Inflammatory and structural biomarkers in acute traumatic spinal cord ...

Article in press - uncorrected proof Clin Chem Lab Med 2011;49(3):425–433 2011 by Walter de Gruyter • Berlin • New York. DOI 10.1515/CCLM.2011.068

Review

Inflammatory and structural biomarkers in acute traumatic spinal cord injury

Brian K. Kwon1,*, Steve Casha2, R. John Hurlbert3 and V. Wee Yong4 1

Combined Neurosurgical and Orthopaedic Spine Program, Department of Orthopaedics, University of British Columbia, Vancouver, British Columbia, Canada 2 Division of Neurosurgery, Dalhousie University, Halifax, Nova Scotia, Canada 3 Department of Clinical Neurosciences and Surgery, University of Calgary Spine Program, University of Calgary, Calgary, Alberta, Canada 4 Hotchkiss Brain Institute and the Department of Clinical Neurosciences, University of Calgary, Calgary, Alberta, Canada

Abstract The paralysis of an acute spinal cord injury (SCI) remains a catastrophic condition for which there are currently no effective treatments. While the diagnosis of acute traumatic SCI is typically quite easy to make, distinguishing the exact degree of severity and prognosticating the extent of neurologic recovery are challenging. Functional neurologic measures are currently used to stratify injury severity and predict neurologic outcome. However, these measures are often impossible to determine in acutely injured patients. Additionally, for patients deemed to be of a specific injury severity, the variability in spontaneous neurologic recovery is high. Both of these issues severely impair the ability to perform clinical trials in novel therapies for SCI. Biomarkers that could more precisely define the severity of injury and better predict neurologic outcome would be extremely valuable. Furthermore, biological surrogate outcomes measures would be very useful in small preliminary clinical trials of novel therapies if they could inform decisions around the therapeutic regimen for subsequent larger clinical trials. This review highlights our ongoing work in establishing biomarkers for SCI using cerebrospinal fluid samples from acutely injured patients. Keywords: biomarkers; cerebrospinal fluid; clinical trial; spinal cord injury. *Corresponding author: Brian K. Kwon, MD, PhD, FRCSC, Associate Professor, Department of Orthopaedics, University of British Columbia, 6th Floor, Blusson Spinal Cord Center, VGH, 818 West 10th Avenue, Vancouver, BC V5Z 1M9, Canada Phone: q604-875-5857, Fax: q604-875-8223, E-mail: [email protected] Received July 18, 2010; accepted October 19, 2010; previously published online December 23, 2010

Introduction Spinal cord injury (SCI) represents one of the most physically and psychologically devastating traumas an individual can suffer. This can result in lifelong paralysis for individuals who were often otherwise previously healthy. Each year, over 10,000 individuals in North America and many thousands more around the world suffer an acute traumatic SCI and are left paralyzed (1). A recent estimate reported that approximately 1.3 million individuals live with chronic spinal cord paralysis in the US alone (2). Once considered to be an injury of the young, the burgeoning elderly population has also dramatically altered the epidemiology of SCI, with the average age of a newly paraplegic or tetraplegic individual rising from 29 years in the mid-1970s to 40 years in 2005 (3). Recent Canadian data estimates the mean age to be even higher, at 51.3 years (4), reflecting the increasing number of elderly sustaining spinal cord injuries in falls. At any age, suffering a spinal cord injury is catastrophic, not only for the individual on a personal level, but also for society as a whole with respect to the enormous costs of acute and chronic care (5, 6). Enormous progress has undeniably been made over the past 30 years in the medical, surgical, and rehabilitative care of individuals with acute and chronic spinal cord injuries (7–10). However, despite considerable global research efforts, the search for a ‘‘cure’’ for spinal cord injury has yet to produce a convincingly efficacious treatment that substantially improves neurologic function in SCI patients. The only neuroprotective treatment option for acute SCI patients is methylprednisolone, and while this was once a widely accepted ‘‘standard of care’’, many physicians have now abandoned it due to concerns about its side effects and skepticism around its purported efficacy (11). Biomarkers have been vigorously studied in chronic neurodegenerative disorders where their incipient onset is often measured in months, and their precise diagnosis may be uncertain for years. In contrast, the nature of an acute traumatic SCI typically leaves little doubt as to the neurologic diagnosis of spinal cord paralysis or the time of onset at the moment the trauma occurred. Furthermore, given that no effective treatments currently exist for patients who have suffered an acute spinal cord injury, biomarkers of spinal cord injury would not currently be useful for helping clinicians select one therapeutic agent vs. another. In this regard, it would be reasonable to question the rationale for seeking biomarkers in acute spinal cord injury. However, biomarkers may provide a useful measure of the extent of primary injury to the spinal cord after acute traumatic injury, and it is in

2010/426

Article in press - uncorrected proof 426 Kwon et al.: Inflammatory and structural biomarkers in acute traumatic spinal cord injury

this context that biomarkers in acute spinal cord may play an important role in the research community’s ongoing search for new and effective neuroprotective treatments.

