CHAPTER 267 Principles of Translation of Biologic Therapies in Spinal Cord Injury
Unique Aspects of Spinal Cord Pathophysiology
The spinal cord connects the brain to the effector and sensor structures of the body and contains extensive integrative neurocircuitry. It is protected by the vertebral column. The normal spinal cord is a stable structure throughout life, but it exhibits modest decrements in myelinated axon1 and neuron numbers with advancing age.2 Its complexity greatly exceeds that of other tissues, and its thin cellular processes can extend long distances, thereby increasing their vulnerability. It has unique vascular barriers and fluid compartments. Its cellular components and the tissues that they form are highly susceptible to injury. Together with the brain, this organ has the least ability to undergo self-repair among human tissues. Increasing evidence indicates that the spinal cord also has distinct differences from the brain in its injury responses.3,4 Its neurons have highly specific connectional complexity with numerous input and output that are extensively modulated by feedback. Neuronal systems are especially vulnerable to injury because of their complexity and specificity of connections. In addition, because of the receptor and membrane specializations that enable chemical and electrical neuronal transmission, there is a high capacity and vulnerability to major ionic shifts. The spinal cord’s component tissues are rarely exposed to inflammatory cells, and the blood–spinal cord barrier restricts the movement of proteins and other molecules. The extent of endogenous repair after injury is limited. Although the spinal cord of teleost fishes and urodele amphibians can regenerate,5 as can that of some fetal mammals,6–8 no adult mammals undergo effective spontaneous regenerative repair. The evolutionary value of the spinal cord’s limited ability to undergo axonal regeneration is not clear but includes the expression of molecules that tightly regulate axonal sprouting to prevent aberrant connections from forming9 and thus facilitate learning and memory consolidation. Until the modern era, SCI was so rapidly fatal that no seriously injured persons would survive long enough for regeneration to occur. The current surgical treatment of SCI can be summarized as follows: prevent further cord injury, maintain blood flow, relieve spinal cord compression, and provide secure vertebral stabilization to allow mobilization and rehabilitation. Because the outcome of SCI, despite standard treatment, is often permanent paralysis, there was previously a pessimistic opinion regarding the potential to improve neurological recovery. Despite this pessimism, spinal cord medicine evolved to achieve major gains in life expectancy and innovative functional restorative treatments that do not require neuronal regeneration, such as tendon transfer.10 In the last 25 years, clinical trials of agents to provide neuroprotection or promote spinal cord repair have been initiated, and vast new experimental knowledge regarding spinal cord pathobiology and repair has been generated. Now that several potential therapies are becoming available, new challenges have become evident. Currently, our field has reached an important evolutionary point where the capability of translating experimental treatments into patient care may be as much a determinant of their success as the biology behind the therapeutics.
Translational Research and Clinical Trials
The resources associated with any human endeavor are finite, and their management and allocation both enable and limit activities. SCI preclinical experimental research is conducted with the result of human application and reduction of the injury burden in mind. Application requires clinical trial testing and thus a translation from studies in laboratory animals to human subjects. Translational research is an encompassing term referring to the iterative process by which clinical problems inform basic experimental research, which then generates potential new therapies that are tested for their safety and efficacy in research subjects. Translational research requires cooperative interaction between industry, academic researchers and clinicians, and regulatory and funding agencies. This type of research is often a subject of media attention and controversy and may be strongly influenced by political advocacy11 and legislation.12 Currently, it is recognized that there is a need to improve the translational process because some studies initiated in the past 2 decades have been hampered for reasons that may reflect inadequate translational preparation. There is a growing consensus among SCI researchers that the body of preclinical evidence for translation of a promising therapy should meet more stringent criteria,13 including verification of efficacy by an independent research laboratory and replication of the effects in an additional species. This chapter reviews several examples and suggests how the translational process, as applied to SCI, can be improved. A “successful” clinical trial is one with a sound biologic and clinical rationale, designed to allow a key hypothesis to be tested with minimal bias and robust statistical power, that recruits sufficient patients, adheres to protocols, acquires data accurately, has relevant outcome measures, maintains high compliance with follow-up, and is conducted with minimal waste. If clinical trials are not based on solid preclinical evidence, their chance of success is marginal. A clinical trial should be based on a plausible mechanism of therapeutic action that can be measured. If this is not the case and the study is negative, little meaningful knowledge can be generated. Rigorous study design, adequate study power, and sensitive valid outcome measures should be used to avoid failure to detect a benefit in patients when one is actually present. There are few perfect clinical trials, but there is usually a relationship between the quality of a trial’s design and conduct and its eventual impact. In the 11 years since the last version of this chapter was written, several early-phase clinical trials have tested the concepts of translational medicine as applied to biologic therapies for SCI.14,15 Important lessons have led to maturation of the “field” of spinal cord repair. An exhaustive review of biologic strategies to promote spinal cord repair would require a separate, substantial textbook, which is a favorable reflection on growth of the field in the last 3 decades.
