R Douglas Fields
https://www.researchgate.net/publication/266624474_Myelination_of_the_Nervous_System_Mechanisms_and_Functions
Myelination of the Nervous System: Mechanisms and Functions
DOI: 10.1146/annurev-cellbio-100913-013101 · Source: PubMed
Myelination of the Nervous System: Mechanisms and Functions (PDF Download Available). Available from:
https://www.researchgate.net/publication/266624474_Myelination_of_the_Nervous_System_Mechanisms_and_Functions [accessed Feb 14 2018].
8
9
11
12
13
14
15
19
20
21
22
PLP
23
(Vallitsevan teorian kehittämisen pioneerit ja suuret nimet merkitty punaisella ja linkitetty, kriteerinä erityisesti, että R. D. Fields on kertonut tutkimuksista aiheen perusteoksissaan, HM, suomalaiset henkilöt ja tutkimukset merkitty sinisellä.)
26
28
31
She’s 17 years old and already helping patients. Meet the winner of one of the country’s most prestigious science fairs.
Indrani Das has been fascinated with brain injuries since her freshman year of high school, when she learned that their effects can be devastating and irreversible.
Later, her fascination evolved into a full-fledged research project. Das, now 17 and a senior at the Academy for Medical Science Technology in Hackensack, N.J., ex-plored how brain damage occurs, examining a process called astrogliosis, which can lead to the excess production of a toxin that can damage neurons. If she and other researchers could better understand how brain damage occurs, perhaps they could figure out how to slow or reverse the process.
“My work centers on repairing the behavior of supporting cells to prevent neuron in-jury and death,” Das said. “It was really that shock of what it can do to a person that pushed me to work” on research involving brain injuries.
Das’s project, which explores the role of brain cells called astrocytes in the death of neurons, was awarded the top prize and $250,000 at the Regeneron Science Talent Search.
Das bested thousands of high school scientists from across the country. The talent search selected 40 finalists, who traveled last week to D.C., where a selection com-mittee grilled them on their work and put them through the wringer, testing their grasp of scientific concepts and their ability to solve problems.
Four of the finalists were from the Washington area: Prathik Naidu of Thomas Jef-ferson High School for Science and Technology in Fairfax County; David Rekhtman of Walt Whitman High School in Bethesda; and Sambuddha Chattopadhyay and Rohan Dalvi, both of Montgomery Blair High School in Silver Spring.
Naidu placed seventh, taking home $70,000 for his project, which created 3-D models of the genetics of cancer cells, using a computer program he built.
The Science Talent Search, previously sponsored by Intel, is one of the best-known and among the most competitive science fairs for young researchers. This year, the talent search gave out $1.8 million to 40 finalists, much of which will go to cover college tuition for the budding researchers.
[These teens are working to cure cancer and solve the mysteries of the universe]
Maya Ajmera, president and chief executive of the Society for Science and the Pub-lic, said the finalists are often well-rounded and driven by their desire to make the world a better place, an altruism that is reflected in extracurricular activities.
Das, who hails from Orendell, N.J., recently became a certified emergency medical technician and is already working with patients, helping to transport them to hospitals. While she is deeply fascinated by research, she also hopes to become a practicing physician so she can work with patients.
“I would say my happiest time is when I’m with my patients,” Das said. “I love connecting with people and understanding how I can help them. It keeps me human.”
She plans to use the prize money to help pay for college and medical school.
Nonsynaptic junctions on myelinating glia promote preferential myelination of electrically active axons
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4439926/
Physiological Function of Microglia
Role of Myelin Plasticity in Oscillations and Synchrony of Neuronal Activity
https://science.howstuffworks.com/life/inside-the-mind/human-brain/remember-birth.htm
Can a person remember being born?
Think back to your earliest memory. Perhaps images of a birthday party or scenes from a family vacation come to mind. Now think about your age when that event oc-curred. Chances are that earliest recollection extends no further back than your third birthday. In fact, you can probably come up with only a handful of memories from be-tween the ages of 3 and 7, although family photo albums or other cues may trigger more.
Psychologists refer to this inability of most adults to remember events from early life, including their birth,as childhood amnesia.Sigmund Freud first coined the term infan- tile amnesia, now more broadly referred to as childhood amnesia,as early as 1899 to explain his adult patients' scarcity of childhood memories [source: Rapaport]. Freud proposed that people use it as a means of repressing traumatic, and often sexual, urgings during that time. To block those unconscious drives of the id, Freud claimed that humans create screen memories, or revised versions of events, to protect the conscious ego. "
Puutaheinää. Oikea selitys täällä:
CG: " More than a century later, researchers have yet to pin down a precise expla-nation for why childhood amnesia occurs. Only in the last 20 years have people in-vestigated children's,rather than adults',memory capabilities in search of the answer. This research has brought with it a new batch of questions about the nuances of young children's memory.
For a long time, the rationale behind childhood amnesia rested on the assumption that the memory-making parts of babies' brains were undeveloped. Then, around age 3, children's memory capabilities rapidly accelerate to adult levels. "
T.: Näinkin voi sanoa. Mutta kielen oppiminen nimenomaan luo myös ne aivojen ka-pasiteetit, joille ihmisen ajattelu perustuu. Tämän huomasi jo 1700-luvulla englanti-lainen psykiatri David Hartley alun perin fyysiä aivoja tutkimalla vastikään kehitetyllä mikroskoopilla.
CC: " However, psychologists have discovered that children as young as 3 months old and 6 months old can form long-term memories. "
T.: Yes. Suoria ehdollisia refleksejä. Niitä ei muisteta sellaisinaan, vaan muiden, kie-lellisten muistojen yhteydessä,joiden muodostumiseen ne vaikuttavat ollen alun perin näiden rakennuspalikkoina, jotka kootaan kielellisesti muistettaviksi muistoiksi.
https://science.howstuffworks.com/life/inside-the-mind/human-brain/remember-birth1.htm
... CC: " The difference comes in which memories stick around. For instance, it appears that babies are born with more intact implicit, or unconscious, memories. "
T.: Ehdottomia refleksejä EI pidä nimittää lainkaan "MUISTOIKSI" eikä tajunnaksi, koska ne eivät ole sitä. Suorat ehdolliset refleksit sellaisenaan eivät myöskään ole tajuntaa. Ne ovat sitä, millä esimerkiksi simpanssi käyttäytyy, ja mekin kävelemme ja juoksemme pystyssä ja pidämme myös polkupyörää ja moottoripyörää pystyssä ajaessamme. (Koska muuten tekoäly oppii ajamaan moottoripyörällä? Sitä ennen ei kannata puhua TIETOISESTA tekoälystä, sillä tuo tehtävä on HELPOMPI.)
CC: " At the same time the explicit, or episodic, memory that records specific events does not carry information over that three-year gap, explaining why people do not remember their births.
But why does this happen, and what changes take place in those first years? And if we can form memories as babies, why don't we retain them into adulthood?
On the next page, we'll take a closer look at a baby's brain to find out the answer.
Memory Encoding in Children
To form memories,humans must create synapses,or connections between brain cells (linkitys lähteen, T.), that encode sensory information from an event into our memory. From there, our brains organize that information into categories and link it to other similar data, which is called consolidation. In order for that memory to last, we must periodically retrieve these memories and retrace those initial synapses, reinforcing those connections.
Studies have largely refuted the long-held thinking that babies cannot encode infor-mation that forms the foundation of memories. For instance, in one experiment invol-ving 2- and 3-month-old infants, the babies' legs were attached by a ribbon to a mo-bile [source: Hayne]. By kicking their legs, the babies learned that the motion caused the mobile to move. Later, placed under the same mobile without the ribbon, the in-fants remembered to kick their legs. When the same experiment was performed with 6-month-olds, they picked up the kicking relationship much more quickly, indicating that their encoding ability must accelerate gradually with time, instead of in one significant burst around 3 years old.
This memory encoding could relate to a baby's development of the prefrontal cortex at the forehead. This area, which is active during the encoding and retrieval of expli-cit memories, is not fully functional at birth [source: Newcombe et al]. However, by 24 months, the number of synapses in the prefrontal cortex has reached adult levels [source: Bauer].
Also, the size of the hippocampus at the base of the brain steadily grows until your second or third year [source: Bauer]. This is important because the hippocampus determines what sensory information to transfer into long-term storage.
https://science.howstuffworks.com/life/inside-the-mind/human-brain/remember-birth2.htm
"... But what about implicit memory? Housed in the cerebellum, implicit memory is essential for newborns, allowing them to associate feelings of warmth and safety with the sound of their mother's voice and instinctively knowing how to feed. Confirming this early presence,studies have revealed few developmental changes in implicit me- mory as we age [source: Newcombe et al]. Even in many adult amnesia cases, impli-cit skills such as riding a bicycle or playing a piano often survive the brain trauma.
Now we know that babies have a strong implicit memory and can encode explicit ones as well, which indicates that childhood amnesia may stem from faulty explicit memory retrieval. Unless we're thinking specifically about a past event,it takes some sort of cue to prompt an explicit memory in all age groups [source: Bauer]. Up next, find out what those cues are. "
T.: Kyse ei ole vain "palautusvirheestä",vaan siitä, että ihmiselle ominainen kielellis-rakenteinen tajunta on vasta muodostumassa ympäristöstä riippumattomalla tavalla ajatuksellisesti mieleenpalautettavissa olevine muistoineen.
CC: " Language and Sense of Self in Memory-Making
Our earliest memories may remain blocked from our consciousness because we had no language skills at that time.A 2004 study traced the verbal development in 27- and 39-month old boys and girls as a measure of how well they could recall a past event. The researchers found that if the children didn't know the words to describe the event when it happened, they couldn't describe it later after learning the appropriate words [source: Simcock and Hayne].
Verbalizing our personal memories of events contributes to our autobiographical memories. These types of memories help to define our sense of self and relationship to people around us. Closely linked to this is the ability to recognize yourself. Some researchers have proposed that children do not develop self-recognition skills and a personal identity until 16 or 24 months [source: Fivush and Nelson].
In addition, we develop knowledge of our personal past when we begin to organize memories into a context.
Many preschool-age children can explain the different parts of an event in sequential order, such as what happened when they went to a circus. But it isn't until their fifth year that they can understand the ideas of time and the past and are able to place that trip to the circus on a mental time line [source: Fivush and Nelson].
Parents play a pivotal role in developing children's autobiographical memory as well. Research has shown that the way parents verbally recall memories with their small children correlates to those children's narrative style for retelling memories later in life. In other words, children whose parents tell them about past events, such as birth- day parties or trips to the zoo, in detail will be more likely to vividly describe their own memories [source: Urshwa]. Interestingly, autobiographical memory also has a cultu-ral component, with Westerners' personal memories focusing more on themselves and Easterners remembering themselves more in group contexts [source: Urshwa].
More detailed explanations exist regarding childhood amnesia. But brain structure, language and sense of self are its foundation. To learn more about amnesia and memory, don't forget to read the links below. ...
Primal Healing
Flying in the face of childhood amnesia research,some people claim to recall prever- bal memories and even recollections from the womb. One form of psychoanalysis, called primal healing, focuses on traumatic early memories similar to Sigmund Freud's theory of repressed and screen memories.Primal therapy links people's pre- sent pain with the pain of birth,taking patients back to the memory of their own birth in a process referred to as rebirthing. However,in spite of anecdotal evidence, no scien- tific study has verified the authenticity of these rebirthing experiences [source:Eisner]. "
Related HowStuffWorks Articles
Dr. Douglas Fields: "Exploring New Frontiers in Neuroscience" | Talks at Google
Nonsynaptic junctions on myelinating glia promote preferential myelination of electrically active axons
Abstract
The myelin sheath on vertebrate axons is critical for neural impulse transmission, but whether electrically active axons are preferentially myelinated by glial cells, and if so, whether axo-glial synapses are involved, are long-standing questions of significance to nervous system development, plasticity and disease. Here we show using an in vitro system that oligodendrocytes preferentially myelinate electrically active axons, but synapses from axons onto myelin-forming oligodendroglial cells are not required. Instead, vesicular release at nonsynaptic axo-glial junctions induces myelination. Axons releasing neurotransmitter from vesicles that accumulate in axon varicosities induces a local rise in cytoplasmic calcium in glial cell processes at these nonsynap-tic functional junctions, and this signalling stimulates local translation of myelin basic protein to initiate myelination.
The surprising discovery of synapses formed on glial progenitors, oligodendrocyte progenitor cells (OPCs, also called NG2 cells) has remained enigmatic for over a decade1.These cells mature to form myelin insulation on axons2,3, and several func- tions for synapses on OPCs have been proposed4.A leading hypothesis is that axon-OPC synapses may stimulate myelination selectively on electrically active axons to increase the speed of impulse transmission through electrically active neural circuits 5,6. This would have significant effects on neural circuit function. Since myelination continues in many brain regions through early life, preferential myelination of electri-cally active axons could enable environmental factors to modify neural circuit development according to functional experience7.
Synapses on OPCs could increase myelination in an activity-dependent manner in several ways,including promoting OPCs to differentiate into mature oligodendrocytes or by increasing OPC survival or proliferation. However,signals from axons must also regulate initiation of myelin wrapping even after OPCs have matured,because mature oligodendrocytes can be associated with axons early in development but not form myelin until much later in prenatal or adult life 8. It has been shown that vesicular release of glutamate from axons stimulates local translation of myelin basic protein (MBP) and stimulates myelin induction9.This signalling could be mediated by synap- tic transmission or by spillover of neurotransmitter from axo-glial synapses activating extrasynaptic glutamate receptors on OPC processes10,11.
