NINDS Workshop on Re-establishing Connectivity in the Damaged Spinal Cord
January 18-19, 2001
Pooks Hill Marriott, Bethesda, MD
Chaired by Albert Aguayo, Arlene Chiu and Josh Sanes
Table of Contents
Introduction
Inducing adult neurons to grow axons again
Changing the environment around regenerating axons
Change response of neurons to non-permissive environment
Reforming connections - Interactions between the neuron and the target
Making the most of what remains after injury
Recommendations from the participants
Collaborations between groups and investigators
Development of needed research tools for the community
Imaging
Injury models and outcome measures
Introduction
Injury to the spinal cord initiates a complex series of events that has devastating consequences such as partial or complete
paralysis, muscle atrophy, compromised ability to breathe, lack of bladder and bowel control, sexual dysfunction and chronic
pain. The quarter of a million patients with spinal cord injury (SCI) in the United States require specialized care to deal
with these problems every day for the rest of their lives. A major cause of dysfunction arises from damage to nerve fibers
passing through the injury site or lesion. We are dependent on long fibers that traverse the length of the cord to coordinate
messages between the brain and other parts of the body. Interruption of these fibers can disrupt the voluntary control of
movement and the experience of sensation. In developing treatments for SCI, a central question is how to restore connectivity
by promoting the regeneration of axons in the environment of the damaged nervous system and the establishment of new and appropriate
synaptic contacts within the spinal cord.
To address this question, this workshop focused on "bridging" damaged tissue, encouraging re-growth of axons across the lesion,
and facilitating reconnection to synaptic targets beyond the site of injury. However, it was recognized that complementary
efforts to limit cell death, reduce inflammation-related damage, optimize the function of surviving neurons, and train intact
local circuits would all contribute to functional recovery from SCI.
Presentations and discussions revolved around the following topics:
Inducing adult neurons to grow axons again
Damaged neurons in the adult mammalian central nervous system (CNS) typically do not regenerate axons in vivo. In experimental
situations, it has been observed that injuring a nerve more than once leads to more robust regeneration after the second injury.
However impractical it may be in a clinical setting, the phenomenon of a pre-injury, termed a "conditioning lesion", seems
to prime the nervous system, and suggests that intracellular conditions can be changed so that neurons are more in a growth
state. Several growth-associated molecules are "switched on" following a conditioning lesion. GAP43 and CAP23 are growth cone
membrane proteins that are up-regulated by peripheral nerve injury and whose expression is correlated with regenerative potential.
Although over-expression of either protein alone does not result in axonal regeneration, both proteins together in transgenic
mice mimic the effects of a conditioning lesion and lead to higher numbers of regenerating axons in the spinal cord. DNA microarrays
have also been used to identify several thousand genes that are regulated in response to injury. The challenge now is to elucidate
the function of these genes, and to identify the ones that are crucial for spinal cord repair.
Top
Changing the environment around regenerating axons
Getting axons to grow across the cavity or "cyst" that often develops at the site of injury is a major challenge in SCI, and
many strategies have been proposed to bridge this gap. Cells and substances that have been tested as bridge material include
Schwann cells, olfactory ensheathing glia, peripheral nerve segments, collagen embedded with neurotrophic or growth factors,
and a variety of stem/progenitor cells. Although many of these bridges and scaffolds have supported axon growth, the axons
tended to stay within the bridge and do not extend beyond it to innervate targets on the other side of the lesion. To entice
axons out of the bridge, it may be possible to introduce or genetically engineer in factors such as neurotrophins into the
distal side of the host cord to lead axons onward.
Scar tissue surrounding the injury site is well known to inhibit axonal growth. Some of this activity is caused by accumulations
of extracellular matrix such as chondroitin sulfate proteoglycans. When growth cones contact this inhibitory barrier, they
remain motile and can persist for months, but make no further progress along the cord. However, the long-term persistence
of these dystrophic endings raises the possibility that axons might resume growing should conditions around them improve and
become more permissive, even at times long after the initial injury.
One direct approach to optimize the environment was to replace the injured tissue with transplants of fetal rat spinal cord.
When these were grafted into newborn rats, axons regenerated through the transplant and into the host cord. When similar studies
were conducted in older rats, axon regeneration through the transplant diminished, even though the same transplant material
(fetal cord) was used, suggesting that the age of the host was responsible for the difference in regeneration. The numbers
of axons growing into the transplant increased however, if exogenous neurotrophic factors such as BDNF, CNTF, NT-3, NT-4 were
supplied to the older rats. Interestingly, if transplantation was delayed for 2 weeks after the initial injury to the cord,
descending axons now were able to traverse completely through the graft and into the host cord at the caudal end. The effects
of the delayed transplantation may have mimicked that of a conditioning lesion. In addition, the developing glial scar may
have been removed during the surgery to insert the graft. The best behavioral recovery was seen when transplantation of fetal
cord was delayed for 2 weeks, and with exogenous neurotrophins. Under this regimen, animals had near normal weight support
and stepping behavior, although coordinated quadrupedal movement was not sustained. Top
Change response of neurons to non-permissive environment
Understanding the neuronal response to inhibitory molecules may provide hints on how to manipulate the intracellular signals
and pathways that regulate this response. We may then develop strategies and design molecules to override the "no-growth"
signals. Over the last decade, numerous molecules have been shown to have inhibitory, or growth-stopping effects on neuronal
growth cones. Some are molecules that are present during development where they act to facilitate correct guidance when pathways
are first laid down. Others appear to be components of myelin and prominent in white matter tracts in the adult nervous system
The semaphorins are a family of proteins, some of which are secreted and others membrane bound. They interact with a receptor
complex made up of neuropilins and plexins. Because family members can act as antagonists by competing for binding to the
same receptor, the activity of semaphorins can be modulated from inhibitory for growth to positive for outgrowth. Intracellular
signaling pathways activated by semaphorins involve rac-1 and rho, small GTPases that control actin polymerization in the
cytoskeleton.
