Synopsis

Conversion of acetoacetate (AcAc) to β-hydroxybutyrate (β-HB) by the mitochondrial enzyme β-hydroxybutyrate dehydrogenase depends upon NADH availability. Previous studies have shown the potential of β-HB-to-AcAc ratio to reflect the redox state in rat hearts and lymphoma cells. Here, we assessed the value of HP [1,3-¹³C]-AcAc in brain. We demonstrated the potential to probe AcAc to β-HB conversion in normal mice and mice with glioblastoma. Significantly higher levels of [1-¹³C]-β-AcAc and [1-¹³C]-β-HB were observed in tumor-bearing mice compared to control mice. Consistent with lower levels of NADH measured in tumors, the [1-¹³C]-β-HB-to-[1-¹³C]-β-AcAc ratio trended towards a decrease compared to normal brain.

Introduction

Material and Methods

Hyperpolarized [1,3-¹³C]-AcAc [1,3-¹³C]-AcAc was produced from [1,3-¹³C]-ethyl-acetoacetate (Sigma) as described previously^[3] and was resuspended in a 1:3 DMSO:water solution, leading to a final concentration of ~2.5M. Samples for polarization were mixed with OX-63 (15mM) and gadolinium (1mM). After polarization and dissolution, HP [1,3-¹³C]-AcAc was injected into an NMR tube and single-pulse ¹³C spectra acquired using an 11.7T INOVA spectrometer (Agilent Technologies) to evaluate T1 and percent polarization (n=2). Spectra were quantified by peak integration using MestRenova (Mestrelab).

Mitochondria studies Following dissolution, 2mL of HP [1,3-¹³C]-AcAc was injected into an NMR tube containing 1mL of cow heart mitochondria (5.5mg/L, Abcam) and 2mM NADH. Dynamic sets of HP ¹³C spectra (TR=3s/FA=13deg) were acquired on an 11.7T INOVA spectrometer. High-resolution proton-decoupled thermal ¹³C spectra were acquired on an 11.7T Bruker spectrometer equipped with a triple resonance cryoprobe. Spectra were analyzed as above (n=2).

In vivo studies 12 female athymic nu/nu mice were used. Tumors were implanted by intracranial injection of 3x10⁵ cells. 3 mice were implanted with IDH1 wild-type U87 (U87wt) GBM cells and 4 with U87 cells engineered to express mutant IDH1 (U87mut). 4 mice served as tumor-free controls. Studies were performed on a 14.1T vertical MR system (Agilent Technologies) equipped with dual-tune ¹H-¹³C coil. Axial T2-weighted images were used to evaluate tumor size and were acquired using a multi-slice spin-echo sequence (TE=20ms/TR=1200ms/in-plane resolution=0.12x0.12mm²/slice thickness=1.5mm/NA=2). For HP studies, 500μL of [1,3-¹³C]-AcAc was injected in the tail-vein over 12s. 10s after beginning of the injection, dynamic ¹³C MR spectra were acquired from a 10mm slab (TR=4s/FA=20deg/NT=10) followed by a final 90deg acquisition. The dynamic data were summed and analyzed by measuring the [1-¹³C]-β-HB and [1-¹³C]-AcAc signal-to-noise ratios using MestRenova. The 90deg acquisition was analyzed in the same way. At end-point, all mice were euthanized. Tumor and normal brain tissues were clamp frozen to quantify NAD⁺ and NADH by colorimetric assay (BioVision).

Statistical analysis All results are expressed as mean±s.d. Paired and unpaired two-tailed Student’s t-test were used.

Results and Discussion

Characterization Following polarization, resonances of HP [1-¹³C]-AcAc and [3-¹³C]-AcAc were detected at 175.5ppm and 211ppm with an enhancement of 21±3% and 19±2% when compared to thermal spectrum (Fig.2A). Additional resonances detected at 182.1, 180.5 and 161ppm were attributed to the natural abundance of HP [1-¹³C]-acetate, an unknown contaminant and [¹³C]-bicarbonate. Similar to a previous report^[6], the T1 of HP [1-¹³C]-AcAc and [3-¹³C]-AcAc were 48±4s and 35±2s (Fig.2B).

Mitochondria studies Injection of HP [1,3-¹³C]-AcAc to a suspension of mitochondria resulted in a detectable build-up of HP [1-¹³C]-β-HB at 181.1ppm (Fig.3A/3C), confirming our ability to monitor the conversion of [1-¹³C]-AcAc to [1-¹³C]-β-HB. Signal from [3-¹³C]-β-HB was also detected but rapidly disappeared, probably due to its shorter T1. High-resolution thermal ¹³C acquisitions confirmed [1-¹³C]-β-HB and [3-¹³C]-β-HB production (Fig.3B/3D).

GBM studies Quantification of NADH and NAD⁺ showed no significant differences between U87wt and U87mut tumors (Fig.4A). However, NADH was significantly lower in U87wt+mut tumor tissues compared to normal brain from control mice or contra-lateral normal-appearing brain in U87wt+mut tumor-bearing mice (Fig.4B). HP acquisitions were performed on tumor-bearing animals when tumors reached a volume of ~0.22cm³ (Fig.5A). Spectra from dynamic acquisitions or 90deg acquisitions showed no difference in [1-¹³C]-β-HB, [1-¹³C]-AcAc or [1-¹³C]-β-HB-to-[1-¹³C]-AcAc ratio between U87wt and U87mut as expected by their comparable NADH and NAD⁺ levels (Fig.5B/C/D). When comparing control and tumor-bearing mice, HP [1-¹³C]-AcAc and [1-¹³C]-β-HB were significantly lower in controls. Consistent with lower levels of NADH in tumors, [1-¹³C]-β-HB-to-[1-¹³C]-β-AcAc ratio trended towards a decrease compared to normal brain, but the difference was not statistically significant (the ratio was not quantified in the 90deg acquisition because [1-¹³C]-β-HB was below detection, Fig.5C/D).

Conclusion

Acknowledgements

References