NINDS Workshops on "New Strategies in Spinal Cord Injury
About 200,000 Americans are confined to wheelchairs because of spinal cord injury. Each year, more than 10,000 new serious spinal cord injuries occur, with two-thirds of the victims being under the age of 30. The National Institute of Neurological Disorders and Stroke (NINDS) has recently initiated a series of six small, focused, multidisciplinary workshops to review recent research advances, explore new mechanisms, and develop novel strategies for the treatment of different aspects of acute and chronic SCI. Each workshop brings together scientists who conduct research on spinal cord injury, experts in related or relevant fields, and patient advocates to focus on a specific topic in the recovery process. Topics of interest will include inflammation and cell responses following spinal cord injury, barriers to regeneration, synapse formation and plasticity in regenerating systems, cell and gene therapy in the damaged spinal cord, the role of rehabilitation and neural prostheses, and the development of improved imaging and clinical outcome measures for spinal cord injury. The first workshop in the series was held in April, the second in June, 2000, and others are planned on a regular basis over the next year and a half.
The strong interest in and enthusiasm for these workshops from numerous public and private agencies made us realize that this series is a unique opportunity to share resources and information. These meetings were therefore organized in partnership with the National Aeronautical and Space Agency (NASA), the Christopher Reeve Paralysis Foundation, the Kent Waldrep National Paralysis Foundation, the Paralyzed Veterans of America and the Eastern Paralyzed Veterans of America.
April 5-6, 2000
Chairman: Brad Stokes, Ohio State University
Organizer: Arlene Chiu, NINDS
Designing effective therapies for spinal cord injury (SCI) has been a challenging problem because spinal cord injuries are heterogeneous in causality, severity, and location of injury. It seems highly likely that different treatments will be required for different types of damage and different stages of recovery. The National Institute of Neurological Disorders and Stroke (NINDS) has organized a series of workshops that bring together specialists in SCI and other fields such as stroke, cell and gene therapy, rehabilitation medicine, and immunology to discuss specific problems in SCI. The goals of these workshops are to review current research results, to highlight areas that need or deserve further investigation, and to identify potential new therapies for functional recovery after SCI.
The first workshop, on the role of the immune system in SCI, was held in Rockville, MD in April 2000, and focused on the early events following injury, when the blood-brain barrier between the circulation and the central nervous system (CNS) is breached, followed by a rapid and robust invasion of inflammatory cells into the damaged nervous system. Inflammation is the physiologic process by which vascularized tissues respond to injury, and can last from minutes to days. Inflammation is characterized by fluid accumulation, and the influx of plasma proteins, neutrophils, T lymphocytes and macrophages. Since these cells include major components of the immune system, a role for an immune response has been suggested. This is a relatively new area of investigation, engaging the two fields of neuroscience and immunology. Studies on the role of the infiltrating immune system following trauma to the CNS are still in their infancy. It was pointed out at the meeting that the first paper in this area, described macrophage invasion in SCI and was published in1985; the second, which analyzed inflammatory damage in SCI, appeared in print in1992.
The CNS is relatively "immune privileged", meaning that under normal conditions there is minimal immune surveillance. This is partly due to the blood-brain barrier, a specialized structure made up of endothelia and astrocytic endfeet that separates the circulatory and the central nervous systems, and tightly regulates passage of molecules and cells between them. Thus, few immune cells are found in the healthy CNS. Upon injury, however, the blood-brain barrier is physically and functionally altered broken, blood vessels become leaky and cells of the immune system invade the CNS, triggering an inflammatory response. The immediate, but transient, appearance of neutrophils characterizes the early immune response. The next phase of the immediate early inflammatory response that follows SCI is mediated mainly by two groups of leukocytes (white blood cells) - T lymphocytes (or T cells) and macrophages. T cells are antigen-specific; each is activated by a specific stimulus (or antigen) that is frequently just a part of a molecule. They reside in the major immune organs, circulate through the body in the lymphatic system, and can "home" to sites of injury where they migrate into the surrounding tissue. If activated by binding to their respective antigen, they increase their release of cytokines and are cytotoxic to target cells. Macrophages, on the other hand, are not antigen-specific. They circulate, but also reside in various tissues. Best known for their phagocytic properties, macrophages also have important functions in antigen presentation, cytokine secretion and cytotoxicity. Quiescent macrophages residing within the CNS are known as microglia.
