Neuron, Vol. 44, 609–621, November 18, 2004, Copyright 2004 by Cell Press Activated CREB Is Sufficient to Overcome Inhibitors in Myelin and Promote Spinal Axon Regeneration In Vivo associated glycoprotein (MAG), and oligodendrocyte- myelin glycoprotein (OMgp) (Filbin, 2003). Interestingly, MAG and OMgp, as well as at least one inhibitory domain of Nogo, interact with the same receptor complex to exert their inhibitory effect, suggesting redundancy Ying Gao, 1 Kangwen Deng, 1 Jianwei Hou, 1 J. Barney Bryson, 1 Angel Barco, 3 Elena Nikulina, 1 Tim Spencer, 1 Wilfredo Mellado, 1 Eric R. Kandel, 2,3 and Marie T. Filbin 1, * 1 Biology Department among the inhibitors. This receptor complex consists Hunter College of a GPI-linked protein named Nogo receptor (NgR) be- City University of New York cause it was originally identified as a binding partner 695 Park Avenue for a 66 amino acid sequence of Nogo, predicted to be New York, New York 10021 extracellular (Domeniconi et al., 2002; Fournier et al., 2 Howard Hughes Medical Institute 2001; Liu et al., 2002; Wang et al., 2002b). The transduc- 3 Center for Neurobiology and Behavior ing component of this receptor complex was identified College of Physicians and Surgeons as the neurotrophin p75 receptor (p75NTR) (Wang et al., of Columbia University 2002a; Wong et al., 2002). Recently, a third component 1051 Riverside Drive of this receptor complex has been described—a trans- New York, New York 10032 membrane protein called Lingo (Mi et al., 2004). As these three inhibitors all interact with at least one common receptor complex, the signal transduction pathway that Summary results in inhibition of axonal growth must also be shared by all inhibitors. It follows, then, that activation of a Inhibitors in myelin play a major role in preventing parallel pathway that blocks the signaling of one inhibi- spontaneous axonal regeneration after CNS injury. El- tor would do the same for all three inhibitors. evation of cAMP overcomes this inhibition, in a tran- Previously, we (Cai et al., 1999) and others (Song et scription-dependent manner, through the upregula- al., 1998) found that when cAMP is elevated, either with tion of Arginase I (Arg I) and increased synthesis of analogs or by priming neurons with a variety of neuro- polyamines. Here, we show that the cAMP effect re- trophins, inhibition and repulsion by both MAG and by quires activation of the transcription factor cAMP re- myelin in general are overcome. Importantly, elevation sponse element binding protein (CREB) to overcome of cAMP by direct injection of db-cAMP into the cell myelin inhibitors; a dominant-negative CREB abol- body of dorsal root ganglion (DRG) neurons is sufficient ishes the effect, and neurons expressing a constitu- to induce regeneration of subsequently lesioned spinal tively active form of CREB are not inhibited. Activation axons that emit from the same cell body as was injected of CREB is also required for cAMP to upregulate Arg (Neumann et al., 2002; Qiu et al., 2002). Without elevation I, and the ability of constitutively active CREB to over- of cAMP, these axons, like all spinal cord axons, fail come inhibition is blocked by an inhibitor of polyamine to regenerate. The question then becomes how cAMP synthesis. Finally, expression of constitutively active induces axons to grow through an inhibitory environ- CREB in DRG neurons is sufficient to promote regen- ment. We showed that both the db-cAMP and neuro- eration of subsequently lesioned dorsal column axons. trophin effect in overcoming inhibition are each tran- These results indicate that CREB plays a central role scription dependent. One gene that is upregulated in in overcoming myelin inhibitors and so encourages response to either treatment is for the enzyme Arginase regeneration in vivo. I(Arg I), a key enzyme in the synthesis of polyamines. Indeed, overexpressing Arg I or addition of the poly- Introduction amine putrescine to neurons in culture is each sufficient to overcome inhibition by MAG and myelin (Cai et al., The lack of CNS regeneration in adult mammals is largely 2002). The steps between elevation of cAMP (activation attributable to the presence of an inhibitory environ- of PKA) and upregulation of Arg I and polyamines are ment. In general, there have been two approaches to not known. encourage regeneration in vivo—change the environ- In many other situations, activation of the transcription ment by blocking/neutralizing inhibitors of regeneration factor cAMP response element binding protein (CREB) (Bradbury et al., 2002; Bregman et al., 1995; GrandPre is a requirement not only for upregulation of a variety et al., 2002), or change the intrinsic growth state of of genes in response to elevated cAMP, but also in the neuron/axon such that it no longer responds to an response to a variety of signal transduction pathways, inhibitory environment (Lehmann et al., 1999; Neumann not necessarily involving cAMP (Lonze and Ginty, 2002). et al., 2002; Qiu et al., 2002). In an injured CNS, an To date, at least six CREB-like molecules have been axon that attempts regeneration faces both inhibitors of identified, and they act by forming homo- or heterodim- regeneration in myelin and a physical barrier presented ers, before binding to CRE sequences to regulate tran- by a glial scar (Filbin, 2003; Schwab and Bartholdi, 1996; scription (Mayr and Montminy, 2001; Montminy, 1997; Silver and Miller, 2004). To date, three inhibitors of re- Shaywitz and Greenberg, 1999). More than 100 genes generation have been identified in myelin-Nogo, myelin- have been reported to be upregulated following activa- tion of either CREB or a closely related protein (Lonze and Ginty, 2002). Of these, different repertoires of gene *Correspondence: filbin@genectr.hunter.cuny.edu