Establishing new treatments for spinal cord injury – the role of biological markers Not unlike other neurologic disorders, the two general elements for establishing new SCI treatments are the: 1. preclinical development and optimization of treatments in the laboratory setting, and 2. clinical evaluation and validation in a human trial of SCI patients. Both of these elements have met with significant challenges. First, the preclinical development of new treatments is dependent largely upon rodent models of SCI, within which a great deal of scientific research has gone into understanding the pathophysiology of acute SCI. In contrast, a great deal less is understood about the biology of the human injury. It is convenient to assume that the pathobiology of human injury will be similar to that in rat or mouse models. However, the observation that numerous efficacious treatments in such animal models of SCI failed to demonstrate convincing efficacy in clinical trials (12) intimates that important biological differences may exist between the human and animal condition. Characterizing the pathophysiology of acute human SCI therefore has considerable translational merit in determining the clinical relevance of our commonly utilized animal models. In comparison to the myriad of histological, biochemical, molecular, and physiological assessments that can be done on the spinal cord of an animal, evaluating the biology of human SCI depends largely upon studies of post-mortem spinal cord tissue (13–15), and serum (16, 17) and cerebrospinal fluid (CSF) samples from live patients (18–20). Such descriptive studies that provide insights into the human neurobiology of injury also represent the source of biomarker discovery. The second element of developing new therapies – their clinical evaluation in human trials – is an excruciatingly time-consuming, labor-intensive, and costly process. The monumental effort needed to do such clinical trials is easily overlooked, particularly when media reports publicize the promise of spinal cord injury ‘‘breakthroughs’’ in the laboratory (21) and spectacular ‘‘cures’’ with clinically unproven treatments (22). A good illustration of this process was the Sygen (GM-1 ganglioside) clinical trial, the last multicenter trial to be completed on a novel acute SCI treatment. Sygen’s investigation began with a small phase I study in Maryland in the late 1980s (23, 24), and ended with a 760-patient phase III multicenter randomized controlled trial published in 2001 (25). In total, almost 14 years passed between the initiation of the Phase 1 and conclusion of the Phase III study (the latter of which was run at 28 neurotrauma institutions throughout North America). The reasons for the excruciatingly slow pace of SCI trials are multiple. First, our current method of determining the extent of paralysis in acute SCI patients is with a manual neurologic assessment, performed according to the American Spinal Injury Association (ASIA) standards. This requires

the SCI patient to demonstrate his/her residual strength in 10 muscle groups in the arms and legs, and to report their sensation to pin-prick and light touch throughout the body, including the peri-anal region. It is often impossible to do this assessment validly, particularly in patients who have multiple injuries, brain injuries, or are intoxicated or sedated pharmacologically (26). Second, even for patients with the same extent of paralysis on the ASIA Impairment Scale (AIS), the variability in spontaneous neurologic recovery is quite high, making it necessary to recruit large numbers of patients in order to have sufficient statistical power to detect a modest (yet meaningful) improvement in neurologic function (27). As an illustration of this, the post-hoc analysis of the variability in spontaneous neurologic recovery in the Sygen multicenter study revealed that in order to detect at 5 point difference in motor score in patients with complete cervical spinal cord injury, one would need to enroll approximately 380 patients (27), a number that was not achieved after 5 years of enrolment for the Sygen multicenter trial (25). Clearly, the dependence of clinical trials on functional neurologic metrics to recruit patients and then interpret the efficacy of the intervention is a major impediment because of these two important issues – the inability to perform such measures in many patients, and the variability in spontaneous recovery in those patients in whom such functional measures can be obtained. Biological markers that 1. reflect the extent of damage to the spinal cord and 2. more precisely predict neurologic recovery than the baseline neurologic assessment would be immensely helpful in conducting such trials. An additional utility for a biological marker in a neuroprotection trial for acute SCI patients would be as a surrogate outcome measure to monitor a physiologic response to the intervention. In addition, biological markers may provide surrogate mechanistic outcomes that are currently poorly addressed. While the ultimate goal for neuroprotective treatments is obviously to promote functional recovery, the ability to demonstrate that the treatment is having a biological effect is far from trivial. Again, because of the high variability in spontaneous neurologic recovery, it may be extremely difficult – if not impossible – to determine in a small Phase 2 study whether the intervention is actually having any functional effect. Decisions about whether to proceed with a large and expensive Phase 3 study would be greatly facilitated if there was some indication of whether the intervention was having the expected biological response in the cord. Such biomarkers of response could be critical in deciding whether to pursue further clinical evaluation of a drug, and if so, in determining important parameters, such as dose, time window of intervention, and monitoring schedule for a more definitive clinical trial. For example, it is conceivable that the doses of such drugs used in previous SCI clinical trials (even ‘‘high-dose’’ steroids) were insufficient to induce the desired biological response within the injured spinal cord (28). Furthermore, it has been argued that given the broad spectrum of secondary injury mechanisms that are initiated following SCI, it is unlikely that a single agent addressing one mechanism would result in adequate neuroprotection to

Article in press - uncorrected proof Kwon et al.: Inflammatory and structural biomarkers in acute traumatic spinal cord injury 427

provide a clinically detectable change. Biomarker outcomes may provide evidence of smaller subclinical benefits, which would inform combination therapies that may be more likely to be clinically efficacious.

