Neuregulin-1 and Myelination

See allHide authors and affiliations

Science's STKE  07 Mar 2006:
Vol. 2006, Issue 325, pp. pe11
DOI: 10.1126/stke.3252006pe11


The trigger for the process of myelin formation during neural development has long been a mystery. Recent studies suggest that this trigger is the growth factor neuregulin-1, expressed on the surface of axons, which signals through ErbB2 and ErbB3 receptors expressed on the surface of myelin-forming glia.

In vertebrates, the nerves of the peripheral nervous system (PNS) and the axon tracts of the central nervous system (CNS) carry two types of axons. Myelinated axons are surrounded by a very large multilayered sheet of specialized plasma membrane that is elaborated by Schwann cells in the PNS and oligodendrocytes in the CNS. These glial cells repeatedly wrap and then tightly compact myelin around axons. In so doing, they electrically insulate the axonal membrane and dramatically increase the speed at which the nerve impulse propagates (1). Unmyelinated axons lack this insulation, and as a consequence, they propagate action potentials at a much slower rate. Myelin is essential to the operation of all higher-order nervous systems, and the phylogenetic appearance of myelin-forming cells represents the last major cellular advance in neural evolution (1). Two recent papers that examine myelination in the PNS (2, 3) demonstrate that a specific isoform of the growth factor neuregulin-1 (Nrg-1) (46), expressed on the surface of axons, binds to ErbB receptor tyrosine kinases expressed by Schwann cells and serves as the inducing trigger for myelination. These studies demonstrate that the strength of ErbB signaling determines whether Schwann cells elaborate myelin, and if so, how much myelin they produce.

Whether a particular axon is myelinated has been known for decades to be tightly correlated with its size. In the mature PNS, axons greater than 1 μm in diameter are almost invariably myelinated, whereas those smaller than this are not myelinated; further, the thickness of the sheath around a myelinated axon is directly related to the size of the axon. Unmyelinated peripheral axons are only loosely embraced by a surrounding Schwann cell cuff. The decision to myelinate is a big one for the myelinating cell. It involves the activation and high-level expression of a set of myelin-specific genes (1), as well as a major metabolic reorganization in which the cell switches to the production of vast quantities of myelin membrane. The process of myelinating a very large axon in the PNS, for example, may result in a several-thousandfold increase in the membrane surface area of the myelinating Schwann cell (1). It now appears that Nrg-1 signaling through ErbB receptors mediates this all-or-none decision.

Nrg-1 was first purified to homogeneity by means of a proliferation bioassay with Schwann cells as the target (7), and recent analyses once again demonstrate that these cells are a primary locus of Nrg-1 signaling during mammalian development. Earlier genetic studies in mice carrying engineered mutations in the Nrg-1 gene, or in the genes encoding the ErbB2 and ErbB3 Nrg-1 receptors, demonstrated that Nrg-1 is required for the proliferation, directed migration, subsequent maturation, and postnatal survival of Schwann cell precursors and premyelinating Schwann cells (811). Mouse knock-outs of the ErbB3 gene, for example, display a nearly complete lack of Schwann cells in the body of peripheral nerves by the time of late embryogenesis (8). The newer work extends the requirement for Nrg-1 signaling to the later process of myelination itself.

The first indication of the importance of Nrg-1 to myelination appears in the work of Garratt and colleagues (12), in which abnormally thin myelin sheaths were observed to form in peripheral nerves with Schwann cells that expressed reduced levels of ErbB2. (ErbB2 and ErbB3 together form a heterodimeric Nrg-1 receptor in Schwann cells.) These investigators used conditional loxP alleles of ErbB2, together with a Krox-20-Cre Schwann cell driver line, to inactivate the mouse ErbB2 gene in peripheral nerve Schwann cells around the time of birth. [This conditional inactivation circumvents the embryonic lethality of conventional ErbB2 knock-outs (13), which die from a failure of muscle trabeculation in the developing heart around embryonic day 10.5 (E10.5).] The more recent work of Michailov et al. (2) also exploited mouse mutants and mouse transgenics, but examined the effects of over- or underexpression of axonal Nrg-1 on the thickness of the myelin sheath elaborated by Schwann cells in contact with the axons; that is, on the overall size of the myelin membrane and the number of times that this membrane is wrapped around the axon. These authors compared the myelin sheaths of wild-type mice to those of Nrg-1+/− heterozygotes and of Nrg-1+/−*ErbB2+/− and Nrg-1+/−*ErbB2+/−*ErbB3+/− compound heterozygotes. They found that myelin thickness (measured in proportion to the caliber of myelinated axons) was equivalently reduced in all sets of mice. This observation suggests that it is axonal Nrg-1 itself that normally limits myelin production. Reductions in myelin thickness were, as expected, associated with reduced peripheral nerve conduction velocities. The authors also performed similar morphometric analyses in transgenic mice carrying a Nrg-1 construct driven by the mouse Thy 1.2 promoter. The peripheral nerves of these transgenic mice contain axons from Thy 1.2+ motor and dorsal root ganglion (DRG) neurons, and these axons therefore overexpress Nrg-1. The bottom line from all of these studies is that more Nrg-1 expression on the surface of axons at the time of myelination results in more robust myelination and thicker myelin, whereas less Nrg-1 results in less myelin (2).

