Introduction

Glutamine is highly consumed by fast- growing cells, and its main cellular role is providing carbon and nitrogen through various metabolic pathways. Glutaminase utilizes glutamine as a substrate and produces glutamate and ammonia. Quantification of glutamine metabolism appears to be crucial since it is involved in multiple diseases including cancer and diabetes.¹ Due to the global relevance of glutaminase in a variety of diseases, we seek to develop an in vivo metabolic probe. In this study, a xenograft murine model of renal cell carcinoma as a glutamine-avid tumor has been used. The goal is to develop a probe for real-time flux in glutamine metabolism.
Hyperpolarized magnetic resonance provides high transient spin-lattice signal. Fast acquisition sequences in magnetic resonance enable us to obtain dynamics fast biochemical processes. Prior studies using commercial [5-¹³C]-L-glutamine confirmed that detecting the carbon-5 resonance of [5-¹³C]-L-glutamate with ¹³C nuclear magnetic resonance (NMR) could be used to monitor dynamics of the reaction in rat model(Fig.1). However, directly-bonded quadrupolar ¹⁴N-nuclei increase signal decay outside the polarizer due to quadrupolar effect at low magnetic field.2 We also learned that T1 can be increased with deuteration of adjacent carbons.³ Here, we provide a modular synthetic strategy for a series of ¹³C-, ²H(D)-, ¹⁵N-isotope-labeled glutamine in high yield. Enriching with ¹⁵N decreases low field quadrupolar relaxation. The presence of deuterium next to side chain carbonyl increases T1 by eliminating dipole-dipole relaxation. In addition, we discovered the presence of ¹⁵N increases T₂.⁴ Following up our on previous studies, we used D₂O phosphate buffer for dissolution resulting in the exchange of protons with deuterium, which increased T₁ and T₂ relaxations. We have characterized the physical chemical properties of these compounds and have showed triple labeling [5-¹³C,4-²H₂,5-¹⁵N­]-L-Glutamine and dissolving in D₂O improved SNR and enabled detection of in vivo glutaminase activity and performed Magnetic resonance spectroscopy with two-dimensional resolution.⁵

Methods

Four labeling patterns of glutamine were synthesized using a four-step synthesis (Fig.2B) adapted from previous work on arginine.4 For hyperpolarized experiments, glutamine was polarized in a SPINLab (GE Healthcare) for >2h (5.0T, 0.8K, 139.960GHz). Inversion-recovery and Carr-Purcell-Meiboom-Gill acquisitions for T₁ and T₂ measurements of the carbon-5 resonance, centered at 178ppm, were performed on a Bruker 14.1T NMR. UOK262 cells (gift from Dr. W.M. Linehan at NIH) were cultured in high-glucose DMEM. 5 million cells were xenografted in female athymic nude mouse and imaged when the tumors were approximately 200-300mm³. Hyperpolarized T1 were measured at low field on a 1T 13C-NMR. For in vivo ¹³C spectroscopy, xenografted mouse was infused with 30mM [5-¹³C,4-²H₂,5-¹⁵N­]-L-Glutamine via tail vein catheter and imaged on a Bruker 3T MRI following an IACUC-approved protocol. A dual-tuned ¹H/¹³C coil was used to acquire anatomic reference images and spectroscopic data. T₂-weighted fast-spin-echo (FSE) images were acquired for anatomic localization, covering the whole tumor (40x40x24-mm FOV, 128x128x12 matrix, TE/TR=69/1800ms). A chemical-shift imaging sequence (2DCSI) was used to acquire HP spectra (40x40x10-mm FOV, 12x12x1 matrix, 5kHz sweep width, 0.1ms dwell time, 320 spectral points, TE/TR 5/70ms, 20° excitation), 25s post-injection derived from the previous dynamic studies.

Results

At 14.1T, the T­₁ of the glutamine carbon-5 resonance increased with ²H-enrichment: [5-¹³C]-L-Glutamine T­1=10.8±0.55s and [5-¹³C,4-²H₂]-L-Glutamine T­₁=12.74±0.69s. ¹⁵N-enrichment provided minimal increase: [5-¹³C,5-¹⁵N­]-L-Glutamine T­1=10.9±0.5s. Triple-labeling resulted in the largest increase: [5-¹³C,4-²H₂,5-¹⁵N­]-L-Glutamine T₁=12.74±0.69s) (Fig.2B). Dissolution in D₂O increased T_­1 in all species, e.g., [5-¹³C,4-²H₂,5-¹⁵N­]-L-Glutamine T­₁=15.58±0.7s. Hyperpolarization decay at 1T showed similar trends: [5-¹³C,4-2H2,5-15N­]-L-Glutamine relaxation in phosphate buffer in D₂O decreased ~3.3-fold resulting in a T₁=72±5s. (Fig.3C) Interestingly, by using slightly acidic phosphate buffer (pH= 3.1), accounting for the NaOH in the prep formulation, no side production of pyroglutamate or glutamate were detected in the final dissolution as highlighted in the hyperpolarized ¹³C dynamic decay (Fig.3D). At 14.1T, the T­₂ of the carbon-5 resonance increased with ¹⁵N-enrichment: [5-¹³C]-L-Glutamine T₂=0.474±0.029s and [5-¹³C, 5-¹⁵N­]-L-Glutamine T₂=4.82±0.23s. ²H-enrichment provided a minimal increase: [5-¹³C,4-²H₂]-L-Glutamine T₂=0.521±0.048s. Triple labeling provided the largest increase: [5-¹³C,4-²H₂,5-¹⁵N­]-L-Glutamine T­2=7.28±0.48s. Dissolution in D₂O increased the T­₂ in all species, e.g., [5-¹³C,4-²H₂,5-¹⁵N­]-L-Glutamine T­₂=12.23±0.36s, ~26-fold enhancement (Fig.3B). Axial 2D CSI ¹³C HP MRS in a mouse injected with hyperpolarized [5-¹³C,4-²H₂,5-¹⁵N­]-L-Glutamine revealed in vivo conversion to hyperpolarized glutamate (Fig.4D,E). The ratio of conversion in tumor voxels (0.324) was higher than in kidney voxels (0.255) (Fig 4C,F).

Discussion

²H-enricment has a substantial impact on extending glutamine carbon-5 resonance T₁ at high or low field. ¹⁵N-enrichment has an insignificant effect on thermal T₁, but reduces quadrupolar relaxation and alleviates rapid depolarization following dissolution when the sample is removed from a strong magnetic field. ¹⁵N-enrichment increased T₂ at all field strengths, thereby improving imaging quality. Conversion of multi-labeled hyperpolarized glutamine can be used for glutaminase activity in vivo. ¹³C spectroscopic data in a xenograft model with hyperpolarized glutamine demonstrated detection of higher metabolic flux in the tumor as compared to neighboring normal kidney tissue, suggesting delivery and metabolic conversion occur rapidly and further hyperpolarized studies should be pursued.

Conclusions

The quantification of glutamine flux through conversion of hyperpolarized [5-¹³C,4-²H₂,5-¹⁵N­]-L-Glutamine to [5-¹³C,4-²H₂]-L-glutamate was detected in vivo in a xenograft model, and it is a promising first step towards using this probe to study glutaminolysis in more disease models. Moving forward, we aim to further optimize our acquisition scheme for higher resolution spectroscopy and rapid imaging

Acknowledgements

This work was supported by NIH/NCI Cancer Center Support Grant P30 CA008748, NIH S10 OD016422 as well as Tow Foundation, Thompson Family Foundation.