Abstract Body:
Background
Stereotactic radiosurgery (SRS) is an effective treatment for various intracranial pathologies. Despite advances in SRS treatment precision, the surrounding healthy brain tissue receives a significant dose of radiation resulting in severe side effects such as radiation necrosis (RN) and neurocognitive deficits1. Detailed characterization of RN and factors that influence its onset and progression is of great interest and many studies has identified neuro-inflammation as a key driver2. The vascular endothelium of the brain plays an essential role in the maintenance of the brain environment and its function are mediated by cell adhesion molecules (CAMs). Particularly the vascular cell adhesion molecule (VCAM-1) is a key mediator of inflammation and has been shown to play a significant role in neurological disease 3,4,5. By using antibody-conjugated microparticles of iron oxides (MPIO) targeting VCAM-1 (MPIOVCAM-1) as published by others6, our goal was to characterize the molecular signature of radiation-induced vascular inflammation of the healthy rat brain.
Methods
RN: Fisher 344 rats were irradiated with a single hemispheric radiation total dose of 80 Gy (Figure) using a Leksell Gamma Knife Perfexion (Elekta AB, Stockholm, Sweden) with a 4 mm collimator7. 5 rats underwent SRS and 3 remained as control.
MRI : At 6 and 12 weeks after irradiation, rats were imaged with a T2*-weighted sequence before MPIOVCAM-1 injection to visualize emergence and evolution of radionecrotic lesions (“pre-injection” scan) and, after injection of MPIOVCAM-1 to image vascular inflammation as described previously8,9. All MRI experiments were conducted on a 210-mm small animal 7T scanner (Varian Inc., Palo Alto, CA) with a dedicated rat head-coil (RAPID MR International, OH) using a T2*-weighted 3D multiple gradient echo sequence (TR = 50 ms, first TE = 5,5 ms, echo spacing = 6.3 ms (5 echoes), flip angle = 15°, data matrix = 192x192x96, field of view = 30x30x15 mm, 2 averages).
Image analysis: Negative contrast quantification (hypointensities) in T2*-weighted images was performed similarly to previously described9. Briefly, a brain template and its segmented areas10 (grouped in 18 areas) were registered (diffeomorphic transform) to each data set using advanced normalization tools (ANTs). A threshold was determined for each brain area as the mean signal minus four standard deviations, and negative contrast volume was determined as the total volume of voxels with a magnitude value under this threshold for each brain region.
Results
Performing T2*-weighted imaging before and after the injection of MPIOVCAM-1 allowed us to differentiate between endogenous (i.e., radionecrosis present at both 6 and 12 weeks) and MPIOVCAM-1 NC. Relative volume of MPIO-induced NC (% of total volume) was calculated for both irradiated and control groups. Since VCAM-1 is expressed endogenously, VCAM-1 expression changes due to irradiation were determined based on the difference between irradiated and control animal groups (DNC) and is reported here as a function of radiation dose (Supplementary Figure). Imaging at 12 weeks post-irradiation revealed an increase in VCAM-1 levels compared with 6 weeks scans, suggesting inflammation ramps up between these time points. NC also increased as a function of dose at 12 weeks, with results at 6 weeks being more ambiguous
Conclusion
The present study shows that VCAM-1 targeted molecular imaging enables the detection of vascular inflammation in the rat brain post SRS. Compared to stroke rat brain inflammation model11, or compared with a similar SRS model in the mouse12, rat inflammatory response to radiation injury seems low. Immunohistological validations are ongoing, but initial results suggest that VCAM-1 is not highly expressed in our SRS rat model.
Author
Universite de Sherbrooke