of PrP-GFP expressing 1.1B4 cells to ferric ammonium citrate (FAC) results in the loss of GFP tag from the plasma membrane within 20 minutes (panels 1–4). (e) Exposure of 1.1B4 cells expressing GFP-tagged DMT-1, ZIP8, or ZIP14 to FAC, on the other hand, shows no change in localization or intensity of GFP fluorescence up to 20 minutes (panels 1–6). Scale bar 20 μm. (f) Probing of deglycosylated lysates from 1.1B4 cells expressing vector, DMT-1, ZIP8, and ZIP14 with 3F4 reveals significant increase in the ratio of C2/FL in cells expressing DMT-1 and ZIP14 (lane 5 vs. 6 & 8; g). (The expression of GFP-tagged transfected proteins was equivalent (at 15–17%) as determined by manual counting of 20 random fields of each at 5×)- (see Supplementary Fig. S3). (g) The ratio of C2 vs. FL after normalization with transfection efficiency and β-actin shows a significant increase in cells expressing DMT-1 and ZIP14. Values are mean + SEM of the indicated n. *p < 0.05. The full-length blots are provided in Supplementary File (Raw data).

Figure 3

PrP mediates iron uptake by pancreatic β-cells: (a) Non-transfected and 1.1B4 cells transfected with scrambled or PrP-specific siRNA were cultured in the absence (−FAC) or presence of ferric ammonium citrate (+FAC) for 16 h, and lysates were processed for Western blotting. Probing with 3F4 shows PrP glycoforms, and down-regulation by PrP-specific siRNA. Probing for ferritin shows significant down-regulation in cells where PrP had been silenced. Exposure to exogenous iron upregulates ferritin in untreated controls and cells transfected with scrambled siRNA, but has minimal effect on cells where PrP is silenced. (b) Quantification of ferritin expression by densitometry after normalization with β-actin. Values are mean ± SEM of the indicated n. *Represents change in ferritin relative to untreated, non-transfected control. ^##Represents change in ferritin relative to FAC exposed non-transfected control. *p < 0.05, ^##p < 0.01. (c) Western blotting shows upregulation of PrP in iron-overloaded C6 PrP^+/+, no signal in C6 PrP^−/− samples, and the expected glycoforms in human brain sample (lane 9). Ferritin is significantly higher in iron-overloaded relative to untreated C6 PrP^+/+ samples and matched C6 PrP^−/− samples. There is no change in ferritin iron overloaded C6 PrP^−/− samples relative to untreated controls. Probing for TfR shows significant reduction in iron-overloaded relative to untreated C6 PrP^+/+ samples and matched C6 PrP^−/− samples. There is minimal change in TfR expression in iron-overloaded C6 PrP^−/− samples relative to untreated controls. (d) Density of protein bands after normalization with β-actin. Values are mean ± SEM of the indicated n. *Represents change in expression relative to untreated C6 PrP^+/+ samples. *p < 0.05. Full-length blots are provided in Supplementary File (Raw data). (e) Immunoreaction of fixed pancreatic sections from the above mice for ferritin shows a positive reaction in mainly β-cell rich endocrine islets (panels 1–4). Iron over-loading increases ferritin reactivity in C6 PrP^+/+ sections, but shows minimal change in C6 PrP^−/− samples (panels 3 & 4). Reactivity for TfR is also localized to the endocrine islets (panels 5–8). Iron-overloading down-regulates TfR expression in C6 PrP^+/+ samples (panel 6 vs. 5), but has minimal effect on C6 PrP^−/− samples (panels 7 & 8). Scale bar 20 μm.

Figure 4

PrP-mediated increase in pancreatic β-cell iron downregulates GLUT2: (a) Pancreatic lysates from control and iron overloaded mice were subjected to Western blotting and probed for GLUT2. Samples from control PrP^−/− samples show significant upregulation of GLUT2 relative to Tg40 PrP samples (lanes 3 & 7 vs. 1 & 5; B). Iron overloading downregulates GLUT2 in Tg40 PrP samples (*star), but has minimal effect on similarly treated PrP^−/− samples (arrowhead) (lanes 2 & 6 vs. 4 & 8; B). (b) Densitometric analysis of GLUT2 expression after normalization with β-actin. Values are mean ± SEM of the indicated n. *p < 0.05, **p < 0.01, ***p < 0.001, ns, not significant. The full-length blots are provided in Supplementary File (Raw data). (c) Immunoreaction of pancreatic sections from the above mice mirrors the results in panel a. Immunoreaction for GLUT2 is higher in PrP^−/− relative to Tg40 PrP samples (panels 1 & 3). Iron overloading reduces reactivity for GLUT2 in Tg40 PrP, but not in PrP^−/− samples (panels 1 vs. 2 & 3 vs. 4). Reaction for PrP is limited to pancreatic islets as for GLUT2, and iron overloading upregulates PrP (panels 5 & 6). PrP^−/− samples show no reactivity for PrP as expected (panels 7 & 8). H&E sections show no obvious toxicity due to iron overloading (panels 9–12). Scale bar 20 μm.