Alternatively to synaptic transmission, other forms of axo-glial communication could signal electrical activity in axons to OPCs. Nonsynaptic release of neurotransmitter operates by both vesicular and non-vesicular release mechanisms. Neurotransmit-ters can be released in the absence of morphological synaptic contacts to activate neurotransmitter receptors on other cells (volume conduction)12. In contrast to synap-tic communication, which is a specialization for rapid (millisecond) and highly point-to-point localized signalling between axons and dendrites,volume transmission could be particularly well suited for communication between axons and myelinating glia13. Vesicle fusion is seen at axonal swellings (varicosities) that lack identifying features of a synapse. Notably missing are the close apposition of pre- and post-synaptic membranes, submembrane thickening caused by cytoskeletal proteins that organize neurotransmitter receptors and intracellular signalling molecules in the postsynaptic apparatus and the focused accumulation of synaptic vesicles docked at the presy-naptic membrane. Neurotransmitter signalling at axonal varicosities along nerve fib-res is characteristic of autonomic transmission in the enteric nervous system14 and cholinergic transmission in neocortex15, but most neurons have similar axon varico-sities. In addition to nonsynaptic vesicular release, neurotransmitters also can be released along axons through membrane channels16.
Another important question is if given a choice,will oligodendroglial cells preferential- ly myelinate electrically active axons? In addition, oligodendrocytes are multipolar cells but it is unknown how different branches of the same oligodencrocyte are instructed by axons to act autonomously and selectively synthesize myelin in those processes that are in contact with active axons. In the present experiments, calcium imaging, electron microscopy and electrophysiology were used to determine the involvement of axon-glia communication in myelination of electrically active axons in vitro. The results indicate a strong preference for oligodendrocytes to myelinate elect-rically active axons via a mechanism dependent on nonsynaptic vesicular release of glutamate but independent of synapses on OPCs.
Results
Preferential myelination of electrically active axons
The hypothesis that axons that are electrically active would be preferentially myelina-ted was tested by co-culturing OPCs with neurons that could release synaptic vesicles together with other neurons in which vesicular release was blocked. Dorsal root ganglion (DRG) neurons were used in these studies because they have several advantages. DRG neurons have no dendrites and thus they are ideal for studying oligodendrocyte interactions with axons. The long central axons of DRG neurons are myelinated by oligodendrocytes, and DRG neurons do not form synapses on them-selves (in vivo or in culture)17,18. DRG neurons do not fire action potentials sponta-neously, and they fire a single action potential in response to a brief electrical stimu-lus; thus, the firing frequency and pattern can be regulated precisely by electrical sti-mulation of neurons in cell culture.In these experiments,half of the neurons were trea- ted with the clostridial neurotoxin, botulinum A (BoNT/A) together with a blue dye to identify these axons.BoNT/A is a potent and highly selective enzyme that cleaves sy- naptosome-associated protein-25 (SNAP-25), the t-SNARE (Target membrane-asso-ciated soluble N-ethylmaleimide-sensitive factor attachment protein receptor) neces-sary for neurotransmitter release from synaptic vesicles. The other half of the neurons were untreated,providing OPCs a choice as to which axons to myelinate once under- going differentiation.After washing out the toxin,which continues to inhibit neurotrans- mitter release for at least 4 weeks19, OPCs were added to neuronal cultures contai-ning normal and BoNT/A-treated neurons to determine whether myelin formed prefe-rentially on axons that release synaptic vesicles in response to electrical stimulation (Fig. 1a,b; Supplementary Fig. 1b–d). Compact myelin was identified by immunocyto-chemistry for MBP 3 weeks after culturing OPCs on DRG axons.
(a) DRG neurons treated with BoNT/A and stained with cell tracker (blue, see red arrow) co-cultured with normal (untreated) neurons (grey, see yellow arrow).
(b) OPCs (green, GCaMP3) were added to the cultures to determine whether exocytosis of neurotransmitter from axons influenced myelination.
(c) Axons were stimulated for 9 s at 10 Hz every 5 min for 10 h and cultured for 3 weeks. Myelin (green, myelin basic protein, MBP) analysed 3 weeks after co-culture formed preferentially on axons releasing synaptic vesicles (purple, neurofilament), and
(d) number of myelin segments/cell were more in normal axons (P<0.001, n=7 cells from four dishes)
(e) myelin segments were also longer in normal axons (P<0.001, n=9 cells from four dishes). Scale bar, 10 μm (a,b); 20 μm (c).
Consistent with the hypothesis, the results showed that when given a choice, oligo-dendrocytes preferentially myelinated axons that could release synaptic vesicles (Fig. 1c). By far, the majority of myelin segments (10 times more) were found on con-trol axons compared with axons in the same culture in which vesicular release was inhibited (Fig. 1d) (P<0.001, t-test n=7 cells from four dishes). When myelin did form on axons that were unable to release synaptic vesicles, individual myelin segments were only 1/8 as long as normal (Fig. 1e, P<0.001, t-test, n=9 cells from four dishes).
Thus,oligodendroglia preferentially myelinate electrically active axons in these expe- riments by a mechanism dependent on exocytosis. This is consistent with the hypo-thesis that synaptic transmission between axons and OPCs promotes the initial events of myelination.However,release of neurotransmitter along axons can also take place in the absence of synapses, providing an alternative explanation for this result.
Axo-glial synaptic transmission in myelination
Preferential myelination of electrically active axons could result from synaptic com-munication stimulating OPC development; however, further experimental results did not support this. Supplementary Fig. 1a shows typical OPC morphology in co-culture for 24 h with DRG neurons. In our conditions, monocultured OPCs did not exhibit in-ward sodium currents until 2–3 days in vitro (d.i.v.) (0 of 8 cells at 0–1 d.i.v. 5 of 10 at 2–3 d.i.v.; Supplementary Fig.2b,c). Interestingly,we observed that contact with axons greatly accelerated the onset of inward sodium current expression in OPCs. Sodium currents were evident in 84% of 91 recorded cells from the first day of co-culture with DRG neurons, independently of the treatment (Fig. 2a,b; Supplementary Fig. 2e). No changes on I–V curves for sodium or potassium currents, input resistance or capaci-tance of OPCs were observed in co-cultures unstimulated and pre-stimulated electrically with or without BoNT/A (Supplementary Fig. 2d–g).
(a) Currents elicited in a recorded OPC held at −80 mV 1 day after plating by voltage steps from +20 to −110 mV. Note the presence of sodium currents (inset).
(b) I–V curve of sodium currents for the same OPC after leak subtraction (inset).
(c) Lack of evoked synaptic currents in an OPC upon DRG axonal stimulation. There was no response in 74 cells tested (unstimulated n=17 cells, from seven dishes; pre-stimulated electrically with (n=15 cells from six dishes) or without (n=42 from 17 dishes) BoNT/A). Individual (grey, 15 sweeps) and average traces (black) are shown. Axonal stimulation time is indicated with arrowheads.
(d) Bath application of the secretagogue ruthenium red (75 μM) in absence of electri-cal DRG axonal stimulation did not evoke any currents in the same OPC. Capacitive currents in response to a test pulse are shown (unstimulated n=10 cells from four dishes, pre-stimulated n=18 cells from seven dishes and pre-stimulated with BoNT/A n=7 cells from four dishes).
(e,f) Spontaneous (e) and miniature (f) synaptic currents in an OPC recorded at 15 postnatal days in acute coronal corpus callosum slices of NG2-DsRed mice (N=2 mice). Miniature synaptic currents were recorded in 1 μM tetrodotoxin (TTX) and 75 μM ruthenium red (Rred). The mean frequency, rise and decay times of spontaneous synaptic activity are 0.51 Hz, t10–90%=332 μs and τ=1.35 ms, respectively (insets, n=7 cells from seven different brain slices). The holding potential is indicated for each trace.
To test for the presence of synaptic currents in OPCs co-cultured with neurons, elect-rical stimulation (1 Hz;Supplementary Fig.2a) was applied through extracellular elect- rodes to depolarize DRG axons that traverse beneath a high-resistance barrier sepa-rating DRG cell bodies and axons in different compartments of three-compartment chambers (Campenot chambers).To test for the presence of evoked synaptic currents in OPCs cultured on axons in the central compartment, we recorded these cells in whole-cell configuration at a holding potential of −80 mV while stimulating DRG neu-rons through the extracellular electrodes. Importantly, evoked AMPAR (α-amino-3- hydroxy-5-methyl-4-isoxazole propionic acid receptor)-mediated currents in OPCs were never detected in co-cultures, regardless of whether OPCs were in contact with axons that were previously unstimulated (n=17), or axons that were pre-stimulated electrically with (n=15) or without (n=42) BoNT/A (Fig. 2c). To corroborate the lack of axon-OPC synapses, we analysed the possible existence of spontaneous synaptic currents in OPCs recorded before and during application of the secretagogue ruthe-nium red, which stimulates a massive release of synaptic vesicles. No spontaneous synaptic currents were detected in OPCs either in control conditions or in response to application of ruthenium red (unstimulated n=10; pre-stimulated n=18; pre-stimulated with BoNT/A n=7) (Fig.2d).It is noteworthy that most OPCs (80%) recorded during ex- tracellular stimulation and/or in ruthenium red had a membrane resistance >300 MΩ, sufficient for the detection of small or distant synaptic currents in these progenitors11 (SupplementaryFig.2f).In addition,confirming previous research,we find that AMPAR-mediated synaptic currents are common in recordings of OPCs in acute brain slices, using the same recording conditions that were used in co-cultures (Fig. 2e,f). Since myelination is promoted by electrical stimulation in co-cultures20,21, our results indi-cate that activity-dependent regulation of myelination does not require AMPAR-medi-ated synaptic communication between axons and OPCs. Therefore the hypothesis that synapses are required for formation of myelin on electrically active axons is not supported.
Nonsynaptic communication in activity-dependent myelination
Next,we wished to determine how OPCs could selectively myelinate axons that were electrically active in the absence of synapses from axons. DRG neurons do not form synapses in monoculture17,18; however, vesicle recycling in monoculture DRG neu-rons occurs along axons in response to electrical stimulation as shown by monitoring the fluorescent indicator FM 4–64 (Supplementary Fig.1e). Since treatment with botu- linum toxin strongly inhibited both vesicle recycling along axons induced by electrical stimulation9 and myelination independent of synaptic transmission (Figs 1 and 2), we hypothesized that nonsynaptic vesicular release of neurotransmitter along axons may stimulate myelination.
To test this possibility, we imaged cytoplasmic calcium responses in OPCs transfec-ted with the genetic calcium indicator GCaMP3 in response to electrical stimulation of axons. The results showed functional communication between axons and OPCs despite the absence of synaptic currents (Fig. 3). Ca2+ transients induced in OPCs by axonal action potentials were typically seen at points where OPC processes were associated with axons (Fig. 3, insets).
Ca2+ increases in processes of OPCs at axonal varicosities.
(a) OPC processes (green, GCaMP3 transfection) form specialized functional junc-tions with DRG axons (blue, neurofilament), containing accumulations of synaptic vesicles containing glutamate (red, vGluT2). Scale bar, 10 μm. After live-cell calcium imaging (b), the cultures were fixed and stained by immunocytochemistry to deter-mine whether that the calcium responses were associated with axo-glial contacts containing glutamatergic synaptic vesicles (inset in a). Note colocalization between axonal varicosity (red, vGluT2) and swellings in OPC process (green, GCaMP3).
Scale bar,2 μm. (b) Time-lapse series showing a local increase in Ca2+ in OPC-axon junctions in response to electrical stimulation of axons. A stimulus-induced, local in-crease in Ca2+ in the same axo-glial junction (yellow square) that is shown in a is re-corded in the OPC transfected with GCaMP3. Time since stimulus onset is shown in each frame (5.5–13 s). The fluorescence intensities at the axo-glial contact (yellow square) and the OPC cell body (yellow circle) are shown. Note the local increase in fluorescence intensity at the axo-glial junction after stimulation but this is not accompanied by an increase in fluorescence intensity in the soma. Scale bar, 5 μm.
The onset latency of electrically induced calcium rise in OPCs was much longer than synaptic transmission; >1 s. In response to pulsed stimulation (0.5 s at 10 Hz every 2 s),the average time to peak of the first response was 25±5.7 s (n=14 dishes),with mul- tiple subsequent calcium responses during stimulation sustained for 200 s (Fig. 4a, b). Application of BoNT/A significantly reduced the amplitude of electrically induced calcium response in OPCs (Fig. 4a,b). Selective inhibitors of neurotransmitter recep-tors indicate that these responses were mediated by both glutamatergic and puriner-gic signalling. A cocktail of glutamate receptor antagonists inhibiting NMDA (N- me-thyl- D-aspartate), mGluR and AMPA glutamate receptors (DAPV (D-(−)-2-amino-5- phosphonopentanoic acid),MCPG((RS)-α-methyl-4-carboxyphenylglycine) and CNQX (6-cyano-7-nitroquinoxaline-2,3-dione), respectively) reduced the amplitude of electrically evoked Ca2+ in OPCs by 80% (Fig. 4d). Suramin inhibition of purinergic receptors, or MCPG inhibition of mGluR receptors alone, reduced the amplitude of responses significantly (Fig. 4d). These results implicate vesicular release of ATP and glutamate from varicosities in signalling to OPC processes.
Ca2+ increase in OPC processes is mediated by glutamate and ATP released in response to action potentials in axons.
(a) Representative Ca2+ traces in OPC processes in response to electrical stimu-lation of axons without (black) and with (blue) BoNT/A treatment to block SNARE-dependent exocytosis.
(b) Averaged Ca2+ traces in OPCs from 14 dishes are shown. Responses were inhibited by BoNT/A treatment. **P=0.001, t-test, peak amplitude, n=14 cells from 14 dishes with no BoNT/A, n=14 cells from 14 dishes with BoNT/A.
(c) Number of Ca2+ responses were reduced significantly by BoNT/A treatment (blocked). The peak Ca2+ response within 200 s of axonal stimulation (10 Hz) were measured. P=0.004, t-test,n=14 cells from 14 dishes with no BoNT/A, n=14 cells from 14 dishes with BoNT/A.
(d) Summary graph shows per cent inhibition of Ca2+ responses following treatment with selective blockers of glutamatergic and purinergic neurotransmitter receptors, including a combination of CNQX, DAPV, MCPG (GluR antagonists), suramin and MCPG. The results implicate both glutamate and ATP neurotransmitter signalling. P=0.3, n=6 cells from six dishes, *1P=0.03, n=4 cells from four dishes, **P=0.008, n=5 cells from five dishes, *2P=0.02, n=6 cells from six dishes, all paired t-test.
