The imaging of insulinomas using a radionuclide-labelled molecule of the GLP-1 analogue liraglutide: a new application of liraglutide

Abstract

Objective: This study explores a new, non-invasive imaging method for the specific diagnosis of insulinoma by providing an initial investigation of the use of 125I-labelled molecules of the glucagon-like peptide-1 (GLP-1) analogue liraglutide for in vivo and in vitro small-animal SPECT/CT (single-photon emission computed tomography/computed tomography) imaging of insulinomas.

Methods: Liraglutide was labelled with 125I by the Iodogen method. The labelled 125I-liraglutide compound and insulinoma cells from the INS-1 cell line were then used for in vitro saturation and competitive binding experiments. In addition, in a nude mouse model, the use of 125I-liraglutide for the in vivo small-animal SPECT/CT imaging of insulinomas and the resulting distribution of radioactivity across various organs were examined.

Results: The labelling of liraglutide with 125I was successful, yielding a labelling rate of approximately 95% and a radiochemical purity of greater than 95%. For the binding between 125I-liraglutide and the GLP-1 receptor on the surface of INS-1 cells, the equilibrium dissociation constant (Kd) was 128.8 ± 30.4 nmol/L(N = 3), and the half-inhibition concentration (IC50) was 542.4 ± 187.5 nmol/L(N = 3). Small-animal SPECT/CT imaging with 125I-liraglutide indicated that the tumour imaging was clearest at 90 min after the 125I-liraglutide treatment. An examination of the in vivo distribution of radioactivity revealed that at 90 min after the 125I-liraglutide treatment, the target/non-target (T/NT) ratio for tumour and muscle tissue was 4.83 ± 1.30(N = 3). Our study suggested that 125I-liraglutide was predominantly metabolised and cleared by the liver and kidneys.

Conclusion: The radionuclide 125I-liraglutide can be utilised for the specific imaging of insulinomas, representing a new non-invasive approach for the in vivo diagnosis of insulinomas.

Figure 1. Saturation curve.

The equilibrium dissociation constant K_d for the binding between ¹²⁵I-liraglutide and the GLP-1 receptors of INS-1 cells was 128.8±30.4 nmol/L, and the maximum binding constant B_max was (1.50±0.23)×10⁵. Error bars represent 1 SD of the mean(N = 3).

Figure 2. Competitive binding curve.

IC₅₀, the half-inhibition concentration of liraglutide, was 542.4±187.5 nmol/L. Error bars represent 1 SD of the mean(N = 3).

Figure 3. Small-animal SPECT/CT imaging in the insulinoma model.

Small-animal SPECT/CT imaging after the injection of ¹²⁵I-liraglutide into the tail veins of nude mice indicated that the tumour could be clearly imaged at 30 min after this injection (A), with the clearest images of the tumour obtained at 90 min after the injection (B). The tumour had begun to fade by 180 min after the injection (C) and had significantly faded by 360 min after the injection (D). In addition, the thyroid, liver, heart, lung, kidney, bladder, and other organs were visible to various degrees in the imaging results. The CT coronal tomographic images correspond to the SPECT/CT images, and the arrow indicates the tumour location.

Figure 4. Histograms of the %ID/g for various organs of nude mice at 90 min and 360 min after an injection of ¹²⁵I-liraglutide.

At 90%ID/g ± SD(N = 3).

Figure 5. Imaging of the nude mouse model of insulinoma 90 min after blocking with liraglutide.

Small-animal SPECT/CT imaging indicated that tumour images significantly faded after blocking with liraglutide (B), while A presents an image obtained prior to blocking. The images of the lungs also faded after blocking. The coronal CT images correspond to the SPECT/CT images, and the arrow indicates the tumour location.

Figure 6. A histogram of %ID/g for various organs in a model mouse 90 min after blocking with liraglutide.

The radioactivity of the tumour(p<0.05) was significantly reduced after blocking, and blocking caused the radioactivity to decrease to varying degrees in the lungs(p<0.05), stomach, pancreas(p<0.05), and other organs. Student t test for significance, p<0.05(statiscally significant p values are marked in brackets). Data are expressed as average %ID/g ± SD(N = 3).