Time to revisit oligodendrocytes in multiple sclerosis

The analysis of autopsy material from individuals with multiple sclerosis with single-cell transcriptomics and 14C carbon dating calls for a reevaluation of mature oligodendrocytes in myelin repair.

Multiple sclerosis (MS) is clinically the most important disease affecting myelin, but its underlying cause is not well understood1. Driven by an autoimmune response against one or more components of myelin in the central nervous system, a gradual loss of axons turns an initially mild and often ‘relapsing–remitting’ course of MS into a severe, progressive disease with persistent disabilities2. Recently, promising immunomodulatory therapies have been developed that strongly reduce the number of attacks. However, it is unclear whether they can stop the clinical course of secondary progressive MS. Hope therefore rests on experimental strategies aimed at stimulating endogenous myelin repair3. Such remyelination by oligodendrocytes, derived from precursor cells (OPCs), is anticipated to protect and reinstate metabolic support for axons in addition to fast impulse propagation. Two recent studies in Nature4,5 reveal an unexpected heterogeneity of oligodendrocytes throughout the MS-affected brain and raise questions about the role of OPCs in permanent lesion repair.

In young laboratory mice and rats, remyelination typically occurs quickly. This has been documented in a range of experimental systems with focal demyelination3. Classical studies in the 1970s documented that myelin repair in these experimental lesions requires resident OPCs to divide and repopulate the area before differentiating into myelinating oligodendrocytes3. These new sheaths are thinner than those made during development, presumably because of weaker promyelinating signals in the adult brain. The situation in humans is different, however, as the repair of MS plaques often fails. Moreover, the occasional finding of ‘shadow plaques,’ which appear pale on histological stainings and comprise thinly myelinated axons, has been taken as evidence for incomplete remyelination in humans, which nevertheless provides proof of principle for myelin repair in MS6.

In their study, Castelo-Branco, Williams, ffrench-Constant and colleagues4 used single-cell transcriptomics in one of the first applications of this technology to neurological disease and human autopsy material. Starting with single-cell nuclei from white matter of normal brains, the authors identified eight clusters of oligodendroglia that confirm their previous finding that there is an unexpected oligodendrocyte heterogeneity in mice7. In human autopsy material from individuals with MS, they analyzed cells from the ‘normal appearing white matter’ (NAWM), which comprises fully myelinated regions thought to be unaffected by focal inflammatory demyelination. They discovered that oligodendrocytes are present in normal numbers, but, remarkably, the molecular signatures and the clusters based on these signatures are shifted compared with oligodendrocytes from non-MS brains, with one cluster of mature cells being nearly absent. Another cluster, defined by moderate gene expression for myelin proteins and elevated expression of cytoskeletal elements, was strikingly small, suggesting that there were fewer cells with transport functions that could be required for metabolic support. Within lesions, there were also signatures of oligodendrocytes specific to the stage of inflammation. On average, myelin gene expression was increased in oligodendrocytes within lesions, leading the authors to speculate that the surviving mature oligodendrocytes contribute to remyelination.

The observations of these authors would not be possible with conventional transcriptomics of cells at the population level — not only because most genetic signatures are not cell-specific, but also because altered frequencies of cellular subtypes in a heterogenous population (as a result of disease) can mask shifts of gene expression in individual cells. Thus, even with the relatively small number of samples assessed so far, the authors can conclude that oligodendrocytes that appear morphologically normal in individuals affected by MS are affected by disease and are in an altered state. This could be the result of an underlying developmental/degenerative defect, or it may be a global effect of inflammation. Indeed, the pathology of MS lesions may be the tip of the iceberg of all cellular changes ongoing in the MS brain as a whole. While that has been suggested by others before, providing direct evidence at the single-cell level is important. Interestingly, this study4 also found that the number of OPCs was reduced both within MS plaques and in the NAWM. The finding was confirmed by in situ hybridization in independent patient samples and could be one cause of poor myelin repair if remyelination were to depend on OPC.

Myelin repair is also at the heart of the study from Frisén and colleagues5, who have previously pioneered a 14C-based birth-dating technique for cellular material in autopsy samples. Yeung et al.5 utilize the globally elevated content of 14C in all biological material synthesized during the peak of atmospheric atomic bomb tests in the early 1960s and its characteristic washout in the years thereafter to determine the age of cellular components. Using this technique, they have shown that oligodendrocytes in white matter tracts of postmortem brains from healthy individuals are virtually as old as neurons; that is, they are largely born during early postnatal life8.

In the present study, the authors applied this technology to autopsy material from individuals with MS and analyzed the age of oligodendrocytes in the NAWM and in shadow plaques. In the NAWM of a subset of patients with MS born in the 1960s and defined by a severe course of disease, the authors discovered a much larger fraction of recently generated OPCs and oligodendrocytes than in the white matter of unaffected individuals. This demonstrates that NAWM is clearly affected by MS and also reveals an unexpected heterogeneity of patients with respect to the response of OPCs to the disease. Furthermore, the authors discovered that the oligodendocytes in shadow plaques, lesions thought to exhibit incomplete myelin repair, are as old as the patient and therefore could not have been generated from newly recruited OPCs. Thus, they do not find that remyelination in the human brain is carried out by newly generated oligodendrocytes, as was predicted from animal models (Fig. 1).

