Reprinted from Manitoba Medicine, Vol. 63, No. 3, 1993, pp. 87-89...

Membrane-associated protein factors which inhibit central nervous system regeneration

KW Cheng, PhD

Spinal Cord Research Centre, Neuroscience Research Program, Department of Physiology, University of Manitoba

In higher vertebrates, lesions in the central nervous system (CNS) are irreversible due to the almost complete lack of regenerative growth from the injured axons. In obvious contrast, axons of the peripheral nervous system (PNS) regenerate well. There seem to be no obvious differences, however, between neurons from the CNS and PNS in their ability to grow back after injury. Aguayo and colleagues1 in Montreal bridged lesions in the spinal cord of rats with a piece from peripheral (sciatic) nerve; after a short time, the sciatic nerve explant was invaded from both sides by regenerating neurites that eventually bridged the injury site, but function was not restored because the neurites stopped growing after re-entering the spinal cord. These experiments showed that motor neurons with cell bodies in the spinal cord will re-establish connections in the periphery but not within the CNS, indicating that CNS neurons could regenerate in a favourable environment.

The poor regenerative capacity of higher vertebrate CNS might be explained in one of two ways: either certain growth-promoting substances are missing, or there are active inhibitors of neurite extension. For example, Schwann cells in peripheral nerves are surrounded by basement membranes containing laminin. This extracellular matrix protein is the most potent substrate for neurite growth, and it acts synergistically with neurotrophic factors. Since laminin is virtually absent from the adult CNS of higher vertebrates, it has been argued that the pattern of expression of laminin might determine whether regeneration can occur. In addition, Schwann cells in peripheral nerves produce a variety of neurotrophic factors and even increase this production after denervation. The hypothesis that there is a difference in trophic factor production between the PNS and CNS being responsible for their different regeneration capabilities seems plausible. However, identified neurotrophic factors such as nerve growth factor, brain-derived neurotrophic factor, ciliary neurotrophic factor, and fibroblast growth factor are present in the adult CNS. Furthermore, increases in laminin as well as neurotrophic factors have been found at central lesion sites. Why, then, are the re-expressed neurotrophic factors unable to trigger functional regeneration as in the PNS? Recent studies by several research groups have suggested that neurite outgrowth may be actively inhibited by the central glial cells via an inhibitory mechanism.




Figure 1 Isolation and characterization of rat central nervous system (CNS) plasma membrane. A Flow chart for the isolation of plasma membranes and myelin from adult rat spinal cord or brain tissues by differential and density gradient centrifugation. B Fractionation of myelin and plasma membranes upon sucrose density gradient centrifugation. C Analysis of total (1,2) and sonication-solubilized (3,4)proteins of purified spinal cord myelin (1,3) and plasma membranes (2,4) by 10% sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE). Myelin basic proteins (MBP) 18K, 17K and 14K are the characteristic proteins of rat CNS myelin

Oligodendrocytes in the CNS Inhibit Neurite Outgrowth

Using cell cultures, Schwab and associates2 at the University of Zurich, Switzerland, observed that cultured neurons from both the CNS and the PNS extended neurites through a sciatic nerve explant, but failed to invade a similar explant from optic nerve. The two explants differed in their myelin producing glial cells, ie, Schwann cells in the sciatic nerve and oligodendrocytes in the optic nerve. It is speculated that oligodendrocytes inhibit regeneration within CNS fibre tracts. Further experiments by co-culturing central or peripheral neurons with astrocytes, immature oligodendrocytes and mature oligodendrocytes showed that both CNS and PNS neurons attached to most cells, except for one type of glial cell. The web of growing neurites soon formed `windows' around the highly branched mature oligodendrocytes. When frozen sections of spinal cords were used as culture substrata, cells adhered mostly to grey matter, indicating that the CNS white matter is a highly nonpermissive substratum. Similar observations have been reported by several other investigators. These findings indicate that oligodendrocytes of the CNS actively inhibit neurite outgrowth by a contact-mediated mechanism.

Myelin-Associated Neurite Growth Inhibitors

Because oligodendrocytes are the myelin-producing cells of the CNS, myelin was examined for its property as a neuronal substratum and as a source of inhibitory components. CNS myelin inhibitory activity is membrane-bound and associated with the protein fraction of CNS myelin. This inhibitory substrate activity could be recovered after separation in sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) as two minor myelin-associated proteins with relative molecular masses (Mr) of 35 kDa and 250 kDa, now called neurite growth inhibitors NI-35 and NI-250, respectively.2 These have been shown to be potent inhibitors of neurite outgrowth; an inclusion of small amounts of these proteins was sufficient to convert a neutral substrate into a nonpermissive one, and addition of one or both of these proteins to favourable substrata, such as peripheral nerve myelin, resulted in inhibition of neurite outgrowth. However, the inhibitory activity could be neutralized by specific antibodies raised against these protein inhibitors.

