In vivo pH imaging can be accomplished by measuring the ratio of bicarbonate and CO2 present in various organs following the injection of hyperpolarized bicarbonate. A major impediment in the development of this imaging modality is the poor polarization of bicarbonate by DNP. One way to get around this is to polarize pyruvate by DNP and use it as a precursor to produce hyperpolarized bicarbonate via decarboxylation by hydrogen peroxide.¹ The uncatalyzed decarboxylation reaction at neutral pH proceeds far too slow relative to the spin-lattice relaxation time of pyruvate to be of much use. We show that the reaction can be sped up by a factor of ~10 by maintaining the pH above 10.3 throughout the duration of the reaction. This allows for maximum production of polarized bicarbonate from pyruvate while minimizing the polarization lost during the conversion. Because the decarboxylation of pyruvate is highly exothermic, kinetic rate constants can be obtained calorimetrically and confirmed using HP 13C NMR.

1mL of 160mM of pyruvate was reacted in a glass test tube with 1mL of 160mM hydrogen peroxide; the pH was varied by adding sodium carbonate to the pyruvate solution before the reaction. The pH of the solution before and after each reaction was measured using an electronic pH-meter (Mettler Toledo). The temperature data was obtained using a resistive thermal device (RTD) connected to a microcontroller. The temperature data was fit to the following system of differential equations to obtain the kinetic rate constants from the temperature vs. time curves. In the first equation, k is the second order reaction rate constant.

$$ \frac{\text{d}[Pyr]}{\text{d}t} = -k[Pyr][H_{2}O_{2}] $$

$$ \frac{\text{d}T}{\text{d}t} = c_{1}\frac{\text{d}[Pyr]}{\text{d}t}-c_{2}(T-T_{0}) $$

An example of the model fitting to the temperature vs time data of a decarboxylation reaction is shown in figure 1. Figure 2 shows several temperature vs time plots at pHs ranging from 6.89 to 10.11; as the steepness of the rising part of the curve decreases, the reaction rate constant decreases accordingly. To confirm that the reactions were proceeding faster at elevated pH, as is seen in figure 3, a series of spectra were obtained from hyperpolarized phantoms at varying pH. For each trial, 57.4mg of [1-13C]pyruvate (Cambridge Isotopes Inc.) was polarized in a Hypersense DNP system and diluted to 160mM using a neutralizing dissolution medium containing 1mM of phenol red to monitor the pH of the solution. 2.5mL of the hyperpolarized pyruvate solution was quickly mixed with 2.5mL of 160mM hydrogen peroxide in a 15mL falcon tube. The pH of each reaction was varied by adding between 0 and 0.40mL of 3.6M NaOH to the solution, depending on the target pH. The tube was inserted into a 9.4T vertical Bruker magnet and imaged. The spectra were acquired every second with 20 kHz bandwidth and α=15°. The spectra obtained were analyzed with a custom routine in MATLAB 2014b (MathWorks Inc.).

Figure 5 shows the k values obtained from one of the calorimetric trials, these trials consistently show that as the pH of the solution increases, the reaction rate goes up. The most dramatic increase in reaction rate occurs around pH 10.3, the pKa for the bicarbonate <-> carbonate reaction. From this it was inferred that the reaction should proceed most rapidly when the pH is kept well above 10. This is confirmed from the % completion data shown in figure 4, as the pH increases, the amount of pyruvate that has been converted to bicarbonate increases accordingly. It is important to note that the pH of the solution does not remain constant over the course of the reaction. For every CO2 produced that converts to bicarbonate, one proton is released which lowers the pH, and that conversion itself is also pH dependent. Figure 5 shows the hyperpolarized spectra obtained for three pH values. It is notable that there is such a significant difference in the % conversion for spectra a and b despite similar initial values, this signifies that spending a little bit more time above pH ~10.3 results in significantly more decarboxylation.We have demonstrated with two separate methods that the rate of decarboxylation of pyruvate by hydrogen peroxide is heavily pH dependent, and by maintaining the pH above 10.3, full conversion can be achieved within ~20s. By altering the pH, any ratio of hyperpolarized pyruvate to bicarbonate can be obtained. This is of special significance to pH mapping research using hyperpolarized 13C-bicarbonate, which requires the rapid production of hyperpolarized bicarbonate from a more easily polarizable precursor, namely pyruvate.