Myeloid-derived suppressor cells (MDSCs) are highly prevalent inflammatory cells that play a key role in tumor development^1,2 and are considered therapeutic targets³. MDSCs promote tumor growth by blocking T-cell-mediated immunosuppression through depletion of the extracellular arginine pool essential for T-cell proliferation^2,4. To deplete arginine, MDSCs express high levels of the enzyme arginase, which catalyzes the breakdown of arginine into urea and ornithine (Figure 1). Based on this knowledge, we developed and characterized ¹³C-labeled arginine ([guanido-¹³C]-arginine) as a new hyperpolarized probe. We show that [guanido-¹³C]-arginine can be hyperpolarized and can be used to detect arginase activity in solution, and in MDSCs, demonstrating its potential for imaging MDSCs and their inhibition in cancer.

Hyperpolarized 13C arginine preparation and characterization 3.4M [guanido-¹³C]-arginine was prepared in 7.5μM Trizma solution in water with 15mM OX63 and 1.5mM Dotarem, and polarized using a Hypersense DNP polarizer (Oxford Instruments) for 90 minutes. After dissolution in a Tris-based buffer (40mM Tris, 3μM EDTA, pH~7.8), single-pulse ¹³C spectra were acquired using an 11.7T INOVA spectrometer (Agilent Technologies) to evaluate T1 and signal-to-noise ratio (SNR) enhancement. T1 was also evaluated on a 3T GE clinical scanner (GE Healthcare). Spectra were quantified by peak integration using MestRenova (Mestrelab).

Enzyme studies Following polarization and dissolution, 3mL of hyperpolarized [guanido-¹³C]-arginine was injected into an NMR tube containing different concentrations of arginase (0 to 2000U/L; n=3/concentration). Dynamic sets of hyperpolarized ¹³C spectra (TR=3s, flip angle=10deg, number of transients=50, spectral width=20kHz, 20000 points), and thermal equilibrium ¹³C spectra at the end of the study (90deg pulse, TR=80s, NT=16), were acquired on the 11.7T scanner and analyzed as above except that for hyperpolarized data contaminants were subtracted.

MDSC studies To generate MDSCs, bone marrow cells isolated from Balb/c mice were treated with colony-stimulating factors (0.1μg/mL G-CSF; 250U/mL GM-CSF, Peprotech) then exposed to interleukin-13 (+IL13, 80ng/mL) to induce arginase expression. Controls were not exposed to IL13 (-IL13)². To probe for arginase activity, 3mL of hyperpolarized [guanido-¹³C]-arginine were injected into an NMR tube containing: (1) Cells in 500μL fresh medium ­– to assess intracellular arginase activity (~1e7 cells, n=3 +IL13/-IL13), or (2) 500μL of growth medium previously exposed to MDSCs ­– to assess extracellular arginase activity (n=3 +IL13/-IL13). Dynamic hyperpolarized spectra followed by thermal equilibrium spectra were recorded and analyzed as for enzyme. Extracellular and intracellular arginase activities were determined using a colorimetric assay (QuantiChrom Arginase Assay Kit, BioAssays, n=3).

Characterization Following polarization, the resonance of hyperpolarized [guanido-¹³C]-arginine was detected at 159.7ppm with an SNR enhancement of 5018±412 compared to the thermal spectrum (n=3; Figure 2). Additional resonances detected at 177.1, 165.5 and 164.2ppm were attributed, respectively, to the natural abundance ¹³C carbonyl of arginine and to ¹³C-urea and [6-¹³C]-citrulline contaminations. The T1 of hyperpolarized [guanido-¹³C]-arginine was 9.9±0.1s at 11.7T (n=3) and 12.3±0.8s at 3T (n=2), indicating little T1 dependence on magnetic field, in contrast to carbonyls⁵.

Enzyme studies Exposure of hyperpolarized arginine to arginase at or above 300U/L resulted in a detectable build-up of hyperpolarized ¹³C-urea at 165.5ppm with a maximum at 14±2s post-maximum arginine signal (Figure 3A). Furthermore, hyperpolarized ¹³C-urea production increased linearly with enzyme concentration (Figure 3B), but no build-up in [6-¹³C]-citrulline could be detected. Thermal ¹³C spectra acquired post-hyperpolarized signal decay confirmed continued production of urea (Figure 3C/3D).

MDSC studies Following injection of hyperpolarized arginine, a build-up in hyperpolarized ¹³C-urea was observed in +IL13 MDSCs but not in –IL13 cells (Figure 4A/4B) resulting in a significantly higher hyperpolarized ¹³C urea-to-arginine signal (Figure 4C), and consistent with the expected induction of arginase expression only in cells exposed to IL13. No hyperpolarized ¹³C-urea was observed in the growth media, indicating that extracellular arginase was below detection (Figure 4C). No build-up in hyperpolarized ¹³C citrulline was observed consistent with the expected absence of iNOS activation. Production of urea only in the +IL13 MDSCs was confirmed on thermal ¹³C spectra. To validate our hyperpolarized findings, we used a colorimetric enzyme assay to determined arginase activity. As expected, the extracellular arginase activity, as well as the intra-cellular arginase activity in –IL13 cells, were below detection level of the hyperpolarized ¹³C method (<260U/L in –IL13 cells and <25U/L in +IL13/-IL13 growth media), whereas the up-regulated intracellular arginase activity in +IL13 MDSCs (576±67U/L) was readily detectable by our hyperpolarized probe and significantly higher than in –IL13 cells (unpaired t-test, p-value=0.004) (Figure 4D).

We demonstrate, to our knowledge for the first time, the feasibility of using [guanido-¹³C]-arginine to monitor arginase activity. This probe could serve as a readout of MDSC function in tumor development and its inhibition following MDSC-targeted immunotherapies.