bHLH Transcription Factor Olig1 Is Required to Repair Demyelinated Lesions in the CNS

See allHide authors and affiliations

Science  17 Dec 2004:
Vol. 306, Issue 5704, pp. 2111-2115
DOI: 10.1126/science.1103709


Olig1 and Olig2 are closely related basic helix-loop-helix (bHLH) transcription factors that are expressed in myelinating oligodendrocytes and their progenitor cells in the developing central nervous system (CNS). Olig2 is necessary for the specification of oligodendrocytes, but the biological functions of Olig1 during oligodendrocyte lineage development are poorly understood. We show here that Olig1 function in mice is required not to develop the brain but to repair it. Specifically, we demonstrate a genetic requirement for Olig1 in repairing the types of lesions that occur in patients with multiple sclerosis.

The oligodendrocyte is the myelinating cell of the central nervous system (CNS) (1, 2). The appearance of these myelinating cells during evolution is thought to have enabled the vertebrate nervous system to grow large and complex by allowing saltatory conduction of nervous impulses (3). A variety of demyelinating diseases exist in humans, wherein the myelin sheaths are lost, usually through the death of mature oligodendrocytes. One principal example, multiple sclerosis (MS), affects roughly two and a half million people worldwide and is one of the most common causes of neurological disability in young adults (46). Typically, patients with MS present with a relapsing/remitting form of the disease, characterized by acute demyelinating episodes followed by the generation of new oligodendrocytes, remyelination, and functional recovery. However, remyelination is an inconsistent event in this disease, and the accumulated load of lesions that fail to remyelinate results in progressive neurological deterioration, in part because the capacity to generate new oligodendrocytes becomes limited (711).

To understand the molecular mechanisms of remyelination and its limitations, and thereby identify potential targets for therapeutic intervention to promote repair, it is instructive to look at the mechanisms that create oligodendrocytes in the developing CNS (12). The basic helix-loop-helix (bHLH) transcription factor Olig2 is necessary for specification of both oligodendrocytes and motor neurons during vertebrate embryogenesis (1315). Olig2-null (Olig2–/–) mice die at birth because of a deficit of motor neurons. A closely related homolog, Olig1, is coexpressed with Olig2 in many cells of the oligodendrocyte lineage, but the biological functions of Olig1 are not well resolved. Gain-of-function analysis suggests a role for Olig1 in oligodendrocyte progenitor development (16, 17); however, Olig1–/– mice are viable and fertile and show only a subtle delay in oligodendrocyte maturation (13).

Both Olig genes continue to be expressed in mature oligodendrocytes, which suggests that they may have functions in the adult brain and spinal cord, independent of their role in development (18). To explore these functions, we generated polyclonal and monoclonal antibodies that specifically recognize Olig1 or Olig2 (fig. S1) and used these reagents for immunohistochemical analysis at several stages of mouse CNS development. Olig2 was localized to the nucleus at all stages examined and in all regions of the CNS (Fig. 1, A and C). In neonatal mice, Olig1 was likewise localized to the nucleus. However, Olig1 proteins were located mostly in the cytoplasm by 2 weeks after birth and were entirely cytoplasmic in the white matter of the adult mouse (Fig. 1A). In vitro studies of neonatal oligodendrocyte cultures similarly demonstrate that the subcellular localization of Olig1 becomes progressively more cytoplasmic as these cells mature into myelin basic protein positive (MBP+) oligodendrocytes (19). Cytoplasmic localization of Olig1 and nuclear localization of Olig2 in the adult mouse brain were confirmed by deconvolution microscopy (Fig. 1B). This pattern of subcellular localization for Olig1 and Olig2 in the adult CNS was reproduced in all areas of white matter analyzed (Fig. 1C). The differential localization of Olig1 and Olig2 is also seen in the adult human brain (Fig. 2D).

Fig. 1.

Subcellular relocalization of Olig1, but not Olig2, during development. (A) Olig1 and Olig2 are both present in the nucleus of oligodendrocytes and their progenitors in the postnatal day 1 (P1) mouse brain. Olig2 continues to be nuclear at all developmental stages, whereas Olig1 is almost completely cytoplasmic in the adult mouse brain. (B) Subcellular localization in the adult brain was verified using deconvolution microscopy. DAPI, 4′,6′-diamidino-2-phenylindole. (C) This pattern of expression of Olig1 and Olig2 is recapitulated in all CNS white-matter regions examined.

Fig. 2.

