The CD4–Th1 model for multiple sclerosis: a crucial re-appraisal