Noninvasive assessment of pancreatic beta-cell function in vivo with manganese-enhanced magnetic resonance imaging

Abstract

Loss of beta-cell function in type 1 and type 2 diabetes leads to metabolic dysregulation and inability to maintain normoglycemia. Noninvasive imaging of beta-cell function in vivo would therefore provide a valuable diagnostic and research tool for quantifying progression to diabetes and response to therapeutic intervention. Because manganese (Mn(2+)) is a longitudinal relaxation time (T1)-shortening magnetic resonance imaging (MRI) contrast agent that enters cells such as pancreatic beta-cells through voltage-gated calcium channels, we hypothesized that Mn(2+)-enhanced MRI of the pancreas after glucose infusion would allow for noninvasive detection of beta-cell function in vivo. To test this hypothesis, we administered glucose and saline challenges intravenously to normal mice and mice given high or low doses of streptozotocin (STZ) to induce diabetes. Serial inversion recovery MRI was subsequently performed after Mn(2+) injection to probe Mn(2+) accumulation in the pancreas. Time-intensity curves of the pancreas (normalized to the liver) fit to a sigmoid function showed a 51% increase in signal plateau height after glucose stimulation relative to saline (P < 0.01) in normal mice. In diabetic mice given a high dose of STZ, only a 9% increase in plateau signal intensity was observed after glucose challenge (P = not significant); in mice given a low dose of STZ, a 20% increase in plateau signal intensity was seen after glucose challenge (P = 0.02). Consistent with these imaging findings, the pancreatic insulin content of high- and low-dose STZ diabetic mice was reduced about 20-fold and 10-fold, respectively, compared with normal mice. We conclude that Mn(2+)-enhanced MRI demonstrates excellent potential as a means for noninvasively monitoring beta-cell function in vivo and may have the sensitivity to detect progressive decreases in function that occur in the diabetic disease process.

Fig. 1.

Mn²⁺-enhanced magnetic resonance (MR) images of mouse pancreas. A: gradient-echo anatomic reference image. B: precontrast inversion recovery image with the pancreas nulled. C and D: inversion recovery images acquired 5 min (C) and 45 min (D) after injection of MnCl₂. Signal intensity in the pancreas initially increased rapidly and reached plateau ∼15 min after injection of MnCl₂. Arrows indicate location of the pancreas.

Fig. 2.

Normalized time-signal intensity curves (TICs) for the pancreas after saline and glucose infusions in normal mice. Mn²⁺-enhanced MR signal in the pancreas (normalized to liver signal) after saline and glucose infusions is plotted vs. time after MnCl₂ injection. Arrow denotes time of injection for saline/glucose, and MnCl₂ was injected at time = 0 min. Data were fit to a sigmoid function, and 2-way ANOVA was performed on the parameters corresponding to sigmoid function plateau height and slope (see research design and methods). Data points represent means ± SE from 5 mice.

Fig. 3.

Intraperitoneal glucose tolerance tests (IPGTTs) after streptozotocin (STZ) injection. C57BL/6J mice were treated with high- or low-dose STZ and then subjected to IPGTTs, as detailed in research design and methods. A: blood glucose levels in normal and high- and low-dose STZ mice after intraperitoneal glucose injection at time = 0 min. B: results of area under the curve analysis of data in A. C: serum insulin levels at 0, 10, and 30 min after intraperitoneal glucose injection in the IPGTT in A. *Statistically different (P < 0.002) compared with normal animals; ^#statistically different (P < 0.002) compared with high-dose STZ animals. Data represent means ± SE of 4 animals per treatment group. ND, not detectable.

Fig. 4.

Normalized TICs for the pancreas after saline and glucose infusions in STZ-diabetic mice. Mn²⁺-enhanced MR signal in the pancreas (normalized to liver signal) after saline and glucose infusions is plotted vs. time after MnCl₂ injection. Arrows denote time of injection for saline/glucose, and MnCl₂ was injected at time = 0 min. Data were fit to a sigmoid function, and 2-way ANOVA was performed on the parameters corresponding to sigmoid function plateau height and slope (see research design and methods). A: high-dose STZ mice. B: low-dose STZ mice. Data points represent means ± SE of 5 high-dose STZ mice and 4 low-dose STZ mice.

Fig. 5.

Plateau heights of normalized pancreatic TICs. Plateau heights for normalized pancreatic TICs are shown after glucose and saline infusions for untreated normal mice, mice given a high dose of STZ, and mice given a low dose of STZ. Two-way ANOVA revealed significant differences between saline and glucose responses in nondiabetic animals (P < 0.01) and between the nondiabetic glucose response and the glucose response in either diabetic group (P < 0.006). Pairwise Tukey tests showed that the glucose response in low-dose STZ mice was significantly different (P = 0.02) from the saline response in the same mice. No statistical difference was observable when comparing the glucose response between diabetic groups. Values are shown as means ± SE.