Natural D-glucose as a biodegradable MRI contrast agent for detecting cancer

Abstract

Purpose: Modern imaging technologies such as CT, PET, SPECT, and MRI employ contrast agents to visualize the tumor microenvironment, providing information on malignancy and response to treatment. Currently, all clinical imaging agents require chemical labeling, i.e. with iodine (CT), radioisotopes (PET/SPECT), or paramagnetic metals (MRI). The goal was to explore the possibility of using simple D-glucose as an infusable biodegradable MRI agent for cancer detection.

Methods: D-glucose signals were detected using chemical exchange saturation transfer (glucoCEST) MRI of its hydroxyl groups. Feasibility was established in phantoms as well as in vivo using two human breast cancer cell lines, MDA-MB-231 and MCF-7, implanted orthotopically in nude mice. PET and contrast-enhanced MRI were also acquired.

Results: Both tumor types exhibited significant glucoCEST signal enhancement during systemic sugar infusion (mild hyperglycemia), allowing their noninvasive visualization. GlucoCEST showed differences between types, while PET and CE-MRI did not. Data are discussed in terms of signal contributions from the increased vascular volume in tumors and especially from the acidic extracellular extravascular space (EES), where glucoCEST signal is expected to be enhanced due to a slow down of hydroxyl proton exchange.

Conclusions: This observation opens up the possibility for using simple non-toxic sugars as contrast agents for cancer detection with MRI by employing hydroxyl protons as a natural label.

Copyright © 2012 Wiley Periodicals, Inc.

Figures

Figure 1. Glucose infusion protocol with blood concentrations measured (a) and MRI protocol (b)

The rate of glucose infusion (black curve) resulted in blood glucose concentrations as shown by the red curve. The blood glucose concentration was measured for two mice (average ± SD). CE: contrast enhanced.

Figure 2. Concentration, pH and radiofrequency (RF) saturation field (B1) dependence of glucoCEST contrast in D-glucose solutions in PBS at 37°C at 11.7T

(a) Z-spectra (top) and MTRasym (bottom) as a function of concentration at pH 7.3 for B1=1.6 µT; (b) Z-spectra (top) and MTRasym (bottom) as a function of pH for 20 mM D-glucose at B1 = 1.6 µT; a slow down of exchange at low pH increases the contrast of several OH protons, especially peaks at 1.2, 2.2 and 2.8 ppm. (c) MTRasym spectra as a function of B1 for 20 mM D-glucose at pH 6.2. (d) glucoCEST contrast (averaged MTRasym over 0.8–2.2 ppm) as a function of concentration at different B1.

Figure 3. GlucoCEST results for a mouse inoculated with MDA-MB-231 and MCF-7 breast tumors (a–i)

(a) Anatomical image for a mouse inoculated with breast tumors. (b,c) GlucoCEST images (average MTRasym % over spectral points between 0.8–2.2 ppm) before infusion (b) and during infusion (c). (d,e) Pre-infusion and Infusion MTRasym spectra from ROIs including the whole tumor for MDA-MB-231 (d) and MCF-7 (e). (f) MTRasym values pre-infusion and during infusion (n = 5; * p < 0.05; *** p < 0.005). (g) GlucoCEST difference map: ΔMTRasym = MTRasym(Infusion) – MTRasym(Pre-infusion). The intensity of the internal body was thresholded out because it contains moving areas (lungs and heart) that have large magnetic susceptibility differences with surrounding tissues, which complicates difference imaging. The ΔMTRasym profiles (h) of MDA-MB-231 and MCF-7 show a profile resembling that of D-glucose in phantoms in Fig. 2. (i) Average ΔMTRasym difference spectrum for 5 mice.

Figure 4. Non-edited 18FDG-PET/CT, glucoCEST and contrast-enhanced (CE) difference images in vivo

(a) 18FDG-PET/CT coronal view obtained one hour after intravenous (i.v.) injection of 18FDG showing accumulation in both tumors. (b) GlucoCEST ΔMTRasym map (Infusion – Pre-infusion). (c) T1-weighted difference image (Injection – pre-injection) showing Gd-enhanced regions, mainly in the edges of tumors. (d–f) Comparison of signal intensities (n = 5) for the three modalities using ROIs comprising the two tumors. Even though some trends appear visible for PET and contrast-enhanced MRI, significant differences (p<0.05; paired student’s t-test) between the tumors could be detected only in glucoCEST. (g) Bar graph showing average glucoCEST contrast for MDA-MB-231 and MCF-7 tumors.

Figure 5. GlucoCEST and CE-MRI data in tumor regions enhanced in MRI perfusion imaging

(a) ΔMTRasym spectra for MDA-MB-231 and MCF-7. Note the distinctive glucose shape for the difference profiles. (b) Quantitative glucoCEST shows a significant difference (N = 5; p<0.05; paired student’s t-test) between tumor types (c) Quantitative CE-MRI shows no significant difference (N = 5; p>0.05; paired student’s t-test) between tumor types.

Figure 6. Overview of rate constants and contrast contributions (darker color = higher contrast; white is negligible contrast) for glucoCEST, 18FDG-PET and contrast enhanced MRI and CT in tumors

For glucoCEST, the glucose concentrations in vascular space and EES are comparable, but due to lower pH the EES has higher signal contribution. Intracellular signal is very small to negligible due to rapid glycolysis. In PET, the signal is predominantly due to trapped intracellular phosphorylated FDG. For contrast enhanced MR and CT, the agents occupy only plasma in blood and while they enter the interstitium, the EES concentration is generally lower than in plasma due to limited Ktrans. Glucose, on the other hand moves freely into the interstitium and the erythrocytes. v = vascular (plasma + erythrocytes), p = plasma.