In the central nervous system, oligodendrocytes wrap axons of certain neuron subtypes with a specialised cell membrane, myelin. The distribution of myelin along axons depends on the specific neuron they belong to, e.g., excitatory axons from deep cortical layers are more myelinated than those found in superficial layers (1), whilst the myelination of inhibitory neurons depends on the specific subtype (2). This diversity suggests that myelination might affect neuronal integrity and function in a subtype-specific manner.

To date, however, myelin manipulations are not neuron subtype-specific, so we have little insight into its specific roles for the individual subtypes. To study this, we are developing a genetic toolkit to modulate the myelination of specific axons of interest in an otherwise intact central nervous system. We have overexpressed the candidate inhibitor JAM2 using cortical organotypic slices and in vivo mice, and its ortholog Jam2a in the spinal cord of zebrafish. JAM2 has previously been shown to prevent the myelination of the soma of specific inhibitory neurons in mice (3).

In cortical organotypic slices, overexpressing JAM2 in the inhibitory parvalbumin interneurons suggests a reduced myelination of 38% and 43% in deep and superficial layers, respectively (N = 3 litters). This modulation is not significant (LMM with treatment and layer as fixed effects and litter/brain/axon nested as random effects; n = 114 axons, 17 inserts, and 3 litters; Type III ANOVA: p = 0.09818, F_{(1, 6.386)} = 3.7453 for treatment), but it has led us to follow it up in vivo. Preliminary results show a significant reduction in myelination (LMM as previously; n = 77 axons, 13 brains, and 4 litters; Type III ANOVA: p = 0.007507, F_{(1, 73)} = 7.5628 for treatment).

However, overexpressing JAM2 in the excitatory pyramidal neurons in slices does not affect their myelination (LMM as previously, nested as litter/insert/brain; n = 19 brains, 8 inserts, and 2 litters; Type III ANOVA: p = 0.5797, F_{(1, 4.7017)} = 0.3534 for treatment), suggesting different regulatory mechanisms than for parvalbumin interneurons.

In addition, overexpressing Jam2a in zebrafish has not affected the myelination of reticulospinal axons at 4dpf (LMM with treatment and diameter as fixed effects and fish as random effects, n = 30 axons and 26 fish, Type III ANOVA: p = 0.877439, F_{(1, 23.7113)} = 0.0243 for treatment), nor the myelination of Mauthner axons at 3dpf (LM with treatment and diameter as fixed effects, n = 20 axons and fish, ANOVA: p = 0.89352, F_(1,17) = 0.0185 for treatment). Both neuron subtypes are excitatory; this is consistent with the mammalian results. We plan to test Jam2a in inhibitory neurons of zebrafish and other candidates in excitatory neurons in both models.

Overall, the development of genetic tools to modulate myelination in individual neuronal subtypes will allow us to characterise the role of myelin in those specific cells, as the rest of the neurons will maintain their normal myelin and function. Future experiments include the assessment of the morphology and function of those neurons in which myelination has been successfully reduced.