The molecular basis of spinal cord inflammation draws from a vast arsenal of known immunologic molecules, including cytokines and chemokines, as well as growth factors, trophic factors, and other agents. Cytokines, of which there are about 60, regulate cell-cell communication between immune cells. They are small proteins that produce local and transient effects. Chemokines are chemotactic molecules that attract immune cells, helping them to "home" to sites of inflammation. Frequently, the cells producing these regulatory molecules also bear receptors for them, participating in a complex network of self-regulating and local interactions that orchestrates the proliferation of immune cells and then the subsequent decline of immune activity.
How do these diverse cell types, and the factors they secrete, interact during the progression of inflammation following SCI? Are the activities of the immune system beneficial or deleterious for subsequent recovery of function? Are there aspects of the immune response that can be manipulated or used to promote protection and repair of the damaged nervous system?
In attempting to understand the cascade of events triggered by the immune system during SCI, the discussions revolved around (A) studies that addressed the actions of cells of the immune system and (B) experiments to understand the CNS response to molecules released by the immune system. Finally, (C) new approaches to the problem were identified.
III.A. Cellular Immune Response
A central theme is to determine in what ways the activity of inflammatory cells is good for neural recovery and regeneration, and in what ways deleterious. Gennadij Raivich presented evidence that the immediate inflammatory response in the CNS seems better suited to provide immune surveillance, to fight possible infection, and to clean up debris rather than to help the repair process. Quiescent microglia normally express few of the receptors to molecules released by the immune system, but move to an "alert" phase with injury, producing more of these receptors, dividing and "homing" to the damaged cells. The cell debris then triggers their transformation into macrophage-like cells: they engulf cell debris, and begin to produce pro-inflammatory cytokines and factors, starting a cascade of action by stimulating and recruiting other bystander microglia. Injury also stimulates resting astrocytes to express cytokines, cell adhesion molecules, trophic factors and glial fibrillary acidic protein (GFAP), a hallmark of injury and a component of gliotic scars in the adult nervous system.
If the inflammatory response is deleterious to recovery, one approach might be to eliminate the immune cells that mediate this response. To test this idea, Lisa Schnell described experiments that attempted to deplete leukocytes in mice following traumatic brain injury. In one study, irradiation was used to eliminate all circulating leukocytes; the breakdown of the blood-brain barrier was reduced, but not secondary cell death. In a second study, mice were treated with rabbit antibodies made against white blood cells (anti-leukocyte antisera). This treatment, given to mice where the spinal cord was transected, resulted in a larger area of cell death compared to control animals that received the injury, but did not receive the antisera. However, treatment of mice with non-immune antisera produced the same result. Since depleting leukocytes did not decrease the amount of cell death, one interpretation of these results is that the inflammatory response is not necessarily detrimental. This experiment also demonstrates that a systemic challenge to the immune system (in this case by injection of foreign antibodies) can have a negative affect on outcome after injury.
Phil Popovich also described cell depletion experiments. He reduced blood-borne macrophages by injection of clodronate liposomes after spinal cord lesion in experimental rats. Infiltration of macrophages into the lesion was reduced. There was also a reduction in the size of the cavity that develops as a common consequence of SCI in rats and man. Furthermore, recovery of hindlimb function, as scored by the Basso-Beattie-Bresnahan (BBB) locomotor rating scale, was improved, paralleled by less degeneration in the gray matter, higher survival of myelinated axons and an increase in axonal outgrowth. Since blood-derived macrophages were specifically ablated, these experiments represent a first step towards teasing out the contributions of specific immunecell types. A tantalizing suggestion is that microglia-derived macrophages, being resident in the neural microenvironment, may turn out to be more beneficial than blood-derived macrophages.