Dynamic localization of Olig1 during remyelination. (A) Demyelination was induced in the brain with the dietary administration of cuprizone, and in the caudal cerebellar peduncle via an injection of ethidium bromide, and analyzed over various time points for cytoplasmic (white arrows) or nuclear (gray arrows) expression of Olig1. dpl, days post lesion. (B and C) Deconvolution microscopy was used to verify the nuclear localization of Olig1 and Olig2 during remyelination. (D) Postmortem tissue from six MS patients was examined for the subcellular localization of Olig1 and Olig2 surrounding demyelinated plaques. The examined tissue contained chronic demyelinated plaques, with varying degrees of inflammatory infiltrate, and varying degrees of remyelination occurring at the lesion edge. The images shown are serial sections from a single donor stained with hemotoxylin and eosin (H&E) or antibodies directed against mature oligodendrocytes and MBP, oligodendrocyte progenitors (NG2), Olig1, or Olig2. As indicated with black boxes in the H&E-stained sections, areas of normal-appearing white matter and a chronic active lesion were identified and maintained constant across the serial sections for imaging analysis. (E) Colocalization of NG2 (brown) and Olig1 (purple) in MS tissue demonstrated that Olig1 is nuclear in oligodendrocyte progenitors surrounding an MS lesion.

Several toxins, including cuprizone, lysolecithin, and ethidium bromide, are widely used to study the mechanisms of remyelination (20, 21). As shown in Fig. 2, A to C, Olig1 proteins are initially located in the nucleus of cells in early remyelinating lesions of the corpus callosum or of brain stem white matter after treatment with cuprizone or ethidium bromide, respectively. Similar observations were made after lysolecithin-induced injury in the spinal cord of adult mice (fig. S2). The differential localization of Olig1 in these rodent models of demyelinating disease was recapitulated in postmortem brain tissue from patients with MS (Fig. 2D). Tissue was analyzed from six cases of MS, three displaying chronic active MS lesions and three displaying chronic silent lesions (22). Cells containing cytosolic Olig1 were present in normal-appearing white matter. Nuclear Olig1 was present at the edges of the active lesions and of all but one of the chronic lesions. Serial sections were analyzed for the presence of mature oligodendrocytes and myelin (MBP), oligodendrocyte progenitors (NG2), and Olig1 and Olig2 protein expression. The distribution of mature and immature oligodendrocytes suggests that the cells containing nuclear Olig1 at the edge of MS lesions are likely to be undifferentiated progenitors (Fig. 2D). Colabeling localizes nuclear Olig1 to the NG2 oligodendrocyte progenitors (Fig. 2E). Because the process of remyelination is more likely to occur in the active lesions (23), this distribution is consistent with a role for Olig1 nuclear translocation in the repair process in MS patients.

Functional insights into the role of Olig1 in CNS repair can be derived from analysis of Olig1–/– mice (13). The ultrastuctural images of the adult CNS show that Olig1 function is not required to develop a healthy, fully myelinated brain or spinal cord (Fig. 3A). Moreover, as noted previously, Olig1–/– mice show no obvious abnormal behavioral phenotype. However, as seen in Fig. 3, B to D, and fig. S3, Olig1 function is required for the remyelination phase of both cuprizone (brain)– and lysolecithin (spinal cord)–induced demyelination. These data indicate a critical role for Olig1 in the repair of the adult CNS.

Fig. 3.

Failure of remyelination in mice lacking Olig1. (A) Morphological analysis of myelinated tracts in the ventrolateral funiculus of the Olig1–/– mouse spinal cord indicates normal myelin formation and compaction. (B) Luxol fast blue stain and light microscopy show that Olig1 is required for efficient remyelination of the corpus callosum in the brains of wild-type (WT) and Olig1–/– mice after demyelination with cuprizone. Slides were scored in a double-blind fashion on a scale of zero to three, where a score of zero indicates complete myelination of all fibers at the midline corpus callosum, and a score of three indicates complete demyelination. (C) Olig1 is also required for efficient remyelination of lysolecithin-induced lesions in the dorsal or ventrolateral funiculi of the spinal cord. At 14 dpl, the WT mice exhibited extensive remyelination as seen by light microscopy (fig. S3A), with 80 to 90% of the axons within the lesion showing remyelination. The new myelin sheaths were very thin at this time, and so their presence was confirmed by means of electron microscopy (EM). This extensive remyelination contrasts dramatically with that demonstrated by the Olig1–/– mice at the same time point, where there was almost no evidence of remyelination as indicated by either light microscopy using standard morphological criteria (29) or by EM. (D) Quantitation of remyelination shown in (C). The extent of remyelination for each animal was assessed with light microscopy and each was given a score (by two independent experimenters) for the number of axons showing remyelination as a percentage of the total number of demyelinated axons. There is a significant difference at both 14 and 28 dpl in the extent of remyelination between WT (open symbols) and Olig1–/– mice (solid symbols), and remyelination in the Olig1–/– mice remains incomplete at 46 dpl (fig. S3, B and C). (E and F) Solochrome cyanine staining was used in all animals to determine the extent of the area of demyelination in the ventrolateral funiculus of the spinal cord. In normal unlesioned white-matter, the density of mature PLP/DM20-expressing oligodendrocytes is indistinguishable between WT and Olig1–/– mice. However, at 14 dpl the lesions in the WT mice contain differentiated oligodendrocytes, whereas there is an almost complete absence of mature oligodendrocytes within the Olig1–/– lesions at this time (t test, ** = P < 0.0007).

