The radiation response of cells from 9L gliosarcoma tumours is correlated with [F18]-EF5 uptake

Abstract

Purpose: Tumour hypoxia affects cancer biology and therapy-resistance in both animals and humans. The purpose of this study was to determine whether EF5 ([2-(2-nitro-1-H-imidazol-1-yl)-N-(2,2,3,3,3-pentafluoropropyl)-acetamide]) binding and/or radioactive drug uptake correlated with single-dose radiation response in 9L gliosarcoma tumours.

Materials and methods: Twenty-two 9L tumours were grown in male Fischer rats. Rats were administered low specific activity (18)F-EF5 and their tumours irradiated and assessed for cell survival and hypoxia. Hypoxia assays included EF5 binding measured by antibodies against bound-drug adducts and gamma counts of (18)F-EF5 tumour uptake compared with uptake by normal muscle and blood. These assays were compared with cellular radiation response (in vivo to in vitro assay). In six cases, uptake of tumour versus muscle was also assayed using images from a PET (positron emission tomography) camera (PENN G-PET).

Results: The intertumoural variation in radiation response of 9L tumour-cells was significantly correlated with uptake of (18)F-labelled EF5 (i.e., including both bound and non-bound drug) using either tumour to muscle or tumour to blood gamma count ratios. In the tumours where imaging was performed, there was a significant correlation between the image analysis and gamma count analysis. Intertumoural variation in cellular radiation response of the same 22 tumours was also correlated with mean flow cytometry signal due to EF5 binding.

Conclusion: To our knowledge, this is the first animal model/drug combination demonstrating a correlation of radioresponse for tumour-cells from individual tumours with drug metabolism using either immunohistochemical or non-invasive techniques.

Figure 1

Surviving fraction of cells dissociated and plated from individual 9L tumours was significantly correlated with the mean EF5 binding of the same cells (n = 22; r² = 0.248; p = 0.019). For each point, a tumour-bearing animal was administered EF5 (30 mg/kg, mixed unlabelled plus ¹⁸F-labelled) 3 h before the tumour was irradiated with 17 Gy (continuous isoflurane anesthesia). The tumour was removed and minced and cells dissociated from a portion of the mince. Some of the resulting cells were plated for colony formation and others were fixed, stained for EF5 adducts, and analysed by flow cytometry using a calibrated scale. In this and subsequent figures, the surviving fraction was not corrected for the average plating efficiency of cells from non-irradiated tumours (15–35%).

Figure 2

(a) Surviving fraction of cells dissociated and plated from individual 9L tumours was significantly correlated with the mean ¹⁸F-EF5 uptake, determined by gamma counts. For each point, a tumour bearing animal was administered EF5 (30 mg/kg, mixed unlabelled plus ¹⁸F-labelled) 3 h before the tumour was irradiated with 17 Gy (continuous isoflurane anesthesia). The tumour was removed and minced cells dissociated from a portion of the mince (as described in Figure 1) and gamma counts determined from another portion of the mince. The gamma counts were normalised by comparing the cpm/mg of tumour tissue versus the cpm/mg of muscle (n = 22; r² = 0.263; p = 0.015). (b) For 6 tumours imaged by the PENN G-PET scanner, there was a significant correlation between the tumour:muscle ratio, determined from the PET image, and tumour:muscle ratio, determined from samples assayed by gamma counts (n = 6; r² = 0.735; p = 0.029).

Figure 3

(a) Surviving fraction of cells dissociated and plated from individual 9L tumours was significantly correlated with the mean ¹⁸F-EF5 uptake, determined by gamma counts. For each point, a tumour-bearing animal was administered EF5 (30 mg/kg, mixed unlabelled plus ¹⁸F-labelled) 3 h before the tumour was irradiated with 17 Gy (continuous isoflurane anesthesia). The tumour was removed and minced cells dissociated from a portion of the mince (as described in Figure 1) and gamma counts determined from another portion of the mince. The gamma counts were normalised by comparing the cpm/mg of tumour tissue versus the cpm/mg of blood (n = 22; r² = 0.371; p <0.003). (b) For the 22 tumours studied, there was a very strong correlation between tumour:blood and tumour:muscle (n = 22; r² = 0.90; p <0.0001). The small differences relative uptake in muscle compared to blood may reflect very modest hypoxia in the muscle, independent of tumour, since tumour:blood ratio provided the best estimate of relative radioresistance (i.e., compare Figure 2a versus Figure 3a).

Figure 4

For the 22 tumours studied, there was no significant relationship between the mean flow cytometry signal, determined by EF5 binding and tumour to muscle (a) or tumour:blood (b) ratios determined by gamma counts measuring ¹⁸F-EF5 uptake.

Figure 5

Sample images from the Philips MOSAIC scanner. Upper panels: planar and orthogonal views (transaxial, sagital, coronal) at 10 minutes following drug injection of an animal whose tumour had a large central necrosis. Lower panels: similar views of the same animal but at 150 minutes following drug injection. Note that partial volume effects limit the ability to determine relatively high uptake that appears around the necrotic zone.