Candidate biological markers for spinal cord injury Given that samples of spinal cord parenchyma are obviously unobtainable from patients after a spinal cord injury, the ‘‘tissue’’ in greatest proximity to the injured cord is cerebrospinal fluid (CSF). Importantly, SCI studies in animal models have validated the utility of CSF as a biological representation of what is occurring within the injured spinal cord. In an animal model of acute contusive SCI, Wang et al. demonstrated that increases in IL-1b concentrations within the spinal cord over the first 72 h post-injury were paralleled by increased IL-1b concentrations within the CSF (29). The specificity of this relationship between the cord and surrounding CSF was supported by the fact that the systemic (serum) concentrations of IL-1b were lower and did not correlate with that in the injured spinal cord. Harrington et al. also demonstrated a similar phenomenon with tumor necrosis factor-a (TNF-a) (30). Again, the TNF-a concentrations in the cord correlated closely with those measured in the CSF, while serum TNF-a concentrations were much lower and did not correlate with cord concentrations. In contrast to acute traumatic brain injury, comparatively little has been studied on CSF biomarkers for acute, traumatic spinal cord injury. The reason for this disparity is largely related to the accessibility of CSF for sampling. Patients with traumatic brain injuries often have intracranial pressure monitoring through extraventricular drains, through which CSF samples can be taken for analysis. Such monitoring is typically not performed in patients with traumatic spinal cord injuries, making it logistically more challenging to obtain such CSF samples. While performing a lumbar puncture to obtain a CSF sample may be relatively easy in the elective, ambulatory clinic setting for patients with chronic neurodegenerative diseases, doing so in an acute SCI patient with a dislocated cervical or thoracolumbar spine and numerous other injuries can be quite difficult. ‘‘Inflammatory’’ biomarkers

Inflammation is considered to play a central role in the pathophysiology of secondary injury after acute SCI. It is almost certain that any therapy administered in the acute spinal cord injury setting will influence or be influenced by the inflammatory response – a notion supported by animal studies of pharmacologic agents (31), growth factors (32), gene therapy vectors (33), and cell transplants (34, 35). The same would quite likely apply to any human SCI clinical trial of an acute intervention. The correlation between mediators of inflammation and severity of spinal cord injury has been investigated by numerous investigators in both animal models and human patients. Animal studies in which the severity of spinal cord

injury can be precisely controlled have confirmed that inflammatory cytokine concentrations after injury correlate with injury severity. Using a rodent thoracic weight drop contusive SCI model in which a 10-g weight was dropped from either 3 cm (mild injury) or 12 cm (severe injury), Yang and colleagues reported that concentrations of IL-1b, IL-6, and TNF-a mRNA and protein in the spinal cord were significantly increased after the severe but not the mild injury (36). In another rodent thoracic contusive SCI study in which the cord was subjected to impactor forces of 100, 150, or 200 kdynes (Infinite Horizon Impactor, Precision System and Instrumentation, Lexington, KY, USA), Knerlich-Lukoschus and colleagues reported increased expression of CCL2 (MCP-1) in the 150 and 200 kdyne injuries as compared to the mild 100 kdyne injury (37). This differential inflammatory cytokine expression was correlated with the neuropathic pain in the most severely injured animals. These preclinical studies support the concept that the extent to which specific inflammatory cytokines are expressed after cord injury is ‘‘titrated’’ according to the severity of the neurologic deficit, and substantiate the approach of evaluating such cytokines as potential biomarkers of injury severity. The evaluation of inflammatory mediators in human spinal cord injury has confirmed that injury severity can influence this process. Nishisho and colleagues conducted a study in which lumbar punctures were performed on six patients with complete SCI and five with incomplete paralysis in order to obtain CSF samples for analysis of leukotriene C4 (LTC4), thromboxane B2 (TXB2), and 6-keto-prostaglandin F1a (6keto-PGF1a) (38). They reported that all of these eicosanoids were increased in the CSF of SCI patients as compared to non-injured controls, and that the LTC4 concentrations were significantly increased in patients with complete as compared to incomplete paralysis. Concentrations of TXB2 and 6-ketoPGF1a were also higher in complete SCI patients, but not statistically higher. A Taiwanese study of seven cervical SCI patients evaluated CSF samples obtained via lumbar puncture in an effort to discern whether the administration of methylprednisolone altered the concentration of inflammatory cytokines within the CSF (19). Three patients were administered MP (one complete and two incomplete SCI), the other four did not receive MP (one complete and three incomplete SCI). The CSF samples were obtained ‘‘1–4 days after SCI’’, but the exact time of acquisition post-injury for each CSF sample was not specified. The CSF samples were analyzed using a human cytokine antibody array, ELISA, and gelatin zymography. The authors observed increased concentrations of IL-8, MCP-1, IL-6, and ICAM in complete SCI patients, and they reported that methylprednisolone appeared to reduce matrix metalloproteinase-2 (MMP-2) and MMP-9 concentrations. However, in general the small numbers of patients precluded any meaningful statistical analysis, and therefore any definite conclusions about the influence of injury severity or methylprednisolone administration on CSF inflammatory cytokines. The study of non-traumatic spinal cord injury is also relevant to the discussion of inflammatory biomarkers. In 25 patients with spinal cord paralysis secondary to transverse


myelitis (a primary local inflammatory disorder), Kaplin et al. demonstrated a 262-fold increase in CSF IL-6 concentrations when compared to that in 16 control patients (39). In addition, they found a strong correlation between IL-6 concentrations and the clinical severity of paralysis. Finally, in patients undergoing thoracoabdominal aortic aneurysm repair, lumbar intrathecal catheters are commonly installed to drain CSF and improve spinal cord perfusion pressure, with the hope of avoiding ischemic damage to the thoracic cord and subsequent paralysis. In a study of 15 patients undergoing such aneurysm repairs, Kunihara et al. found that the occurrence of ischemic paralysis was associated with increased concentrations of IL-8 in the CSF (40). In contrast, serum IL-8 concentrations were not associated with ischemic paralysis, illustrating the ‘‘specificity’’ of the CSF for cord pathology. ‘‘Structural’’ biomarkers