Transmission electron microscopy showed specialized contacts (arrows) between OPC processes (OPC) and axon varicosities (Ax) containing intracellular vesicles (small arrows in a) and mitochondrion, but no synapses were detected. Three examples for each condition are shown. Insets (right column) show these junctions at higher magnification. Such contacts were evident in cultures stimulated for 9 s at 10 Hz every 5 min for 10 h (Stim) before plating OPCs, and unstimulated cultures (b), but such contacts were not found in stimulated cultures treated with BoNT/A before adding OPCs (c). Scale bar, 1 μm.
OPC processes preferentially contact electrically active axons releasing synaptic vesicles and form nonsynaptic axo-glial junctions.
(a,b) Images showing OPCs transfected with GCaMP3 construct (green) and vGluT2 immunocytochemistry (red) to identify glutamate-containing vesicles in axons. OPCs were plated on axons either previously treated (b) or not treated (a) with BoNT/A. (a,b) Magnified views of glutamate-containing vesicles and OPC processes. Yellow arrows indicate vGluT2 stained puncta in close opposition to GCaMP3 processes.
(c) The number of vGluT2 puncta >0.5 μm in diameter was reduced significantly on axons previously treated with BoNT/A (P=0.0005, t-test, n=7 cells from seven dishes with no BoNT/A, n=8 cells from eight dishes with BoNT/A).
(d,e) Images showing OPC transfected with GCaMP3 construct (green) and immuno-cytochemistry for neurofilament (blue, an axon marker). (f) Summary graph showing that the fraction of all OPC processes (d,e) in individual OPCs forming parallel asso-ciations with axons was significantly reduced when axons were previously treated with BoNT/A (P=0.0002,t-test, n=7 cells from seven dishes with no BoNT/A, n=8 cells from eight dishes with BoNT/A). Scale bars, 10 μm (a, b,d,e); 2 μm (a,b).
Interestingly, OPC processes were less closely associated with axons in which exocytosis was inhibited by BoNT/A treatment. Rather than adhering to the axon and tracking along it, OPC processes frequently intersected and crossed over axons rather than running parallel together with the axon (Fig. 6d–f; P=0.0002, t-test, n=15 cells from 15 dishes).
Formation of specialized junctions between axons and oligodendrocytes, which were described as ‘spot welds' in the earliest electron microscopic study of central nervous system myelination,is the first event in myelin formation23.Importantly,these junctions shown here in cell culture and previously in vivo23,24 lack ultrastructural specializa-tions characteristic of synaptic junctions. In line with these anatomical observations, our findings show that specialized nonsynaptic axon-OPC junctions are functional and signal via vesicular release from axon varicosities that induces Ca2+ increases in OPCs and may serve to promote myelination.
Autonomous oligodendrocyte processes in activity-dependent myelination
Oligodendrocytes are multipolar cells that can myelinate up to 50 different axons. Axonal factors are known to induce myelination, but it is unclear whether individual cell processes of an oligodendrocyte can act autonomously to myelinate different axons, or whether oligodendrocyte commitment to myelin induction is a cell-wide event. Local translation of myelin basic protein is stimulated in OPCs by glutamate released from axons acting on mGluR and NMDA receptors on oligodendrocytes9 but it is unknown whether this occurs selectively in those processes of an OPC that are in contact with electrically active axons. To answer this question, OPCs were transfected with kikume MBP-3′-untranslated region, a photo-activated fluorophore that enables identification of newly synthesized protein by a change in fluorescence from red to green after photoactivation. Local translation of MBP was found to be pre-ferentially induced in OPC processes in contact with electrically active axons compa-red with OPC processes from the same cell in contact with axons in which vesicular release had been inhibited by prior treatment with BoNT/A (Fig 7). Thus, different cell processes of an oligodendrocyte act independently in myelin induction and thus are able to compartmentalize signals.We conclude that one of the axonal factors determi- ning whether different branches of an oligodendrocyte myelinate an axon is the activity-dependent exocytosis of vesicles containing neurotransmitter along the fibre.
Local translation of MBP occurs preferentially on OPC processes contacting electrically active axons releasing synaptic vesicles.
(a) Newly synthesized MBP (green). Axons were stimulated electrically at 10 Hz for 10 min and local translation of MBP was monitored using kikume MBP fluorescence after 40 min. White arrows indicate new MBP translation in contrast yellow arrow showing no new MBP translation.
(b) Total MBP (red).
(c) Bright field showing cell morphology of co-culture.
(d) Axons treated previously with BoNT/A to block synaptic vesicle release (blue).
(e) Combined images from a–d.
Differential interference contrast and fluorescence to identify axons treated (blue, yel-low arrow) and not treated with BoNT/A. Yellow arrow shows axon and OPC interac-tion without MBP translation (notice yellow arrow in a with lack of new MBP green puncta indicating no new MBP translation).
(f) Quantification shows significantly more local translation of MBP in OPC processes that were in contact with axons releasing synaptic vesicles (grey axons) (0.11±0.024 versus 0.028±0.011 puncta per micrometre length of axon not treated or previously treated with BoNT/A respectively, P<0.001, t-test, n=24 cells from non-blue, n=23 cells from blue axons, five dishes). Scale bar, 10 μm.
Discussion
The experiments provide several novel findings: When provided a choice, OPCs pre-ferentially myelinate electrically active axons releasing synaptic vesicles. This activi-ty-dependent communication between axons and OPCs is mediated by nonsynaptic intercellular junctions that signal through intracellular calcium. Myelin forms normally and is stimulated by electrical activity in axons in culture even though there is no evi-dence of synapses by electrophysiological and electron microscopic analysis. Indivi-dual cell processes of oligodendrocytes act autonomously, initiating myelination in response to electrical activity in axons in contact with the oligodendroglial process, which stimulates local translation of MBP in that subcellular domain.
In conclusion, myelin is formed preferentially on electrically active axons releasing vesicles, but synapses on OPCs are not required for activity-dependent stimulation of myelination. Local calcium signalling produced by vesicular release of glutamate is well suited to local subcellular control of myelin induction in individual OPC proces-ses, by activating glutamate receptors localized in the fine processes closely associ-ated with axons9. ATP release can also occur from synapses and mediate discrete signals, but purinergic signalling appears to be more relevant to modulating differen-tiation and proliferation of OPCs, in part because these signalling pathways activate global intracellular calcium responses in the cell9. This is possibly associated with wider spread release of ATP from axons through volume-regulated anion channels 16, and differences in localization of glutamate and purinergic receptors in the cell.
Preferential myelin induction on electrically active axons would have profound effects on circuit function by the resulting increased conduction velocity,and thus provide another mechanism of plasticity complementing synaptic plasticity7. In human brain imaging studies, white matter structure is affected by learning25, and recent studies show that social isolation impairs myelin formation in the forebrain of mice 26,27; the present findings could provide a cellular mechanism participating in the effects of en-vironmental experience on myelination. These new findings may also have implica-tions for disease, including psychiatric illness and impaired remyelination after conduction block in multiple sclerosis.
Methods
Cell culture
All experiments were conducted in accordance with animal study protocols approved by the NICHD Animal Care and Use Committee. DRG neurons were dissected from the spinal cords of embryonic day-13.5 mice as described9 Neurons were grown for ∼ 4 weeks in MEM media supplemented with N3 containing 100 ng ml−1 of nerve growth factor and 5% heat-inactivated horse serum in the side compartments of three-compartment chambers equipped with stimulating electrodes28 or on cover-slips that were coated with poly-L-lysine and collagen. Mitosis of non-neuronal cells was inhibited by a 4-day treatment with 13 μg ml−1 fluoro-2′-deoxyuridine beginning 1 day after plating. These cultures can be maintained indefinitely with half-volume changes of media every 3 days. Primary cultures of OPCs were obtained from cereb-ral cortices of embryonic day-19 P2 rats and plated into 75-cm2 tissue culture flasks.
The resulting cultures were maintained at 5% CO2 and 37 °C in media containing 10% fetal bovine serum (Life Technologies, Carlsbad,CA, USA). After 11 days in cul- ture, the flasks were shaken at 37 °C for 1 h to kill non-glial cells and remove micro-glia, then the media was changed and the flasks were shaken overnight to lift OPCs from the flask.To enrich for OPCs,the cell suspension was pelleted,resuspended and incubated in an uncoated culture dish for 30 min.Contaminating cells, primarily endo- thelial cells, astrocytes,macrophages and microglia adhere strongly to the plastic and can be separated out by this panning method. For myelinated co-cultures, purified OPCs (>90% OPCs) were counted and plated on 3–4-week-old DRG cultures at a density of 40,000 cells per side compartment.After 3–4 h,the media was removed and replaced with N1 differentiating media. Action potentials were induced in DRG axons by a 200-μs 5-V biphasic pulse through platinum electrodes in three-compartment chambers28. Axons growing into the central compartment beneath the high-resistant barriers are stimulated but cells in the lateral compartments are not depolarized. For 3-week co-culturing myelination experiments, stimulation was applied in 0.5 s bursts at 10 Hz every 2 and 15 s bursts at 10 Hz every 5 min, using biphasic pulses (200-μs 6 V, square-wave pulse) for 5 h (ref. 9). For Ca2+ imaging, field stimulation (15 s at 10 Hz, 30–50 V) was delivered through two 1-cm-long horizontal platinum electrodes in cultures of DRG neurons grown on 25-mm coverslips.
Intracellular Ca2+ imaging
OPCs transfected with GCaMP3 were imaged during electrical stimulation using a confocal microscope equipped with a 40 × (1.3 numerical aperture, NA) objective lens, excitation at 488 nm by scanning laser, and emission light filtered through a HQ528/50 splitter/filter. Quantification of images was performed using Image J software (NIH).
Vesicle recycling
Recycling synaptic vesicles in DRG axons were labelled with FM 4–64 (Life Techno-logies). To stain the total pool of recycling vesicles, axons were loaded by electrical stimulation (10 Hz,30 V, 5 ms, biphasic square wave) for 20 s in media containing 2.5 μM FM 4–64 dye. The dye was allowed to remain on the cells for 60 s after cessation of the stimulus to permit complete compensatory endocytosis, and was subsequently removed during a 7-min period with eight complete solution changes.
Pharmacology
Drugs were added directly to the culture dish before experiments. Botulinum toxin A, kindly provided by E. A. Johnson, was added to cultures at a final concentration of 3 nM for at least 18 h before experiments to block vesicular release19. Immunoblotting confirmed cleavage of SNAP-25 in DRG neurons treated with BoNT/A. Previous stu-dies show that block of neurosecretion from DRG neurons in cell culture occurs with-in 4 h of BoNT/A and last at least 4 weeks19. MCPG (500 μM), DAPV (50 μM), CNQX (20 μM), and suramin (50 μM) were obtained from Tocris (Ellisville, MO, USA). Cell Tracker Blue CMAC Dye is from Life Technologies (NY, USA).
Confocal microscopy
All images were acquired on a Zeiss 510 NLO confocal microscope (Carl Zeiss MicroImaging, Inc. Thornwood, NY,USA) equipped with both 40 × (1.3 NA) and 63 × (1.4 NA) oil-immersion lenses using appropriate laser lines and excitation/emission filters. For live-cell imaging, coverslips were mounted in an imaging chamber and continuously superfused with sterile-filtered saline. Photoactivation of Kikume was performed with ultraviolet lamp exposure for >10 min.
Antibodies
The antibodies used were as follows: vGluT2 (1:1,000, Cat. AB2251, Millipore, Bille-rica, MA, USA), neurofilament (1:5,000, Cat. NFH, Aves Labs, Tigard, Oregon), MBP (1:1,000, Cat. SMI-99, Sternberger Monoclonals), synaptophysin 1 (1:1,000, Cat. 101 011, Synaptic Systems, Gottingen, Germany), Olig2 (1:250,Cat. 18953, IBL Co., LTD, Japan) and NG2 chondroitin sulfate proteoglycan (1:500, Cat. MAB5384, Chemicon International, Billerica, MA, USA).
Electron microscopy
Cell cultures were fixed with 2% paraformaldehyde/2% glutaraldehyde in 0.13 M sodium cacodylate buffer, pH 7.4 at 37 °C and postfixed in 2% OsO4 for 2 h at 4 °C. Samples were dehydrated in a graded series of alcohol, infiltrated in propylene oxide / Spurr's and embedded in Spurr/s resin. Ultrathin sections were cut with a diamond knife, stained with uranyl acetate and lead citrate and examined by transmission electron microscopy.
Electrophysiological recordings
Patch-clamp recordings of OPCs from monocultures or DRG-OPC co-cultures were performed in voltage-clamp mode 1–3 days after plating at room temperature and using an extracellular solution containing (in mM):126 NaCl, 2.5 KCl, 1.25 NaH2PO4, 26 NaHCO3, 20 glucose, 5 pyruvate, 2 CaCl2 and 1 MgCl2 (95% O2 and 5% CO2). The intracellular solution contained (in mM): 120 K-gluconate, 5 NaCl, 3 MgCl2, 0.2 EGTA, 10 HEPES, 0.3 Na-GTP,4 Na-ATP and 10 Na-phosphocreatine (pH≈7.3, 295 mOsm). To investigate the presence of fast inward AMPAR-mediated synaptic currents in OPCs of DRG-OPC co-cultures, DRG axons were stimulated while OPCs were recorded in voltage-clamp mode at a holding potential of −80 mV in co-cultures previously unstimulated and pre-stimulated electrically with or without BoNT/A. Hol-ding potentials were corrected by a junction potential of −10 mV. Extracellular stimu-lation of DRG axons was performed using a bipolar electrode placed in a side com-partment (0.5–5 V, 1–10 ms of duration). Single pulses or train of stimuli at a rate of 1, 10 and 100 Hz were applied for each neuron (Supplementary Fig. 2a).
