A significant increase of dynein protein level was detected in PLP-tg mouse optic nerves compared to PLP-wt mice. Similar results with elevated dynein levels have been observed in PLP-null mice that also show disturbed axonal transport. It has been suggested that raised dynein protein levels might reflect an accumulation of dynein-linked retrogradely moving organelles. This is in line with our finding that juxtaparanodal axonal swellings with morphologically-altered mitochondria were located predominantly on the distal side of the node of Ranvier. Accumulation of mitochondria or other cell organelles in axonal abnormalities are often a correlate for disturbed axonal transport thus suggesting a link between the impaired retrograde axonal transport and the formation of the juxtaparanodal swellings and increase in dynein protein levels. These features were nearly restored in the absence of RAG-1 in PLP-tg mice suggesting a central role of adaptive immune cells in axonopathic changes in the respective mutants. Since in our study, juxtaparanodes appeared to be most susceptible for changes mediated by the cytotoxic T-lymphocytes, we investigated the possible association of these structures. Indeed, T-lymphocytes in direct vicinity to juxtaparanodes were detectable and, most interestingly, preferentially displayed a spindle-like shape at this location. We do presently not know the significance of this constant observation, but it is worthwhile to speculate that increased adhesion to a possible target structure might lead to the maximal cell contact extension between lymphocytes and the juxtaparanode resulting in the elongated shape of the lymphocytes. Interestingly, this phenomenon has been observed also in perforinand WY 14643 granzyme B-deficient myelin mutants suggesting that this cellcell interaction is independent of the cytotoxic features of the lymphocytes. Whether typical axo-glial molecules, which can even serve as antigens in multiple sclerosis, are functionally involved in this interaction, remains to be resolved. Our basic finding that cytotoxic lymphocytes with their respective cytotoxic agents are essential for impaired retrograde axonal transport is reminiscent of the pathomechanisms described in Theiler’s virus-related model of demyelination. However, in contrast to the Theiler’s virus model, in PLP transgenic mice MHC class I immunoreactivity is only detectable on the mutant oligodendrocytes. This might have substantial consequences for the respective pathomechanisms. For the Theiler’s virus model, it is assumed that the virus-mediated demyelination exposing the MHC class I-positive axolemma is an important prerequisite for the formation of an immunological synapse between lymphocytes and MHC class I-positive axons and, thus eventually, for the immune attack by the cytotoxic Tcells. In the PLP mutant, in which MHC I-restricted T-cell receptors play a crucial role for neural damage, cytotoxic T-cells might only be able to attack the MHC class Ipositive oligodendrocytes, since the molecule is not detectable on axons. Paradoxically, in the PLP mutants, myelin often remains intact over wide stretches of the internodes whereas the axons show shrinkage or swelling causing periaxonal vacuoles and axonal enlargements, respectively. This scenario resembles pathological features seen in MOG-EAE, where the immunological attack against the target cell, the oligodendrocyte, is transmitted to the axon without major local damage to myelin. How the glial-related attack is transmitted to the axonal partner and whether the spindle-shaped lymphocytes directly attaching to the juxtaparanodes are the cytotoxic mediators is presently not clear. The primary defect of PLP transgenic mice clearly resides in mutant oligodendrocytes.