Synopsis

This study compared the capability of 3D magnetic resonance spectroscopic imaging (MRSI) of hyperpolarized [1-¹³C]pyruvate metabolism and ¹⁸FDG-PET to detect early responses to a TRAIL agonist (Medimmune MEDI3039) in Colo205 colorectal cancer and MDA-MB-231 triple-negative breast cancer xenograft models expressing luciferase and mStrawberry fluorescent protein. 24 hours after treatment, bioluminescence and hyperpolarized lactate/pyruvate ratio decreased significantly in all treated animals, preceding decreases in tumor volume. Conversely, ¹⁸FDG-PET did not detect treatment response. This suggests that for some tumors hyperpolarized [1-¹³C]pyruvate may be an improvement on ¹⁸FDG-PET and RECIST for the early detection of treatment response.

Introduction

We compared the capability of PET imaging of 2-deoxy-2-(18F)fluoro-D-glucose (18FDG) and 3D magnetic resonance spectroscopic imaging (MRSI) of hyperpolarized [1-13C]pyruvate metabolism to detect early responses to a TRAIL agonist (Medimmune MEDI3039) in Colo205 colorectal cancer and MDA-MB-231 triple-negative breast cancer xenograft models.

Methods

Colo205 and MDA-MB-231 cells (ATCC) were transfected with luciferase and mStrawberry reporter genes. 5 ×10⁶ cells were implanted subcutaneously into the right flank of BALB/c nude mice and tumors grown until they were 0.8 ± 0.4 cm3. All mice underwent bioluminescence and fluorescence imaging, hyperpolarized [1-¹³C]pyruvate MRSI and ¹⁸FDG-PET imaging in a single 2 hour imaging session repeated before and 24 hours after treatment with MEDI3039. Bioluminescence images were acquired 5 min after injection of 150 mg/kg D-luciferin. ¹⁸FDG (13.08±1.7 MBq) and hyperpolarized [1-¹³C]pyruvate (15 ml/kg 75mM) were injected intravenously. A spiral 3D spectral-spatial pulse sequence was used to excite alternately the [1-¹³C]pyruvate and [1-¹³C]lactate resonances. Images of each metabolite had a spatial resolution of 1.25 ×1.25 × 2.5 mm and a temporal resolution of 1 s.¹ Following hyperpolarized MRI a 15 min static PET acquisition began at 90 min post-18FDG administration using a NanoPET/CT (Mediso). Images were analyzed and co-registered in VivoQuant (InviCRO) and Matlab (Mathworks). Ex vivo analysis of tumors included ¹H and ¹³C NMR of tumor extracts following [1,6-¹³C]glucose infusion, western blotting, histology, immunohistochemistry and FACS sorting of tumors following ¹⁸FDG administration.

Results

In Colo205 tumors, treatment with MEDI3039 resulted in a mean decrease of 71±31% in tumor bioluminescence (P = 0.005). This was attributed, at least in part, to a mean decrease of 61% in ATP measurements made ex vivo (P = 0.0008). Hyperpolarized MRSI was completed in 10 animals, six undergoing treatment with MEDI3039 and four controls. A decrease in the lactate/pyruvate ratio was observed in all MEDI3039-treated tumors, the mean pre-treatment ratio of 2.4±0.5 decreasing to 1.3±0.3 after treatment (P = 0.01). This was consistent with significant increases in TUNEL and cleaved caspase-3 (CC3) staining of tumor sections of 137% and 168%, respectively (P = 0.007 and 0.0002, respectively). ¹⁸FDG-PET was completed in 13 animals (7 treated and 6 controls) and did not detect a response to MEDI3039 treatment, with a mean difference in SUV_max of 0.03 (P = 0.8) and SUV_mean of 0.2 (P = 0.6). mStrawberry fluorescence intensity and tumor volume also did not change 24 hours after treatment (P = 0.22 and 0.35, respectively). Across all modalities control treatment did not result in significant changes in the measured signals. Similar observations were made in MDA-MB-231 tumors, with bioluminescence and hyperpolarized MRSI detecting response in all seven tumors treated with MEDI3039. Tumor bioluminescence decreased by 60±29% and the hyperpolarized lactate/pyruvate ratio decreased from 2.2±0.5 to 1.4±0.3. There was no significant change in FDG SUV_max after treatment (P = 0.09). ¹⁸FDG autoradiography of Colo205 tumors confirmed the in vivo PET findings that tumour FDG uptake was unchanged after treatment. Western blot analysis showed a significant decrease in expression of GLUT1 and GLUT3 after treatment. Following [1,6-¹³C]glucose infusions into Colo205 tumor-bearing mice, ex vivo ¹³C NMR measurements on tumor extracts showed similar glucose labelling in tumors before and after treatment but treatment resulted in a large decrease in all downstream metabolites.

Discussion

Response to treatment with a novel TRAIL agonist, MEDI3039, was detected in Colo205 and MDA-MB-231 tumors just 24 hours after treatment using both BLI and ¹³C MRSI with hyperpolarized [1-¹³C]pyruvate, which preceded changes in tumor volume. Conversely, ¹⁸FDG-PET and mStrawberry fluorescence did not detect response at this time. Ex vivo analysis of Colo205 tumors suggest that despite a decrease in tumor glycolysis, after treatment with MEDI3039 it is still possible for glucose and ¹⁸FDG to accumulate in the tumor, resulting in ¹⁸FDG-PET being unable to detect treatment response.

Conclusions

This result suggests that for some tumors hyperpolarized [1-¹³C]pyruvate may be an improvement on ¹⁸FDG-PET and RECIST for the early detection of treatment response.

Acknowledgements

References