Purpose

In glioblastoma (GBM), the most common and deadly primary brain cancer, tumor invasion into adjacent brain tissue can make it difficult to distinguish between normal and cancerous tissue. However, 90% of GBMs exhibit an excessive buildup of the fluorescent molecule protoporphyrin IX, the precursor to heme. High levels of protoporphyrin IX can be induced in GBM by administering 5-aminolevulinic acid (5-ALA), the metabolic precursor in heme biosynthesis, allowing the cancer to be targeted by photodynamic therapy and fluorescent imaging. We have begun exploring whether the same abnormal heme synthesis pathway can be used for targeted hyperpolarized metabolic imaging of GBM using ¹³C-labeled 5-ALA. In this initial work, we demonstrate that [4-¹³C]5-ALA can be polarized to moderate levels using dynamic nuclear polarization and that its relaxation time is amenable to hyperpolarized imaging.

Methods

A. HyperSense DNP: An 88-µL sample was prepared by dissolving 50 mg of [4-¹³C]5-ALA hydrochloride (HCl) (gift from Cambridge Isotopes) in a mixture of sodium hydroxide (NaOH), H₂O, DMSO, and OX063Me trityl radical (AH111501 Na salt) with final concentrations of 5-ALA and radical at 3.4 M and 15 mM, respectively. Using a HyperSense polarizer (Oxford Instruments, Abingdon, UK), we measured the polarization buildup with different amounts of Gd (Dotarem). B. SpinLab DNP: [4-¹³C]5-ALA HCl was synthesized via the scheme in Figure 1 and used to prepare an 80-µL sample, as described in (A), and doped with 10% v/v of a 1:50 dilution of Dotarem. It was then polarized for 28 h in a GE SpinLab polarizer (5.0 T, 0.9 K) and dissolved in 4 mL of water containing 0.1 g/L EDTA to a final 5-ALA concentration of 70 mM. The absolute level of liquid state polarization, i.e., polarization after dissolution, was determined by measuring the signal decay of the hyperpolarized sample using a pulse-and-acquire sequence (5° flip angle, 3-s repetition time (TR)) in a GE clinical 3T MR scanner. The signal was extrapolated to the time of dissolution and compared to the signal intensity of the sample measured at thermal equilibrium after adding 30 µL of Dotarem to shorten the T₁ (90° flip angle, TR = 9 s, 200 averages).

Results

The buildup time constant at 3.35 T and 1.4 K was estimated to be ~4 h, with the best polarization achieved by doping with 10% v/v of a 1:50 Dotarem dilution (Fig. 2a). Extrapolation and comparison with pyruvate polarization curves gave an estimated maximal polarization of 8-10%. The polarization buildup time constant for the same sample at 5.0 T and 0.9 K was estimated to be 6-8 h. The T₁ of the [4-¹³C] resonance at 205.9 ppm was 20 s (Fig. 2b) and the polarization was 13% at the time of dissolution. The resonance of natural abundance [1-¹³C]5-ALA was also detected, and its T₁ was estimated to be about 24 s.

Discussion

Our initial results are very promising in light of the ability of Rodrigues, et al.¹ to follow the ¹³C label in [U-²H, U-¹³C]glucose (in vivo T₁ of 9 s, ~15% polarization) through the glycolytic pathway to lactate in an EL4 tumor–bearing mouse in vivo. We chose to focus on [4-¹³C]5-ALA, because although [1-¹³C]5-ALA has a longer relaxation time, the chemical shift of its downstream metabolites are within a few ppm, making them difficult to resolve in vivo. Moreover, 75% of [1-¹³C]5-ALA is metabolized into CO₂, whereas 100% of [4-¹³C]5-ALA is incorporated into protoporphyrin IX (Fig. 3).

Conclusion

5-Aminolevulinic acid shows promise as a substrate for hyperpolarized metabolic imaging. Systematic optimization of DNP conditions will likely increase the polarization above the 13% achieved in this preliminary trial. Rates of cellular uptake and metabolism have yet to be determined.

Acknowledgements

This work was supported by NIH grant R21 CA202694. We also thank Dr. Murali Cherukuri from the National Cancer Institute for use of the HyperSense polarizer.

References

1. Rodrigues TB, Serrao EM, Kennedy BWC, et al. Magnetic resonance imaging of tumor glycolysis using hyperpolarized ¹³C-labeled glucose. Nature medicine, 2014;20:93–97.