Target Audience

Results will be of interest to researchers working on technical aspects of magnetic resonance spectroscopy complexed with hyperpolarization techniques. It has special application to work at very high excitation bandwidths.

Introduction

Hyperpolarization offers tremendous signal enhancements and enable measurement of ¹³C spectra with temporal resolution in cell suspensions,¹ perfused organs,^2–4 and in humans.^5,6 In experiments involving hyperpolarization, small flip angles are routinely used in order to preserve the spin polarization across multiple scans and to avoid receiver overflow. However, this results in lower sensitivity to metabolites generated by the system under study since they are a small fraction of the injected substrate. In this work, we demonstrate binomial excitation schemes to detect HP dihydroxyacetone (DHA) metabolism in perfused mouse livers.

Methods

Fasted C57/BL6 mice were used as models for all experiments. Animals were handled in compliance with the University of Florida IACUC regulations. Livers were perfused through the portal vein with Krebs-Henseleit buffer containing 1.5%(w/v) bovine serum albumin, mixed fatty acids, lactate and pyruvate (lac:pyr ratio was 10:1). Perfusate was constantly oxygenated using 95%/5% O₂/CO₂ mixed gas. After 30 min of perfusion, 8 mM or 16 mM HP [2-¹³C] dihydroxyacetone was injected directly into the liver and ¹³C NMR spectra were acquired in real time. All experiments were carried out in a NMR spectrometer (14.1 T magnet) equipped with an 18 mm broadband probe (Doty Scientific, SC). ¹³C spectra were collected either following a 30° pulse (16 mM HP DHA) or a binomial excitation scheme (45° {x, -x} pulses; 8mM HP DHA). For the displayed spectra, ¹H decoupling (WALTZ64) was turned on for the square pulse excitation, but turned off for the binomial. The DHA resonance was set directly on resonance for the excitation. The Bloch simulator was written in MatLab (Natwick, MA).

Results and Discussion

A Bloch simulation of the effect of a 45° {x, -x} binomial pulse at a B₁ field strength of 4.16 kHz produced a broad excitation profile with no excitation of the parent DHA resonance (Figure 1). A maximal consumption of the Z-magnetization of the peaks in the region of the DHA hydrate (-96 ppm) is 30% per excitation. The excitation profile is broad, and covers the bandwidth from PEP (~151 ppm) to lactate (~68 ppm) with nearly equivalent flip angles. Figure 2 shows the result from using binomial excitation to observe DHA metabolism. The DHA resonance is minimally excited in the experiment (inset). Note the DHA hydrate signal at ~96 ppm is essentially equivalent for the two excitation modes. Since DHA signal does not overflow the receiver, it was possible to use a higher receiver gain with the binomial. It is notable that despite the predicted frequency dependent phase differences (Figure 1), the spectrum is easily phased. As a direct consequence, it was possible to record ¹³C spectra with excellent S/N of various metabolites generated by the liver. While the higher gain used for the binomial excitation produces a larger noise baseline, the signal between the two excitation modes is dramatically increased by the binomial. With the binomial, the peak amplitudes of the downstream metabolites of DHA are approximately 7 times greater, even without the ¹H decoupling that simplifies the spectra acquired without the binomial pulse. Preservation of the HP magnetization of the precursor substrate is essential for producing high signal intensities, a fact generally acknowledged by researchers in HP methods. Excitation schemes based on shaped pulses are robust, but can be difficult to implement when power handling is a concern. The binomial pulse is robust and easily implemented. The only precondition is an exact knowledge of the resonance frequency prior to the experiment. Binomial sequences with more than 2 pulses back to back are less suitable, as they tend to produce more modulation of the excitation profile as a function of frequency.

Conclusion

We demonstrate the application of a 45° {x, -x} binomial pulse for the detection of HP [2-¹³C]dihydroxyacetone metabolism in a perfused liver. Even operating at 14.1 T (150 MHz ¹³C frequency) the binomial pulse produces a broad excitation profile suitable for exciting the large bandwidth of frequencies that encompass the metabolites of DHA. The method is extremely robust, and small changes in B₁ produce minimal changes in bandwidth and flip angle. Due to the shortness of the pulses used here, the method is not suitable for integration with a gradient pulse to produce a slice selective excitation. The results indicate that binomial excitation is suitable for significantly enhancing the resonance amplitudes for metabolites downstream of the precursor substrate.

Acknowledgements

The authors acknowledge funding from NSF DMR 1157490 and NIH 8P41-EB015908 and P41122698, , U24DK097209, and R01s DK105346, HD087306, DK112865.