Introduction

Gliomas constitute the majority of malignant primary brain tumors and are a significant clinical problem with long-lasting physical and cognitive effects. Recent advances in genomics have uncovered a variety of clinically relevant mutations in gliomas¹. There is an urgent need to identify methods of non-invasively imaging these oncogenic events, thereby integrating this genomic information into patient management.

A critical oncogenic event in cancer, including in low-grade oligodendrogliomas, is the acquisition of mutations in the promoter of telomerase reverse transcriptase (TERT), leading to reactivation of TERT expression². TERT is responsible for the synthesis of telomeres, which are chromosomal structures that are essential for cell proliferation³. TERT is, therefore, a hallmark of tumor proliferation³. Since TERT is silenced in normal cells and specifically reactivated in tumor cells, TERT is also a therapeutic target^{2, 4}. Imaging TERT can differentiate “true” tumor burden from pseudo-progression or pseudoresponse and allow assessment of response to emerging TERT inhibitors. Therefore,the goal of this study was to identify hyperpolarized ¹³C-magnetic resonance spectroscopy (MRS) agents that can detect TERT expression in low-grade oligodendrogliomas.

Methods

Cell culture: We examined immortalized normal human astrocytes without (NHA_control) and with TERT (NHA_tert) expression. TERT expression in these models has been confirmed⁵.

¹H-MRS: Metabolites were extracted by methanol-chloroform extraction and ¹H-MRS spectra acquired on an 11.7T Bruker spectrometer (90°flip angle, 3s relaxation delay, 256 scans)⁶.

Thermally-polarized ¹³C-MRS: Cells were incubated with medium containing 5mM [2-¹³C]-glucose, metabolites extracted and proton-decoupled ¹³C-MRS spectra acquired on an 11.7T Bruker spectrometer (30°flip angle, 3s relaxation delay, 2048 acquisitions)⁶. Glucose flux to the pentose phosphate pathway (PPP) was calculated from the relative proportions of [4-¹³C]-glutamate and [5-¹³C]-glutamate as described⁷.

Hyperpolarized ¹³C-MRS in cells: [U-¹³C, U-²H]-glucose or [1-¹³C]-alanine were prepared as described^{8, 9}, polarized, dissolved in isotonic buffer and injected into live cells in a 10mm NMR tube to a concentration of 10mM. ¹³C spectra were acquired every 3s for 300s on a Bruker 14.1T spectrometer and ratios of area under the curve (AUC) for product to substrate quantified.

In vivo hyperpolarized ¹³C-MRS: For slab studies, dynamic ¹³C spectra were acquired following intravenous injection of hyperpolarized [U-¹³C, U-²H]-glucose or [1-¹³C]-alanine from a 12mm (glucose) or 15mm (alanine) slab through the brain every 3s using a flyback spectral-spatial RF pulse¹⁰. For imaging studies, a 2D spectral-spatial echo-planar spectroscopic imaging (EPSI) pulse¹⁰ was used (spatial resolution=231mm³, temporal resolution=3s). ¹³C slab spectra were analyzed using Mnova. EPSI data was analyzed using in-house Matlab codes. Color heatmaps were produced by interpolating the data using a Lanczos-2 kernel and normalizing to noise.

Statistical analysis: Results are expressed as mean±SD and statistical significance assessed using an unpaired Student’s t-test (** = p<0.01, *** = p<0.005).

Results

Using ¹H-MRS we established that TERT expression induces unique metabolic reprogramming, leading to elevated levels of reduced glutathione (GSH), NAD(P)H, aspartate and AXP in NHA_tert cells relative to NHA_control (Fig.1A-1B). Since ¹H-MRS cannot distinguish between NADPH and NADH, we used spectrophotometric assays to establish that both NADPH and NADH were elevated in NHA_tert cells (data not shown in the interest of clarity). Glucose flux through the PPP is a major producer of NADPH, which, in turn maintains GSH in the reduced state¹¹. Using thermally-polarized ¹³C-MRS to measure the relative proportions of [4-¹³C]-glutamate and [5-¹³C]-glutamate produced from [2-¹³C]-glucose (Fig.1C), we established that PPP flux was elevated in NHA_tert cells (Fig. 1D).

In order to leverage this TERT-induced increase in PPP flux for non-invasive hyperpolarized ¹³C-MRS, we examined metabolism of hyperpolarized [U-¹³C,U-²H]-glucose (Fig. 2A). The flux of hyperpolarized [U-¹³C,U-²H]-glucose to the PPP metabolite 6-phosphogluconate (6-PG, 179.4ppm) was significantly elevated in NHA_tert cells relative to NHA_control (Fig. 2B).

Next, we examined the feasibility of imaging TERT status in vivo using hyperpolarized [U-¹³C, U-²H]-glucose. Dynamic ¹³C slab spectra acquired after intravenous injection of hyperpolarized [U-¹³C, U-²H]-glucose showed build-up of [1-¹³C]-6-PG only in rats bearing orthotopic NHA_tert tumors and not tumor-free controls (Fig.3A-3B). Importantly, 2D EPSI imaging indicated that 6-PG production clearly differentiated tumor from normal brain in orthotopic NHA_tert tumors (Fig. 3C).

We then exploited the observation that TERT expression increased NADH. Hyperpolarized [1-¹³C]alanine has been used to measure intracellular NADH in rat abdomen⁹ and liver^12. Hyperpolarized [1-¹³C]-alanine is first converted to [1-¹³C]-pyruvate and further conversion of [1-¹³C]-pyruvate to [1-¹³C]-lactate is dependent on NADH^{13, 14}. Lactate production from hyperpolarized [1-¹³C]-alanine was higher in NHA_tert cells relative to NHA_control (Fig.4A-4B).

To determine the feasibility of imaging TERT status in vivo using hyperpolarized [1-¹³C]-alanine, we acquired dynamic ¹³C spectra from a 15mm slab after intravenous injection of hyperpolarized [1-¹³C]-alanine. We observed [1-¹³C]-lactate only in NHA_tert tumor-bearing rats (Fig.5A) and not in tumor-free controls (Fig.5B). Importantly, 2D EPSI imaging of hyperpolarized [1-¹³C]-alanine metabolism in orthotopic NHA_tert tumors revealed pronounced differences in lactate production between tumor and normal brain (Fig.5C).

Conclusions

We show, for the first time, that hyperpolarized [U-¹³C,U-²H]-glucose and hyperpolarized [1-¹³C]-alanine can non-invasively monitor TERT expression in vivo in low-grade oligodendrogliomas. Since TERT expression is a hallmark of tumor proliferation^{2, 3}, our biomarkers have the potential to differentiate tumor from areas of pseudoprogression and pseudoresponse, which are currently major challenges in glioma imaging¹⁵. Our biomarkers can also monitor response to therapy, including novel TERT inhibitors currently under development^{2, 4}.

Acknowledgements

This work was supported by the following grants: NIH R01CA172845, NIH R01CA197254, UCSF Brain Tumor Loglio Collective, NICO, NIH P41EB013598 and NIH P50CA97257.