A total of 22 young adult female (242 ± 34 g) Sprague-Dawley rats (Charles River, Sulzfeld, Germany) were used. Animals were randomly distributed into three groups: (I) 8 rats were irradiated with carbon ions, (II) 8 rats were irradiated with photons, and (III) 6 rats served as non-irradiated controls. Animals were kept under the standard laboratory conditions at the German Cancer Research Center (DKFZ, Heidelberg, Germany). All experiments were approved by the governmental review committee on animal care.
The rat cervical spinal cord was irradiated either with carbon ions or photons. For irradiation, rats were anesthetized with a mixture of 4 Vol% Sevoflurane (Abbott, Wiesbaden, Germany) and oxygen at 2 l/min and animals were irradiated in a hanging position in a customized holding device (7). The rat spinal cord was irradiated from ventral to dorsal with a horizontal beam (field size of 10 × 15 mm2) that covered the cervical segments C1-C6 with C3 being in the center and C1 and C6 being located at the field edge.
Carbon ion irradiations were performed at the Heidelberg Ion-Beam Therapy Center (HIT, Heidelberg, Germany). The cervical spinal cord was positioned in the middle of a 1 cm spread-out Bragg peak (SOBP) with a dose-averaged LET of 91 keV/µm (80–104 keV/µm) and was irradiated with a total dose of 23 Gy given in 6 fractions (3.83 Gy/fraction) on consecutive days. The dose was selected to reach 100% complication probability (8). Photon irradiations (6 MeV, Artiste, Siemens, Erlangen) were performed with an iso-effective total dose of 61 Gy given in 6 fractions (10.167 Gy/fraction) on consecutive days to also reach 100% complication probability (8). The iso-effective dose was determined based on the RBE for the endpoint radiation-induced myelopathy measured in a previous study (8).
Animals underwent MRI before irradiation and were then imaged on a monthly basis to detect morphological alterations. As soon as an alteration occurred, the respective rat was monitored in shorter time intervals (Fig. 1).
MRI measurements were performed on a 1.5 T whole-body MR-scanner (Symphony, Siemens, Erlangen, Germany) using an in-house built small animal coil. Rats were anesthetized by inhalation anesthesia (Isoflurane, Baxter, 2.5 Vol% in 1.2 l/ min oxygen). During each MRI session, sagittal and axial T2-weighted images (4000/109 ms [repetition time/ echo time], voxel size 0.3 × 0.3 × 1.0 mm) were acquired per animal to assess morphological changes such as edema formation. Additionally, T1-weighted images (600/14 ms [repetition time/ echo time], voxel size 0.2 × 0.2 × 1.0 mm) were recorded before and after contrast agent injection to examine the onset of the blood spinal cord barrier (BSCB) disruption. The contrast agent (0.2 mmol/kg Magnevist®, Bayer, in 0.9% NaCl) was injected intravenously as a bolus via the tail vein and animals were imaged 5 min after injection.
Follow up and endpoints
After irradiation, rats were checked weekly for neurological alterations, weight, and general condition. The final biological endpoint was radiation-induced myelopathy (paresis II) within 300 days meaning that both forelimbs showed signs of paralysis (7). As soon as an animal reached paresis II, a final MRI measurement was performed and the animal was sacrificed. Before animals reached the final endpoint, they showed various morphological alterations (see “Image evaluation”) and paresis I (at least one forelimb showed signs of neurological dysfunction (7)). The latency times, defined as the time between the day of the first fraction of irradiation and the occurrence of the respective morphological and neurological alterations after irradiation, were recorded for each animal.
Two blinded readers (T.W., M.S.) performed qualitative visual inspections of the MR-images with a consensus reading of pathologies to determine the latency times for the morphological alterations. For this, the axial and sagittal T2-weighted images as well as the sagittal T1-weighted images (native and post contrast agent) were used.
Native T1-weighted images were used to identify bone marrow conversion. To examine the onset of loss of BSCB integrity, native and enhanced T1-weighted images were compared. The T2-weighted images were used to identify edema and the dilatation of the canalis centralis (syrinx). All latency times, i.e. the time between the first fraction and the occurrence of MR visible morphological alterations, were recorded.
When animals reached paresis II they were sacrificed by an overdose of i.p. injected Xylazine/ Ketamine (100 µl/kg each). Animals were perfused with a mixture of 4% paraformaldehyde (PFA) in phosphate buffered saline before the cervical spinal cord C1-C6 were dissected and fixated in PFA. After dehydration in ethanol the segments C1-C6 were embedded in paraffin with an axial orientation.
All histological stainings were performed on 9 µm thick paraffin sections that were deparaffinized and rehydrated before staining. For all immunohistochemistry stainings antigen retrieval in sodium citrate buffer (pH 6) was performed for 30 min at 95 °C to unmask antigen sites. At the end, sections were dehydrated and mounted in Eukitt.
For qualitative examination of the extent of demyelination, paraffin sections were stained with Luxol fast blue which binds to the lipoproteins of myelin (21) in combination with hemalaun/eosin (H&E). Sections were incubated in 0.1% Luxol blue at 50 °C over night, and washed in 95% ethanol, and differentiated in 0.05% lithium carbonat and 70% ethanol the next day. Afterwards, sections were counterstained with hematoxylin and eosin.
To study the degree of blood vessel permeability, extravagated endogenous serum albumin was immunohistochemically visualized. Endogenous peroxidase activity was blocked with 3% H2O2. After antigen retrieval, sections were incubated overnight at 4 °C with the primary antibody against albumin (Acris, 1:6000 diluted in 3% bovine serum albumin in TBS-T) followed by incubation with the secondary antibody (Abcam, 1:500, horse raddish peroxidase) for 30 min at room temperature. 3,3′-diaminobenzidine was used as chromogen (Vector Laboratories, Burlingame, CA, US). In the end, sections were counterstained with 1% Nissl in sodium acetate buffer.
To assess the degradation of the blood spinal cord barrier, sections were stained for the endothelial barrier antigen (EBA), a protein that is exclusively expressed on the luminal surface of non-fenestrated central nervous system blood vessels (22), in combination with the endothelial marker CD34. Autofluorescence was quenched by 30 min incubation with 0.1% Sudan Black B (Sigma-Aldrich®, Missouri, US). After blocking in 10% normal goat serum, sections were incubated with the primary antibodies monoclonal mouse anti-EBA SMI 71 (1:500, Biolegend, San Diego, Ca, US) and recombinant rabbit anti-CD34 antibody (1:1000, abcam, Cambridge, UK) in 3% bovine serum albumin in TBS-T at 4 °C over night. For detection, sections were incubated with the fluorescent-coupled secondary antibodies goat anti-rabbit Alexa Flour 555 (1:1000, Molecular Probes®, Invitrogen AG, Carlsbad, USA) and goat anti-mouse Alexa Fluor 488 (1:1000, Molecular Probes®) at room temperature for 30 min each. Sections were counter stained with DAPI and mounted with Fluoromount-G.
The difference in the latency times per morphological and neurological alterations between the irradiation modalities was tested by a two-sided student t-test with a significance level of 0.05. The longitudinal difference of the latency times within one irradiation modality was tested by one way repeated measure ANOVA. All statistical testing was performed in SigmaPlot 14.0 (Systat Software Inc., San Jose, CA, USA).