The search for biomarkers is obviously not limited to aspects of the inflammatory response. Structural proteins that reflect injury to neural tissue are perhaps more obvious choices for biomarkers of SCI, and have been investigated extensively in traumatic brain injury (41). Increased concentrations of tau, a microtubule associated protein, have been measured in the CSF of TBI patients, with worsened long-term outcome correlated with higher tau concentrations (42, 43). In the study by Ost et al., while total tau concentrations in the CSF on Day 2 or 3 post-injury correlated with poor outcome at 1 year, the serum total tau concentrations did not change and therefore were not helpful as a biomarker of injury severity. S100b is a calcium binding protein found in astroglial and Schwann cells, in addition to adipocytes, chondrocytes, and melanocytes. It is arguably the most extensively studied marker after TBI, and has been assessed both in the serum and CSF (44–48). Controversy exists about serum S100b concentrations as a measure of injury severity, as its release from adipose tissue may confound the interpretation of increased concentrations (49), and it may in fact have some beneficial/reparative effects (50–52). Glial fibrillary acidic protein (GFAP) released from injured glial cells and the light subunit of neurofilament protein (NFL) released from axons after spinal cord injury has been evaluated after spinal cord injuries by Guez et al. in a small series of six patients who underwent lumbar punctures to obtain CSF (18). The authors reported that increased concentrations of GFAP and NFL correlated with severity of paralysis, but this was based on an evaluation of only three patients in whom CSF samples were obtained at an early time point (1 day post-injury), while the rest were obtained at 3 weeks post-injury. Structural biomarkers have also been evaluated in the CSF of patients who suffer ischemic damage to the spinal cord during or after surgical repairs of their thoraco-abdominal aortic aneurysms. CSF samples in these circumstances are obtained through intrathecal catheters that are routinely installed to monitor CSF pressure. Such studies have suggested that increases in S100b, GFAP, and NFL may also reflect damage to the spinal cord induced by ischemic insult (40, 53, 54). These findings support the rationale for further studying

these ‘‘structural’’ biomarkers as measures of injury severity in traumatic spinal cord injury.

Clinical trial of CSF pressure monitoring and biochemical analysis In 2006, Kwon and colleagues at the University of British Columbia initiated a prospective randomized clinical trial of CSF drainage for patients with acute SCI, with the rationale that lowering intrathecal pressure might improve spinal cord perfusion pressure in these patients (55). In this trial, a lumbar indwelling intrathecal catheter was inserted prior to surgery (within 48 h of injury) to monitor CSF pressure and/or drain CSF for 72 h. During this time, CSF samples were collected every 6–8 h for analysis and subjected to biochemical analysis using a Multiplex device or standard ELISAs. Blood samples were collected simultaneously in order to compare CSF and serum concentrations of the various proteins of interest. Additionally, control ‘‘non-injury’’ CSF samples were collected via a single lumbar puncture from patients undergoing hip or knee surgery during their spinal anesthetics, or intra-operatively from patients undergoing lumbar spine fusions. The multiplex analysis included: TNF-a, TNF-R1, IL-1b, IL-2, IL-2R, IL-4, IL-5, IL-6, IL-7, IL-8, IL-10, IL-12p40, IL-13, IL-15, IL-16, IL-17, IP-10, MCP-1, interferon-a (IFN-a), IFN-g, eotaxin, GM-CSF, monokine induced by interferon (MIG), macrophage inflammatory protein-1a (MIP-1a), MIP-1b, RANTES, brain-derived neurotrophic factor (BDNF), fibroblast growth factor (FGF)-basic, glial cell line-derived neurotrophic factor (GDNF), and vascular endothelial growth factor (VEGF). Standard ELISAs were also performed for the structural proteins: total tau, S100b, and GFAP. In this study, the majority of the cytokines and all of the growth factors on the multiplex kit were not measurable in the CSF using this technology (20). We documented substantial early post-injury elevations in IL-6, IL-8, MCP-1, TNF-R1, IL-16, and IP10 (Figure 1). Additionally, we saw increases in tau, S100b, and GFAP. In general, the CSF concentrations of these cytokines and structural proteins far exceeded that of normal ‘‘non-SCI’’ control patients. Additionally, the CSF concentrations far exceeded that of the serum samples taken from the SCI patients at the same time, often by many orders of magnitude (Figure 2). Importantly, at 24 h post-injury, an injury-severity dependent relationship was observed in the CSF concentrations of IL-6, IL-8, MCP-1, tau, GFAP, and S100b, such that patients with complete, AIS A spinal cord injuries (complete motor and sensory paralysis) appeared different from those with AIS B injuries (complete motor paralysis but incomplete sensory loss), and different from those with AIS C injuries (incomplete motor and sensory loss) (Figure 3). Using this observation, we were able to establish a ‘‘biochemical prediction model’’ that utilized the 24-h post-injury concentration of IL-8, S100b, and GFAP, to classify injury severity with an accuracy of 89%. Furthermore, the prediction model


