Kinetic Analysis of Hepatic Metabolism Using Hyperpolarized Dihydroxyacetone

Abstract

Hyperpolarized carbon-13 magnetic resonance (HP-MR) is a new metabolic imaging method the does not use ionizing radiation. Due to the inherent chemical specificity of MR, not only tracer uptake but also downstream metabolism of the agent is detected in a straightforward manner. HP [2-13 C] dihydroxyacetone (DHA) is a promising new agent that directly interrogates hepatic glucose metabolism. DHA has three metabolic fates in the liver: glucose production, glycerol production and potential inclusion into triglycerides, and oxidation in the tricarboxylic acid cycle. Each pathway is regulated by flux through multiple enzymes. Using Duhamel's formula, the kinetics of DHA metabolism is modeled, resulting in estimates of specific reaction rate constants. The multiple enzymatic steps that control DHA metabolism make more simplified methods for extracting kinetic data less than satisfactory. The described modeling paradigm effectively identifies changes in metabolism between gluconeogenic and glycogenolytic models of hepatic function.

Probing hepatic metabolism of [2-13C]dihydroxyacetone in vivo with 1H-decoupled hyperpolarized 13C-MR

“…The metabolite-to-substrate ratios measured in vivo at 20 s after injection of hyperpolarized [2-13 C]DHAc were compared with those estimated at the same time point from the ex vivo results obtained in perfused livers (Fig. 4) [23]. The estimated ratios are relatively insensitive to the shape of the input function, except for lactate-C2 (see Fig.…”

“…The mean ratio is shown in grey with the standard deviation indicated by an error bar. The diagonal striped bars indicate the calculated ratios at 20 s mark after the start of DHAc infusion based on the analysis described by Kirpich et al [23] and G3P dehydrogenase (which is dependent on NADH and therefore redox status). It is likely that the flux from DHAc to Ga3P is lower in the in vivo experiment due to differing experimental conditions.…”

“…Estimated metabolite-to-substrate ratios were calculated at t = 20 s based on the ex vivo results and model of Kirpich et al [23]. The rate constants and relaxation time constants included in the model were not adjusted for field strength or in vivo conditions.…”

“…Briefly, a gamma function (Eq. 11 in reference [23]) was used, with α = 1 and β = 1, and the time courses of the 14 metabolites included in the model were simulated with an ordinary differential equation solver at t = 20 s. This was then repeated with α = 2/β = 2 and α = 3/β = 3 to generate the data in Fig. S1, showing that for many metabolite ratios the input function does not perturb the result greatly.…”

“…S1, showing that for many metabolite ratios the input function does not perturb the result greatly. The simulation is again repeated with α = 1/β = 1 and 11 values of k drawn randomly from a normal distribution with the mean and standard deviation values from Kirpich et al [23] This is repeated 1000 times with any negative values of rate constants discarded and replaced. The mean (bar value) and standard deviation (error bar) from these 1000 simulations are plotted as the modelled data in Fig.…”