Plasma Membrane Associated Growth Cone Collapsing and Neurite Growth Inhibitory Proteins

Similar, but not identical, membrane-bound inhibitory proteins have been identified in neural and other tissues, and reported to induce collapse of axonal growth cones and thus retraction of neurites. Using time-lapse video analysis, Kapfhammer and Raper3 of the Max-Planck Institute, Germany, have observed that a few minutes after contact between the filopodia of a growing retinal axon and a sympathetic axon, the retinal growth cone thickened and shortened its filopodia and the neurite retracted. Plasma membranes from chick embryonic brain have been found to contain components that cause collapse of growth cones of dorsal root ganglion (DRG) neurons in culture. Furthermore, membranes prepared from the chick posterior optic tectum have been shown to collapse growth cones of axons from temporal retina explants. Two membrane glycoproteins of Mr 48 kDa and 55 kDa from chick embryonic posterior sclerotome have been observed to induce collapse of axonal growth cones of DRG neurons in culture. Results of all these studies by various investigators are consistent with the hypothesis that growth cone motility is inhibited by specific membrane-associated proteins in the developing nervous system.


Figure 2 Effects of precoating sonication-solubilized proteins from adult rat spinal cord plasma membranes on primary cultures of fetal rat spinal cord neurons. Dissociated fetal rat spinal cord neuronal cells were plated onto dishes precoated with solubilized plasma membrane proteins A or control dishes B, and cultured for two weeks in minimal essential medium containing 5% heat-inactivated horse serum

In parallel with these studies of membrane-bound neurite growth inhibitory proteins, we have recently demonstrated that the plasma membrane of embryonic chick spinal cord undergoes a developmental transition from permissive to nonpermissive substrates for neuritogenesis, and that the transition period occurs around embryonic day 13 of the 21-day developmental period.4 Cell surface plasma membranes were prepared from homogenates of embryonic chick spinal cord segments by our established procedure for rat CNS tissues, as outlined in Figure 1A, and fractionated from myelins by sucrose density gradient centrifugation (Figure 1B). The plasma membrane proteins were solubilized by ultrasonication and subjected to an in vitro assay using clonal NG108-15 cells to monitor permissive and nonpermissive substrates. The chick spinal cord of early embryonic days (eg, embryonic day 10) was highly permissive, and the permissiveness decreased with development as the spinal cord and brain of late embryonic chicks became highly nonpermissive.4 Recently, in our experimentation on mammalian CNS, we have observed that plasma membranes from the brain and spinal cord of newborn and adult rats were highly nonpermissive substrates for cell adhesion and neurite outgrowth. When cultured on dishes precoated homogeneously with solubilized proteins from adult rat spinal cord plasma membranes, primary fetal rat spinal cord neuronal cells remained very loosely adhesive, forming large `windows' between neuronal aggregates (Figure 2A), compared with the control of a cell monolayer covering the whole surface (Figure 2B). This substrate inhibitory activity in plasma membranes has been observed to be substantially higher than that of the myelin fraction, suggesting that CNS cell-surface plasma membrane is a significant cellular source of neurite growth inhibitory proteins. Our finding of plasma membrane-associated inhibitory proteins appears to differ from that of Schwab and associates2 on the myelin-associated inhibitors NI-35 and NI-250, which are found in the myelin fraction. The protein content of our rat spinal cord plasma membrane preparation was characterized by SDS-PAGE to contain major proteins of Mr 40 to 70 kDa (Figure 1C), significantly different from that of the myelin, which is characterized by the small Mr 14 to 18 kDa myelin basic proteins. In addition, from our preliminary data, this plasma membrane inhibitory principle appears to be an acidic protein of Mr 50 to 70 kDa. However, its molecular structure and biological role in neuritogenesis remain to be elucidated.


Recent findings indicate that a fine balance between neurite outgrowth stimulators and inhibitors is essential for the rate and direction of neurite extension. The plasma membrane is not simply a passive surface, but functions together with other growth factors and cell adhesion molecules to serve as a permissive `gating' mechanism for the function of contact-dependent interactions. Myelin-associated neurite growth inhibitors, plasma membrane-associated growth cone collapsing and nonpermissive substrate molecules could play important roles in the spatial restriction of growth and plasticity in the adult brain and spinal cord, thereby exerting a stabilizing function for the differentiated CNS. In view of these recent findings of various neurite outgrowth inhibitors, it is not difficult to speculate a family or families of cell type-specific axonal growth-inhibiting regulators. Ongoing research on molecular mechanisms modulating growth cone behaviour and neurite extension is essential not only to understand the process of neuronal differentiation, but also to define the essential conditions necessary for the regeneration of nerve fibres after spinal cord injury.


1. David S Aguayo AJ Axonal elongation into peripheral nervous system `bridges' after central nervous system injury in adult rats Science 1981214 931-3 

2. Schwab ME Myelin-associated inhibitors of neurite growth and regeneration in the CNS Trends Neurosci 199013 452-6   

3. Walter J Allsopp TE Bonhoeffer F A common denominator of growth cone guidance and collapse? Trends Neurosci 1990 11 447-52 

4. Ethell DW Steeves JD Jordan LM Cheng KW Developmental transition by spinal cord plasma membranes of embryonic chick from permissive to restrictive substrates for the morphological differentiation of neuroblastoma X glioma NG108-15 cell Dev Brain Res 1993 72 1-8  

SCRC WWW administrator:
Reprinted from Manitoba Medicine, Vol. 63, No. 3, 1993, pp. 87-89.
Copyright © 1993, Faculty of Medicine, University of Manitoba.
Back to articles about SCRC