Figure 1 Inflammatory and structural biomarkers in human CSF. The concentrations of CSF biomarkers are plotted over time according to the baseline severity of paralysis (ASIA A, B, or C) in 27 patients recruited in Vancouver. The blue inset within each graph is the average CSF concentration in 12 non-SCI, control individuals, with the minimum and maximum listed below in parenthesis. (Reprinted from Kwon et al., J Neurotrauma, 2010, with permission.)

Figure 2 CSF concentrations of IL-6, IL-8, MCP-1, Tau, S100b, and GFAP at 24 h post-injury are correlated with baseline injury severity (ASIA A, B, or C). This Figure illustrates the injury-severity dependent pattern of expression for various inflammatory and structural proteins, analyzed in 27 SCI patients in Vancouver. Statistically significant differences are indicated (*) (p-0.05, Wilcoxon signed-rank test). (Reprinted from Kwon et al., J Neurotrauma, 2010, with permission.)


Figure 3 CSF concentrations of IL-6, IL-8, MCP-1, Tau, S100b, and GFAP far exceed serum concentrations. Blood samples were collected at the same time as CSF samples to compare CSF and serum concentrations. Here, the ratio between CSF and serum concentrations of IL-6, IL-8, MCP-1, Tau, S100b, and GFAP at 24-h post-injury illustrate that the CSF concentrations represent a CNS-specific process. (Reprinted from Kwon et al., J Neurotrauma, 2010, with permission.)

utilizing the CSF concentrations of IL-8, S100b, and GFAP at 24 h post-injury was able to predict motor recovery at 6 months post-injury in the cervical SCI patients, as well as (if not slightly better than) the functional ASIA classification. This reflects the fairly intuitive concept that a marker of biological damage to the cord might be a better reflection of the severity of injury (and ultimately neurologic outcome) than the gross functional measures. Validation of these biomarkers with an independent cohort of patients is ongoing. Additionally, further work is needed to evaluate CSF at earlier time points to establish whether the pattern of injury severity dependent expression is maintained. The 24-h postinjury time point was chosen in our studies because this represented a time point where we felt that most patients could realistically have a CSF sample collected in the future. It is recognized, however, that patients enroled in neuroprotective trials (such as the minocycline trial discussed in the next section) may have CSF samples collected prior to this time point, and so the utility of biomarkers at earlier time points needs to be established.

Clinical trial of minocycline and CSF pressure monitoring In 2004, in Calgary we initiated a prospective randomized controlled single center trial of minocycline for acute SCI

(ClinicalTrials.gov identifier: NCT00559494) (56). Minocycline, a tetracycline antibiotic with various potential neuroprotective properties (57), was evaluated against placebo. Lumbar intrathecal catheters were installed to measure CSF pressure and to obtain CSF samples for both pharmacokinetic studies and biomarker studies. This study represents a prototype trial for how changes in inflammatory mediators may be utilized to monitor the effect of the drug. CSF from 27 subjects was assayed by ELISA for several inflammatory and structural markers: IL-1b, MMP-9, MCP1, IP-10, heme oxygenase-1 (HO-1), neurofilament heavy chain (NfH), oxidation products of nitric oxide (NOx) and neural cell adhesion molecule (NCAM). Cumulative concentrations over seven days were determined and compared among motor complete (ASIA A or B) and motor incomplete (ASIA C or D) subjects to examine the relationship to injury severity. While all molecules with the exception of NOx demonstrated increased concentrations with motor complete injury, cumulative concentrations of IL-1b, MMP-9 and HO-1 showed a statistically significant correlation with injury severity (58). In addition, cumulative concentrations of IL-1b, MMP-9 and IP-10 over the first 7 days correlated negatively with neurological recovery at 12 months. With minocycline treatment, early CSF NfH concentrations (days 1–3) were reduced, suggesting an effect on secondary injury mechanisms during that time. However, minocycline also blunted a rise in HO-1 at 7 days, possibly implicating


reduced HO-1 mediated neurotoxicity among the human targets of minocycline action. Notably, a convincing effect of minocycline in reducing IL-1b, MMP-9 and NOx concentrations was not seen in the human study, despite these molecules being reduced by minocycline treatment in animal models (57, 59). This work illustrates possible differences in the importance of specific secondary injury mechanisms between animal and human SCI and the possible utility of biomarkers in identifying those mechanisms more prominent in the human.

ined reliably. Our combined experience of over 80 patients with lumbar punctures and indwelling intrathecal drains indicates that CSF can be safely obtained in acute SCI patients. Given the nature of the unique biological insights garnered from CSF in this setting, we hope that the field will continue to expand in the near future. Recognizing the challenges of establishing new therapies for spinal cord injury, we feel that there is great utility to pursuing the development of biomarkers that can be used to facilitate this important process.