Recordings were made without series resistance compensation. Series resistances were monitored during recordings and cells showing a change of >30% were discar-ded. Whole-cell recordings were obtained using Multiclamp 200B amplifier (Molecu-lar Devices),filtered at 5 kHz and digitized at 10 kHz.Digitized data were analysed off-line using pClamp 10.1 software (Molecular Devices). Current densities for steady-state I–V relationships of OPCs were obtained by dividing the measured current am-plitudes by cell capacitance. The amplitudes of inward sodium currents were mea-sured at different voltage steps from +20 to −110 mV after leak subtraction using a MATLAB custom routine (MATLAB, MathWorks, Inc). The ohmic leak current for each potential was estimated by scaling the responses at most hyperpolarizing pulses. Means and data distributions of steady-state and inward sodium current densities, membrane resistance and capacitance were not different among treatments (P>0.05) (Supplementary Fig. 2d–g).
Statistical analysis
Statistical significance was tested by analysis of variance in experimental designs in-volving multiple groups, followed by Dunnet's post hoc test for evaluating differences with respect to a control group and Fischer's comparison test for designs comparing differences among all groups. Two-sided Student's t-test was used for analysis of experimental designs with two groups, and a paired t-test was used in experiments in which repeated measures were made in the same sample. Data are displayed as means and s.e.m. For electrophysiological data, each data group was first subject to D'Agostino and Pearson normality test. According to the data structure (non-normally or normally distributed), the Kruskal–Wallis or two-way analysis of variance test was used for comparisons. Multiple Kolmogorov–Smirnov test was used to compare distributions of data. All statistical analysis and plotting were performed with Minitab version 12 (State College, PA, USA), SigmaPlot and GraphPad Prism 5.00 software (GraphPad Software Inc., USA).
Additional information
How to cite this article: Wake, H. et al. Nonsynaptic junctions on myelinating glia promote preferential myelination of electrically active axons. Nat. Commun. 6:7844 doi: 10.1038/ncomms8844 (2015).
Supplementary Material
Acknowledgments
We thank Andrea Helo for her help on building the MATLAB custom routine used for leak subtraction. This work was supported by NICHD funds for intramural research. M.C.A. is part of the ENP-Ile-de-France network and was supported by grants from Agence Nationale de la Recherche (ANR, R14193KK), Fondation pour l'aide à la recherche sur la Sclérose en Plaques (ARSEP) and IDEX-Cité Paris Sorbonne. FCO is supported by a post-doc fellowship from Fondation pour la Recherche Médicale (FRM).
Footnotes
Author contributions All authors contributed to the data analysis,experimental plan-ning and writing the paper. R.D.F. ideated and directed the project. H.W., D.H.W. and P.R.L. performed the non-electrophysiological experiments and F.C.O. performed the electrophysiological recordings.
References
- Bergles D. E., Roberts J. D., Somogyi P. & Jahr C. E. Glutamatergic synapses on oligodendrocyte precursor cells in the hippocampus. Nature 405,187–191 (2000). [PubMed] [Google Scholar]
- Emery B. Regulation of oligodendrocyte differentiation and myelination. Science 330, 779–782 (2010). [PubMed] [Google Scholar]
- Fields R. D. Chapter in Neuroglia 3rd edn eds Kettenmann H., Ransom B. R. 573 – 585Oxford Univ. Press (2013). [Google Scholar]
- Mangin J. M. & Gallo V. The curious case of NG2 cells: transient trend or game changer? ASN Neuro 3, e00052 (2011). [PMC free article] [PubMed] [Google Scholar]
- Ziskin J. L., Nishiyama A., Rubio M., Fukaya M. & Bergles D. E. Vesicular release of glutamate from unmyelinated axons in white matter. Nat.Neurosci.10,321 – 330 (2007). [PMC free article] [PubMed] [Google Scholar]
- Kukley M.,Capetillo-Zarate E.&Dietrich D.Vesicular glutamate release from axons in white matter. Nat. Neurosci. 10, 311–320 (2007). [PubMed] [Google Scholar]
- Fields R. D. Neuroscience. Change in the brain's white matter. Science 330, 768 – 769 (2010). [PMC free article] [PubMed] [Google Scholar]
- Back S. a., Luo N. L., Borenstein N. S., Volpe J. J. & Kinney H. C. Arrested oligo-dendrocyte lineage progression during human cerebral white matter develop-ment: dissociation between the timing of progenitor differentiation and myelinoge-nesis. J. Neuropath. Exp. Neurol. 61, 197–211 (2002). [PubMed] [Google Scholar]
- Wake H., Lee P. R. & Fields R. D. Control of local protein synthesis and initial events in myelination by action potentials. Science 333, 1647–1651 (2011). [PMC free article] [PubMed] [Google Scholar]
- Maldonado P. P. & Angulo M. C. Multiple modes of communication between neurons and oligodendrocyte precursor cells. Neuroscientist 21, 266–276 (2015). [PubMed] [Google Scholar]
- Velez-Fort M., Maldonado P. P., Butt A. M., Audinat E. & Angulo M. C. Postnatal switch from synaptic to extrasynaptic transmission between interneurons and NG2 cells. J. Neurosci. 30, 6921–6929 (2010). [PMC free article] [PubMed] [Google Scholar]
- Vargova L. & Sykova E. Extracellular space diffusion and extrasynaptic transmission. Physiol. Res. 57, S89–S99 (2008). [PubMed] [Google Scholar]
- Fields R. D. Volume transmission in activity-dependent regulation of myelinating glia. Neurochem. Int. 45, 503–509 (2004). [PubMed] [Google Scholar]
- Burnstock G. Non-synaptic transmission at autonomic neuroeffector junctions. Neurochem. Int. 52, 14–25 (2008). [PubMed] [Google Scholar]
- Zhang Z. W., Kang J. I. & Vaucher E. Axonal varicosity density as an index of local neuronal interactions. PLoS ONE 6, e22543 (2011). [PMC free article] [PubMed] [Google Scholar]
- Fields R. D. & Ni Y. Nonsynaptic communication through ATP release from volume-activated anion channels in axons. Sci. Signal. 3, ra73 (2010). [PMC free article] [PubMed] [Google Scholar]
- Ransom B. R., Neale E., Henkart M., Bullock P. N. & Nelson P. G. Mouse spinal cord in cell culture. I. Morphology and intrinsic neuronal electrophysiologic properties. J. Neurophysiol. 40, 1132–1150 (1977). [PubMed] [Google Scholar]
- Ransom B. R., Christian C. N., Bullock P. N. & Nelson P. G. Mouse spinal cord in cell culture. II. Synaptic activity and circuit behavior. J. Neurophysiol. 40, 1151 – 1162 (1977). [PubMed] [Google Scholar]
- Welch M. J., Purkiss J. R. & Foster K. A. Sensitivity of embryonic rat dorsal root ganglia neurons to Clostridium botulinum neurotoxins. Toxicon 38, 245–258 (2000). [PubMed] [Google Scholar]
- Stevens B., Porta S., Haak L.L., Gallo V. & Fields R. D. Adenosine: a neuron-glial transmitter promoting myelination in the CNS in response to action potentials. Neuron 36, 855–868 (2002). [PMC free article] [PubMed] [Google Scholar]
- Ishibashi T. et al. Astrocytes promote myelination in response to electrical impul-ses. Neuron 49, 823–832 (2006). [PMC free article] [PubMed] [Google Scholar]
- Peters A. & Palay S. L. The morphology of synapses. J. Neurocytol. 25, 687–700 (1996). [PubMed] [Google Scholar]
- Luse S. A. The fine structure of the morphogenesis of myelin. Prog. Neurobiol. 4, 59–95 (1959). [PubMed] [Google Scholar]
- Alix J.J.,Dolphin A.C. & Fern R. Vesicular apparatus, including functional calcium channels, are present in developing rodent optic nerve axons and are required for normal node of Ranvier formation. J. Physiol. 586, 4069–4089 (2008). [PMC free article] [PubMed] [Google Scholar]
- Zatorre R. J., Fields R. D. & Johansen-Berg H. Plasticity in gray and white: neuro-imaging changes in brain structure during learning. Nat. Neurosci. 15, 528–536 (2012). [PMC free article] [PubMed] [Google Scholar]
- Makinodan M., Rosen K. M., Ito S. & Corfas G. A critical period for social experi-ence–dependent oligodendrocyte maturation and myelination. Science 337, 1357 – 1360 (2010). [PMC free article] [PubMed] [Google Scholar]
- Liu J. et al. Impaired adult myelination in the prefrontal cortex of socially isolated mice. Nat. Neurosci. 15, 1621–1623 (2012). [PMC free article] [PubMed] [Google Scholar]
- Fields R. D., Yu C., Neale E. A. & Nelson P. G. In Electrophysiological Methods for In Vitro Studies in Vertebrate Neurobiology eds Kettenmann H., Grantyn R. 67–76Alan R. Liss (1992). [Google Scholar]
Tässä on pääpaino mekanismin muodostumisella mm. geneettisesti eikäm eikä sen toiminnalla:
" The Brain Prize 2023
Pioneering work on molecular mechanisms of brain development and plasticity awarded with the world’s top prize in neuroscience
Professors Michael Greenberg, Christine Holt and Erin Schuman have revolutionized our understanding of how neurons regulate the thousands of different proteins – the building blocks of life, that are needed to support brain development, plasticity and maintenance. They have revealed crucial molecular mechanisms that sustain the development and function of the healthy brain and also provided key insights into the causes of neurodevelopmental and neurodegenerative diseases.
Copenhagen, Denmark - The Lundbeck Foundation has announced the recipients of The Brain Prize 2023, the world’s largest award for outstanding contributions to neuroscience. This year’s award recognizes the pioneering work of three leading neuroscientists - Professor Michael Greenberg at Harvard Medical School, Professor Christine Holt at University of Cambridge, and Professor Erin Schuman at the Max Planck Institute for Brain Research.
The Brain Prize 2023 worth DKK 10 million (€1.3 million) is awarded to:
- Christine Holt (UK)
- Michael Greenberg (USA)
- Erin Schuman (Germany)
A profound aspect of our nervous system is that during development and adulthood our brains are subject to extensive change, known as neural plasticity. Such plasticity requires that the complement of neural proteins - the neural proteome, be dynamically regulated in space and time. An international group of three neuroscientists, Michael Greenberg, Christine Holt, and Erin Schuman have each revealed the fundamental principles of how this is mediated at the molecular level – from activity-dependent gene transcription to the local translation of mRNA into new proteins in dendrites and growing axons.
Their findings have provided spectacular new insights into the cellular and molecular mechanisms that guide growing axons during brain development, and that enable the developing and adult brain to be shaped by experience. Theirs is a beautiful discovery story in fundamental neuroscience that also provides clues to the aetiology of neurodevelopmental and neurodegenerative diseases of the brain. For their work, the three neuroscientists are awarded the world’s largest prize for brain research – The Brain Prize.
Professor Richard Morris, Chair of The Brain Prize Selection Committee explains the reasoning behind this year’s award.
The Brain Prize winners of 2023, Michael Greenberg, Christine Holt, and Erin Schuman have revealed the fundamental principles of how this enigmatic feature of brain function is mediated at the molecular level. Together, the Brain Prize 2023 winners have made ground-breaking discoveries by showing how the synthesis of new proteins is triggered in different neuronal compartments, thereby guiding brain development and plasticity in ways that impact our behavior for a lifetime.’’
“On behalf of the Lundbeck Foundation, I am delighted that The Brain Prize 2023 is awarded to these three outstanding neurobiologists,” said Lene Skole, CEO of the Lundbeck Foundation.
“Their pioneering research has broken new ground and provided deep insights into the molecular mechanisms of neural development and plasticity. Their work also provides vital new insights into the causes and mechanisms of some of the most devastating disorders of the brain. The awarding of this year’s Brain Prize is thoroughly well-deserved.”
About the Brain Prize
The Brain Prize is the world’s largest neuroscience research prize, and it is awarded each year by the Lundbeck Foundation. The Brain Prize recognises highly original and influential advances in any area of brain research, from basic neuroscience to applied clinical research. Recipients of The Brain Prize may be of any nationality and work in any country in the world. Since it was first awarded in 2011 The Brain Prize has been awarded to 44 scientists from 9 countries.
Brain Prize recipients are presented with their award by His Royal Highness, The Crown Prince of Denmark, at a ceremony in the Danish capital, Copenhagen.
About the Lundbeck Foundation
The Lundbeck Foundation is an enterprise foundation encompassing a comprehensive range of commercial and philanthropic activities – all united by its strong purpose; Bringing Discoveries to Lives. The Foundation is the long-term and engaged owner of several international healthcare companies – Lundbeck, Falck , ALK, and Ferrosan Medical Devices – and an active investor in business, science and people through its commercial investments in the financial markets; in biotech companies based on Danish research and through philanthropic grants to science talents and programmes in Danish universities. The Foundation’s philanthropic grants amount to more than DKK 500m annually primarily focusing on the brain – including the world’s largest personal prize awarded in neuroscience, The Brain Prize. "
***
https://www.biorxiv.org/content/10.1101/2024.06.02.591488v2.full.pdf
" 1
A myelinic channel system for motor-driven organelle transport
Katie J. Chapple 1 , Tabitha R.F. Green 1, Sarah Wirth 1 , Yi-Hsin Chen 1 , Ulrike Gerwig 2, Marie
Louise Aicher 2 , Yeonsu Kim 2 , Lina Komarek 1&2 , Angus Brown 3 , Colin L. Crawford 1,
Rebecca Sherrard Smith 1 , Jeff Lee 4 , Luis Pardo-Fernandez 1, Rebecca E McHugh1 , Celia
M. Kassmann2 , Hauke B. Werner 2 , Ilan Davis 4, Matthias Kneussel 5 , Euan R Brown6, Sandra Goebbels 2, Klaus-Armin Nave** 2, Julia M. Edgar* 1&2
1 School of Infection & Immunity, College of Medical Veterinary and Life Sciences, University of Glasgow, Glasgow, G12 8TA
2 Dept of Neurogenetics, Max Planck Institute for Multidisciplinary Sciences City Campus, Hermann-Rein-Strasse 3, D-37075 Goettingen, Germany
3 School of Life Sciences, University of Nottingham, University Park, Nottingham NG7 2UH
4. School of Molecular Biosciences, College of Medical Veterinary and Life Sciences, University of Glasgow, Glasgow, G12 8TA
5 Department of Molecular Neurogenetics, Center for Molecular Neurobiology, ZMNH, University Medical Center Hamburg-Eppendorf, 20251, Hamburg, Germany
6 School of Engineering and Physical Sciences, Institute of Biological Chemistry, Biophysics and Bioengineering, Heriot Watt University, Edinburgh EH14 4AS, UK.
