Portal de Periódicos da CAPES

Nogo Domains and a Nogo Receptor:Implications for Axon Regeneration

2001; Cell Press; Volume: 30; Issue: 1 Linguagem: Inglês

10.1016/s0896-6273(01)00258-6

1097-4199

Perry A. Brittis, John G. Flanagan,

Neurogenesis and neuroplasticity mechanisms

Diagnosis: You should say of him: "One having a crushed vertebra in his neck; he is unconscious of his two arms, his two legs, he is speechless. An ailment not to be treated." –Edwin Smith Surgical Papyrus, c. 2500–1600 B.C. Unfortunately, little has changed clinically in the several thousand years since this anonymous author gave us the first known descriptions of the tragic consequences of central nervous system (CNS) injury, and the limited therapeutic options (Breasted 1930Breasted J.H Edwin Smith Surgical Papyrus. University of Chicago Oriental Institute Publications, Chicago1930Google Scholar). Today, it is estimated that more than 250,000 Americans have spinal cord injuries, with 11,000 new cases every year (Berkowitz et al. 1998Berkowitz M O'Leary P.K Kruse D.L Harvey C Spinal Cord Injury. Demos Medical Publishing, New York1998Google Scholar). When axonal connections are damaged in the adult brain or spinal cord, they show an extremely limited ability to regenerate, even though axons can grow and regenerate efficiently in the embryonic CNS, and in the adult peripheral nervous system. What are the reasons for this selective shutdown of regeneration in the adult CNS? Factors that may play a role can be grouped in two categories: intrinsic properties of CNS neurons that may make them incapable of regenerating, or extrinsic factors in the environment that may influence axon growth positively or negatively (reviewed by Horner and Gage 2000Horner P.J Gage F.H Nature. 2000; 407: 963-970Crossref PubMed Scopus (661) Google Scholar). Here, we focus on extrinsic negative factors, particularly recent studies of the inhibitory molecule Nogo. After many years of remarkable studies showing effects on axon growth in vitro and even partial recovery of function in rodent models of spinal cord injury after treatment with anti-Nogo antibodies (Bregman et al. 1995Bregman B.S Kunkelbagden E Schnell L Dai H.N Gao D Schwab M.E Nature. 1995; 378: 498-501Crossref PubMed Scopus (637) Google Scholar), Nogo was cloned last year by three groups, in the labs of Martin Schwab, Stephen Strittmatter, and Frank Walsh (Chen et al. 2000Chen M.S Huber A.B van der Haar M.E Frank M Schnell L Spillmann A.A Christ F Schwab M.E Nature. 2000; 403: 434-439Crossref PubMed Scopus (323) Google Scholar, GrandPre et al. 2000GrandPre T Nakamura F Vartanian T Strittmatter S.M Nature. 2000; 403: 439-444Crossref PubMed Scopus (986) Google Scholar, Prinjha et al. 2000Prinjha R Moore S.E Vinson M Blake S Morrow R Christie G Michlovich D Simmons D.L Walsh F.S Nature. 2000; 403: 383-384Crossref PubMed Scopus (532) Google Scholar). A surprising aspect of these initial studies is that the different groups identified inhibitory activity in two entirely different domains of Nogo. A recent study from the Strittmatter group now extends the characterization of these two domains and for the first time identifies a receptor that can mediate Nogo activity (Fournier et al. 2001Fournier A.E GrandPre T Strittmatter S.M Nature. 2001; 409: 341-346Crossref PubMed Scopus (917) Google Scholar). These studies bring a promising new molecular focus to the study of CNS regeneration. The idea that factors in the CNS environment could prevent regeneration dates to the early 20th century. Tello showed in 1911, as later described by Ramon y Cajal 1928Ramon y Cajal S Degeneration and Regeneration of the Nervous System. Hafner, New York1928Google Scholar, that the inability of adult CNS neurons to extend axonal processes could be overcome by giving them the permissive environment of a peripheral nerve. This observation was extended by impressive studies showing that retinal neurons can form long projections in peripheral nerve grafts (David and Aguayo 1981David S Aguayo A.J Science. 