Acknowledgments Summary of findings between Vancouver and Calgary in human SCI While the clinical trials conducted in Vancouver and in Calgary were both on acute SCI patients, they differed slightly in their time window for enrolment (48 h in Vancouver, 12 h in Calgary) and the nature of the intervention (CSF drainage in Vancouver, minocycline in Calgary). The analysis of CSF was conducted with both multiplex and standard ELISAs in Vancouver, and with single target ELISAs in Calgary. The use of single target ELISA in Calgary allowed for the detection of IL-1b, while the concentrations reported were at or below the detection level for the multiplex analyses. In Vancouver, baseline injury severity (ASIA A, B, or C) was correlated to the 24 h post-injury concentrations of IL-6, IL-8, MCP-1, tau, GFAP, and S100b (i.e., single time point), while in Calgary, baseline injury severity was related to the 7 day cumulative concentrations (area under the curve) of IL-1b, MMP-9, MCP-1, IP-10, HO-1, and NfH, which was higher in motor complete vs. incomplete SCI. In Vancouver, the 24 h post-injury concentrations of IL-8, S100b, and GFAP predicted segmental motor outcome at 6 months post-injury, while in Calgary, the cumulative IL-1, MMP-9 and IP-10 concentrations over 7 days correlate negatively with recovery at 6 months. Collectively, we have thus identified structural (NfH, S100b, tau, GFAP), inflammatory (IL-1b, IL-8, MMP-9, MCP-1, IP-10) and enzymatic (HO-1) biomarkers in acute SCI, and have initiated further validation studies of these.

Conclusions The field of cerebrospinal fluid biomarker research in spinal cord injury is relatively small, and the recent reporting of inflammatory and structural proteins within the CSF of acute SCI patients provides only an early perspective of the potential for this line of research. Much validation is still required to establish which biomarkers are best suited to classifying injury severity, predicting neurologic recovery, and monitoring biological outcome in early therapeutic interventions. At the current time, the validation of biomarkers is tied closely to the evaluation of neurologic function using the ASIA standards; however, in time, it is conceivable that biomarkers could, by themselves, provide accurate information about the extent of neurologic injury in patients who cannot be exam-

Dr. Kwon holds a New Investigator Award from the Canadian Institutes for Health Research. Dr. Yong is a Canada Research Chair (Tier 1) in Neuroimmunology.

Conflict of interest statement Authors’ conflict of interest disclosure: The authors stated that there are no conflicts of interest regarding the publication of this article. Research funding: None declared. Employment or leadership: None declared. Honorarium: None declared.

References 1. van den Berg ME, Castellote JM, Mahillo-Fernandez I, de Pedro-Cuesta J. Incidence of spinal cord injury worldwide: a systematic review. Neuroepidemiology 2010;34:184–92. 2. Christopher, Dana Reeve Foundation. One Degree of Separation; Paralysis and Spinal Cord Injury in the United States. 2009. 3. National Spinal Cord Injury Statistical Center. Spinal Cord Injury, Facts and Figures at a Glance. 2008. 4. Couris CM, Guilcher SJ, Munce SE, Fung K, Craven BC, Verrier M, et al. Characteristics of adults with incident traumatic spinal cord injury in Ontario, Canada. Spinal Cord 2010;48: 39–44. 5. French DD, Campbell RR, Sabharwal S, Nelson AL, Palacios PA, Gavin-Dreschnack D. Health care costs for patients with chronic spinal cord injury in the Veterans Health Administration. J Spinal Cord Med 2007;30:477–81. 6. New PW, Jackson T. The costs and adverse events associated with hospitalization of patients with spinal cord injury in Victoria, Australia. Spine 2010;35:796–802. 7. Thibault-Halman G, Casha S, Singer S, Christie S. Acute management of nutritional demands following SCI. J Neurotrauma 2010. 8. Casha S, Christie S. A systematic review of intensive cardiopulmonary management after spinal cord injury. J Neurotrauma 2010. 9. Furlan JC, Noonan V, Cadotte DW, Fehlings MG. Timing of decompressive surgery of spinal cord after traumatic spinal cord injury: an evidence-based examination of pre-clinical and clinical studies. J Neurotrauma 2010. 10. Donovan WH. Donald Munro Lecture. Spinal cord injury – past, present, and future. J Spinal Cord Med 2007;30:85–100.