The novel concept of "protective autoimmunity" was presented by Michal Schwartz. She tested the relative benefits of 3 types of activated T cells in a contusion model of SCI: (i) T cells with specificity for CNS neural antigens ( myelin basic protein), (ii) T cells to "self" antigens that are not restricted to the CNS (heat shock proteins), and (iii) T cells to non-self antigens (ovalbumin). All three types of T cells , accumulate at the site of lesion, regardless of antigenic specificity, but only T cells with neural specificity were able to reduce the spread of secondary cell death and enhance the sparing of adjacent neural tissue. In experiments to date, only T cells against myelin-associated proteins (proteolipid protein, myelin associated glycoprotein, and myelin basic protein) are effective. One possible mechanism of action is that myelin-specific T cells are re-activated by their respective antigens at the site of injury, leading to increased secretion of criticalcytokines and neurotrophic factors, that ultimately have a local neuroprotective effect. It is important to emphasize that the action of T cells results in protection of surrounding tissue, rather than in promoting regeneration of the directly damaged tissue. The protection does not depend on the virulence of the T cells, since both T cells that cause autoimmune disease and T cells that do not (being directed against non-antigenic epitopes or regions of myelin basic protein) were equally effective. Obviously, therefore, efforts should be made to select safe epitopes that will induce immune neuroprotection without the accompanying threat of autoimmune destruction of the nervous system.
III.B. Molecular Immune Response
Pharmacological intervention will depend upon the identification of appropriate molecular targets. Many molecules known to be involved in immune function are present at the site of spinal cord lesions. Levels of cytokines, growth factors, and chemokines have all been shown to rise and fall following injury, with a time course specific to each individual molecule.
Cytokine-mediated inflammatory processes and their underlying molecular substrates can be studied in the normal brain using transgenic mice. Iain Campbell's laboratory has generated transgenic mice with astrocyte-targeted expression of IL-6, IL-3, IL-12, TNF-a , and IFN-a. These animals develop specific neurologic defects that are first detected at 3-6 months and are the result of distinct glial cell alterations in function as well as in the expression of many key host response and inflammatory genes. The cytokine actions were found to involve complex and distinct signaling pathways with neuroinflammatory consequences. Several of these conditions mimic human neurological diseases highlighting the direct pathogenic potential of chronic low-level cytokine production in the CNS.
In the brain, astrocytes are major producers of several chemokines, which function as chemo-attractants for immune cells, and modulate T-cell function as well. Astrocytes can be activated by diverse mechanisms; Tika Beniveniste showed that engagement of the cell surface molecules Fas and ICAM-1 stimulated astrocytes to secrete chemokines. The binding of Fas to the Fas ligand has traditionally been thought to play a role in mediating cell death; this occurs in microglia. ICAM-1 is a cell adhesion molecule that plays an important role in leukocyte extravasation as well as in the activation of T cells. Since activated T cells express both Fas and ICAM-1, it is conceivable that after SCI, their presence could in turn activate astrocytes, which would further propagate inflammatory responses.
The timing of these interactions is a particularly important issue. Tumor necrosis factor (TNF), a cell death-inducing molecule, is known to increase after spinal cord injury. John Bethea showed that when the anti-inflammatory cytokine interleukin-10 (IL-10) was injected intraperitoneally into experimental rats 30 minutes after SCI, it reduced the expression and release of TNF-a by macrophages. This single injection of IL-10 also resulted in improved recovery of hindlimb function, as measured by the BBB locomotor rating scale, and delayed the onset of grooming behavior that is indicative of pain. However, if rats were injected twice, at 30 minutes and 3 days post-injury, with IL-10, release of TNF was reduced but no improvement in functional recovery was measured.. These results suggest that suppression of an early inflammatory response may have great benefits in neuroprotection, however a later response may be important for other aspects of healing.
Paul Patterson explained why leukemia inhibitory factor (LIF) is likely to be an important cytokine following damage to the nervous system. In studies using mutant mice unable to produce LIF, he finds no up-regulation of the glial marker GFAP when the brain or spinal cord is injured. At the same time, far fewer neurons are lost following injury or ischemia in these animals, suggesting that LIF may play a role in cell death under these traumatic circumstances. Thus, LIF and related cytokines are possible targets for therapeutic intervention.
