Dissolution dynamic nuclear polarization (dDNP) is a powerful technique that enhances the magnetic resonance signal of nuclear spins by several orders of magnitude. DNP relies on the principle of cross-relaxation by electron spins driven out of equilibrium to enhance nuclear polarization. When coupled to magnetic resonance spectroscopy/imaging (MRS/I), infusion of 13C-labeled tracers hyperpolarized via DNP in vivo provides real-time estimates of metabolic fluxes, physiological parameters and/or tissue perfusion.
The present work focuses on fundamental aspects of DNP, in particular how sample composition affects the DNP of low-gamma nuclei, as well as on in vivo dDNP applications targeting renal metabolism and physiology.
The DNP properties of the SA-BDPA radical were studied in a sample of 13C-urea. SA-BDPA, a water-soluble derivative of BDPA, is characterized by a small g-anisotropy and unresolved hyperfine coupling to protons. As a result, at 6.7 T, 1.1 K its ESR linewidth is much smaller than the 13C Larmor frequency, which enabled the observation for the first time of 13C DNP via the solid effect and pure thermal mixing, the latter defined as the process where the electron non-Zeeman reservoir alone provides the energy required for triple-spin flips.
The effect of solvent deuteration on 6Li DNP and radical ESR properties was studied in 6LiCl water:glycerol solutions doped with either the nitroxide radical TEMPOL or the trityl radical OX063. Unexpectedly, the TEMPOL-doped samples polarized better than the trityl-doped ones. Across all samples, the relationship between the degree of solvent deuteration with the buildup time constant and polarization level was notably different from what has been reported for 13C. This behavior is indicative of DNP via a combination of the cross effect and thermal mixing mechanisms.
The uptake and metabolism of hyperpolarized L-[1-13C]alaninamide was investigated in the rat kidney in vivo. This probe of aminopeptidase N, which can play a role in tumor growth, was previously studied in vitro. Our study showed that alanine production from alaninamide also occurs in vivo, however, with spectral overlap of substrate and product. Alaninamide, having a pKa of 7.9, proved to be sensitive to local pH. Three spectral peaks, corresponding to at least three environments with different pH values, could be observed in the kidney. The two peaks at higher pH were assigned to the blood extra- and (partially) intracellular compartments, while the third one was mainly in the inner part of the kidney. Finally, alaninamide was shown to also be sensitive to dissolved CO2, with the rapid formation of a carbamate adduct following infusion.
The renal metabolism of D-[1-13C]alanine by D-amino acid oxidase (DAO) was also studied. Conversion of hyperpolarized D-alanine to pyruvate and further metabolism to lactate and bicarbonate was observed in the kidney only when DAO was not inhibited. DAO activity could also be detected in blood, where leukocytes express the enzyme, but not in the brain and liver, in line with their lower DAO activity.
Overall, this thesis provides additional insight into how the experimental conditions can favor a particular DNP mechanism over another. It also significantly expands the scope of in vivo dDNP applications, showing that additional enzyme-catalyzed processes can be detected, along with the potential of amino-acid based hyperpolarized 13C sensors for physiological studies.
EPFL2021