1981; 214: 391-393Crossref Scopus (1412) Google Scholar). These results show that failure to regenerate is not purely an intrinsic deficit of CNS neurons, and is blocked by the CNS environment. Two main sites have been considered for the location of CNS factors that might inhibit axon regeneration (Figure 1). One site is the scar that forms at the region of injury. Following CNS injury the central area of necrosis is infiltrated by glia and other nonneuronal cells, and a fibrous scar forms. Axons do not extend through the scar and appear to be inhibited by it, with the axon tips forming club-like structures that can remain in place for months or even years. Molecular components that may contribute to this inhibitory activity include chondroitin sulfate proteoglycans (CSPG), tenascin, and semaphorin-3A, which are upregulated in the region of scarring and are inhibitory to axon growth in culture (Letourneau et al. 1994Letourneau P.C Condic M.L Snow D.M J. Neurosci. 1994; 14: 915-928Crossref PubMed Google Scholar, Davies et al. 1999Davies S.J.A Goucher D.R Doller C Silver J J. Neurosci. 1999; 19: 5810-5822Crossref PubMed Google Scholar, Pasterkamp et al. 1999Pasterkamp R.J Giger R.J Ruitenberg M.J Holtmaat A De Wit J De Winter F Verhaagen J Mol. Cell. Neurosci. 1999; 13: 143-166Crossref PubMed Scopus (267) Google Scholar, and references therein). Moreover, glial scar tissue placed in culture can be converted to a permissive substrate by enzymatic removal of glycosaminoglycans, supporting the idea that CSPG is an important component. It is worth bearing in mind that proteoglycans bind many other molecules, and their role could be as a scaffold that presents molecular cues to the responding cell. The other main proposal is that inhibitors would be broadly distributed in the myelin that ensheaths axons in white matter tracts of the adult CNS. Supporting this idea, the loss of regeneration potential during development correlates roughly with the onset of myelination. Moreover, myelin carpets or oligodendrocytes, the cells that produce CNS myelin, are poor substrates for axon outgrowth in vitro. Important experiments by Schwab and colleagues showed that myelin can be converted to a more permissive in vitro axon substrate by addition of the monoclonal antibody IN-1, providing persuasive evidence for inhibitors in myelin (Caroni and Schwab 1988Caroni P Schwab M.E Neuron. 1988; 1: 85-96Abstract Full Text PDF PubMed Scopus (744) Google Scholar). IN-1 was raised against a purified 250 kilodalton protein named NI-250, a neurite inhibitory activity of CNS myelin, and also recognizes a smaller inhibitory protein, NI-35, found in rat but not bovine or human myelin. Most excitingly, IN-1 treatment of rats with spinal cord injury caused an improvement in regeneration of corticospinal axons, as well as enhanced recovery of behavioral function (Bregman et al. 1995Bregman B.S Kunkelbagden E Schnell L Dai H.N Gao D Schwab M.E Nature. 1995; 378: 498-501Crossref PubMed Scopus (637) Google Scholar). Molecular characterization of the components of myelin inhibitory activity has been a major goal. One such component may be myelin associated glycoprotein (MAG), which is an inhibitor of neurite outgrowth in culture, although there have been varying reports of the effect of MAG gene disruption on regeneration in vivo (reviewed by Filbin 1996Filbin M.T Mol. Cell. Neurosci. 1996; 8: 84-92Crossref Scopus (39) Google Scholar). In view of the actions of IN-1 antibody in spinal cord regeneration, molecular cloning of the key molecule or molecules it recognizes was eagerly awaited. An important advance last year came with the description of cDNA clones encoding a protein, termed Nogo, by three groups (Chen et al. 2000Chen M.S Huber A.B van der Haar M.E Frank M Schnell L Spillmann A.A Christ F Schwab M.E Nature. 2000; 403: 434-439Crossref PubMed Scopus (323) Google Scholar, GrandPre et al. 2000GrandPre T Nakamura F Vartanian T Strittmatter S.M Nature. 2000; 403: 439-444Crossref PubMed Scopus (986) Google Scholar, Prinjha et al. 2000Prinjha R Moore S.E Vinson M Blake S Morrow R Christie G Michlovich D Simmons D.L Walsh F.S Nature. 