III.C. New Approaches
The potential role of spreading depression in SCI was raised by Maiken Nedergaard. Spreading depression is a transient, slowly propagating wave of tissue depolarization that can be evoked by mechanical stimulation, and has been associated with ischemia or traumatic brain injury. Each wave lasts for 1-2 minutes; waves of spreading depression have been observed to occur for several days after injury. In a model of ischemia, these waves have been shown by MRI to precede neural damage, with infarct volume increasing by up to 23% after each wave. Spreading depression is of non-neuronal origin, being mediated by astrocytes, and neurons are only secondarily affected. Two different, interacting mechanisms help to propagate spreading depression from cell to cell: a gap junction-mediated pathway and an ATP-purinergic pathway. While most of the work on spreading depression has been done in the brain, recent studies have shown that spinal cord astrocytes can mediate spreading depression via the ATP purinergic pathway after edema.
Changes in electrical activity following SCI may also play a role in regulating interactions with the immune system. Harmut Wekerle described the induction of expression of major histocompatibility (MHC) class II proteins in hippocampal slices when neuronal activity was abolished by tetrodotoxin, and the slice treated simultaneously with interferon. This finding is significant since T cells are activated only by the combination of specific antigen together with MHC proteins on an antigen-presenting cell. In a search for activity-dependent factors released by neurons that can suppress MHC expression, Wekerle also found that neurotrophins, and in particular nerve growth factor, were effective. These preliminary data suggest that neurotrophins may indeed have some anti-inflammatory properties.
Larry Steinman discussed the use of gene chip microarrays to probe for gene expression at different times and sites following injury. Microarray profiling of spinal cord tissue from animal models will certainly yield insights into the inflammatory process and pathophysiology of SCI.
IV. Workshop Conclusions and Recommendations
IV.A. Basic Science
The role of inflammation following SCI is complex, with cells from both the immune system and the nervous system responding, interacting, and changing over time. Evidence was presented for both beneficial and detrimental aspects of the inflammatory response following SCI. However, answers to fundamental questions concerning the mechanisms, time course, and effects of inflammation are still incomplete.
Areas of basic science that should be given research priority include:
IV.B. New Animal Models
Although the rat has been the animal of choice for studies on SCI, development of a mouse model will allow researchers to make use of the numerous important mouse mutants that have been produced in the past decade. Brad Stokes described great progress in a new spinal cord contusion model that has been developed for the mouse, at Ohio State University. Characterization of the physiological response in the murine model will provide a baseline so that maximal information may be gained from the many lines of transgenic mice that have already been generated. In one respect, however, the rat is a more faithful model for the human condition: they develop cavities in the injured cord similar to those seen in patients with SCI, while mice show little no cavitation in a variety of strains injury. Select strains of transgenic rats should be developed, therefore so that the enormous database on rat SCI may be fully utilized. Molecules that regulate the infiltration of immune cells into and out of the injured spinal cord may be prime candidates for this approach.
Etty N. Benveniste
John R Bethea
Iain L Campbell
Arlene Y. Chiu
Samuel David
Gregory J del Zoppo
Gerald D Fischbach
William Heetderks
Maiken Nedergaard
Paul H. Patterson
Phillip G Popovich
Gennadij G Raivich
Michal Schwartz
Lisa Schnell
Lawrence Steinman
Oswald Steward
Bradford T Stokes
Harmut Wekerle
Other Participants
Susannah Chang
Lisa Cash
Katherine Woodbury Harris
National Institute of Neurological Disorders and StrokeSusan Howley
Aimee Hunnewell
Tom Jacobs
Melinda Kelly
David Liskowsky
John Marler
Ralph Nitkin
National Institute of Child Health and Development
61E/2A03 6100 Executive Blvd
Bethesda, MD 20892-7510
Phone: 301 402-4206
Fax: 301 402-0832
email: rn21e@nih.gov
Holly Patton
Audrey Penn
Lillian Pubols
Victor Schneider
Bette Siegel
Robert Zalutsky
Last updated September 15, 2008