br Results br Discussion The most relevant overall finding
Discussion The most relevant overall finding was that DDR1 is present in myelin and is probably synthesized by oligodendrocytes. This result was supported by three different approaches. First, we detected presence of DDR1 mRNA in myelin by ISH. Second, the pattern of DDR1 immunostaining using an antibody against the C-terminal domain of the receptor is indistinguishable from, and co-localized with, MBP. Finally, we found by RT-qPCR a strong positive association between the Cy3-UTP levels of DDR1 and the expression levels of MAG and OLIG2 suggesting that the three proteins are expressed in oligodendrocytes. OLIG2 is a transcription factor implicated in oligodendrocyte cell fate determination and maturation (Ligon et al., 2006), whereas MAG is a glycoprotein synthesized by mature oligodendrocytes that has been implicated in the formation and maintenance of the myelin sheaths (Schachner and Bartsch, 2000). The specific role of DDR1 in myelin cannot be elucidated from our results, although we can speculate about several possible functions. First, DDR1 may participate in oligodendrocyte process outgrowth. Processes are snail-like extensions of the oligodendrocyte cytoplasmic membrane, each of which forms a segment of sheathing around an axon. A recent study demonstrated an association between DDR1 and non-muscle myosin heavy chain IIA (NMHC-IIA), which regulates fibroblast cell spreading and motility (Huang et al., 2009). Oligodendrocytes also express NMHC-IIA, which has been shown to regulate process formation (Song et al., 2001). Therefore, DDR1 anchored in the plasma membrane of oligodendrocyte processes may bind to NMHC-IIA and assemble to filaments once activated by an external stimulus (for instance, collagen). From this initial contact, the oligodendrocyte process would be promoted to extend further. Second, DDR1 could participate in axon–oligodendrocyte contact, which is important in commencing the ensheathment. Third, DDR1 may be involved in the oligodendrocyte–oligodendrocyte plasma membrane interactions, which are important in the myelin compacting process. No gross anatomical or histological alterations have been observed in the brains of DDR1-deficient animals (E. Vilella and W.F. Vogel, unpublished results). Hearing loss is the only defect observed in the sensory nervous system of DDR1-deficient mice (Meyer zum Gottesberge et al., 2008). Meyer zum Gottesberge and co-authors (2008) also showed that the axons of the spiral ganglion of these mice are myelinated but that their myelin structure is perturbed. As such, perhaps DDR1 is not essential for myelin synthesis. However, one can speculate that its absence results in myelin that is less compact and more susceptible to external insult or challenge. Similarly, oligodendrocytes in MBP- and MOBP-deficient mice (Matthieu, 1981, Yool et al., 2002) undergo initial myelination steps, but the myelin subsequently fails to become compact. An in-depth examination of myelin sheaths in DDR1-deficient mice may elucidate the role of this receptor in the myelin assembly process. Apart from classical myelin-related diseases such as multiple sclerosis and leukodystrophies, recent compelling data suggest that myelin defects may underlie certain psychiatric diseases (Davis et al., 2003, Dwork et al., 2007, Walterfang et al., 2006). Recently, a strong genetic association was found between CNPase and OLIG2 (two oligodendrocyte markers) and schizophrenia (Georgieva et al., 2006, Peirce et al., 2006). In addition, we found a significant association between the DDR1 gene and schizophrenia (Roig et al., 2007). Hence, it is plausible that mutations in these genes may alter the myelin structure and consequently predispose an individual to psychotic behavior. In contrast to the C-terminal antibody, an antibody directed against the N-terminal of DDR1 does not produce a signal in myelin. This observation suggests either that DDR1 undergoes a shedding process in myelin and subsequently remains anchored to the myelin membrane as a cleaved receptor or that the epitope for the N-terminal antibody (amino acids 51–60) is masked. Shedding of DDR1 induced by ligand binding has been previously reported (Slack et al., 2006, Vogel, 2002, Flynn et al., 2009). Although the expected 125kDa full-length receptor was detected by western blot in this study, a 62kDa band was also observed, suggesting that DDR1 ectodomain shedding may occur in the human brain.