Unique Translational Aspects of Spinal Cord Injury
SCI is a relatively uncommon condition that has profound lifelong systemic effects. Because it is not nearly as common as stroke or Alzheimer’s disease, the development of new complex therapies for SCI is not especially attractive commercially. In addition, because no single therapies have yet been proved effective and because it is recognized that multiple therapies may be necessary to achieve clinically substantial effects, the investment risk scenario is often prohibitive. As an incentive, the Food and Drug Administration (FDA) designation of “orphan status” to SCI allows a company an extended period for marketing without competition (e.g., 7 years). Another important problem for SCI translational studies is the ability to detect the effect of a therapy in patients. If it is true that a combination of several therapeutics16 may be needed to obtain substantial neurological recovery, there may not be outcome measures sufficiently sensitive to detect the effects of individual components of combination therapies when tested alone in patients. This may require pragmatic shifts in regulation, such as not requiring demonstrated efficacy of each individual component and a focus on the efficacy and safety of the combined components. Because of the difficulties inherent in industry-supported translation, there has been increasing emphasis on noncommercial consortia. Examples include the North American Clinical Trials Network (NACTN—http://www.christopherreeve.org), the European Multicenter Study about Spinal Cord Injury (EMSCI), and the Chinese SCI network (ChinaSCINet—http://www.chinascinet.org/). These networks may allow testing of drugs or other therapies that are not of interest to industry because of being off patent or not patentable. In addition, these networks can develop an influential consensus on translational principles. A clinical trial of riluzole started in 2010 within the NACTN. Riluzole is an anticonvulsant that antagonizes sodium channels and inhibits calcium-dependent glutamate release.17 Currently, riluzole is administered to patients with amyotrophic lateral sclerosis (ALS), in whom it has an established safety profile and has been found to prolong life for several months.18 Riluzole has been shown to promote neurological recovery in rodent models of SCI19–21 through neuroprotective tissue-sparing effects. The fact that riluzole has a well-established safety profile at therapeutic human doses has simplified the translational process. Lithium, another off-patent drug with an extensive history of human use, is being tested in the ChinSCINet22 because of evidence of favorable effects on neural precursor cells after spinal cord transplantation.23 The National Institutes of Health Rapid Access to Interventional Development (NIH-RAID) program can provide support for early therapeutics development and animal toxicology testing for promising newly discovered therapeutics.
What Is A “Biologic?”
This chapter focuses on “biologic strategies” for spinal cord repair as opposed to strategies that use drugs or physical methods such as cooling. All neurorestorative strategies are inherently biologic because their effects involve and alter living processes. The distinction here is to focus on strategies that use tools similar to the protein or cellular components of living beings. Biologic therapies include products such as proteins, antibodies, enzymes, and cells that are prepared through processes such as cell culture or recombinant protein technology.24 Production of biologics creates potential product risks that differ from those in chemical processing, including alteration of cellular genetic material and infection by adventitious viruses, which could adversely affect the recipient.25 These products are produced under specific rigorous conditions to ensure their safety.26 Regulation of their production and human testing in the United States is the responsibility of the Center for Biologics Research and Evaluation (CBER), an FDA center regulated by the Public Health Service Act (sections 351 and 361). CBER publishes regulations in the Code of Federal Regulations (CFR) that apply to all entities developing new therapies to be used in patients in the United States. Increasing understanding of similarities between the translational processes established for drugs, devices, and tissue products and those evolving for neurorestorative biologics is influencing how experimental spinal cord research is being conducted. It is increasingly being recognized that experimental studies aimed at translation need to include (1) dose-response efficacy/safety and maximum tolerated dose studies; (2) animal studies with safety end points related to risk for adverse outcomes such as aberrant plasticity, pain, or tumor formation; (3) product potency assays; (4) stringent product characterization and release criteria; and (5) surrogate markers that allow comparison of biologic effects between in vitro, experimental animal, and clinical studies. Such pivotal studies should be acquired in compliance with good laboratory practice (GLP) and good manufacturing process (GMP) standards. These standards emphasize consistency and documentation but can increase research costs substantially. The biologics used in pivotal animal studies should be prepared exactly as the intended clinical product. An understanding of the FDA regulations, the investigational new drug (IND 21 CFR:312) approval process, and key data required can improve the design of preclinical translational studies by reducing the occurrence of ineligible studies that may need to be repeated. Cellular biologic materials that have undergone extensive in vitro manipulation are a new and complex form of therapeutic.27 Provision of regulatory guidance becomes more complex when the regulators are confronted by therapies for which no regulatory precedent exists,28 such as embryonic stem cells. Approval tends to be slower with an emphasis on enrolling small numbers of patients in phase I studies and allowing sufficient time to elapse for detection of complications before permitting the initiation of phase II studies. In addition, the cost of biologic therapies can be great, and there is an ethical and societal imperative to establish the cost-effectiveness of expensive therapies.29 If biologic therapies will ultimately be paid for by insurance companies, they will become widely used and consistently available only if they are shown to have efficacy and safety. At the phase III study level, such efficacy is required to be evident in the form of functional outcome measures such as the spinal cord independence measure (SCIM).30
Delivery of Biologic Therapies
Only a few classes of biologic therapy for SCI can be delivered through the systemic intravenous route, and none are suitable for oral delivery. This limitation means that most biologic therapies require relatively invasive technical delivery methodologies such as direct implantation or delivery into cerebrospinal fluid (CSF).31,32 The risks associated with the delivery method may exceed those associated with the therapy itself. In addition, established methods for monitoring dose distribution and biologic half-life, such as pharmacokinetics, are unavailable unless invasive monitoring methods such as CSF sampling are used.33 This fact places even more importance on comprehensive preclinical data in relevant animal models. Although delivery by simple lumbar puncture is theoretically attractive, there are concerns that delivery of cells based on homing within the neuraxis may lead to seeding of cells into normal loci with unpredictable effects. However, several studies have shown that cell homing to sites of lesions is robust.34–36 Evolving techniques such as percutaneous37 and endoscopic38 transplantation may reduce the magnitude of surgical exposure associated with transplantation. However, these minimally invasive techniques must be established to have safety equivalent or superior to that of open techniques.
Outcome Measures and Surrogate Markers
There is concern that the currently accepted SCI outcome measures may be insufficiently sensitive to detect small changes induced by therapeutics. The detection of any actual change is important to foster optimization of therapy. There are two approaches to this problem: surrogate measures and more sensitive outcome tests. The term surrogate marker (end point) refers to indirect measures that can confirm the activity and effect of a therapeutic that may correlate with an important clinical end point. Serum cholesterol is a classic biomarker for coronary artery disease. No surrogate markers have yet been established for treatment of SCI with biologics.39 Spinal cord imaging techniques show promise for this purpose40,41 and can detect changes in signal that would be of concern, such as increased edema, cavitation, and hemorrhage. The use of magnetic resonance imaging (MRI) as a surrogate marker has been explored extensively for multiple sclerosis (MS).42 Somatosensory evoked potentials and motor evoked potentials have not proved to be more sensitive than clinical examination for outcome assessment,42a although detection of perceptual thresholds may be a very sensitive indicator of sensory changes at the thoracic level.43 Brain motor control assessment, a multichannel electromyographic technique, may detect subclinical changes44 in muscle function.
Spinal Cord Injury Models
The basis of animal testing is the assumption that a related mammalian system will show similar responses to a therapeutic agent as the target human disease or injury.45 Thus, human therapeutics testing is based on proof-of-concept preclinical studies showing robust effects that are predicted to also occur in patients. Animal models of SCI are, however, only an approximation of human pathophysiology, and animals differ across species in their responses to injury.46 Dorsally delivered spinal cord contusion is popular and reproducible,47 as is laterally applied (clip) compression.48 Models of balloon compression have been used for several decades and can be performed without laminectomy; however, these models tend to give more variable injuries.49 Standardized contusion devices have allowed highly reproducible experimental injuries to be created.50 This has several important consequences. It increases the precision of experiments by reducing variability and allows experiments to be conducted with fewer animals. It also allows more meaningful comparisons between laboratories and within the scientific literature. The most common injury used today is a 12.5-g/cm dorsal contusion with the NYU piston impactor.47,51 This magnitude of injury causes an incomplete SCI that typically shows recovery to approximately 12 of 21 on a broadly used recovery scale.52 It is important to realize that most human SCIs result from anteriorly applied forces that substantially damage the anterior white matter and anterior spinal artery complex in a manner that differs from animal injury methods. Animal models do not reproduce bone fracture, dislocation, disk herniation, and spinal nerve root injury, and few evaluate the ongoing compression that is frequently present.53 Animal SCI is performed under anesthesia, which may affect secondary injury mechanisms.54 No models thus test the usual sequence of injury events: acute injury, resuscitation, administration of various drugs, surgical decompression and fixation, common complications such as pneumonia or urinary tract infection with fever, and treatment variables such as rehabilitation. Animal models typically do not assess the effect of concomitant medications commonly administered to patients with SCI, such as heparin/enoxaparin. Furthermore, rodent models of SCI have significant “scale” and therapy-targeting limitations when translated to clinical studies55; although there is an abundant literature on scale-up of orally and systemically administered drugs from rodents to humans,56 much less is known on how to scale localized treatments such as cell transplants. Besides the large differences in tissue volumes, there are also differences in CSF flow57 and kinetics58,59 that may be important. Thus, the data generated in these models have limited predictive strength. An especially difficult problem in SCI research is that correlation between currently accepted clinical outcome measures and experimental measures in animals has never been rigorously established. It is assumed that, for example, an improvement in the sensorimotor aspects of animal gait predicts improved ambulation in patients, but the absence of successful examples of translation in SCI has precluded validation of these assumptions. To partially address some of the foregoing issues, larger animal models may be necessary to more fully emulate human application.60,61
Key Considerations in Spinal Cord Repair
Plasticity and Regeneration within the Spinal Cord
The spinal cord is essentially a relay and compensation system that manages signal transmission and integration in concert with control centers in the brain, cerebellum, and brainstem, as well as peripheral sensory information. The spinal cord contains local integrative networks. When completely interrupted, ascending and descending input and output are lost and result in the well-known deficits in motor, sensory, and autonomic functions and dysregulation of various reflex functions. The severity of an SCI is the primary determinant of its potential to be repaired. Most currently promising therapies have been developed in models of incomplete SCI. A partially injured spinal cord has substantial endogenous repair capability that can be augmented and amplified to recover some signal transmission and integration by adaptive change within spared circuits.62,63 The adaptive changes that occur as a result of modifications in residual connectivity are called plasticity. Anatomically complete injuries pose different challenges. Although the isolated caudal spinal cord is capable of substantial intrinsic plasticity, the potential to achieve functional reconnection between the brain and spinal cord is limited. Recovery of transmission would require substantial regeneration through scarred areas containing non-CNS tissue and then regrowth within a complex environment that is undergoing changes, including wallerian degeneration, axonal dieback, and segmental plasticity. Even though many studies have reported a modicum of regeneration, for the most part this has occurred over just a few millimeters, and no strategies currently planned for clinical translation are targeted to cause long-distance regeneration in the human spinal cord. The isolated caudal spinal cord is capable of modification by training,64–66 drugs, and other techniques. Some experimental treatments seek to directly regulate its activity,66a and ultimately this may be very useful. Although motor output can be activated from the distal isolated cord, the cerebellar and brainstem connections that regulate balance are absent. Another interesting strategy is to directly activate denervated muscles through embryonic neural precursors implanted outside the CNS within nerves.67 Muscle fibers can be reinnervated by the transplanted neurons, and the externalized neurons can be activated.
Endogenous Reparative Processes
The CNS is composed of cells with variable sensitivity to traumatic injury, neurons68 and oligodendroglia being most vulnerable.69 SCI often causes cavitation of a major portion of the injury epicenter as a result of extensive cell death, mechanical injury, ischemia,70 and secondary processes, including excitotoxicity and inflammation. The result of these processes is often complete destruction of segmental gray matter, including the death of motor and sensory neurons, with a residual peripheral shell of white matter, essentially a hole in the spinal cord. After SCI, numerous precursor cells are activated with the result being a partial endogenous reparative tissue response.71 All known components of the spinal cord respond to injury, and the considerable recovery that follows incomplete SCI is dependent on these processes. Severe injuries, however, overwhelm this repair capability. Among the endogenous responses to SCI, most focus has historically been on the impediment that astrocytic gliosis represents to axonal regeneration.72,73 This inhibition of axon growth was first described by Ramón y Cajal in 1928. The postinjury inflammatory macrophage mass is invaded by astrocyte precursors.74 Other important endogenous responses include an initial wave of angiogenesis that appears to regress because of inflammatory factors promoting cavitation.75 Several studies have confirmed that new cells are born around the central canal from ependymal precursors.76,77 This response is modest in comparison to that in regenerative animals such as urodeles, where the ependymal mass drives regeneration. Other stem cells born after SCI78,79 actively replace depleted cells, especially oligodendroglia.80,81 Newly born cells can be increased with exercise,82,83 functional electrical stimulation,84 and agents such as Sonic Hedgehog85 delivered intravenously or, alternatively, by transduced transcription factors.86 Other treatments such as methylprednisolone may attenuate the normal endogenous response87 that is linked to the presence of inflammatory cells.74 Axonal repair includes terminal sprouting and collateral axonal branch elaboration from intact fibers.88 In-migrating Schwann cells may effect functional myelin repair of CNS axons.89
Effects of Repetitive Training on Plasticity
To be most effective, biologic therapies for SCI should amplify the beneficial forms of plasticity that enhance circuits providing connection from the spinal cord to the brain, cerebellum, and brainstem90 and attenuate adverse plasticity (e.g., neuropathic pain, severe spasticity). Ultimately, it is the stable connections that persist in the damaged spinal cord and the integrity of effectors such as muscle that can limit the extent of neurological recovery. In the last 10 years, great progress has been made in understanding how activity and exercise can be used to modulate and amplify plasticity in residual circuits after SCI. This is built on the knowledge that repetitive activity in neural circuits modifies synaptic function,91 receptor organization, and transmitter release and structure,92 which are forms of learning. Experimental studies have confirmed structural,93 biochemical, and neurophysiologic plasticity94 in response to exercise activity.62 Exercise can induce trophic effects in the brain that vary with the specific form of exercise.95,96 Neural circuits within the spinal cord are capable of self-sustaining patterns of locomotor output97–99 in response to afferent sensory information. These central pattern generators (CPGs) can function independent of supraspinal input100 and show use-dependent plasticity.101 A CPG has been identified in humans with SCI.65,102,103 Spontaneous reorganization of propriospinal circuits, as well as spared descending fibers, occurs spontaneously after incomplete SCI and can lead to functional recovery104 that can be enhanced with locomotor training.105 In adult cats with complete spinal transection, regular training resulted in significant recovery of hind limb function.104 The effects of training appear to be task specific (e.g., swim training does not correlate with improved walking106), and key afferent input is required.107 A sufficient level of excitability of the spinal circuitry through use-dependent mechanisms occurs during partial weight support–assisted locomotion104 to cause entrainment108,109 and improved function.110 The effects of this exercise-induced plasticity also include changes in muscle properties111 and function and metabolism.112 In the past decade it has been encouraging to find augmentation of recovery after SCI through the combination of activity and therapeutics in some112–114 but not all studies.115,116
Chronic Spinal Cord Injury
Acute injuries evolve to chronicity, but there is no universal definition of the specific time of onset of chronic injury. One useful definition may be the time at which there is no further neurological recovery, the neurological plateau. By definition, no further substantial recovery is expected, and thus the effect of a treatment may be tested against this baseline. There are several reasons why the treatment of chronic SCI is more challenging than the treatment of acute SCI. First, there has been segmental sprouting of axons, plasticity, and alterations in neuronal excitability leading to reorganization of the residual spinal cord systems117,118 and the formation of new synapses and patterns of transmission. In addition, there are extensive adaptive changes in the cortex119–123 and transneuronal and retrograde neuronal degeneration.124,125 The clinical sequelae associated with these changes include spasticity, autonomic dysreflexia, and neuropathic pain. To some extent, chronic SCI-targeted interventions may require reversal of this new connectivity. The mature chronic lesion and residual neuronal function may be influenced by other pathologic changes such as syringomyelia, myelomalacia, and spinal cord tethering and even by chronic medications.126,127 In as many as 30% of patients with SCI a syrinx may develop and cause delayed neurological dysfunction, such as ascending paralysis, brainstem symptoms, and pain.128 It is hoped that new acute SCI treatments will lead to a chronically injured state with fewer complications and greater ability to benefit from ongoing interventions.
Important Examples of Attempts to Translate Biologic Therapies
Several previous well-designed clinical trials failed to meet the primary study end points.129,130 These trials were expensive, and in several instances the impetus to further test the therapeutic in SCI has been abandoned. These examples have increased the concern of industry about the financial risk of SCI trials. On retrospective examination of some clinical studies it is clear that the necessary translational issues were not strongly established in rigorous and relevant preclinical models before clinical testing, and thus the clinical testing had a weak basis. In fairness, there has been substantial improvement in SCI experimental models and translational insight subsequent to these studies. A second issue is the perception that although many promising experimental therapies exist, their progress to clinical testing is slow. If the preclinical pathway toward a clinical trial were better defined, the process might be more coherent and efficient.
GM1
The first biologic to undergo extensive clinical testing with the aim of neurological recovery in patients with acute SCI was the monosialoganglioside GM1. The GM1/Sygen multicenter study is the largest clinical trial yet performed for SCI.130 Gangliosides are a group of glycosphingolipids with a high concentration in the outer membranes of nervous tissue. Abundant aberrant neuron sprouting is observed in patients with gangliosidoses, and gangliosides induce trophic effects on cultured neurons.131 Exogenous administration of gangliosides was found to promote neural repair and functional outcome in some animal models of neural toxicity132,133 and spinal cord transection.134,135 Many of the studies that showed potent in vitro effects of GM1 were published by researchers at the Fidia research laboratories, which produces GM1 for clinical use. The Maryland GM1 study conducted in 1990136 tested the efficacy of GM1 (monosialotetrahexosylganglioside sodium salt) versus placebo in patients with cervical and thoracic SCI. The results from this study demonstrated efficacy in motor recovery of the lower extremities.137 Another study claiming a similar effect in chronically injured patients138 was met with controversy.139 These results led to the much larger randomized double-blind Sygen Multicenter Acute SCI Trial. This trial enrolled 797 patients, all of whom received a loading dose of 30 mg/kg of methylprednisolone followed by 5.4 mg/kg per hour for 23 hours as standard of care according to recommendations of the second National Acute Spinal Cord Injury Study (NASCIS II).129 Patients were then randomized to receive high-dose GM1, low-dose GM1, or placebo for 56 days, beginning within 72 hours. The primary outcome measure, defined as a two-grade improvement in the American Spinal Injury Association (ASIA) impairment score (AIS) from baseline, was not achieved. However, patients did show more rapid recovery and had improved bowel/bladder function, sacral sensation, and anal contraction.130 Because at the time of the GM1 study methylprednisolone was considered a standard of care, the impact of GM1 alone was not determined, thus illustrating the serious scientific limitations that can arise from assigning standard of care importance to a treatment. In addition, the primary outcome measure, a two-grade change in the AIS, may have been an unrealistically large expected improvement. The AIS is an outcome measure that combines sensory and motor function and relies on integrity of the laminated spinal cord long tracts subserving anorectal sensation. It is a complex outcome measure and perhaps not as suitable as a single measure such as a motor score. Data from the Sygen trial have proved very beneficial in evaluating the natural history140 and outcome measures after SCI,141 and the design and execution of the trial were excellent. However, examination of the preclinical basis for GM1 testing in SCI is very instructive. In the first study that tested GM1 in a contusive rodent model (3 years after the first trial was reported), no acute effect of GM1 on contusion size was found, and when GM1 was combined with methylprednisolone, the tissue-sparing effect of methylprednisolone was eliminated.142 In retrospect, it is clear that GM1 was not studied systematically in SCI preclinical models before its use in clinical trials. Remarkably, no published study has examined the effects of GM1 at the clinical dosages used in a contusive or compressive SCI model to substantiate the long-term delivery clinical protocol. This fact strengthens the argument that sufficient preclinical efficacy and safety testing in animal models is a necessary basis for translation to human testing. It is assumed that because GM1 had been tested in other clinical settings of neurological disease, it was considered safe and extensive animal efficacy testing was not required for approval of the study by the FDA. However, had the preclinical testing been performed with the currently available models and outcome measures, the trial might never have taken place. GM1 is now rarely used in patients with SCI.
Activated Macrophages: Transplantation of Autologous Macrophages
Macrophage transplantation was the first multicenter cell transplant study for SCI approved in the United States. Macrophages play a key role in peripheral nerve regeneration.143,144 They remove myelin debris and have been shown to synthesize protein factors that may aid in the growth of axons after SCI.145,146 CNS myelin damage exposes axon growth inhibitors such as Nogo, MAG, and MOG.147 Inadequate clearance of myelin debris by activated microglia and macrophages may be a factor in the poor axonal regenerative response of the CNS after injury.148,149 Based on success in an animal study of complete spinal cord transection150 and other work,151,152 including a transected optic nerve model, a company was founded for commercial preparation of activated autologous macrophages, Proneuron, with the product being designated ProCord. This product’s clinical testing has partially completed phase II human trials. Macrophages are isolated from the patient’s own blood and exposed to autologous skin biopsy samples, and the “activated” macrophages are injected into the injured spinal cord at several points below the injury contusion153 up to 14 days after injury. Phase I data suggested safety and a high spontaneous AIS conversion rate, but there were a number of serious medical complications, possibly related to undergoing a second surgical procedure153 and not caused by the macrophage product. Phase II clinical trial data recently presented at the 2009 ASIA Annual Meeting in Dallas, Texas, by study investigator Dr. Dan Lammertse have shown no difference between the transplanted and control, nontransplanted group for the primary outcome measure, the ASIA motor score. In fact, there was a strong trend for less recovery in the transplanted patients. Substantial criticism has been directed at this trial because there was only one published preclinical study150 at the time when the trial began enrolling patients. A similar paradigm conducted in dogs with SCI later failed to show a benefit.154 A subsequent sponsor study of autologous macrophages activated by ex vivo exposure to skin155 or peripheral nerves and injected into the contused cord led to recovery of motor function and reduced postinjury cyst formation in rodents.155 These issues raise several important questions regarding preclinical evidence and translation, including the fact that the most clinically relevant preclinical model, currently accepted to be spinal cord contusion, should be used in pivotal studies, independent replication of the study should be undertaken by a reputable laboratory lacking a conflict of interest, and testing should be conducted in more than one species.13 It is argued that if a therapy cannot be independently replicated, its prospect for success in human translation is unfavorable. It is also argued that if an effect can be demonstrated only in a single species or even a single strain of rodents, for example, its prospect for meaningful effects in humans is limited. Had these criteria been applied and found to not support the index study, the trial would not have proceeded. However, these criteria have major implications for industry, including increased expense and intellectual property sharing. Working toward consensus principles to guide translational research is currently a major focus of spinal cord expert consortia such as the Spinal Cord Outcomes Partnership Endeavour (http://www.scopesci.org).
Cethrin
The first trial of an engineered protein to support neurological recovery after SCI involved the rho kinase inhibitor BA-210. The adult CNS environment is inhibitory to regeneration. As mentioned previously, there are component proteins of CNS myelin that inhibit neurite growth.156 For example, cultured neurons fail to enter the optic nerve but rapidly regenerate into the sciatic nerve.157 Biochemical studies of CNS myelin have identified several inhibitory proteins, the preeminent one being Nogo-A.158 Identification of Nogo led to discovery of the Nogo receptor, and studies of downstream signaling have identified RhoA and ROCK kinase as important mediators of Nogo receptor inhibition.159 The Rho-ROCK pathway is an important convergence point for multiple signaling proteins, which makes it an attractive target to block the inhibitory effect of multiple pathways. By applying this concept it has been demonstrated that inactivation of Rho enhances CNS regeneration in the optic nerve160 and after SCI.161 The initial blockade of Rho was achieved with C3 transferase, a protein toxin produced by Clostridium botulinum.160 A recombinant version of C3 transferase that has a transport sequence to enhance delivery was commercially developed under the name BA-210 (Cethrin). In a human phase I/IIa (dose finding) nonrandomized open-label multicenter clinical trial, BA-210 at a dose of 0.3 to 9 mg was mixed with Tisseel (Baxter Westlake Village, CA) and applied to the dura during spinal decompression surgery. Unpublished reports of this Cethrin study phase showed a 27% ASIA impairment grade conversion rate.162 Patients converted from ASIA A to B, C, and D. A phase IIb randomized double-blind multicenter trial has been suspended, apparently because of financial issues. In 2008, the developers of BA-210 published a study that demonstrated an effect in a contusion mouse SCI model and showed evidence of an optimal dose and therapeutic window,163 as well as efficacy. This situation poignantly illustrates the reality of clinical translation in SCI and other diseases. Clinical trials are extremely expensive, especially in their later stages. Investors take a calculated risk, and their risk-benefit perception can be affected by many issues, some of which are due to neither scientific nor clinical factors. This issue is important to progress in the SCI field because there must be a reasonable financial payback to companies and investors to sustain their interest.
ATI-355
The first trial of antibody therapy for SCI involves ATI-355 and is supported by Novartis. In this multicenter open-label phase I/IIA (different doses) study, humanized anti–Nogo-A antibody is being infused into the subarachnoid space for 28 days after complete acute SCI.162 Patients have been recruited over the past 3 years in Europe and, more recently, in Canada. Subarachnoid delivery of monoclonal antibodies directed against Nogo-A in rodent164 and primate models of incomplete SCI165–167 is associated with increased axonal sprouting and regeneration. Primates showed anatomic plasticity and behavioral recovery, although the injury model, C7 hemisection,167 is optimized to show this effect and is not similar to human contusive injury. Continuous localized CSF administration to uninjured primates for 7 days was associated with strong local antibody labeling at C6.168 In the same study, antibodies were found to have been internalized within oligodendrocytes, and Nogo-A receptor expression was downregulated. The biologic role of Nogo-A signaling has been tested by other laboratories, one of which has focused on a 66–amino acid loop domain of Nogo receptor, a key ligand in oligodendrocyte-neuron interactions.169 This team developed a peptide, NEP1-40, that antagonizes this Nogo domain and could be infused subcutaneously for a total of 14 days with a delay of up to 7 days after injury. Marked sprouting and elongation of corticospinal tract and serotonergic axons were observed,170 highly reminiscent of studies by Bregman and collaborators.171 This and other studies appeared to powerfully confirm the clinical potential of Nogo-A receptor blockade to increase post-SCI plasticity.172 An effect has also been shown in stroke, which more substantially confirms the biologic importance of the Nogo receptor in suppressing plasticity.168 For SCI, it is accepted that preclinical testing is more relevant if performed in contusive models, which are more relevant than sharp transection models.173 Applying this principle, the number of studies that support anti–Nogo-A receptor antibody therapy after spinal cord contusion is limited.174 Recognizing the frequent observation that independent laboratories often have difficulty confirming the results of another group, the National Institutes of Health launched a Facilities in Research Excellence—Spinal Cord Injury (FORE-SCI) program to “address key factors that have hampered the translation of research on exciting new therapeutic strategies from the laboratory to clinical trials, to replicate novel experimental therapeutic studies in SCI models, and to compare the efficacy of different treatments in a standardized environment with a minimum of variability in surgery, animal care, outcome evaluation and cellular analyses” (www.ninds.nih.gov/funding/areas/repair_and_plasticity/fore_sci.htm). FORE-SCI investigators were unable to replicate the effects reported by Li and coworkers170 and Steward and colleagues.175 Translational researchers are currently grappling with the significance of independent replication and whether this is a realistic requirement for initiation of clinical trials. Theoretically, replication should be a keystone that establishes the reproducibility of a therapy, but it is clearly difficult to accomplish this. The extent of experimental research supporting the current ATI-355 trial is vast, and the biologic demonstration that Nogo-A receptor antagonism with monoclonal antibodies is associated with structural plasticity is nearly indisputable. Still, translation of this therapy relies on practical issues; the method of delivery—intrathecal antibody infusion—has not been used before in an SCI trial. Because the delivery system is not internalized, there is a risk for infection, and such systems are susceptible to catheter complications. It is also likely that the proplasticity effects may be more prominent in incomplete than in complete SCI patients.
Other Biologics for Which Clinical Trials Have Been Proposed
Chondroitinase: Modification of the Glial Scar to Promote Plasticity
A consequence of SCI is the proliferation of reactive astrocytes and the development of an astroglial scar, which is a physical and molecular barrier to regeneration.70 The glial scar has been the principal target of numerous experimental therapies as diverse as proinflammatory proteins (Piromen)176 and therapeutic irradiation.177 The astrocytes that make up this scar secrete a number of growth inhibitory components into their basal lamina known as chondroitin sulfate proteoglycans (CSPGs). The CPSGs consist of a core protein surrounded by glycosaminoglycan side chains that act as molecular cues to inhibit sprouting and regeneration of axons.178 Modulation of the glial scar has been accomplished experimentally via the degradation of CSPGs by using the CSPG-degrading enzyme chondroitinase ABC (chABC). Injection of chABC has been found to improve motor and sensory function after SCI.178,179 It is possible that this therapy may be used in combination with other restorative therapies to increase the capacity for regeneration and plasticity.180–182 It has also been suggested that CSPGs may signal through the Rho-ROCK pathway and that antagonism of the Rho-ROCK pathway may favorably modulate the glial scar.162 A clinical formulation of chondroitinase is under development by Acorda Therapeutics. This application will also have to solve the delivery issues discussed earlier.
Antibody Therapies to Reduce Spinal Cord Inflammation
Antibodies and partial antibody peptides are important biologics that can be commercially purified. These biologics are especially relevant when they can be delivered via the intravenous route. After SCI, the vascular endothelium upregulates the transmembrane protein intercellular adhesion molecule-1 (ICAM-1).183 The CD11b integrin on leukocytes, especially neutrophils, binds to ICAM-1, and transendothelial migration occurs. This is the major mechanism by which neutrophils enter the spinal cord and induce damage.183 Several studies have assessed the potential of anti-integrin antibodies to reduce spinal cord inflammatory damage by blocking the entry of neutrophils.184,185 Even though this is clearly an important therapeutic target, concerns have been raised regarding suppression of neutrophil function after acute SCI, during which infections such as pneumonia and urinary tract infection are common. Efalizumab, a recombinant humanized anti-CD11b monoclonal antibody tested for psoriasis, was withdrawn from the market because of cases of viral and bacterial infection, including meningitis, fungal infection, and progressive multifocal leukoencephalopathy. For SCI, only a brief period of treatment would be needed, and thus the immunosuppressive risks would be limited. Neutrophil binding of an anti–α4 integrin antibody that could block the entry of neutrophils into the spinal cord does not markedly reduce neutrophil function.186 Careful human safety testing will probably proceed if industry support can be obtained.
Neurotrophins
Given the important role that neurotrophic molecules have played in experimental SCI research, the ability to produce large purified quantities through recombinant technology, and the potential for CSF delivery, it is remarkable that neurotrophins have not been tested in a human SCI clinical trial. Several experimental studies have shown their potent effects in improving SCI; for instance, CSF infusion of epidermal growth factor (EGF) and fibroblast growth factor-2 (FGF-2) increased the mitogenesis of ependyma-derived precursor cells.187 In other CNS diseases, the most extensive clinical experience has thus far involved delivery of glial-derived neurotrophic factor (GDNF) for the treatment of Parkinson’s disease.188,189 In Parkinson’s disease, the mechanistic trophic link between GDNF and dopaminergic cells is clear, whereas diffuse CSF delivery after SCI could have numerous effects. CSF infusions of nerve growth factor (NGF) have been associated with extensive remodeling of intact sensory neurons190 and new pain.191 The greater emphasis in SCI has been on using transplanted cells as localized sources of neurotrophin release to provide neuroprotection,192 which may be inherently safer.
Cellular and Tissue Transplantation for Spinal Cord Injury
Because SCI leads to tissue loss, there has been interest in physical replacement of the substance of the spinal cord with the use of fetal tissue, nerve grafts, cells, and biopolymer scaffolds. There are several rationales for such transplantation, including establishment of a physical substrate for axonal regeneration, axonal remyelination, axonal plasticity, and novel connectivity; replacement of neurons and glia; promotion of angiogenesis; and integrated tissue formation to reduce cavitation. It is thought that extensive cavitation may be a risk factor for the development of progressive syringomyelia. A wide variety of cells have been transplanted, including embryonic193 or fetal neurons,194,195 astrocytes, oligodendrocytes, Schwann cells, olfactory ensheathing glia, transfected autologous fibroblasts, bone marrow stromal cells, and more recently, several types of neural stem cells, both embryonic and adult derived.196 There is a strong experimental basis and evidence for the use of cells, tissues, and biopolymers to support tissue remodeling after SCI. Current concepts in cellular therapies can be considered to have five general aims: (1) neuroprotection and tissue sparing; (2) promotion of endogenous repair through mechanisms such as sustained release of neurotrophin; (3) replacement of lost cells (oligodendrocytes, neurons) with functional integration; (4) support of axonal regeneration, plasticity, and remyelination; and (5) restoration of tissue continuity to provide a vascularized substrate for new axons and neurons. The goal is to reestablish the axonal conductivity of motor, sensory, and autonomic tracts and to restore segmental function, especially in the cervical and lumbar spinal cord. Fetal tissues were especially interesting because it was theorized that novel synaptic connections might occur within them and permit intermediate relay networks to form between the disconnected supraspinal and caudal cord structures.197–199 Fetal tissue transplantation was performed safely in a cohort of patients with chronic SCI.200,201 Some considerations, however, make transplantation of fetal tissue relatively problematic, including the need for pooling of multiple donor sources, which amplifies the potential risk for transmission of disease, the need for immune suppression, direct cord implantation through a surgical incision, and the ethical issues associated with the use of donor tissue from therapeutic abortions. There have been some reports of adverse effects such as tumor formation.202 Key issues associated with cell transplantation are survival, migration, integration, eventual phenotype, and the tissue-remodeling effects that transplanted cells induce. In addition, it is thought that cells may have an advantage over biomaterials that, for example, release trophic factors because integrated cells may be regulated by contact and soluble factors.