Figure 5

PrP-mediated increase in IC iron downregulates glucose transporters in the brain, neuroretina, and the liver: (a) Probing of Western blots of brain lysates from Tg40 PrP and PrP^−/− mice for GLUT3 shows a significant increase in PrP^−/− relative to Tg40 PrP samples (lanes 2 & 4 vs. 1 & 3). (b) Density of protein bands after normalization with β-actin. Values are mean ± SEM of the indicated n. *p < 0.05. (c) Immunoreaction of M17 cells for GLUT3 shows increased reactivity in vector-transfected relative to PrP^C over-expressing cells (panels 1 & 3). Exposure to iron downregulates GLUT3 significantly more in PrP^C relative to vector controls (panels 2 & 4). Scale bar 25 μm. (d) Western blotting of lysates from control and iron-exposed PrP^C and vector expressing cells mirrors the results in panel c. (e) Density of protein bands after normalization with β-actin. Values are mean ± SEM of the indicated n. *p < 0.05, ***p < 0.001, ns, not significant. (f) A similar evaluation of retinal lysates shows upregulation of GLUT1 in PrP^−/− relative to Tg40 PrP samples (lanes 1 & 3). Iron overloading downregulates GLUT1 in Tg40 PrP samples relative to untreated controls, but has minimal effect on similarly treated PrP^−/− samples (lanes 2 & 4). (g) Density of protein bands after normalization with β-actin. Values are mean ± SEM of the indicated n. *p < 0.05, **p < 0.01, ns, not significant. (h) Probing of Western blots of liver lysates for GLUT2 mirrors the results in pancreatic, brain, and neuroretinal lysates (lanes 1–4). (i) Density of protein bands after normalization with β-actin. Values are mean ± SEM of the indicated n. *p < 0.05, **p < 0.01, ns, not significant. The full-length blots are provided in Supplementary File (Raw data).

Figure 6

PrP-mediated increase in β-cell iron decreases insulin: (a) Probing of Western blots of pancreatic lysates from control and iron overloaded Tg40 PrP and PrP^−/− mice with antibodies specific for insulin dimer and pentamer (upper and lower panels) shows decreased expression in Tg40 PrP relative to PrP^−/− samples (lanes 1 & 5 vs. 3 & 7; b). Iron overloading decreases insulin levels in Tg40 PrP (lanes 1vs. 2 & 5 vs. 6; b), but has minimal effect on PrP^−/− samples (lanes 3 vs. 4 & 7 vs. 8; b). (b) Density of protein bands after normalization with β-actin. Values are mean ± SEM of the indicated n. *p < 0.05, ***p < 0.001, ns, not significant. The full-length blots are provided in Supplementary File (Raw data). (c) Immunohistochemistry of pancreas shows relatively higher reactivity for insulin in PrP^−/− relative to Tg40 PrP sections (panels 1 & 3). Iron overloading decreases insulin reactivity in Tg40 PrP but not in PrP^−/− samples (panels 1 vs. 2 & 3 vs. 4). Reaction for glucagon is higher in iron overloaded Tg40 PrP relative to control samples (panels 5 & 6). The difference is minimal in PrP^−/− samples (panels 7 & 8). (d) Immunostaining of 1.1B4 cells for insulin shows decreased reactivity after 2 and 4 hours of exposure to FAC (panels 1–3). (e) Silencing of PrP with siRNA abrogates iron-mediated decrease in insulin reactivity (panels 1–4). Scale bar 20 μm.

Figure 7

PrP-mediated increase in IC iron blunts the systemic response to peripheral challenge with glucose and insulin: (a) GTT in C6 PrP^+/+ and C6 PrP^−/− mice shows an initial spike in blood glucose in both mouse lines after 15 min, followed by a gradual decline to control levels by 180 min. Blood glucose is significantly higher in C6 PrP^+/+ relative to C6 PrP^−/− mice at the 60 min time point. (b) Iron overloading decreases glucose tolerance in C6 PrP^+/+ relative to C6 PrP^−/− mice at all-time points tested. (c) ITT shows a hypoglycemic response in C6 PrP^−/− mice relative to that C6 PrP^+/+ controls. (d) Iron overloading increases insulin resistance in C6 PrP^+/+ mice, but has minimal effect on C6 PrP^−/− mice. The data represent mean ± SEM of values from two different strains of mice. *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 8

Graphical representation of PrP^C-mediated modulation of glucose homeostasis through iron. (1 & 2) PrP^C facilitates cellular uptake of Tf-Fe³⁺ and non-Tf-bound iron (Fe³⁺) by functioning as a ferrireductase partner for DMT-1 and ZIP14. (3) The increase in intracellular (IC) iron down-regulates HIF1α, resulting in the downregulation of GLUT2 in pancreatic β-cells and hepatocytes (black), GLUT 3 in neuronal cells (red), and GLUT1 at the blood-retinal barrier (blue). (4) Reduced uptake of glucose downregulates insulin, resulting in hyperglycemia. (5) Down-regulation or deletion of PrP^C decreases IC iron, resulting in the upregulation of HIF1α and glucose transporters. (6) Increased uptake of glucose stimulates insulin secretion with resultant hypoglycemia.