11. Hurlbert RJ, Hamilton MG. Methylprednisolone for acute spinal cord injury: 5-year practice reversal. Can J Neurol Sci 2008;35:41–5. 12. Tator CH. Review of treatment trials in human spinal cord injury: issues, difficulties, and recommendations. Neurosurgery 2006;59:957–82. 13. Fleming JC, Norenberg MD, Ramsay DA, Dekaban GA, Marcillo AE, Saenz AD, et al. The cellular inflammatory response in human spinal cords after injury. Brain 2006;129:3249–69. 14. Kakulas BA. The applied neuropathology of human spinal cord injury. Spinal Cord 1999;37:79–88. 15. Norenberg MD, Smith J, Marcillo A. The pathology of human spinal cord injury: defining the problems. J Neurotrauma 2004; 21:429–40. 16. Bao F, Bailey CS, Gurr KR, Bailey SI, Rosas-Arellano MP, Dekaban GA, et al. Increased oxidative activity in human blood neutrophils and monocytes after spinal cord injury. Exp Neurol 2009;215:308–16. 17. Hayes KC, Hull TC, Delaney GA, Potter PJ, Sequeira KA, Campbell K, et al. Elevated serum titers of proinflammatory cytokines and CNS autoantibodies in patients with chronic spinal cord injury. J Neurotrauma 2002;19:753–61. 18. Guez M, Hildingsson C, Rosengren L, Karlsson K, Toolanen G. Nervous tissue damage markers in cerebrospinal fluid after cervical spine injuries and whiplash trauma. J Neurotrauma 2003;20:853–8. 19. Tsai MC, Wei CP, Lee DY, Tseng YT, Tsai MD, Shih YL, et al. Inflammatory mediators of cerebrospinal fluid from patients with spinal cord injury. Surg Neurol 2008;70(Suppl 1):S1–24. 20. Kwon BK, Stammers AM, Belanger LM, Bernardo A, Chan D, Bishop CM, et al. Cerebrospinal fluid inflammatory cytokines and biomarkers of injury severity in acute human spinal cord injury. J Neurotrauma 2010. 21. CNN. Same blue dye in M&Ms linked to reducing spine injury. July 28, 2009; Available at: http://www.cnn.com/2009/ HEALTH/07/28/spinal.injury.blue.dye/index.html. 22. Layden T. Kevin Everett, The Road Back. Sports Illustrated 107w24x, 56-67. 12-17-2007. 23. Geisler FH, Dorsey FC, Coleman WP. GM1 gangliosides in the treatment of spinal cord injury: report of preliminary data analysis. Acta Neurobiol Exp (Wars) 1990;50:515–21. 24. Geisler FH, Dorsey FC, Coleman WP. Recovery of motor function after spinal-cord injury – a randomized, placebo-controlled trial with GM-1 ganglioside. N Engl J Med 1991;324:1829–38. 25. Geisler FH, Coleman WP, Grieco G, Poonian D. The Sygen multicenter acute spinal cord injury study. Spine 2001;26:S87– 98. 26. Burns AS, Lee BS, Ditunno JF Jr, Tessler A. Patient selection for clinical trials: the reliability of the early spinal cord injury examination. J Neurotrauma 2003;20:477–82. 27. Fawcett JW, Curt A, Steeves JD, Coleman WP, Tuszynski MH, Lammertse D, et al. Guidelines for the conduct of clinical trials for spinal cord injury as developed by the ICCP panel: spontaneous recovery after spinal cord injury and statistical power needed for therapeutic clinical trials. Spinal Cord 2007;45:190– 205. 28. Hawryluk GW, Rowland J, Kwon BK, Fehlings MG. Protection and repair of the injured spinal cord: a review of completed, ongoing, and planned clinical trials for acute spinal cord injury. Neurosurg Focus 2008;25:E14. 29. Wang CX, Olschowka JA, Wrathall JR. Increase of interleukin1beta mRNA and protein in the spinal cord following experimental traumatic injury in the rat. Brain Res 1997;759:190–6.

30. Harrington JF, Messier AA, Levine A, Szmydynger-Chodobska J, Chodobski A. Shedding of tumor necrosis factor type 1 receptor after experimental spinal cord injury. J Neurotrauma 2005;22:919–28. 31. Gorio A, Madaschi L, Di Stefano B, Carelli S, Di Giulio AM, De Biasi S, et al. Methylprednisolone neutralizes the beneficial effects of erythropoietin in experimental spinal cord injury. Proc Natl Acad Sci USA 2005;102:16379–84. 32. Bouhy D, Malgrange B, Multon S, Poirrier AL, Scholtes F, Schoenen J, et al. Delayed GM-CSF treatment stimulates axonal regeneration and functional recovery in paraplegic rats via an increased BDNF expression by endogenous macrophages. FASEB J 2006;20:1239–41. 33. Abdellatif AA, Pelt JL, Benton RL, Howard RM, Tsoulfas P, Ping P, et al. Gene delivery to the spinal cord: comparison between lentiviral, adenoviral, and retroviral vector delivery systems. J Neurosci Res 2006;84:553–67. 34. Yan J, Welsh AM, Bora SH, Snyder EY, Koliatsos VE. Differentiation and tropic/trophic effects of exogenous neural precursors in the adult spinal cord. J Comp Neurol 2004;480:101–14. 35. Pearse DD, Marcillo AE, Oudega M, Lynch MP, Wood PM, Bunge MB. Transplantation of Schwann cells and olfactory ensheathing glia after spinal cord injury: does pretreatment with methylprednisolone and interleukin-10 enhance recovery? J Neurotrauma 2004;21:1223–39. 36. Yang L, Jones NR, Blumbergs PC, Van Den HC, Moore EJ, Manavis J, et al. Severity-dependent expression of pro-inflammatory cytokines in traumatic spinal cord injury in the rat. J Clin Neurosci 2005;12:276–84. 37. Knerlich-Lukoschus F, Juraschek M, Blomer U, Lucius R, Mehdorn HM, Held-Feindt J. Force-dependent development of neuropathic central pain and time-related CCL2/CCR2 expression after graded spinal cord contusion injuries of the rat. J Neurotrauma 2008;25:427–48. 38. Nishisho T, Tonai T, Tamura Y, Ikata T. Experimental and clinical studies of eicosanoids in cerebrospinal fluid after spinal cord injury. Neurosurgery 1996;39:950–6. 39. Kaplin AI, Deshpande DM, Scott E, Krishnan C, Carmen JS, Shats I, et al. IL-6 induces regionally selective spinal cord injury in patients with the neuroinflammatory disorder transverse myelitis. J Clin Invest 2005;115:2731–41. 40. Kunihara T, Sasaki S, Shiiya N, Miyatake T, Mafune N, Yasuda K. Proinflammatory cytokines in cerebrospinal fluid in repair of thoracoabdominal aorta. Ann Thorac Surg 2001;71:801–6. 41. Kovesdi E, Luckl J, Bukovics P, Farkas O, Pal J, Czeiter E, et al. Update on protein biomarkers in traumatic brain injury with emphasis on clinical use in adults and pediatrics. Acta Neurochir (Wien) 2010;152:1–17. 42. Ost M, Nylen K, Csajbok L, Ohrfelt AO, Tullberg M, Wikkelso C, et al. Initial CSF total tau correlates with 1-year outcome in patients with traumatic brain injury. Neurology 2006;67:1600– 4. 43. Kay AD, Petzold A, Kerr M, Keir G, Thompson E, Nicoll JA. Alterations in cerebrospinal fluid apolipoprotein E and amyloid beta-protein after traumatic brain injury. J Neurotrauma 2003; 20:943–52. 44. Wiesmann M, Steinmeier E, Magerkurth O, Linn J, Gottmann D, Missler U. Outcome prediction in traumatic brain injury: comparison of neurological status, CT findings, and blood levels of S100B and GFAP. Acta Neurol Scand 2010;121:178–85. 45. Rainey T, Lesko M, Sacho R, Lecky F, Childs C. Predicting outcome after severe traumatic brain injury using the serum S100B biomarker: results using a single (24h) time-point. Resuscitation 2009;80:341–5. 46. Nylen K, Ost M, Csajbok LZ, Nilsson I, Hall C, Blennow K,