2000; 403: 383-384Crossref PubMed Scopus (532) Google Scholar). All three relied on peptide sequences derived from the purified bovine homolog of NI-250 (Spillmann et al. 1998Spillmann A.A Bandtlow C.E Lottspeich F Keller F Schwab M.E J. Biol. Chem. 1998; 273: 19283-19293Crossref PubMed Scopus (135) Google Scholar), allowing isolation of clones from cDNA libraries, or identification of nucleic acid sequences in the Genbank database. The cDNAs identified in these studies encode three Nogo isoforms, designated Nogo-A, -B, and -C, presumed to be generated by alternative RNA splicing or promoter usage of a single gene (Figure 2). Nogo-A appears to be NI-250, and, although it has not been directly tested, the size of Nogo-B suggests it might potentially be NI-35. In addition to unique regions, the A, B, and C isoforms share a common carboxy-terminal region. No conventional amino-terminal secretion signal peptide was found in any of the isoforms. However, the common region does have two long hydrophobic stretches that are likely to form transmembrane domains. This common region is homologous to a family of three proteins called reticulons, because of their prominent localization in the endoplasmic reticulum (ER). Although Nogo is likewise found prominently in the ER, it is also present on cell surfaces. The functions of the previously described reticulon proteins are unknown, although it will now be interesting to see if they have biological roles analogous to Nogo. Several properties of Nogo confirm its role as a nerve outgrowth inhibitor in CNS myelin. It is found in CNS white matter, including the inner and outer leaflets of myelin, and in cultured oligodendrocytes, and it is recognized by antibody IN-1. It causes growth cone collapse and inhibition of neurite outgrowth from neurons that are sensitive to CNS myelin, and new antibodies raised against Nogo-A neutralize much of the in vitro inhibitory activity of CNS myelin. Thus, in vitro gain- and loss-of-function experiments, as well as expression patterns, all support a role for Nogo as an important component of CNS myelin inhibitory activity. Other important features of Nogo were unexpected and are still not well understood. Two distinct inhibitory domains have been described. Somewhat confusingly, the initial reports on cloned Nogo focused on actions of different domains. Inhibitory activity was found in a recombinant fragment of Nogo-A amino-terminal to the first hydrophobic domain (amino-Nogo) (Prinjha et al. 2000Prinjha R Moore S.E Vinson M Blake S Morrow R Christie G Michlovich D Simmons D.L Walsh F.S Nature. 2000; 403: 383-384Crossref PubMed Scopus (532) Google Scholar). Furthermore, much of the in vitro inhibitory activity of myelin could be blocked by an antibody raised against a peptide from the sequence unique to Nogo-A (Chen et al. 2000Chen M.S Huber A.B van der Haar M.E Frank M Schnell L Spillmann A.A Christ F Schwab M.E Nature. 2000; 403: 434-439Crossref PubMed Scopus (323) Google Scholar). These results implicated the unique amino-terminal domain of Nogo-A as an inhibitory region. But in the other report, the short 66 residue region (Nogo-66) located between the two hydrophobic domains was shown to have inhibitory activity, implicating an entirely distinct region that is found in all three Nogo isoforms (GrandPre et al. 2000GrandPre T Nakamura F Vartanian T Strittmatter S.M Nature. 2000; 403: 439-444Crossref PubMed Scopus (986) Google Scholar). The latest study from the Strittmatter group directly compares soluble recombinant amino-Nogo and Nogo-66 fragments, confirming that both do have independent inhibitory activity (Fournier et al. 2001Fournier A.E GrandPre T Strittmatter S.M Nature. 2001; 409: 341-346Crossref PubMed Scopus (917) Google Scholar). The study also begins to address differences between them. Soluble amino-Nogo with an epitope tag required antibody clustering for activity, whereas Nogo-66 tagged with glutathione S-transferase (GST) did not. Although this suggests a possible molecular difference (Fournier et al. 2001Fournier A.E GrandPre T Strittmatter S.M Nature. 2001; 409: 341-346Crossref PubMed Scopus (917) Google Scholar), GST tags can dimerize, so it remains uncertain whether Nogo-66 can act as a free monomer. In tests of specificity for cellular targets, in addition to effects on neurons, amino-Nogo inhibited fibroblast spreading, whereas Nogo-66 did not. Since amino-Nogo acts on diverse cell types and has an unusual amino acid composition rich in prolines and negative charge, it might act by a mechanism independent of a specific receptor, or alternatively it may have a receptor located on diverse cell types. Another question arising from the initial studies is membrane topology. In oligodendrocytes or transfected cells, the Nogo-66 region was detected on intact cells, whereas the amino- and carboxy-termini were detected in transfected cells only after permeabilization, suggesting a model where only the short Nogo-66 loop is exposed on the surface (GrandPre et al. 2000GrandPre T Nakamura F Vartanian T Strittmatter S.M Nature. 