*Correspondence: [email protected]
**Correspondence: [email protected]
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted August 26, 2024.;
https://doi.org/10.1101/2024.06.02.591488doi:bioRxiv pr
2
Summary
Myelin sheaths comprise compacted layers of oligodendroglial membrane wrapped spirally around axons. Each sheath, if imagined unwrapped, has a cytoplasm-filled space at its perimeter, linking it to the oligodendrocyte soma via a short process. By electron microscopy (EM), this space, which we term the ‘myelinic channel system’ contains microtubules and membranous organelles, but whether these are remnants of development or serve a function is unknown. Performing live imaging of myelinating oligodendrocytes expressing fluorescent reporters, we found that the myelinic channel
system serves microtubule-dependent organelle transport. Further, the intra-myelinic movement of peroxisomes was modulated by neuronal electrical activity in these mixed neural cell cultures. Loss of oligodendroglial Kif21b or CNP in vivo led to apparent stasis of myelin organelles and secondary axon pathology. This suggests that oligodendrocytes require motor transport in myelin to maintain axonal integrity.
Keywords: microtubule-dependent transport; live imaging; ageing; neurodegeneration
3
Introduction
In the nervous system, the unusual size and architecture of neurons with long axons requires an intracellular transport system for moving cargo between the somata and distant axonal terminals. Axonal transport has been studied in detail since the 1960s (Allen et al., 1982; Austin et al., 1966; Brady et al., 1982; Grafstein, 1967; Lasek, 1967; Livett et al., 1968; Lubinska, 1964; Ochs et al., 1967; Weiss and Holland, 1967), including its molecular mechanism involving motor protein transport on microtubular tracts (Brady, 1985; Cason and Holzbaur, 2022; Hirokawa et al., 2010; Vale et al., 1985; Vargas et al., 2022; Zahavi and Hoogenraad, 2021). Here we show that a related logistical challenge exists at a much smaller scale with respect to the vital interaction of process-bearing oligodendrocytes and the axons they myelinate. Compacted myelin decreases the capacitance and increases the electrical resistance across the myelinated fibre (Cohen et al., 2020; Hartline and Colman, 2007). However, in so doing, myelin separates spatially the oligodendrocyte soma from the internodal axon-glial junction (Fig. 1A), a 20 nm wide synapse-like gap (Stys, 2011) thought to mediate the bi-cellular exchange of metabolites through monocarboxylate transporters (Fünfschilling et al., 2012; Lee et al., 2012).
In the central nervous system (CNS), each myelin sheath derives from a lamellipod that extends from the oligodendrocyte soma. Growth of the prospective sheath involves actin remodelling (reviewed in Brown and Macklin, 2020) at the lamellipod’s leading edge (future ‘inner tongue’), which advances in a myelin basic protein- (MBP-) dependent fashion around the axon, underneath accumulating numbers of membrane wraps (Nawaz et al., 2015; Snaidero et al., 2014; Zuchero et al., 2015). Simultaneously, the lamellipod grows along the axon’s length (Snaidero et al., 2014) in response to calcium signalling in the oligodendrocyte (Baraban et al., 2018; Krasnow et al., 2018; Iyer et al., 2024), requiring nucleation of microtubules (Fu et al., 2019).
Membrane compaction occurs by fusion of opposing intracellular surfaces (Aggarwal et al., 2011; Harauz et al., 2009) combined with tethering of opposing extracellular surfaces of the wrapping lamellipod (García-Mateo et al., 2018), culminating in tightly wrapped concentric layers of membrane. A cytoplasm-filled space remains at the perimeter of the sheath, if imagined unwrapped, suggesting a direct connection to the intracellular contents of the cell soma (Butt and Ransom, 1989) via the cell process (Peters, 1960; Sternberger et al., 1978) (Fig. 1A).
This cytoplasm filled space forms the ‘inner and outer tongues’ at the innermost and outermost edges of the wrapped sheath, and arrays of ‘paranodal loops’ at either end (Fig. 1A). The latter are tightly tethered to the axon by septate-like junctions (Bhat, 2003). Other cytoplasm-filled spaces in CNS myelin include transient openings of the compacted myelin
(Snaidero et al., 2017; Velumian et al., 2011), rare Schmidt-Lanterman incisure-like openings on large fibres (Blakemore, 1969) and other conformations not yet fully defined (Edgar et al., 2021; Weruaga-Prieto et al., 1996). Myelin’s cytoplasm-filled spaces are generated and maintained, at least in part, by 2/ ,3/ -cyclic nucleotide 3/ -phosphodiesterase (CNP), which limits the membrane ‘compacting’ function of MBP (Snaidero et al., 2017; Trapp et al., 1988).
This myelinic channel system contain microtubules, endoplasmic reticulum, vesicle-like structures, multivesicular bodies, peroxisomes and Golgi outposts (Edgar and Griffiths, 5 2013; Frühbeis et al., 2020; Nakamura et al., 2021; Richert et al., 2014; Stassart et al., 2018). Since all these compartments have been implicated in developmental myelination, it remains unknown whether they exist as remnants of myelinogenesis, captured in the mature sheath, or play a functional role. We previously speculated that in the adult, the myelinic channel system provides a route for soluble oligodendroglial metabolites to reach the glial-axonal junction (Edgar et al., 2009; Meschkat et al., 2020; Nave, 2010a; Saab et al., 2013). Furthermore, since active transport processes can be visualised in cytosolic spaces of flattened myelin-like sheets in oligodendrocyte monolayer cultures (Ainger et al., 1993; Carson et al., 1997; Kachar et al., 1986; Herbert et al., 2017, Song et al., 2003), we additionally hypothesized that myelinic channels might serve as a route for the movement of membranous cargo to the glial-axonal junction, analogous to axonal transport to synaptic terminals.
Here we demonstrate microtubule-dependent organelle transport in established myelin sheaths and show that ablation from oligodendrocytes of kinesin 21B (Kif21b), a processive motor and modulator of microtubule dynamics (Ghiretti et al., 2016; Muhia et al., 2016; van Riel et al., 2017), leads to late onset secondary axon degeneration in vivo.
This coincides with reduced levels of monocarboxylate transporter 1 (MCT1) in biochemically enriched myelin. Consistent with this, CNP-deficient mice, that lack myelinic channel integrity (Snaidero et al., 2017) and exhibit axonal degeneration (Edgar et al., 2009; Lappe-Siefke et al., 2003), and similarly in aged mice with structural myelin defects, the level of MCT1 in myelin is markedly diminished. As MCT1 resides at the glial-axonal junction and is thought to mediate the transfer of monocarboxylates between 6 oligodendrocytes and axons (Fünfschilling et al., 2012; Lee et al., 2012), our data suggest myelinic transport is vital for oligodendrocyte-mediated axon support.
7
Results
The myelinic channel system and its organelle content.
To define a putative transport infrastructure in the myelinic channel system (yellow in Fig.1A), we used transmission electron microscopy (EM) of adult mouse optic nerve cross sections. As reviewed previously (Edgar and Griffiths, 2013), some paranodal loops of myelin contained microtubules (Fig. 1B), which were also observed in inner and outer tongues (Supplementary Fig. 1A and B).
Multivesicular bodies (Fig. 1C and see Frühbeis et al., 2020) and single membrane structures resembling peroxisomes (Fig. 1D) were found in inner and outer tongues, paranodal loops, and very rarely, in cytosolic pockets of otherwise compact myelin (Supplementary Fig. 1C). Some membrane-bound structures in inner tongues (Fig. 1E) and paranodal loops had a pale core similar to the axoplasm, suggesting these are invading axonal sprouts (Steyer et al., 2023). Of the 2348 optic nerve axon cross-sections examined, 0.82 % (± 0.45 SD; n = 4 optic nerves of 12-month-old
mice) of inner tongue processes contained structures resembling peroxisomes or multi-vesicular bodies. Based on (i) EM tissue sections being ~50 nm thick, (ii) the average optic nerve myelin sheath length being ~76 μm (Young et al., 2013) and the assumption that (iii) organelles are distributed randomly and (iv) spherical in shape, being ~0.3 μm in diameter (Fig. 1C and D), our data suggest that the inner tongue of each optic nerve myelin sheath contains ~2 such organelles (as paranodal loops are encountered only infrequently in EM images, we were unable to independently determine their organelle content using this method). In addition to myelin organelles, we observed invaginations of glial membrane into the axon (Fig. 1F and Supplementary Fig. 1D-F), similar to previous reports in peripheral nerves (Spencer and Thomas, 1974). Very rarely, the myelin membrane even formed fusion-like profiles with the peri-axonal space (Supplementary 8
Fig. 1G), suggesting direct exchange of materials between the two compartments.
As glutamate, which is released from electrically active axons (Micu et al., 2016), enhances the release of exosomes from oligodendrocytes in vitro (Frühbeis et al., 2020), we next asked if axonal electrical activity influences the localisation of organelles to myelin’s inner tongue, which faces the axon (Fig. 1A). To address this, we used an ex vivo optic nerve preparation in which we could precisely control action potential firing using stimulating and recording suction electrodes (Supplementary Fig. 1H), as described (Stys et al., 1991). One of each pair of adult mouse optic nerves was unstimulated while the contralateral nerve was stimulated to fire constitutively at 7 or 50 Hz for 20 minutes. By EM of nerve cross sections, the stimulated nerve had more inner tongues with organelles (average 0.65% ± 0.28 SD: ~1.7 organelles per sheath) compared to the unstimulated contralateral nerve (average 0.32 % of fibres ± 0.34 SD; ~0.8 organelles per sheath),
although with 5 nerve pairs analysed, the differences were not significant (Supplementary Fig. 1I-K). While organelle numbers could also decrease due to fusion, these data led us to hypothesize that organelle transport to or from the inner tongue might be modulated by axonal electrical activity.
Next, we focussed on peroxisomes using a transgenic mouse in which oligodendroglial peroxisomes are labelled with the photoconvertible fluorescent protein mEOS2 (Richert et al., 2014). To demarcate the myelinic channel system, we additionally labelled a proportion of oligodendrocytes with a transgenically-expressed, tamoxifen-inducible fluorescent protein tdTomato. In longitudinal sections of optic nerve, mEOS2-labelled peroxisomes were observed in the oligodendrocyte soma; at paranodes, as indicated by axonal 9 expression of Caspr1; and in internodal myelin (Fig. 1G; Supplementary Fig. 1M). This was confirmed in spinal cord (Supplementary Fig. 1L). However, in the densely packed white matter it was difficult to resolve individual myelin sheaths, even with sparse labelling.
Therefore, for subsequent observations, we used a murine spinal cord-derived myelin-forming co-culture of neurons and glia (“myelinating cell culture”), in which axons are enwrapped in compact myelin (Bijland et al., 2019; Thomson et al., 2008). In nascent sheaths, Mbp mRNA, which is translated locally in myelin (Wake et al., 2011) could be observed in the CNP +ve cytoplasm filled spaces that are eventually remodelled to form the myelinic channel system (Fig. 1H; Supplementary Video 1). In established sheaths, identified by intensely tdTomato positive paranodal regions, oligodendrocyte peroxisomes were observed in tdTomato rendered somata, processes, and sheaths (Fig. 1I), including inner and outer tongues (Fig. 1J and K; Supplementary Fig. 1N) and 48.2 % (± 7.4 SD) of paranodes examined (Fig. 1L; Supplementary Fig. 1O).
11
A. Schematic depiction of an oligodendrocyte (OL) soma and two myelin sheaths.
Cytoplasm (yellow) fills the myelinic channel system including inner and outer tongues (IT,OT) that entwine the myelin sheath on its inner and outer surfaces, respectively, and the paranodal loops that are tethered to the axon at the paranode. Myelin organelles are represented in green. Two dashed lines depict approximately where the cross sections in the corresponding EM images are located.
B-F. Micrographs of myelinated axon cross-sections including paranodal loops (B, D), the inner tongue (C, E), and an invagination of myelin membrane into the axon (F). In B, the white arrows indicate microtubules in paranodal loops, which encircle the axon as indicated by magenta arrows.
In C, a multivesicular body occupies the inner tongue and
in D, a dense organelle resembling a peroxisome resides in a paranodal loop.
In E, three membrane-bound structures reside in myelin’s inner tongue. These have a pale core (asterisks), similar to the axoplasm, suggesting they are axonal sprouts.
In F, an invagination of the glial cell membrane (arrows) surrounds an abnormal-appearing axonal organelle (asterisk), as shown previously in the PNS (Spencer and Thomas, 1974).
G. Maximum intensity projection (MIP) from a confocal z-stack of a longitudinal section of the optic nerve in which oligodendrocytes and myelin are labelled with tdTomato (pseudo coloured yellow), oligodendrocyte peroxisomes are labelled with mEOS2 and paranodes are immunostained with anti-Caspr1, which is located on the axon. Peroxisomes are observed in somata (in this case, obscured by tdTomato) and myelin sheaths (yellow arrows), including paranodes (magenta arrows). A higher magnification view of individual channels in shown in Supplementary Fig. 1M. are shown Some peroxisomes reside in non-tdTomato rendered cells. Inset: two Caspr1 stained paranodes separated by a node of Ranvier. By inference, the two peroxisomes are located in paranodal and internodal
myelin, respectively.
H. MIP from a confocal z-stack of an early myelinating oligodendrocyte in vitro in which
Mbp mRNA is labelled using single molecule fluorescence in-situ hybridisation (smFISH)
and CNP is labelled using immunocytochemistry. Mbp mRNA can be observed in the cell
body (OL), processes and nascent myelin sheaths. The cytoplasm filled spaces that stain
with anti-CNP are remodelled over time, giving rise to the myelinic channels system
depicted in the other images in this figure.
