NMR visibility of deuterium-labeled liver glycogen in vivo

Abstract

Purpose: Deuterium metabolic imaging (DMI) combined with [6,6'-² H₂ ]-glucose has the potential to detect glycogen synthesis in the liver. However, the similar chemical shifts of [6,6'-² H₂ ]-glucose and [6,6'-² H₂ ]-glycogen in the ² H NMR spectrum make unambiguous detection and separation difficult in vivo, in contrast to comparable approaches using ¹³ C MRS. Here the NMR visibility of ² H-labeled glycogen is investigated to better understand its potential contribution to the observed signal in liver following administration of [6,6'-² H₂ ]-glucose.

Methods: Mice were provided drinking water containing ² H-labeled glucose. High-resolution NMR analyses was performed of isolated liver glycogen in solution, before and after the addition of the glucose-releasing enzyme amyloglucosidase.

Results: ² H-labeled glycogen was barely detectable in solution using ² H NMR because of the very short T₂ (<2 ms) of ² H-labeled glycogen, giving a spectral line width that is more than five times as broad as that of ¹³ C-labeled glycogen (T₂ = ~10 ms).

Conclusion: ² H-labeled glycogen is not detectable with ² H MRS(I) under in vivo conditions, leaving ¹³ C MRS as the preferred technique for in vivo detection of glycogen.

Keywords: DMI; deuterium; glycogen.

Figure 1 -

DMI of rat liver during infusion of [6,6’-²H₂]-glucose. (A, C) DMI-based maps of the glucose signal region (3.2 – 4.2 ppm) overlaid on an axial MRI through the liver (outlined by the red line) during (A) intravenous (IV) or (B) intraperitoneal (IP) infusion of [6,6’-²H₂]-glucose. (B, D) Global, non-localized ²H MR spectra acquired before and during the acquisition of the DMI data shown in (A, C). Time stamps (min) indicate timing of DMI or ²H MRS acquisition after start of infusion and represent the midpoint of data acquisition.

Figure 2 -

NMR visibility of mouse liver glycogen. (A, B) ¹H, (C, D) ¹³C and (E, F) ²H NMR of rat liver glycogen (A, C, E) before and (B, D, F) after the administration of glucosidase. ¹H and ¹³C studies were performed in standard NMR buffer (95/5% H₂O/D₂O), whereas ²H studies were performed in ²H-depeleted water. (A, C) With ¹H and ¹³C NMR the glycogen resonances are readily detectable as relatively narrow lines (~10 Hz for ¹H, ~35 Hz for ¹³C), whereby (B, D) glucosidase-catalyzed conversion to glucose confirmed the assignment and presence of glycogen. (E) With ²H NMR the glycogen signal is difficult to recognize directly. A spectral fit of the experimental data (bottom two traces) reveals a broad line (width = ~190 Hz) for glycogen at circa 3.8 ppm with no contribution from deuterated glucose. (F) Glucosidase-catalyzed conversion to [6,6’-²H₂]-glucose confirmed the presence of [6,6’-²H₂]-glycogen. The signals indicated with asterisks (*) are assigned to the non-enriched (natural abundance) positions (3.2 – 5.2 ppm) in glucose and to natural-abundance triglycerides (0.8 – 1.5 ppm).

Figure 3 -

Theoretical T₁ and T₂ relaxation curves for homonuclear dipolar (¹H), heteronuclear dipolar (¹³C-¹H) and homonuclear quadrupolar (²H) relaxation. Experimentally measured values for the H1 or C1 position of glycogen are shown by diamonds and values for the ¹H6, ¹³C6 or ²H6 position by circles. Open symbols represent theoretically predicted values for ²H-glycogen. Solid and dashed curves indicate T₁ and T₂ relaxation, with green, blue and red colors representing ¹H, ¹³C and ²H relaxation, respectively. The experimental T₁ and T₂ relaxation times, rotation correlation times and scaling factors are summarized in Table 1. The solid and dashed curves are calculated with scaling factors that represent the best match to the experimental values for the ¹H6, ¹³C6 and ²H6 positions of glycogen.