2000; 403: 439-444Crossref PubMed Scopus (986) Google Scholar). This model would probably be the simplest to reconcile with the presence of two transmembrane domains, since cotranslational translocation of the growing polypeptide could be initiated by the first transmembrane domain and then terminated by the second. If the amino-Nogo domain were to be translocated to the surface, this would presumably require an unconventional mechanism such as a cryptic internal signal sequence. Alternatively, release of amino-Nogo-A may require lysis of oligodendrocytes, which could occur at sites of CNS injury. Based on the current evidence, the two domains might have different functions, even intracellular functions, or they might act synergistically. Further studies of Nogo topology, for example by determining which sites are exposed for glycosylation in the ER lumen, should help to understand the actions of amino-Nogo and the other Nogo domains. An important advance in understanding the mechanism of action, and the cellular targets, of any ligand is the identification of a cognate receptor. Work on Nogo-66 therefore took a big step forward with the recent identification by Strittmatter's group of a Nogo-66 receptor (Fournier et al. 2001Fournier A.E GrandPre T Strittmatter S.M Nature. 2001; 409: 341-346Crossref PubMed Scopus (917) Google Scholar). They set out to identify this receptor by the alkaline phosphatase (AP) fusion protein approach (Flanagan and Cheng 2000Flanagan J.G Cheng H.J Meth. Enzymol. 2000; 327: 198-210Crossref PubMed Scopus (66) Google Scholar). In initial studies, an AP-Nogo-66 fusion protein was demonstrated to give saturable high-affinity binding to neurons, with a dissociation constant of 3 nM. This fusion protein also acted as a growth cone collapse agent, with a half-maximal response at 1 nM, consistent with the binding affinity. The AP-Nogo-66 fusion protein was then used to screen pools of a mouse brain cDNA expression library transfected into COS cells, resulting in identification of a cDNA that conferred high-affinity binding activity. The encoded protein was named Nogo-66 receptor (NgR; Figure 2). To assess whether Nogo-66 and NgR interact directly, myc-tagged NgR was tested for binding to GST-tagged Nogo-66 in cell extracts, confirming that the two form a protein complex, although the presence of additional components is not ruled out. Moreover, early embryonic chick retinal neurons, which are not normally sensitive to Nogo-66, become responsive upon infection with a viral vector encoding NgR cDNA. Taken together, these results provide convincing evidence that NgR is a functional cell surface receptor that can mediate inhibitory effects of Nogo-66. The cDNA sequence for NgR encodes a protein 473 amino acids in length, with a conventional amino-terminal translocation signal sequence. This is followed by eight leucine-rich repeat (LRR) motifs, and an LRR carboxy-terminal motif (LRRCT); sequence motifs found in a variety of cell surface and secreted molecules (Figure 2). A likely human ortholog of mouse NgR was found, but so far no other closely homologous sequences that would suggest a family of related receptors. At the C terminus is a signal for addition of a glycosylphosphatidylinositol (GPI) lipid, which was confirmed by enzymatic cleavage to anchor NgR in the membrane. By analogy with other GPI-anchored receptors, it is likely that NgR associates with a separate transmembrane signal-transducing polypeptide, but this remains to be addressed in future studies. Is NgR expressed in a pattern consistent with a role in inhibition of CNS regeneration? A Northern blot survey of tissues showed expression predominantly in brain, consistent with the in vitro assays showing selective action of Nogo-66 on neurons and not fibroblasts. By in situ RNA hybridization, NgR is expressed in a wide variety of CNS neurons, including cerebral cortical neurons, hippocampal neurons, cerebellar Purkinje cells, and pontine neurons. This includes the cerebral cortex pyramidal neurons whose regeneration is enhanced by IN-1 treatment of injured rat spinal cord and the cerebellar Purkinje neurons whose sprouting is