I. MIP of a myelinating oligodendrocyte in vitro labelled with tdTomato and anti-Ermin, a
cytoskeletal molecule that first appears during late myelination stages (Brockschnieder et
al., 2006). The soma is marked with an asterisk and a paranode with P. mEOS2-labelled
peroxisomes can be observed in the processes (white arrowheads) and myelin sheaths
(green arrows).
J. MIP of a second tdTomato-rendered oligodendrocyte in vitro showing peroxisomes
(green arrows) in myelin sheaths including in the triangular shaped thickening at the
junction of the process and the sheath delineated by the dotted line. The individual
channels are shown in Supplementary Fig. 1N.
K. High magnification view of sheath delineated by dotted line in I. Peroxisomes reside in
both outer and inner tongues (OT and IT).
12
L. Paranodal region from an oligodendrocyte in vitro labelled with anti-CNP (myelin) and anti-Caspr1. Two mEOS2-labelled peroxisomes are visible; by inference, the lower one is located just inside the paranodal region, marked by axonal Caspr1 expression, and the other at the juxtaparanode, between the internode and the paranode.
Myelinic organelle transport is microtubule dependent.
The optic nerve stimulation experiment suggests that myelin organelles are motile. To explore this further, we used live imaging of mEOS2 + oligodendrocyte peroxisomes in myelinating cell cultures in which oligodendrocytes wrap axons with appropriately thick compact myelin sheaths (Edgar et al., 2021; Thomson et al., 2008). As in other mammalian cell types (Wali et al., 2016) most myelin peroxisomes were apparently stationery or underwent only short-distance, Brownian-like motion. However, within a 10-minute imaging window, 16.2% (± 4.07 SD) of myelin peroxisomes exhibited movement (traversing distances ≧ 3 μm; Fig. 2A and B) and covering distances up to ~70 µm. Travel was punctuated by periods of stasis (Fig. 2A), such that the overall velocity over 10 minutes was approximately 10× less than when motile; the latter being 0.18 µm/sec (± 0.02 SD), on average. Some peroxisomes travelled predominantly in one direction (Fig. 2A-C), while others reversed direction (indicated by cyan circles in Supplementary Fig. 2A) and/or crossed paths with another (indicated by orange crosses in Supplementary Fig. 2A). Often, peroxisomes in the outer tongue (where it could be determined definitively) travelled towards the paranode (Fig. 2B and C; Supplementary Videos 2 and 3). In Supplementary Videos 4 and 5 a peroxisome travels to the paranodal loops of myelin via the presumptive outer tongue, traverses 4 paranodal loops and travels back in the direction of the cell body via the presumptive inner tongue.
In other eukaryotic cell types, peroxisomes are transported on microtubules (Rapp et al., 1996 and reviewed in Covill-Cooke et al., 2021). To determine if this is also true in the myelinic channel system, we treated myelinating cell cultures with nocodazole, which interferes with the polymerization of microtubules (Kesarwani et al., 2020). As axons also contain microtubules, we aimed to minimise confounding effects on myelinic transport (caused by axon dysfunction) by using a concentration and incubation time that did not
markedly alter the microtubule network in neurons. In these cell dense cultures, 3-hours in 20 μM nocodazole led only to minor alterations in anti-β-tubulin 3 staining (Supplementary Fig. 2B), indicating that neuronal microtubules were only partially disrupted over this time period. Furthermore, detyrosinated tubulin, a non-cell type specific marker of stable microtubules, appeared similar to controls (Supplementary Fig. 2B).
Using this protocol, nocodazole caused a significant reduction in the percentage of motile myelinic peroxisomes (Fig. 2D) without altering their velocity during motile phases (Fig. 2E). Nonetheless, the average velocity of individual organelles (including motile and stationary phases) within the 10-minute imaging period, shifted towards slower speeds (Fig. 2F). Consequently, the maximum distance travelled by a single peroxisome in 10 minutes in this experimental series was reduced from 55.0 μm in controls to 23.8 μm in nocodazole-treated cells, likely reflecting the drug’s perturbation of microtubular tracks. Indeed, in DMSO-treated control cultures, β-tubulin IV (which is specific in the CNS to oligodendroglia) stained continuous structures within myelinic channels in 94 % (± 2.0 SD) of tdTomato positive cells (Fig. 2G, Supplementary Fig. 2B, Supplementary Video 6; n = 33 cells; 2 independent cell cultures), but in only 45 % (± 7.0 SD) of tdTomato positive cells after nocodazole-treatment (Fig. 2H, Supplementary Fig. 2B; n = 17 cells; 2 independent cell cultures). Notably, in both control and nocodazole-treated cultures, the amount of polymerised β-tubulin IV varied considerably from sheath to sheath and from cell to cell. In summary, myelinic channels enable an active, microtubule-dependent transport of peroxisomes, and presumably other membrane-bound cargo.
Fig. 2. Myelinic peroxisomes transport is microtubule-dependent.
A. Overview of two tdTomato rendered oligodendrocytes (somata are labelled with asterisks) from which the timelapse images in B and C were acquired.
B and C. The schematics on the left illustrate the totality of the movement of a single myelin peroxisome in each of the timelapse movies depicted on the right. White lines represent cytoplasm-filled spaces within (solid lines) and outside (dashed lines) the focal plane. In B, the timelapse images show a motile peroxisome (white arrows) move down the presumptive outer tongue towards the paranodal loops (see overview), slowing or stalling at 30, 110 and 160 second time points. At 140 seconds, the organelle moves out of the focal plane and the path (small white arrow) is inferred from the movie. Most peroxisomes (coloured arrows) were stationery. In C, paranodal regions of adjacent sheaths, are separated by a node of Ranvier (see A). A motile peroxisome (white arrow) moves down the outer tongue process (determined from the relationship to the node of Ranvier) and traverses the presumptive paranodal loop nearest the node of Ranvier, before moving out of the focal plane. Other peroxisomes (coloured arrows) were stationary.
D. Nocodazole (20 μM, 3-hour incubation) reduced the percentage of myelin peroxisomes that were motile during the 10-minute period. Coloured symbols represent individual oligodendrocytes (n = 22 and 17, control and nocodazole, respectively), black symbols represent experimental medians (n = 3 each control and nocodazole); bars indicate mean of the median values ± SD. An unpaired Student’s t test was used to compare the experimental means. ** p < 0.01.
E. Nocodazole appeared not to alter the velocity of individual myelin peroxisomes when motile. Coloured symbols each represent one motile peroxisome (n = 28 and 29, control and nocodazole respectively) and black symbols represent experimental medians (n = 2 each, control and nocodazole); bars indicate the mean of the experimental median values ± SD.
F. Nocodazole caused a shift to slower average velocities, reflecting motile and stationary phases, of individual peroxisomes during the 10-minute imaging period, likely reflecting a reduction in available transport tracks.
G-I. β−tubulin IV staining of tdTomato labelled oligodendrocytes, following extraction of free tubulins (see also Supplementary Video 6), treated with DMSO (vehicle; G) or nocodazole (20 μM, 3 hours; H and I). White arrows indicate where processes from the oligodendrocyte attach to the myelin sheath and magenta arrows indicate regions of continuous β−tubulin IV staining. The sheath delineated by a white rectangle in G is shown at higher magnification in the inset (individual channels are shown in Supplementary Fig. 2B). The sheath in I is from a different cell from that in H (individual channels are shown in Supplementary Fig. 2B).
16
Myelin microtubules harbour post-translational modifications
Our data suggested that some myelin microtubules are more resistant to nocodazole than others, suggesting that myelin contains mixed populations of “dynamic” and “stable” microtubules (Jansen et al., 2023). Microtubule stability is encoded in post-translational modifications (PTMs) of tubulins, and the presence of differentially modified tubulins is crucial in the assembly, disassembly and rearrangement of the microtubule cytoskeleton (Borys et al., 2020; Janke and Bulinski, 2011). Recently, genetic sensors were developed that identify tyrosinated microtubules (AlaY1; Kesarwani et al., 2020), which are generally considered to be rather labile; or stable microtubules (StableMARK; Jansen et al., 2023), which are generally acetylated and/or detyrosinated. We observed both genetic sensors in oligodendrocyte cell bodies, processes, and myelin sheaths (Fig. 3A and B). However, as StableMARK can by itself stabilise microtubules when overexpressed, and as AlaY1 also labels free tubulins, we next used antibody staining, following extraction of free tubulins (Jansen et al., 2023), to identify PTMs on microtubule polymers. Anti-tyrosinated tubulin partially labelled one or more myelin sheaths in 83% of all tdTomato positive oligodendrocytes (n = 12 cells; 3 independent cell cultures) (Fig. 3C and D; Supplementary Fig. 3). Anti-detyrosinated tubulin was not observed in myelin sheaths (Fig. 3E; Supplementary Fig. 3). Acetylated tubulin was observed in some sheaths (Fig. 3F and G; Supplementary Fig. 3), as reported in vivo previously (Kusch et al., 2017; Werner et al., 2007). Because the antibodies to tyrosinated and acetylated tubulin are raised in the same species, we were unable to determine if different PTMs co-exist in the same sheaths.
A and B. Nine hours after oligodendrocyte transfection, genetic sensors labelling tyrosinated tubulins (Ty tub; A) and stable microtubules (stable MTs; B) in myelin, were visible. In A, arrowheads point to continuous-appearing structures suggesting they are microtubule polymers. In B, the cell body labelling is saturated, and processes are intensely labelled therefore a higher magnification view of an associated myelin sheath (delineated by the broken line), is shown on the right, stained with anti-CNP and illustrated for clarity in the schematic. The abnormal appearance of the somata and processes is likely because StableMARK itself stabilises microtubules if overexpressed.
C-F. Following extraction of free tubulins, immunocytochemistry was used to identify PTMs on myelin microtubules. Anti-tyrosinated tubulin labelled some myelin sheaths (magenta arrows) of tdTomato labelled cells, and the cell bodies (asterisk) and processes of other cell types (C). In general, anti-tyrosinated tubulin only partially labelled sheaths (magenta arrows; D). Anti-detyrosinated tubulin did not colocalise with tdTomato positive sheaths but labelled other cell processes (yellow arrows). Where magenta and yellow coincide, the tubulin stain does not follow the contours of the sheath, suggesting it is the processes of overlapping cells (E). Anti-acetylated tubulin labelled a few myelin sheaths (magenta arrows) although it was prevalent in other cellular processes (yellow arrows). The oligodendrocyte soma is labelled with an asterisk in the overview image on the left (F).
Anti-acetylated tubulin only partially labelled sheaths (G). Individual channels from C, D, E
and F are shown in Supplementary Fig. 3, for clarity.
Together these data demonstrate tubular tracks within the myelinic channel system are required for active translocation of organelles. Interrogation of our databases of purified myelin proteins, identified by label-free mass spectrometry (Gargareta et al., 2022; Jahn et al., 2020), provided evidence for the presence in myelin of molecular motor proteins, including dynein and kinesin (Supplementary Table 1A). Tubulins, microtubule binding proteins and tubulin modifiers were also identified in biochemically isolated ‘myelin’ (Supplementary Table 1B).
Modulators of neuronal activity influence the motility of myelin peroxisomes Myelin-associated peroxisomes are essential for the long-term integrity of axons, particularly faster spiking axons, which are most vulnerable to peroxisomal defects (Kassmann et al., 2007). Furthermore, axo-myelin glutamatergic signalling and axonal electrical activity trigger neuroprotective exosome release and oligodendroglial metabolic support (Frühbeis et al., 2020; Saab et al., 2016). As a simple cellular model to examine the response of myelin-associated peroxisomes to changes in neuronal firing, we used myelinating cell cultures in which neuronal electrical activity and neuronal energy consumption can be enhanced with the GABA receptor blocker picrotoxin (PTX; 100 μM) or diminished with tetrodotoxin (TTX; 1 μM) (Fig. 4A and B).
Live imaging of oligodendrocyte processes over a 10-minute timeframe (beginning 5-minutes post drug application) showed that modulating neuronal activity had no effect on retrograde or anterograde transport of peroxisomes between the myelin sheath and the oligodendrocyte soma (Supplementary Fig. 4A, Supplementary Video 7). This concurs with our observation that peroxisomes entered the ‘myelin sheath’ even when formed on inert fibres (Fig. 4C). Nonetheless, in the presence of neurons, myelin-associated peroxisomes responded to PTX or TTX by becoming more or less motile, respectively (Fig. 4D), demonstrating that neuronal activity influences myelinic transport processes locally.
Despite this, the number of peroxisomes at paranodes, where myelin is tethered to the axon, did not change in response to changes in neuronal activity, at least in the short term (Supplementary Fig. 4B).
Fig. 4. Neuronal activity locally modulates the movement of myelin peroxisomes.
A. In myelinating cell cultures, PTX or TTX enhance or block neuronal activity, respectively. Graphs show firing frequency of individual neurons recorded in current clamp at baseline (T0) and 5 minutes (T5) after the application of drug. A paired Student’s t test was used to compare pre- and post-drug values. Results are from 9 and 5 independent cell cultures, respectively **** p = <0.0001; ** p = <0.01.
B. In neuron-enriched cultures, oxygen consumption rate (OCR) and extracellular acidification rate (ECR) changed in response to PTX or TTX indicating a general increase and decrease, respectively, in energy consumption. Symbols represent the mean of 3 independent experiments ± SD.
C. Widefield epifluorescence micrograph of peroxisomes in the MBP positive myelin- like structures formed on inert fibres, demonstrating that neuronal activity is not necessary for peroxisomes to enter the sheath during myelinogenesis.
D. Graphs showing the proportion of myelin peroxisomes per cell that are motile during a 10-minute imaging period. Coloured dots represent individual cells and black dots represent experimental medians. Bars represent mean of experimental medians ± SD. PTX or TTX significantly increased or decreased, respectively, the percentage of myelin peroxisomes that were motile during a 10-minute imaging period. An unpaired Student’s t test was used to compare pre- and post-drug values between the median values of the independent experiments. Results are from 5 or 6 independent cell cultures. *p = <0.05.
