Chronic Multiple Sclerosis Lesions: Characterization with High-Field-Strength MR Imaging

Appendix E1

Participants

Patients were 18-65 years old (inclusive) and had been free of clinical relapse for at least 3 months at the time of the study. Patients with an EDSS score greater than 6.5 were not included because of physical difficulty in positioning on the MR imaging table. The exclusion criteria included the following: (a) serious medical conditions; (b) pregnancy or breastfeeding; (c) inability to provide informed consent; (d) permanent tattooed makeup (eg, eyeliner, lip) or general tattoos; (e) any nonorganic implant or any other device (eg, cardiac pacemaker; insulin infusion pump; implanted drug infusion device; cochlear, otologic, or ear implant; transdermal medication patch; any metallic implants or objects [body piercing(s), bone or joint pin, screw, nail, plate, wire sutures, surgical staples, shunts]); (f) cerebral aneurysm clips, shrapnel, or other metal embedded in the body (as from war wounds or accidents); (g) previous work in metallurgic fields or with machines that may have left any metallic fragments in or near the eyes; (h) severe auto accident in the past with uncertainty as to whether a metal object might still be present in the patient’s body; and (i) any psychologic contraindications to MR imaging (eg, claustrophobia).

Image Registration

In each patient, 3.0-T images (T1-weighted spin-echo, T2-weighted spin-echo, and FLAIR images) were registered to the 7.0-T gradient-echo magnitude image by using the coregistration algorithm implemented in Statistical Parametric Mapping, version 8. This algorithm is used to register a single volume (object volume) to another volume (target volume). The advantage of this algorithm is that the modality of these two volumes does not need to be the same. Also, coregistration determines a transformation between an image pair (target-object) that can then be applied to a group of volumes. In our study, registration involved the following steps. First, the 3.0-T T1-weighted spin-echo (T1SE) and 3.0-T FLAIR images were registered to the 3.0-T T2-weighted spin-echo (T2SE) images, generating two new registered image data sets: r-T1SE and r-T2SE. Second, the 3.0-T T2-weighted spin-echo image was registered to the corresponding 7.0-T gradient-echo image. The 7.0-T gradient-echo image (echo time, 30 msec) was used as a reference image. Third, the translation-rotation matrix obtained in the second step was applied to the r-FLAIR and r-T2SE images to generate the final registered image data sets. The 7.0-T R2* and phase images were self-registered to the magnitude image and thus needed no additional registration. In this way, all the corresponding 3.0-T and 7.0-T images in the same patient were registered. Careful inspection after every registration step was performed to ensure successful registration.

Immunohistochemistry

After MR imaging, the brain was embedded in paraffin. Small segments (2 × 4 × 3 mm³ each) were cut into 10-mm-thick serial slices with a microtome (Carl Zeiss MicroImaging) and were mounted onto Superfrost glass slides (Thermo Fisher Scientific). Slices were deparaffinized prior to histologic staining.

Iron Staining

Perls iron staining was performed as follows: Slices were incubated with a freshly prepared 1: l mixture of 5% potassium ferrocyanide and 5% hydrochloride for 2 hours at 37°C and were then washed three times in phosphate-buffered saline (PBS) for 5 minutes each. Subsequently, a few drops of 3,3'-diaminobenzidine (Sigma, St Louis, Mo) stock solution (a mixture of 1 mg 3,3'-diaminobenzidine per milliliter of 0.075% H₂O₂) were added to each slice for about 30 minutes. Slices were washed with distilled water, dehydrated in a series of graded alcohol (70%, 95%, 100%) solutions and xylene, air dried, and mounted. As a control, extra slices were incubated with potassium ferrocyanide solution, but the hydrochloride was replaced by distilled water. No staining was observed in the control slices.

Myelin Staining

A standard Luxol fast blue method was used for myelin staining. Slices were incubated overnight at 56°C in 0.1% Luxol fast blue solution (Luxol fast blue MBS: 0.1 mg; 95% ethyl alcohol: 100 mL; glacial acetic acid: 0.5 mL). To enable us differentiate the contrast between white matter and gray matter, slices were immersed in 0.05% lithium carbonate solution for 15 seconds and then in 70% ethyl alcohol for 30 seconds. Slices were then washed with tap water, dehydrated, air dried, and mounted.

Immunochemical Staining for Ferritin

Endogenous peroxidase activity was blocked by incubating slices in PBS containing 3% H₂O₂ for 10 minutes at room temperature. Slices were washed three times in PBS for 3 minutes each. Slices were then incubated with the block solution (2% normal goat serum, 1% bovine serum albumin, and 0.1% Triton X-100 in PBS) for 10 minutes at room temperature to prevent nonspecific binding and were washed. Rabbit anti-ferritin antibody (Sigma) as a primary antibody was used at a 1:100 dilution in the PBS containing 1% bovine serum albumin and 0.1% Triton X-100. Slices were incubated with primary antibody for 1 hour at room temperature and were washed. Subsequent incubation was performed with Envision+ liquid (Dako, Glostrup, Denmark) for 30 minutes at room temperature. Last, peroxidase activity was demonstrated by soaking the slices in 3,3'-diaminobenzidine solution (0.5 mg 3,3'-diaminobenzidine per milliliter of PBS containing 0.03% H₂O₂) for 5 minutes, which leads to a brown reaction product. Slices were then washed with tap water, dehydrated, and mounted. Control staining was done by replacing the rabbit anti-ferritin antibody with normal rabbit immunoglobulin G (Santa Cruz Biotechnology, Santa Cruz, Calif). No staining was observed in these control slices.