Why is remyelination so inefficient in the CNS of Olig1–/– mice? As shown in Fig. 4A, the initial event in remyelination is an appearance of oligodendrocyte progenitor cells marked by NG2, Olig1, Olig2, and Nkx2.2 (a homeodomain transcription factor) (24, 25). A double immunostain with polyclonal (Olig2) and monoclonal (Olig1) antibodies shows that Olig1 and Olig2 are coexpressed in these progenitor cells (Fig. 4B). As shown in Fig. 4, C and D, and fig. S4, loss of Olig1 function has no obvious effect on the recruitment of these progenitor cells. Expression of Olig2 is maintained in these progenitors, ruling out the possibility that Olig1 is required to maintain Olig2 expression. Analysis of oligodendrocyte maturation in Olig1–/– animals after demyelination (Fig. 3, E and F) demonstrates that these progenitors are impaired in their ability to differentiate. Therefore, Olig1 has an essential role in oligodendrocyte differentiation and consequent remyelination in the context of white matter injury.

Fig. 4.

Olig1 is critical for the maturation of oligodendrocyte progenitors. (A) After demyelination with cuprizone, cells positive for NG2, Nkx2.2, and both of the Olig genes are present in the corpus callosum. (B) Olig1 and Olig2 are coexpressed in normal and lesioned white matter. White arrows indicate colocalization of Olig1 and Olig2 in the nucleus of cells in a remyelinating lesion (after 1 week of recovery) of the corpus callosum; gray arrows indicate representative cells in which Olig1 is predominantly cytoplasmic and Olig2 is nuclear. (C) The loss of Olig1 has no obvious effect on the recruitment of NG2+/Nkx2.2+/Olig2+ progenitors after a cuprizone-induced demyelinating injury; however, the induction of MBP is delayed in the Olig1–/– mouse during remyelination. Images shown in C are from mice after 1 week of recovery (7 weeks total) from cuprizone-induced demyelination. (D) Quantitation of the data shown in (C) for recruitment of NG2+ progenitors. Analysis was performed in at least three mice per time point, and cell counts for NG2+ progenitors were restricted to the midline corpus callosum at the level of the fornix.

Another facet of the response to demyelination in both humans and rodents is the proliferation of microglia and astrocytes (reactive gliosis). As shown in fig. S5, neither reactive response is perturbed in the Olig1–/– mice. Thus, the role of Olig1 in the response to injury appears to be confined to oligodendrocytes and their progenitors. This separation of reactive gliosis and remyelination is in accord with several lines of evidence indicating that oligodendrocytes and astrocytes arise from separate progenitor populations (13, 14).

We have shown that intracellular localization of Olig1 protein is dynamic and changes with the developmental state of oligodendrocyte lineage cells. Demyelinating injuries to the adult brain evidently create an environment that recapitulates the immature brain, allowing nuclear localization of Olig1. Because nuclear localization is important for transcription factor function, we deduce that the critical phase of Olig1 activity in oligodendrocyte differentiation occurs before it becomes localized to the cytoplasm.

With respect to the human disease, the lesions that define MS are now thought to arise from a number of different mechanisms, mostly immunologic but not always linear or cell-mediated (26, 27). Accordingly, MS may never be entirely preventable, and new therapeutic approaches must focus on the repair (remyelination) process. At the level of basic science, one important unresolved issue is why remyelination is so limited in patients with MS even though endogenous oligodendrocyte progenitors are often present in abundance (9, 28). Studies shown in this paper indicate that signals regulating the subcellular localization and/or activity of Olig1 during development may play an additional and critical role in activating oligodendrocyte progenitors in the adult CNS. Our data show that requirements for Olig1 function are subtle during development, yet striking during the repair of a demyelinating lesion. This would suggest that cell-intrinsic activity of Olig1 can be compensated for during myelination but not remyelination. Further insights into the molecular mechanisms of Olig1 function during development may have practical overtones for future therapeutic interventions in MS.

Supporting Online Material

Materials and Methods

Figs. S1 to S5


References and Notes

Stay Connected to Science

Navigate This Article