The ephrins, a second family of proteins with inhibitory activity, play an important role in organizing the retinotopic projection
to the tectum. Ephrin receptors are tyrosine kinases and activate downstream signaling cascades that are also linked to rac
and rho pathways. Reverse signaling has been documented for ephrins where the ligand-bearing cell responds as well as the
receptor-bearing cell.
The netrins and slits are two other classes of guidance molecules. In the developing spinal cord, netrin acts as an attractant
cue guiding commissural axons to the floor plate. However, netrin activity can be modulated from being an attractant to being
a repellent. In nematodes, netrin functions as a repellent when associated with another protein such as UNC5. Like netrin,
Slit is found at the midline where it acts to keep axons that have crossed over from returning. Its receptor, Robo, is present
on axons.
While these molecules play importance guidance roles in the developing nervous system, they may have a deleterious function
during regeneration. The expression of molecules such as semaphorins, ephrins, netrins, and slits after SCI needs to be explored,
as well as the expression of their respective receptors. If any are expressed after injury, their actions could be inactivated
by selective use of antagonists and thus allow regeneration of certain axonal tracts or pathways in the cord.
One consequence of trauma is the disruption of oligodendrocytes with the concomitant release of myelin proteins. Many of these
proteins exert a highly inhibitory effect on axonal growth. Some of myelin's inhibitory activity may reside in a recently
identified protein, NOGO that is expressed only by mature oligodendrocytes. Different fragments of NOGO cause growth cone
collapse in a classic in vitro test of motility inhibiting effects. MAG is a second inhibitory molecule present in myelin.
The response of neurons to MAG is modulated by a cAMP dependent pathway. If cAMP levels are raised, protein kinase A is activated
and inhibition by MAG is blocked. Intracellular levels of cAMP can be increased by exposure to neurotrophins. Other agents
include the nonhydrolyzable cAMP analog dibutyrl-cAMP, and rollipram, an inhibitor of the phosphodiesterase that inactivates
cAMP; all overcome inhibition by MAG and myelin.
How does elevated cAMP promote neurite outgrowth even when the extracellular environment is hostile to growth? One enzyme
activated by cAMP is arginase-1, which converts arginine to ornithine which in turn is converted by ornithine decarboxylase
to putrescine. Treating neurons with putrescine alone overcame inhibition by MAG.
Rho, another downstream target of the cAMP signaling cascade, is a GTPase that regulates actin polymerization. Neurite outgrowth
is reduced when Rho is activated. Rho activity can be blocked by C3, a bacterial toxin from clostridium botulinum. Treatment
of neurons in tissue culture with C3 blocks inhibition by MAG. The application of C3 to the site of a dorsal hemisection in
adult mice resulted in good regeneration and behavioral recovery. Although these results seem encouraging, it was puzzling
that recovery seemed too rapid (measurable after 24 hours) to be due to regeneration.
In summary, manipulating different stages of the cAMP pathway appears effective in overcoming the inhibitory actions of myelin
components and may provide an entrée into intracellular mechanisms to change the response of the neuron. Top
Reforming connections - Interactions between the neuron and the target
Re-establishing connections with the appropriate targets is the next critical hurdle to functional recovery. Unfortunately,
the rules governing synapse formation are not well understood, and those for synapse re-formation after injury are even less
clear. What is clear is that a great deal of communication occurs between potential pre- and post-synaptic partners when synapses
first form during development. Some studies suggest that the initial contact may come from the postsynaptic cell. When Mauthner
neurons extend the first axons down the zebrafish spinal cord, they grow past their target motor neurons in each segment without
a pause or change in their rate of extension. Target motor neurons, however, actively explore the environment with their dendritic
protrusions. Dendrites contacting a Mauthner axon becomes stabilized, and synaptic specializations can form within 30 minutes
of the initial contact. Complex interactions between the presynaptic and postsynaptic cells have also been observed in Drosophila
development.
Top
Making the most of what remains after injury
While the focus of the workshop was on regeneration of axons and reformation of synapses, awareness of the need to make the
most of remaining neurons and circuits after injury was high. One approach is to identify ways to facilitate endogenous mechanisms
for recovery.
Unlike central nerves, peripheral nerves face fewer barriers, and because they do regenerate in the adult nervous system,
investigators have turned to this model system to explore ways that encourage appropriate regrowth following injury. Brief
electrical stimulation of the cut femoral nerve greatly enhanced the speed of axonal regeneration. With electrical stimulation,
all motor neurons regenerated axons after 3 weeks, rather than requiring 10 weeks.
In the nervous system, how well a circuit functions depends on the levels of synaptic communication between the contiguous
neurons making up the circuit. Synaptic function can be enhanced by regulating the presence of neurotransmitter receptors
as well as their function. For example, AMPA receptors mediate excitatory synaptic transmission in the spinal cord, and recent
studies find that phosphorylation potentiates their activity by increasing open channel probability. Receptor function can
also be regulated by mobilization and insertion of receptors into active zones. As the intracellular pathways involved in
receptor trafficking start to be understood, it may be possible to change and optimize synaptic function. For example, reduction
of these receptors after injury may prevent excitotoxicity. Conversely their upregulation may enhance residual synaptic function.
Top
Recommendations from the participants
Last updated July 24, 2008