Understanding the role of astrocyte-derived factors during CNS myelination

This dissertation describes the influence of astrocyte-derived factors during CNS myelination using
native in vitro and ex vivo myelinating cultures as well as a novel developed computer-vision algorithm
for reliable myelin sheath quantification. Miniaturized immunopanned retinal ganglion celloligodendrocyte
precursor cell (RGC-OPC) co-cultures in a higher throughput multi-well plate format
in vitro, suitable to test small compound libraries, were established. Maintaining the reciprocal
interaction of vital axons and OLs ensured compact myelin formation and allowed the rapid study
and manipulation of all the steps of myelination. A novel computer-vision algorithm tailored to
quantitate myelin sheaths reliably, permit for the first time analysis at the single cell level and with
enough sensitivity to catch dose-responses of myelinating compounds. The robustness and efficacy
of these combined experimental and technical advances were demonstrated with published promyelinating
compounds BQ3020 and XAV939. This method was the starting point for the study of
myelin affecting astrocyte-derived factors from primary human and rat astrocytes.
A large number of bioactive proteins and lipids was identified from the astrocyte conditioned media
(ACM) and their promoting effect on myelination shown in vitro. On the hunt for astrocyte-derived
factors, brevican was identified as a promoter of late-stage myelination in both in vitro and ex vivo
cultures. Furthermore, brevican emerged to be particularly important during developmental myelin
formation. Brevican knockout (BCAN KO) mice showed profoundly decelerated myelin formation
in vivo at 14 days of age compared with wild-type (WT) mice, but a normal myelin phenotype in
adulthood. At present, it remains uncertain whether brevican acts directly as a ligand of neuronal
or glial receptors or attracts molecules to the OL/axon interphase. Nevertheless, new putative OLderived
interactors of brevican such as contactin 1 (CNTN1) via tenascin-R (TN-R) linkage, are
proposed in this work.
In addition, the changes in astrocytic protein and lipid efflux in response to proinflammatory stimuli
were investigated, and the conserved expression patterns in rats and humans were identified.
The shift toward a reactive, inflamed, and hypertrophic astrocyte phenotype resulted in particular
conserved changes in lipid and protein efflux that ultimately hampered myelination. Furthermore,
brevican was significantly downregulated by reactive astrocytes, suggesting a possible mechanistic
link. However, brevican was not solely responsible for this reduced myelination. In a disease
state, the totality of astrocytic expression changes likely causes multiple disruptions simultaneously, some of which aid myelin homeostasis and some of which hinder it; several of these are
discussed in this work. Overall, reactive astrocytes play a dual role, and concerning myelination
the dominant harmful role could be determined by the interplay of secreted molecules with the
surrounding extracellular milieu or the state of OPCs/OLs.
In summary, a valuable workflow to study the steps of myelination in a reliable and automated
manner in vitro was established. Moreover, this work provides important insights into the conserved
spectrum of astrocyte-derived secretion cues as well as the alterations in protein and lipid
efflux caused by astrocyte reactivity. Furthermore, brevican was introduced as a new relevant factor
that regulates CNS myelination. Future work on the complex crosstalk between the ECM,
astrocytes and myelination in the brain will yield a better understanding of the mechanistic cascade
connecting brevican to myelin regulation in health and disease. Finally, these results may open new
avenues for future relevant astrocyte-specific therapies that directly target remyelination.