Deleting Kif21b from oligodendrocytes leads to secondary axon loss.
In the CNS, β-tubulin IV is specific to oligodendrocytes and localises to the somata, processes and myelin (Edgar et al., 2021; Terada et al., 2005 and Fig. 3 this manuscript), leading us to hypothesize that β-tubulin IV could play a role in cargo transport, sheath formation and axon maintenance. However, targeting the Tubb4 gene selectively in oligodendrocytes in Tubb4flox/flox::Cnp1+/Cre mice did not lead to overt changes in myelinated CNS fibres, even at 14 months of age, except that the g-ratio tended to be slightly higher than in controls (Supplementary Fig.5). Thus, Tubb4 function is likely
compensated for by other β-tubulins.
Next, we turned to a motor protein, Kif21b that we identified as a putative candidate for a function in the myelinic channel system because polymorphisms in KIF21B have been associated with increased susceptibility to multiple sclerosis (MS). Moreover, KIF21B is upregulated ~10 fold in MS white matter irrespective of genotype (Goris et al., 2010; Kreft et al., 2014) and expression of Kif21b is differentially regulated in different populations of mature oligodendrocytes in spinal cord of experimental autoimmune encephalomyelitis (Zheng et al., 2023). Kif21b is both a plus end directed progressive motor protein and a modulator of microtubule dynamics (Ghiretti et al., 2016; Hooikaas et al., 2020; Muhia et al., 2016; van Riel et al., 2017) that in neurons, is enriched in dendrites and growth cones at neurite tips (Labonté et al., 2014; Marszalek et al., 1999). To determine if Kif21b is expressed in oligodendrocytes in the optic nerve, we compared by western blotting optic nerve lysates from Kif21b conditional knockout (cKif21b KO) mice (Cnp1 Cre/+ ::Kif21bflox/flox) with two controls (Cnp1+/+::Kif21bflox/flox and Cnp1+/Cre::Kif21b+/flox or+/+). Whilst Cnp1Cre can recombine ectopically (Genoud et al., 2002), retinal ganglion cells (whose axons traverse the optic nerve) do not recombine, and in the optic nerve, all recombined cells are oligodendrocytes (Saab et al., 2016; Späte et al., 2024). Kif21b was not detected by western blotting of whole optic nerve lysates from cKif21b KO mice but was readily detectable in control optic nerves (Fig. 5A), demonstrating in this white matter tract, Kif21b is predominantly expressed by Cnp1-expressing oligodendrocytes.
To determine if ablating Kif21b from oligodendrocytes had consequences for the myelinated axon, we examined optic nerve cross sections by EM (Fig. 5B-D). Across all genotypes, most myelinated axons appeared morphologically normal at 1 year of age (Fig. 5Bi; 5C i,iii and 5D i, ii). However, in cKif21b KO mice, inner tongues often appeared distended and contained membranous material (Fig. 5D iii, iv) and occasionally otherwise healthy-appearing axons contained mitochondria lacking normal-appearing cristae (Fig. 5D iv). Accumulations of axonal organelles (magenta arrows, Fig. 5D vi), indicative of axonal
transport stasis (Edgar et al., 2004), were observed in cKif21b KO mice from 6 months of age (Fig. 5E). Importantly, these pathological features were seen also at 4 months of age but largely “lost” by age 9-12 months; the decline in absolute axon number by 9-12 months suggesting these pathologically affected axons had degenerated at this age (Fig. 5E).
While similar pathological axonal changes were seen occasionally also in heterozygous Cnp1+/Cre controls at 9-12 months of age, there was no comparable loss of optic nerve axons suggesting that Kif21b-dependent transport in oligodendrocytes (and/or microtubule destabilisation) is required for
long-term axonal integrity.
Fig. 5. Depletion of Kif21b from oligodendrocytes leads to secondary axon loss.
A. In the optic nerve, a pure white matter tract lacking neuronal cell bodies and dendrites, Cnp1Cre-dependent deletion of Kif21b led to depletion of Kif21b in whole nerve lysate, demonstrating this motor protein is expressed in oligodendrocytes
B. Electron micrographs of optic nerve axons from 9-12 month Cnp1+/+::Kif21bflx/flx controls.
Note the exceptional finding of periaxonal electron dense material (blue outline, Bii).
C. In Cnp1Cre/+ control optic nerves, myelinated axons appeared largely normal.
Periaxonal electron dense organelles in Ci (white arrows) are also seen in WT mice (Fig. 1).
Rarely, axons appeared electron dense and degenerating (asterisk Cii) or myelinic channels were abnormally extended (blue outline in Civ).
D. In Kif21b conditional KO nerve (cKif21b KO), most myelinated axons appeared morphologically normal (Di and ii). Myelinic organelles were observed in myelin’s inner tongue (Di) and paranodal loops (Dii) as in WT (Fig. 1). In Di, the yellow arrow points to a microtubule. However, myelinic channels were occasionally locally enlarged, impinging on the axon (delineated in blue in Diii). Redundant myelin and/or abnormal electron dense material were observed in some swollen inner tongues (delineated in blue in Div). Extracellular myelin whorls (blue outlines in Dv and vi) were not more frequent than in controls. Rarely, axons contained abnormal-appearing mitochondria (blue arrow on the left in Dv in comparison to the two mitochondria indicated by blue arrows on the right). Axonal organelle accumulations were also a feature of mutant mice (magenta arrows in Dvii.)
E. Quantitation of axonal organelle accumulations and total axon numbers in cKif21b KO optic nerves and controls at age 6-7 (left) and 9-12 months (right). Data were analysed by one-way ANOVA followed by Tukey’s multiple comparisons test. * p < 0.05; ** p < 0.01. A: axon; P: paranode.
A role for CNP1 and Kif21b in maintaining monocarboxylate transporter levels
We hypothesized that axonal integrity is impaired when metabolic support by oligodendrocytes is compromised. Transport processes in myelinic channels could contribute to overall myelin protein turnover, including that of MCT1 at the glial-axon junction. To determine if oligodendroglial Kif21b depletion impacts specific protein levels in myelin, we performed western blotting of spinal cord myelin from mice at postnatal day 60 (P60) and 1 year. As expected, Kif21b itself was strongly reduced in purified myelin from cKif21b KO mice (Fig. 6A), and to a lesser degree in total spinal cord homogenate containing both white and grey matter (Supplementary Fig.6). However, the Cnp1Cre knock-in allele (resulting in reduced Cnp1 gene dosage) was likely to contribute to these effects as CNP acts as a strut that maintains the integrity of myelinic channels (Snaidero et al., 2017) and Cnp1 heterozygosity leads to myelin dysfunction at advanced age (Hagemeyer et al., 2012).
Indeed, in myelin from 1 year old Cnp1+/Cre littermate controls, Kif21b levels became remarkably variable (Fig. 6A). For comparison, myelin oligodendrocyte glycoprotein (MOG), a transmembrane protein on the outer myelin layer, was unaltered in cKif21b KO mice compared to either control. In contrast, levels of MCT1, which is required at the glial-25 axonal junction for the transfer of pyruvate/lactate to the periaxonal space (Fünfschilling et al., 2012; Lee et al., 2012) were significantly reduced in 1 year old mutants, but not yet in their P60 counterparts (Fig. 6A). Indeed, MCT1 was also reduced in 1 year old Cnp1+/Cre controls, confirming that both Kif21b and CNP are required for maintaining normal MCT1 levels at the glial-axon junction. Importantly, there was no significant change in the levels of β-tubulin IV in purified myelin from cKif21b KO mice, suggesting there was no lack of
tubular tracks. We thus hypothesize that the loss of myelinic channel integrity (with reduced CNP1 and absence of oligodendroglial Kif21b) causes a slow attrition in delivery of MCT1 to the glial-axonal junction.
The role of CNP1 in organelle transport through myelinic channels
To test our hypothesis further, we compared myelin composition and myelinated fibre morphology in heterozygous (Cnp1+Cre) and homozygous (Cnp1-/- ) mutants. In optic nerves from Cnp1 heterozygous mice, at older age (14 months), membranous structures (arrows) and dense cytoplasm (asterisks) were occasionally observed in enlarged inner tongues (Fig. 6B), resembling changes in the aged CNS (Peters and Kemper, 2012). In homozygous Cnp1 mutants, with myelinic channel collapse (Snaidero et al., 2017), axonal changes in myelinated fibres begins as early as P5 (Edgar et al., 2009) and likely underlie premature death, which occurs before one year of age (Lappe-Siefke et al., 2003). When analysing homozygous Cnp1 mutant mice by western blotting, we found that MCT1 was strongly reduced in a myelin-enriched brain fraction by at least P75 while transcription of the encoding gene (Slc16a1) was similar in mutant and control brains (Fig. 6C). This suggests the absence of CNP affects the delivery of MCT1 to the glial-axonal junction and/or its turnover. In support of impaired intra-myelinic transport, we found by EM of optic nerve, that in some fibres in homozygous Cnp1 mutants, organelles appeared trapped within the myelinic channel system (Fig. 6D).
Fig. 6. Monocarboxylate transporter 1 is reduced when myelinic transport is
impaired.
A. Western blots of spinal cord myelin-enriched fractions of P60 and 1 year old mice.
Levels of Kif21b were reduced in Kif21b conditional knockout (cKif21b KO) mice, albeit that the reduction was not significant at P60 due to intra-group variation. Unexpectedly, levels in the Cnp1Cre/+ controls (Cre cont) were highly variable. Levels of CNP1 were significantly reduced in both the Cnp1Cre/+ and cKif21b KO mice compared to the Cnp1+/+ ::Kif21bflx/flx (WT) control, at P60 and 1 year. Levels of MOG were similar in all three genotypes, at both ages. At P60, MCT1 levels were similar in all three genotypes, but at 1 year of age, levels were significantly reduced in the Cnp1Cre/+ mouse compared to
WT control and in the cKif21b KO compared to both WT and Cnp1Cre/+ controls. Data were analysed using a one-way ANOVA followed by Tukey’s multiple comparisons test. *: p < 0.05; **: p < 0.01; ***: p<0.001
B. Electron micrographs of myelinated optic nerve axons from heterozygous Cnp1+/Cre mice at age 14 months.
On the left, a normal appearing paranode containing a peroxisome-like organelle (white arrow). On the right, membranous structures (arrows) and dense cytoplasm (asterisks) appear to accumulate in inner tongues in a small number of fibres. Inner tongues appear grossly swollen and impinge upon the axon.
C. Western blotting of myelin purified from homozygous Cnp1 knockout mice shows a dramatic reduction in steady state levels of MCT1. In contrast, RT-PCR of brain lysates demonstrates no significant change in Slc16a1 mRNA levels, when examined in relation to Slc16a1 mRNA levels in myelin from P20 WT mouse brain, in agreement with a post-transcriptional MCT1 deficit. Data were analysed using a two tailed, unpaired Student’s t-test.
D. Left: Electron micrographs of myelinated optic nerve axons from homozygous Cnp1 knockout mice, taken at (i) P40, (ii) P60 and (iii) P120, with undefined organelles (arrows) residing in the inner tongues (yellow overlay). Right: Quantitation demonstrates these organelles accumulate from age P60, i.e. long before clinical symptoms emerge at age 6 months (Lappe-Siefke et al., 2003). Bars represent mean ± SD. Data were analysed using a two tailed, unpaired Student’s t-test. Data points represent mean values from electron micrographs of optic nerves of individual animals. *: p < 0.05; **: p < 0.01.
White matter ageing is associated with impaired intra-myelin transport
Advanced age is associated with altered structural integrity of myelin in white matter tracts and intracortical myelinated fibres (Mueller, 2024; Peters, 2002), and with axon vulnerability as assessed by magnetic resonance imaging (MRI) in humans (Burzynska et al., 2024). We hypothesised that myelin changes include alterations in the myelinic channel system, similar to its destabilization by loss of CNP. This is suggested by a peculiar interaction of Cnp1 heterozygosity and advanced brain age on behavioural symptoms in mice (Hagemeyer et al., 2012). To directly determine if age-related white matter changes impact intra-myelinic transport processes, we compared myelin morphology and biochemistry in wildtype mice at age 6 and 24 months. We first measured myelin inner tongue sizes across axon diameters, observing a volume increase in the aged mice, most notably in small diameter fibres. This was accompanied by a twofold increase in fibres with organelles in the inner tongue (Fig. 7A). Furthermore, these changes were observed in the presence of an almost two-fold reduction in the overall density of myelinated axons at 24 months (corpus callosum) and a doubling of axonal degeneration profiles (Fig. 7B), raising the possibility that morphologically more severely affected fibres had been lost at this age.
When brain sections were stained for amyloid precursor protein (APP) as a marker of axonal transport impairment, we found a significant increase in APP positive swellings in the corpus callosum of 24-month-old mice (Fig. 7C), but interestingly not in the cortex.
These changes were accompanied by increased densities of Iba1 and Mac3 positive microglia/macrophages. Notably, by western blotting, steady-state level of MCT1 in myelin enriched fractions was approximately 50% reduced in the 24-month-old mice compared to 6-month-old controls (Fig. 7D). Taken together, these data suggest that aging-associated changes of myelin integrity impairs normal transport processes in the myelinic channel system, having a profound impact on MCT1 levels, and on oligodendroglial metabolic support of axonal integrity 29
Fig. 7. White matter aging leads to myelinic channel changes, decreased MCT1 levels and axon degeneration.
A. Electron micrograph of optic nerves from a 24-month-old mouse preserved by high pressure freezing (HPF), demonstrating that inner tongues (highlighted in yellow) are occasionally enlarged and are more likely to contain electron dense organelles (black arrowheads). The graphs on the right shows that inner tongue areas tended to be increased in 24-month-old animals compared to 6-month-old mice across all axon diameters (180-230 axons per nerve were examined), reaching significance in smaller diameter fibres. Data were analysed using multiple unpaired t-tests, corrected for multiple comparisons using Holm-Šídák method (n = 5 or 6; * p<0.05). A higher proportion of myelinated fibres have organelles in the inner tongue (IT; n = 3). Data were analysed using an unpaired two-tailed Student’s t test. * p < 0.05.