47. 48. 49.

50.

51.

52.

53.

et al. Serum levels of S100B, S100A1B and S100BB are all related to outcome after severe traumatic brain injury. Acta Neurochir (Wien) 2008;150:221–7. Sen J, Belli A. S100B in neuropathologic states: the CRP of the brain? J Neurosci Res 2007;85:1373–80. Cotena S, Piazza O, Storti M. The S100B protein and traumatic brain injury. J Neurosurg 2006;104:435–6. Chatfield DA, Zemlan FP, Day DJ, Menon DK. Discordant temporal patterns of S100beta and cleaved tau protein elevation after head injury: a pilot study. Br J Neurosurg 2002;16:471–6. Kleindienst A, Meissner S, Eyupoglu IY, Parsch H, Schmidt C, Buchfelder M. Dynamics of S100B release into serum and cerebrospinal fluid following acute brain injury. Acta Neurochir Suppl 2010;106:247–50. Kleindienst A, Hesse F, Bullock MR, Buchfelder M. The neurotrophic protein S100B: value as a marker of brain damage and possible therapeutic implications. Prog Brain Res 2007; 161:317–25. Kleindienst A, Ross BM. A critical analysis of the role of the neurotrophic protein S100B in acute brain injury. J Neurotrauma 2006;23:1185–200. Winnerkvist A, Anderson RE, Hansson LO, Rosengren L, Estrera AE, Huynh TT, et al. Multilevel somatosensory evoked potentials and cerebrospinal proteins: indicators of spinal cord injury in thoracoabdominal aortic aneurysm surgery. Eur J Cardiothorac Surg 2007;31:637–42.

54. Khaladj N, Teebken OE, Hagl C, Wilhelmi MH, Tschan C, Weissenborn K, et al. The role of cerebrospinal fluid S100 and lactate to predict clinically evident spinal cord ischaemia in thoraco-abdominal aortic surgery. Eur J Vasc Endovasc Surg 2008;36:11–9. 55. Kwon BK, Curt A, Belanger LM, Bernardo A, Chan D, Markez JA, et al. Intrathecal pressure monitoring and cerebrospinal fluid drainage in acute spinal cord injury: a prospective randomized trial. J Neurosurg Spine 2009;10:181–93. 56. Casha S, Zygun D, McGowan D, Yong VW, Hurlbert RJ. Neuroprotection with minocycline after spinal cord injury: results of a double blind randomized controlled pilot study. Congress of Neurological Surgeons, Annual Meeting. 2009. New Orleans, LA. 57. Yong VW, Wells J, Giuliani F, Casha S, Power C, Metz LM. The promise of minocycline in neurology. Lancet Neurol 2004; 3:744–51. 58. Casha S, Rice T, Stirling DP, McGowan D, Silva C, Smekal M, et al. CSF inflammatory changes correlate with injury severity and recovery after human spinal cord injury. Congress of Neurological Surgeons, Annual Meeting. 2009. New Orleans, LA. 59. Wells JE, Hurlbert RJ, Fehlings MG, Yong VW. Neuroprotection by minocycline facilitates significant recovery from spinal cord injury in mice. Brain 2003;126:1628–37.