The most recent work, by Taveggia and colleagues (3), suggests that axonal Nrg-1—specifically, the membrane-associated type III isoform (6)—controls not only the extent of myelin deposition but, more fundamentally, whether the process is even initiated in the first place. These authors first examined myelination in explant cultures of either DRGs or superior cervical ganglia (SCGs), prepared from either wild-type mice or from mice that specifically lacked the Nrg-1 type III isoform (14). The axons of DRG neurons from wild-type mice were readily myelinated when cocultured with wild-type Schwann cells, but the axons of DRG neurons from Nrg-1 type III–deficient mice were not, even when the mutants DRGs were cocultured with a fivefold excess of Schwann cells. Lentiviral expression of Nrg-1 type III in the mutant DRGs rescued myelination. (Many DRG axons are normally myelinated in vivo.) The axons of SCG neurons are normally smaller than those of DRG neurons and are not myelinated, either in vivo or when cultured in vitro together with Schwann cells. Taveggia et al. further demonstrated that lentiviral transduction of Nrg-1 type III onto the surface of SCG axons was sufficient to convert this nonmyelinated population of axons to one that is myelinated. Consistent with the observations of Michailov et al., Taveggia et al. also observed that the axons that are myelinated in vivo in mice that are heterozygous for the Nrg-1 type III mutation have abnormally thin myelin sheaths and reduced peripheral nerve conduction velocities. An important conclusion that may be made from the studies of both Michailov et al. and Taveggia et al., which is consistent with a large body of literature that extends back for a century, is that all Schwann cells are competent to form myelin, but it is axons that determine whether they will do so. The decision to myelinate is therefore one that is imposed on Schwann cells by the axons with which they interact, and it appears as though the level of Nrg-1 type III expression on the surface of axons is the key variable.

Together with an even larger body of earlier work, the results summarized above lead to the remarkable conclusion that Nrg-1, a protein originally purified as a mitogen for cultured Schwann cells (7), in fact drives nearly every feature of the forward differentiation of these cells (Fig. 1). These include the initial segregation of the Schwann cell lineage from other cell lineages that arise from the neural crest, the proliferation of immature Schwann cells as they migrate down and populate peripheral nerves, the directed migration of these cells along peripheral axons, and finally their ultimate differentiation as myelin-forming cells. At this point, it appears as though each of these events is triggered through activation of an ErbB2/ErbB3 heterodimer (Fig. 1). Clearly, there must be some element of developmental specification, presumably involving the expression or activation of specific intracellular transducers, that leads to different signaling outcomes at different points in Schwann cell differentiation. Little is known about these potentially distinct pathways of signal transduction, although myelination has been shown to depend on the activation of phosphatidylinositol 3-kinase, one of the "usual suspects" downstream of receptor protein tyrosine kinase activation, in Schwann cells (3).

Fig. 1.

Multiple roles for Nrg1 in Schwann cell development. Early neural crest–derived Schwann cell progenitors and committed immature Schwann cells both require axonally supplied Nrg-1 for proliferation, survival, and migration down peripheral axons. After birth, high levels of Nrg-1 are required to trigger the process of myelination. Lower levels, present on the surface of small sensory axons, lead to loosely ensheathed, unmyelinated axons.

The association of axon caliber with myelination (large, myelinated; small, unmyelinated) may also reflect the action of Nrg-1, because Taveggia and colleagues demonstrate that SCG neurons normally express appreciably lower levels of Nrg-1 than do DRG neurons. We can imagine that the yes-or-no decision with respect to myelination, which mouse Schwann cells make immediately after birth, may require a critical or triggering threshold of Nrg-1-ErbB signaling. And further, that a higher Nrg-1 density on the surface of DRG versus SCG axons, to which Schwann cells of a more-or-less constant size are exposed, results in a crossing of this threshold in the presence of DRG but not SCG axons. In this way, Nrg-1 signaling intensity, integrated across the area of contact between Schwann cells and axons, would, if above the threshold, trigger myelination, and at increasingly higher levels above this threshold, drive the elaboration of an increasingly thicker myelin sheath. It is possible that this sort of Nrg-1-ErbB signaling mechanism operates during myelination in the CNS, because oligodendrocytes also respond to Nrg-1. However, this remains to be seen.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
View Abstract

Stay Connected to Science Signaling

Navigate This Article