B. HPF EM overview showing that in 24-month-old WT mice, the density of myelinated axons is significantly reduced compared to 6-month-old mice, and the percentage of fibres with a degenerate profile is significantly increased (860-1300 axons per nerve were examined). Data were analysed by unpaired, two-tailed Student’s t-tests (n = 5 or 6; *** p < 0.001, **** p<0.0001).
C. Micrographs and graphs showing that APP positive axonal swellings and Iba1-labelled microglia are increased in 24-month-old mice compared to their 6-month-old counterparts in the corpus callosum (cc) but not in the motor cortex (mCtx), Bregma 0.74 mm. Mac3
positive microglia were significantly increased in both cc and mCtx in aged compared to younger mice. Data were analysed by unpaired, two-tailed Student’s t-tests for each region of interest (n=3; * p < 0.05, ** p<0.01, **** p<0.0001)
D. Western blot showing that MCT1 is about 50% decreased in a “myelin-enriched” brain fraction from 24-month-old mice, compared to 6-month-old mice (top), using fast green (FG) as an internal standard. Data were analysed by an unpaired, two-tailed Student’s t-test (n = 5; ** p < 0.01)
Discussion
The physiological function of compacted myelin in axonal insulation and impulse conduction is well established, but a distinct role for “non-compacted” myelin has been obscure. What is recognized by EM as non-compacted myelin was identified already at the light microscopic level by Rio de Hortega et al (1919), but suspected an experimental artifact (Edgar et al., 2021). Before the application of high-pressure freezing and freeze-substitution techniques to reduce EM dehydration artifacts (Edgar et al., 2021; Möbius et al., 2016), the relative volume of inner and outer tongues was grossly underestimated.
While microtubules and membranous structures were observed in non-compacted myelin (Edgar and Griffiths, 2013), single EM cross-sections failed to show whether these are stationary remnants of development, myelin turnover or part of a vital transport system, analogous to axonal transport, between the oligodendrocyte soma and the innermost cytoplasmic myelin layer facing the axon.
Here, by combining EM, mouse genetics, live imaging of peroxisomes and pharmacological interventions, we have established the concept of motor-driven, microtubule-dependent organelle transport in established myelin sheaths, being critical for long-term axonal integrity. Moreover, in vitro findings demonstrating that neuronal electrical activity affects myelinic transport suggests that
myelinated neurons themselves can control, at least in part, aspects of oligodendrocyte behaviour.
Our interest in these phenomena was triggered by the observation that oligodendrocytes and axons are metabolically coupled (Fünfschilling et al., 2012) through MCT1, a monocarboxylate transporter localized at the glial axonal interface (Lee et al., 2012). While it is unclear currently whether myelin turnover in adult mice rests entirely on the incorporation of new membranes at the innermost myelin layer (Meschkat et al., 2020), i.e. reflecting developmental myelination (Snaidero et al., 2014), the necessary turnover of MCT1 and its functional incorporation at the periaxonal myelin membrane would require vesicular transport to the inner tongue. We thus hypothesized that MCT1 levels in mature myelin depend on the integrity of the intra-myelinic transport route. Our data support this
suggestion by demonstrating a severe depletion of MCT1 and increased frequencies of organelles in myelin from CNP-deficient mice in which the myelinic channel system is structurally impaired (Snaidero et al., 2017); findings replicated in aged animals. In particular, the ‘trapping’ of myelinic organelles is reminiscent of organelle accumulations in axons upon perturbation of axonal transport (reviewed in Stassart et al., 2018).
To model and visualize transport of such nuclear-encoded membrane-bound cargoes in the myelinic channel system, we chose transgenic Cnp-EOSSKL-mice that harbour fluorescent peroxisomes in oligodendrocytes (Richert et al.,2014). These single membrane organelles are abundant in myelinating glia and while used here as a surrogate for myelin organelles more generally, it is noteworthy that oligodendroglial peroxisomes are essential for long-term axonal integrity (Kassmann et al., 2007).
The existence of a transport system operating in mature fibres is plausible because the myelinic channel system is developmentally derived from subcellular compartments of oligodendrocytes in which transport processes contribute directly to myelin outgrowth (Herbert et al., 2017; Iyer et al., 2024; Lyons et al., 2009; Nawaz et al., 2015; Snaidero et 1996; Song et al., 1999) with the notable difference that these processes remain short and deliver new membranes to a flat bilayer structure that, in vivo, grows spirally around axons (Snaidero et al., 2014). Indeed, in vivo, normal myelin sheath growth requires nucleation of microtubules, that is mediated by tubulin polymerization promoting protein (TPPP) (Fu et al., 2019; Lehotzky et al., 2010; Tokési et al., 2010). Further, mRNAs encoding myelin proteins (See Supplementary Fig.1) and presumably the oligodendroglial translation
machinery, are transported on microtubules, as demonstrated in oligodendrocyte monolayer cultures (Brophy et al., 1993; Carson et al., 1998; Colman et al., 1982; Holz et al., 1996; Kursula et al., 2001; Quraishe et al., 2016; Trapp et al., 1987).
Considering our findings reported here, the importance of transport processes for developmental myelination and myelin maintenance is illustrated by the spontaneous taiep mutant rat, which harbours a point mutation in Tubb4a, encoding β-tubulin IV. This increases the ratio of “minus-end-distal” to “plus-end-distal” microtubules in the fine processes of developing oligodendrocytes (Song et al., 1999), which leads to an abnormal distribution of various myelin gene products (O’Connor et al., 2000) and hypomyelination with subsequent demyelination (Lunn et al., 1997).
Remarkably, the Tubb4a null mutation had no overt effect on myelination, indicating compensatory functions by other tubulins, and a dominant-negative effect of the taiep mutation. Nonetheless, the oligodendrocyte-specific expression of this tubulin isoform in the CNS (Schaeren-Wiemers et al., 1995) suggests that its known function as a tubulin stabilizer (Renthal et al., 1993) is important for myelinating oligodendrocytes.
34
A growing body of evidence has shown that neuronal electrical activity can stimulate developmental myelination (reviewed in Taylor and Monje, 2023). Less is known about axon-myelin signalling in the mature fibre, due in part to the inaccessibility of the glial-axonal junction on the inside of the myelin sheath. Our finding that the mobility of peroxisomes in myelinic channels is higher in electrically active networks, which, being a rapid response, suggests it involves direct axon-myelinic signalling.
While mechanistic studies await the analysis of corresponding mouse mutants, we note that mature
oligodendrocytes respond to axonal spiking activity (Yamazaki et al., 2010), at least in part through an NMDA receptor-mediated calcium rise in the myelin sheath (Micu et al., 2018) that also activates the intracellular transport of the glucose transporter GLUT1 in oligodendrocytes (Saab et al., 2016).
Moreover, potassium transients in the periaxonalspace of fast spiking fibres induce Ca 2+ signals that stimulate glycolysis in oligodendrocytes (Looser et al., 2024).
The novelty of our findings is not the “myelinic channel system” as such, because we and others already hypothesized a “continuity” of non-compacted myelin as a cytosol-filled rim that flanks compacted myelin, with obvious reference to its developmental origin (Edgar et al., 2009, 2021; Hirano and Dembitzer, 1967; Nave, 2010b; Snaidero et al., 2014) and as a route for diffusing energy-rich metabolites for axonal support (Saab and Nave, 2017).
Moreover, the same cytosolic spaces were previously visualized ex vivo by dye injections in both the CNS and PNS (Balice-Gordon et al., 1998; Butt and Ransom, 1989; Velumian et al., 2011). Rather, we provide the first experimental evidence that this system serves active transport processes, utilizing motor proteins and tubular tracts, exemplified by peroxisome movement. While this first description of “myelinic transport” (in analogy to “axonal transport”) is necessarily incomplete with respect to the nature of the motor proteins and tubulin isoforms involved, it already helps to explain aspects of the obscure secondary axonal degeneration of myelin-specific disorders. Adult CNP-deficient mice, modelling a rare human neurological disease (Al-Abdi et al., 2020), lack integrity of this myelinic channel system, as CNP normally prevents (MBP-dependent) myelin hyper-compaction and local “channel collapse” (Snaidero et al., 2017). Consequently, nuclear encoded MCT1 of mature oligodendrocytes probably fails in the motor-driven transport via the myelinic channel system to replenish supplies at the inner tongue of myelin as suggested by western blotting of a myelin-enriched tissue fraction. Loss of MCT1 in turn predicts poor metabolic interaction with axons (Lee et al., 2012), progressive axonal degeneration and the previously described premature death of the animal (Lappe-Siefke et al., 2003).
Clinically more relevant are myelin diseases, such as multiple sclerosis (MS), where immune-mediated attack on the myelin sheath causes progressive axon loss. We recently found that in human MS and its mouse models, a major contributor to secondary axonal degeneration is myelin injury (Schäffner et al., 2023). We hypothesise this has the same effect on MCT1 turnover as CNP deficiency, explaining the delayed axonal degeneration occurring months after the primary lesion (Schäffner et al., 2023).
The known neuroprotective function of oligodendrocytes involves the transfer of materials from the oligodendrocyte to the axon (Chamberlain et al., 2021; Frühbeis et al., 2020; Fünfschilling et al., 2012; Lee et al., 2012; Mukherjee et al., 2020). Thus, the oligodendrocyte acts locally, at the level of each myelinated internode, to support the axon 36 (Edgar and Garbern, 2004; Edgar et al., 2004), which, being encased in lipid-rich myelin, is physically isolated from the extracellular milieu (Nave, 2010a). At the inner tongue of normal-appearing fibres in wild-type animals, we observed multivesicular bodies, as reported previously (Frühbeis et al., 2020) and fusion profiles (Supplementary Fig.1G) that could empty exosomes and soluble content, respectively, into the periaxonal space for uptake by the axon.
The brain’s white matter, which contains densely packed, parallel arrays of myelinated axons, is vulnerable during ageing in humans, and small diameter myelinated fibres are particularly vulnerable (Marner et al., 2003; Tang et al., 1997). If one assumes axon energy insufficiency is a pathological driver, this seems at first counter intuitive because thin axons are predicted to require less energy than thicker ones (Perge et al., 2012).
A possible explanation relates to the fact that oligodendrocytes that myelinate thin axons extend many more processes than those that myelinate thick axons (reviewed in Edgar et al., 2021). As oligodendrocytes in the ageing brain can develop inclusion-containing bulbous swellings (presumably sites of transport stasis) along their processes (Peters, 2002), it becomes apparent that thin myelinated fibres might be more susceptible due to a greater risk of transport impairment in oligodendrocytes that extend numerous processes to small diameter axons.
Limitations of the study
Direct visualisation of organelle transport in myelin was conducted in an in vitro model of the myelinated CNS, in which compact myelin with an appropriate g-ratio is formed (Thomson et al., 2008), and in which multiple paranodal loops abut the axon at the paranode (Edgar et al 2021 and Supplementary Videos 4 and 5). Like the CNS in vivo, these myelinating cell cultures also contain astrocytes, microglia and oligodendrocyte progenitor cells. Nonetheless, our findings should be confirmed in vivo.
A second limitation is that we cannot directly correlate myelin peroxisomal transport with the firing rate of the same axon. However, neurons exhibited a range of firing frequencies, corresponding to the wide variability in the proportion of myelin peroxisomes that exhibited movements at the time of imaging. Notably, myelin peroxisome motility was not completely blocked in the presence of TTX, but never reached the mobility seen in cell cultures with intrinsic neuronal electrical activity. Additionally, the number of organelles quantifiable in the inner tongues of myelin in vivo is small (estimated at ~2 per sheath) which limits our ability to quantify changes following electrophysiological stimulation of optic nerve. Finally, the observed loss of MCT1 in a ‘myelin-enriched’ spinal cord lysates by western botting provides no direct evidence that the protein is reduced at the inner lip of the myelin sheath.38
Acknowledgements
We are grateful to Professors Ueli Suter and Pierre Chambon, and Dr Daniel Metzger for gift of the PLP‐Cre‐ERT2 mice. Dr Katherine Kusch (Nave lab) and Professor Ori Peles, Weizmann Institute of Science, kindly provided antibodies to MCT1 and Caspr1, respectively. Professor Lukas Kapitein, Utrecht University generously provided the plasmid encoding StableMARK and mNeonGreen, the latter provided by Allele Biotechnology & Pharmaceuticals. TagRFP-T_A1aY1 (158751) was provided to Addgene by Professor Minhaj Sirajuddin, Institute for Stem Cell Science & Regenerative Medicine.
EM processing and sectioning of conditional Kif21b and Tubb4a knockout mice was performed with outstanding technical assistance from Mrs Margaret Mullin in the Glasgow Imaging Facility; We are grateful to colleagues in Glasgow, i.e. Dr Paul Montague for performed genotyping; Dr Fatma Kok and Mr David Kerrigan in Dr Rhaidhri Carmody’s lab for providing Midi-preps; John Cole for checking calculations for peroxisome densities; also to Dr Michaela Schweizer and Dr Mary Muhia (Hamburg) for assisting in respect to the conditional Kif21b knockout mice. J.M.E. and E.R.B. were supported by grants from the UK MS Society (Grants 58 to JME; Grant 127 to JME and ERB). J.M.E. was also
supported by funds from the Chief Scientist’s Office, Scotland and the University of Glasgow (SPRINT MND/MS PhD studentship) and the National Centre for the Replacement, Refinement and Reduction of Animals in Research (NC3Rs). K.A.N. was supported by the European Research Council (AdvGrant MyeliNANO), the Dr. Myriam and Sheldon Adelson Medical Foundation (AMRF) and the German Research Council (DFG-TRR274). SG was supported by the Deutsche Forschungsgemeinschaft GO 2463/1-1.
Authors‘ contributions
Conceptualization JME, KAN, SG, ERB. al., 2014; Yergert et al., 2021). Analogous to neurons with growth cones, oligodendrocytes in monolayer culture have microtubules within growing cellular processes (Barry et al.
Kommentit