enhanced by antibodies to Nogo-A. An antibody to NgR detects the protein on axons in immunolocalization studies of cultured embryonic spinal neurons. And finally, the antibody detects NgR prominently in cultures of late embryonic (day 13) chick DRG neurons, which are responsive to Nogo, but little or none in DRG or retinal neurons of earlier embryonic stages, which are not responsive. These expression patterns seem to fit well with expectations for a receptor mediating Nogo inhibition of axon outgrowth. Since NgR is a functional receptor for Nogo-66, these studies also support the idea that the Nogo-66 domain is a strong candidate as an inhibitor of axon growth in the CNS. The identification of Nogo and NgR represent important advances in understanding axon inhibition by CNS myelin and provide a new molecular basis to study regeneration. In vivo gain- and loss-of-function studies, including gene targeting, of Nogo isoforms and NgR will be critical to unravel their functions. In addition to testing effects on recovery from injury, it will be fascinating to investigate the normal function of Nogo, which has been speculated to prevent excessive neural plasticity in the adult CNS. The relative contributions to regenerative failure of the glial scar versus myelin inhibitors has been a long-standing subject of debate (see Figure 1). On the one hand, it is clear that IN-1 antibody can block myelin inhibitory activities in culture and can promote recovery in lesioned animals. Moreover, impressive regeneration of up to 50% of corticospinal fibers was recently reported by David and colleagues in a novel approach involving vaccination of mice with myelin (Huang et al. 1999Huang D.W McKerracher L Braun P.E David S Neuron. 1999; 24: 639-647Abstract Full Text Full Text PDF PubMed Scopus (269) Google Scholar). These studies suggest a key role for myelin inhibitors of regeneration. On the other hand, elegant experiments by Silver and colleagues have shown that when adult neurons are carefully microtransplanted without creating a glial scar, their axons can regenerate long distances over heavily myelinated adult CNS tracts. Although it might be argued that the transplanted axons did not come in contact with the inhibitory inner myelin leaflet, they can even grow through areas of Wallerian degeneration and stall only after encountering a gliotic barrier. These results suggest that general myelin inhibitors may not be as potent in preventing regeneration as previously thought (Davies et al. 1997Davies S.J.A Fitch M.T Memberg S.P Hall A.K Raisman G Silver J Nature. 1997; 390: 680-683Crossref PubMed Scopus (654) Google Scholar, Davies et al. 1999Davies S.J.A Goucher D.R Doller C Silver J J. Neurosci. 1999; 19: 5810-5822Crossref PubMed Google Scholar, and references therein). Although these various studies may initially seem contradictory, one possibility is that rather than promoting only direct regeneration of damaged pathways, neutralization of myelin inhibitors may promote functional recovery by freeing up CNS plasticity and thus allowing formation of new pathways that can compensate for the lost ones (Figure 1C; Z'Graggen et al. 2000Z'Graggen W.J Fouad K Raineteau O Metz G.A.S Schwab M.E Kartje G.L J. Neurosci. 2000; 20: 6561-6569PubMed Google Scholar). Alternatively, the studies might be reconciled if some of the same molecules are general myelin inhibitors and also glial scar inhibitors. The new molecular tools should help answer these questions. Moving toward therapeutic approaches for CNS injury is a complex challenge. In addition to the known inhibitory components, molecules initially identified as developmental guidance factors such as semaphorins, ephrins, netrins, and slits probably play a role in the adult CNS. Especially considering the large number of candidate inhibitors, rather than blocking them extracellularly, it may ultimately be more effective to modulate intracellular pathways of axon responsiveness. Successful regeneration depends on additional factors such as survival of injured neurons and establishment of functional connections. On the positive side, even limited functional improvement may be valuable clinically. While even this is likely to take years, the new molecular studies give grounds for cautious hope. We may at last be in an era when CNS injury will become an ailment to be treated.