Fig. 1: Mature oligodendrocytes are heterogeneous in multiple sclerosis.

Frisén and colleagues provide evidence that some mature oligodendrocytes may have a role in remyelination. Castelo-Branco, Williams, ffrench-Constant and colleagues4 find that in MS brains, even in regions previously thought of as ‘unaffected’, oligodendrocytes have a different molecular funion. A schematic view of oligodendrocytes in the MS brain is shown, with NAWM (blue) and two demyelinated plaques. Multicolored circles represent a molecularly heterogeneous population of oligodendrocytes and their precursors, as revealed by single-cell transcriptomics. Interestingly, their genetic signatures differ from controls, even in morphologically unaffected areas4, confirming that MS is a global brain disease. In shadow plaques (magnified), axons remain thinly myelinated, which is often interpreted as limited myelin repair. Unexpectedly, remyelinating cells are preexisting oligodendrocytes5, suggesting that local remyelination can occur without precursor recruitment. Alternatively, shadow plaques are incompletely demyelinated, possibly when oligodendrocytes are under chronic inflammatory stress.

There are two possible scenarios that could explain the work of Frisén and colleagues. First, it is possible that human postmitotic oligodendrocytes that survived the immune attack and demyelination are involved in remyelination, which is compatible with their gene expression signature4. While preexisting mature oligodendrocytes appear to be not involved in remyelination in rodents9, there is recent evidence for participation of adult oligodendrocytes in remyelination in cats10. It is also possible that the shadow plaques (which were defined by neuropathologists at the participating brain banks) are not thinly remyelinated, but rather partly demyelinated lesions, in line with the older literature. Both interpretations are against current thinking6. Notably, additional analyses by the authors provide evidence against the argument that a differentiation of preexisting OPCs (without cell division) might explain their result.

Until this point, all histological evidence for remyelination in MS builds on the similarity with model organisms, in which remyelination leads to thinner myelin. But this evidence cannot be used to conclude that thinner myelin is the result of impaired remyelination in humans. However, there is indirect clinical evidence for remyelination in patients with MS, such as an improvement of central motor conduction time, measured 5 months after treatment with rhEPO11, or faster propagation of visually evoked potentials after receipt of clemastine12, both of which are drugs promoting myelin repair in experimental autoimmune encephalitis (EAE) mice. This leaves as a third possibility that remyelination occurs in MS, but may be unstable and partially lost when autopsy is performed decades later. After acute lesion repair, it is even plausible that oligodendrocytes remain stressed in the chronic inflammatory milieu and later fail to maintain normal myelin sheath thickness. Such an effect of cellular stress on mature oligodendrocytes is well known from mice overexpressing the myelin protein PLP1.

These two papers have implications for MS as a ‘global’ brain disease, the requirement of OPCs for lesion repair and the possible contribution of mature oligodendrocytes in this process. While these studies suggest that remyelination of brains is a more challenging task than previously thought, conceptually they are major step forward.


  1. 1.

    Trapp, B. D. & Nave, K.-A. Annu. Rev. Neurosci. 31, 535–561 (2008).

    Article  Google Scholar 

  2. 2.

    Reich, D. S., Luccinetti, C. F. & Calabresi, P. A. N. Engl. J. Med. 378, 169–180 (2018).

    CAS  Article  Google Scholar 

  3. 3.

    Franklin, R. J. M. & ffrench-Constant, C. Nat. Rev. Neurosci. 18, 753–769 (2017).

    CAS  Article  Google Scholar 

  4. 4.

    Jäkle, S. et al. Nature (2019).

  5. 5.

    Yeung, M. S. Y. et al. Nature (2019).

  6. 6.

    Lucchinetti, C. et al. Ann. Neurol. 47, 707–717 (2000).

    CAS  Article  Google Scholar 

  7. 7.

    Marques, S. et al. Science 352, 1326–1329 (2016).

    CAS  Article  Google Scholar 

  8. 8.

    Yeung, M. S. et al. Cell 159, 766–774 (2014).

    CAS  Article  Google Scholar 

  9. 9.

    Crawford, A. H. et al. Am J. Pathol. 186, 511–516 (2016).

    CAS  Article  Google Scholar 

  10. 10.

    Duncan, I. D. et al. Proc. Natl Acad. Sci. USA 115, E11807–E11816 (2018).

    CAS  Article  Google Scholar 

  11. 11.

    Ehrenreich, H. et al. Brain 130, 2577–2588 (2007).

    Article  Google Scholar 

  12. 12.

    Green, A. J. et al. Lancet 390, 24812489 (2017).

    Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Klaus-Armin Nave.

Ethics declarations

Competing interests

The authors declare no competing interests

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Nave, KA., Ehrenreich, H. Time to revisit oligodendrocytes in multiple sclerosis. Nat Med 25, 364–366 (2019).

Download citation

Further reading


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing