A comparative study of CSF neurofilament light and heavy chain protein in MS

Transcription

1 482374MSJ1.1177/ Multiple Sclerosis JournalKuhle et al. 213 Research Paper MULTIPLE SCLEROSIS JOURNAL MSJ A comparative study of CSF neurofilament light and heavy chain protein in MS Multiple Sclerosis Journal () 1 7 The Author(s) 213 Reprints and permissions: sagepub.co.uk/journalspermissions.nav DOI: / msj.sagepub.com Jens Kuhle 1,2,6, Kim Plattner 1, Jonathan P Bestwick 3, Raija L Lindberg 1, Sreeram V Ramagopalan 2,4, Niklas Norgren 5, Ahuva Nissim 6, Andrea Malaspina 2, David Leppert 1, Gavin Giovannoni 2a and Ludwig Kappos 1a Abstract Background: There is a lack of reliable biomarkers of axonal degeneration. Neurofilaments are promising candidates to fulfil this task. We compared two highly sensitive assays to measure two subunits of the neurofilament protein (neurofilament light (NfL) and neurofilament heavy chain (NfH)). Methods: We evaluated the analytical and clinical performance of the UmanDiagnostics NF-light enzyme-linked immunosorbent assay (ELISA) in the cerebrospinal fluid (CSF) of a group of 148 patients with clinically isolated syndrome (CIS) or multiple sclerosis (MS), and 72 controls. We compared our results with referring levels of our previouslydeveloped CSF NfH SMI35 assay. Results: Exposure to room temperature (up to 8 days) or repetitive thawing (up to 4 thaws) did not influence measurement of NfL concentrations. Values of NfL were higher in all disease stages of CIS/MS, in comparison to controls (p.1). NfL levels correlated with the Expanded Disability Status Scale (EDSS) score in patients with relapsing disease (r s =.31; p =.2), spinal cord relapses and with CSF markers of acute inflammation. The ability of NfL to distinguish patients from controls was greater than that of NfH SMI35 in both CIS patients (p =.1) and all MS stages grouped together (p =.35). Conclusions: NfL proved to be a stable protein, an important prerequisite for a reliable biomarker, and the NF-light ELISA performed better in discriminating patients from controls, compared with the ECL-NfH SMI35 immunoassay. We confirmed and expanded upon previous findings regarding neurofilaments as quantitative markers of neurodegeneration. Our results further support the role of neurofilaments as a potential surrogate measure for neuroprotective treatment in MS studies. Keywords Cerebrospinal fluid, multiple sclerosis, clinically isolated syndrome, neurodegeneration, neurofilament, neurofilament heavy chain, neurofilament light chain, relapse, disability, immunoassay, biomarker, study design Date received: 9th November 212; accepted: 21st February 213 Introduction Axonal injury is increasingly being recognised as the cause of permanent disability in multiple sclerosis (MS). 1,2 Neurofilaments (Nfs) are major structural elements of neurons that are specifically expressed in axons and dendrites. They are heteropolymers composed of four subunits: the triplet of the Nf light (NfL), medium (NfM) and heavy (NfH) chain, and either α-internexin in the central or peripherin in the peripheral nervous system. NfL and NfH have emerged as promising biomarkers for neurodegeneration in a range of neurological disorders. 3 5 NfH is the most extensively phosphorylated protein of the human brain, with regulatory influences on cell structure homeostasis and axonal transport; 6 8 while NfL is the most 1 Blizard Institute, Barts and the London School of Medicine and Dentistry, Queen Mary University of London, UK. 2 Department of Neurology, University Hospital Basel, Switzerland. 3 Wolfson Institute of Preventive Medicine, Barts and the London School of Medicine and Dentistry, Queen Mary University of London, UK. 4 Wellcome Trust Centre for Human Genetics, University of Oxford, UK. 5 UmanDiagnostics, Umeå, Sweden. 6 Bone and Joint Research Unit, John Vane Science Centre, Queen Mary University of London, UK. a These authors contributed equally Corresponding author: Jens Kuhle, Blizard Institute, Barts and the London School of Medicine and Dentistry, 4 Newark Street, Queen Mary University of London, London, UK. j.kuhle@qmul.ac.uk

2 2 Multiple Sclerosis Journal () abundant and essential component of the Nf core, acting as the backbone to which NfM and NfH copolymerise The UmanDiagnostics NF-light assay uses two highly specific, non-competing monoclonal antibodies to quantitate NfL in cerebrospinal fluid (CSF) samples. 12 In parallel, we recently developed a highly sensitive electrochemiluminescence (ECL)-based solid-phase sandwich immunoassay for detecting NfH SMI35 in CSF. 13 Here, we compare CSF levels of NfL (as measured with the UmanDiagnostics NF-light assay) with referring results of NfH SMI35 in a well-characterised group of 148 clinically isolated syndrome (CIS) or MS patients, and 72 controls. 14 We also evaluated the analytical and clinical performance of the UmanDiagnostics NF-light assay and the stability of its analyte. 15 Patients and methods Patients and CSF samples CSF samples were collected in the Department of Neurology of the University Hospital Basel, in the course of routine diagnostic measures, as indicated by the treating physicians and after patient informed consent. The sample collection procedure, clinical measurement methodology and immunomodulatory treatment were previously described in detail. 14,16 We included 86 patients with definite MS and 62 patients with CIS. MS patients were sub-classified by a trained neurologist as having clinically definite relapsing remitting MS (RRMS; n = 38), secondary progressive MS (SPMS; n = 25), or primary progressive MS (PPMS; n = 23). 14,17 The control group consisted of 72 patients who, based on extensive diagnostic evaluation, had no objective clinical nor paraclinical signs of a neurological disease. Due to a lack of enough CSF sample, we were unable to assay one CIS, one RRMS and one control patient s sample that was used in NfH SMI35 testing, for NfL (Table 1, available only online). 14 The investigators who conducted the NfL measurements had no access to the clinical data. UmanDiagnostics NF-light ELISA and ECL- NfH SMI35 assay The UmanDiagnostics NF-light enzyme-linked immunosorbent assay (ELISA) was performed according to the ELISA kit instructions. Our ECL-NfH SMI35 assay and basic CSF analysis is described in detail elsewhere. 13 Precision of the NF-light ELISA and stability of NfL We evaluated the reproducibility (intra-assay variability) and repeatability (inter-assay variability) of the NF-light ELISA, and the stability of NfL at room temperature (RT), 4 C and during freeze-thawing cycles, as described previously. 13 We conducted the study according to the current regulations on research with patient samples and on documentation of disease-related information of the University Hospital Basel. Approval was granted by the Common Institutional Review Board of the Cantons of Basel. Statistical evaluation Continuous variables are described by their median and interquartile range (IQR), and categorical variables by numbers and percentages. Comparison of basic quantitative CSF parameters across groups was performed using the Kruskal- Wallis test, and pair-wise post-hoc comparisons using the Mann-Whitney U test. Comparisons of categorical variables were done using the chi-square test. CSF levels of NfL, Q alb and other basic CSF parameters were log-transformed to achieve a normal distribution for subsequent analyses. Yet, for simplicity of notation, we used the original terms of CSF parameters when reporting and discussing results. To control for age as a potential confounding factor, we performed an analysis of covariance with age as a covariate and disease stage group as a fixed factor. Group-specific levels of NfL and other biomarkers were expressed as geometric means with 95% confidence intervals (CIs). For log-normal variables, the geometric mean equals the median. Partial correlations adjusted for age were computed by first regressing the two variables on age, and then determining the Spearman rank correlation coefficient (r s ) of the respective residuals. Receiver operating characteristic (ROC) curves were derived from logistic regression (with age as a covariate), to compare the discriminatory power of NfL and NfH SMI35 between CIS or different stages of MS, and healthy controls. We calculated the area under the curve (AUC) for NfL and NfH SMI35 and compared them using the method of DeLong et al. 18 A 2-sided p-value <.5 was considered significant. We adjusted the p-values of post hoc comparisons using a Bonferroni correction. We prepared all statistical analyses and graphs with SPSS (Version 15. SPSS, Chicago, IL) and Graph Pad Prism 5.2 for Windows (GraphPad Software, San Diego, CA). Results Analytical performance of the NfL assay and stability of the analyte. The mean coefficients of variation (CV) of duplicates within given assays were 5.6% (568 pg/ml), 3.1% (564 pg/ml), 5.5% (242 pg/ml) and 3.% (156 pg/ml). Inbetween-assay variation was 8.9% (568 pg/ml), 7.3% (564 pg/ml), 11.3% (242 pg/ml) and 13.5% (156 pg/ml). There was no significant change in the measured concentration in CSF samples that were stored at room temperature (RT) nor at 4 C for up to 8 days (RT, day 8: 1.4 ±.53;

3 Kuhle et al. 3 Room temperature Four degrees S hours 3 hours 1 Day 4 Days 8 Days hours 3 hours 1 Day 4 Days 8 Days S S hours 3 hours 1 Day 4 Days 8 Days hours 3 hours 1 Day 4 Days 8 Days hours 3 hours 1 Day 4 Days 8 Days hours 3 hours 1 Day 4 Days 8 Days Figure 1. Stability of the NfL protein in CSF samples kept at RT (left) and 4 C (right) prior to testing. Three CSF samples (S. 1 3) were thawed on day ( hours, the reference); 3 hours in advance of measurement; and 1, 4 and 8 days before the experiment; and then stored at either RT or 4 C until testing. There was no significant change in the measured concentration of NfL in samples that were stored at RT and at 4 C, for up to 8 days (RT, day 8: 1.4 ±.53 (mean normalised ratio between day and day 8 ± SD), p = 1.; and 4 C, day 8: 1.5 ±.51, p =.5). Mean calculated pg/ml of duplicates (SD) are displayed. CSF: cerebrospinal fluid; NfL: neurofilament protein type light; RT: room temperature; S: sample; SD: standard deviation. a mean normalised ratio between day and day 8 ± SD; p = 1. and 4 C, day 8: 1.5 ±.51; p =.5) (Figure 1). Similarly, there was no significant effect of freeze-thawing the CSF up to 4 times on the measured concentrations of NfL in three samples (4 freeze-thawing cycles: 1.3 ±.26; p =.25) (Figure 2). NfL levels in CSF as a function of disease features and age. The CSF NfL levels were increased (F 4, 215 = 26.89; p <.1) in all forms and stages of MS, as compared to controls (p <.1 for all comparisons). A strong correlation with age was seen for NfL in the controls (r =.61, p <.1), while this association was absent in CIS (r =.6; p =.778), RRMS (r =.11; p =.417), SPMS (r =.13; p =.444) and PPMS (r =.8; p =.694). Subsequent analysis with age as a covariate confirmed the previous highly significant group differences of CIS and all stages of MS, in comparison to the controls (F 4, 214 = 26.5; p <.1; p =.1 for SPMS; p <.1 for CIS, RRMS and PPMS). Moreover, this analysis also revealed a difference between RRMS and SPMS (p =.25) (Figure 3). Correlations of NfL with CSF markers of inflammation. In CIS and RRMS, the levels of NfL correlated with the CSF cell count (r s =.27; p =.16 and r s =.43; p =.1, respectively) and with qalb in CIS, RRMS and SPMS (r s =.28; p =.25; r s =.49; p =.2; and r s =.67; p <.1, respectively). Patients with CIS and RRMS with a CSF cell count > 5 cells/mm 3 had almost twice higher NfL concentrations, as compared to patients with normal CSF cytosis (> 5 cells/mm 3 ; n = 47: 1252 pg/ml ( ), versus 5 cells/mm 3 ; n = 53: 684 pg/ml ( ); p =.78). In contrast, no such correlation could be observed in patients with progressive MS. Correlations of NfL with disability and disease activity. Agecorrected NfL levels correlated with the Expanded Disability Status Scale (EDSS) score in patients with relapsing disease (CIS and RRMS: r s =.31; p =.2), but not in progressive stages of MS (SPMS and PPMS: r s =.18; p =.218). Patients with a relapse at the time of their lumbar puncture tended to have higher NfL values (n = 61: 17 pg/ml ( )) than those in remission (n = 64: 734 pg/ml (598 9)); p =.54). Similarly, the levels of NfL in patients with relapses due to spinal cord pathology was nearly double the levels seen in patients with relapses due to cerebral lesions (n = 16: 1728 pg/ml ( ) versus n = 45: 96 pg/ml ( ); p =.37). Discriminatory power of NfL and NfH SMI35 to distinguish CIS, MS and controls; and relationship of NfL and NfH SMI35. The discriminatory power of NfL was greater than that of

4 4 Multiple Sclerosis Journal () S. 1 S. 2 S Number of Freeze-thaws of CSF sample Figure 2. Stability of NfL protein in CSF during freeze-thawing cycles. Three CSF samples (S. 1 3) underwent 1, 2, 3 or 4 freeze-thawing cycles; and then the measured concentrations were normalised to the sample that was freeze-thawed once. We observed no relevant effect of freeze-thawing (4 freeze-thaw cycles: 1.3 ±.26; p =.25). Mean calculated pg/ml of duplicates (SD) are displayed. CSF: cerebrospinal fluid; NfL: neurofilament protein type light; S: sample; SD: standard deviation Controls p<.1 p=.1 p<.1 p<.1 CIS p=.25 Figure 3. NfL levels in the controls, patients with CIS and patients with MS. Geometric mean and 95% CI are displayed (box and whiskers: median and 1 9% percentile). CIS (765.8 pg/ml), RRMS (12.8 pg/ml), SPMS (784.6 pg/ml) and PPMS (17. pg/ml) showed higher CSF NfL levels than the control patients (271.9 pg/ml). RRMS had higher NfL values than SPMS (p =.25). Dots represent individual samples. P values were adjusted for age and corrected by the Bonferroni method. CI: confidence interval; CIS: clinically isolated syndrome; CSF: cerebrospinal fluid; MS: multiple sclerosis; NfL: neurofilament protein type light; PPMS: primary progressive MS; RRMS: relapsing remitting MS; SPMS: secondary progressive MS. RRMS SPMS PPMS NfH SMI35 in CIS patients (Figure 4, left) and RRMS, SPMS and PPMS together (Figure 4, right) (CIS: AUC.83 versus.67; p =.1 and all MS patients (RRMS, SPMS, PPMS): AUC.91 versus.85, p =.35). There was a highly significant correlation of NfH SMI35 and NfL in controls (r =.4; p <.1), CIS (r =.44; p <.1), RRMS (r =.57; p <.1) but not SPMS (r =.23; p =.163) or PPMS (r =.4; p =.61). After age correction, this relationship was no longer observed in controls (r s =.58; p =.627); conversely, in CIS (r s =.46; p <.1) and RRMS (r s =.56; p <.1) it remained strong, whereas in SPMS (r s =.31; p =.128) and PPMS it was absent (r s =.41; p =.54). Discussion The primary findings of this comparative study were that NfL proved to be a stable analyte, and the assay system used here for NfL is more sensitive than that for NfH SMI35. NfL is considered to represent the most abundant and also the most soluble subunit, but there have been concerns about its susceptibility to proteases, especially in the protease-rich CSF or blood. 15 Several groups, including ours, have therefore previously concentrated on NfH as a biomarker for axonal damage, as its phosphorylated state is assumed to be more stable In the present study, CSF NfL levels were demonstrated to be stable up to 8 days at RT and for up to four freeze-thaw cycles. 23,24 We concluded that there is no basis to prefer NfH over NfL as a biomarker of axonal damage due to concerns of sample stability. Persistent neurological deficits in MS likely emerge as a consequence of the accumulating nerve injury, starting in the very early phases of the disease. Interestingly, differences between CIS patients and controls were more pronounced for NfL than for NfH SMI35, while differences in the other disease stages were more similar and strong (Figure 3). 14 The higher abundance of NfL and/or better performance of the two monoclonal antibodies included in the UmanDiagnostics NF-light assay seem to outweigh the known high sensitivity and higher dynamic range of the ECL technology used in the NfH SMI35 assay. 12 In our previous study, NfH SMI35 showed a strong correlation with age in controls and in patients with a CIS; the correlation was weaker in RRMS and absent in SPMS and PPMS. 14 In the present results, for NfL this correlation was even more pronounced in the controls (r s =.61; p <.1); conversely, it was absent in all the disease stages. Both these NfH SMI35 and NfL findings are well in line with a recent report on CSF NfH and NfL levels and age correlation in CIS patients. 25 More pronounced as compared to NfH SMI35, the disease-related neurodegenerative processes paired with higher assay sensitivity seem to outweigh physiologic, age-related changes of NfL clearance, even in the earliest disease stages of MS. Similarly to the above-mentioned study, our results also suggested that CSF NfL levels do not only reflect chronic

5 Kuhle et al. 5 1 CIS 1 MS 8 8 Detection rate (%) p=.1 NfL AUC=.8282 NfH SMI35 AUC= False-positive rate (%) Detection rate (%) p=.35 NfL AUC=.986 NfH SMI35 AUC= False-positive rate (%) Figure 4. ROC plots of NfL and NfH SMI35 in CIS and MS patients, versus controls. The discriminatory power of NfL was greater than that of NfH SMI35 in CIS patients (AUC.83 versus.67; p =.1) and in all MS patients (RRMS, SPMS, PPMS) grouped together (AUC.91 versus.85; p =.35). AUC: area under the curve; CIS: clinically isolated syndrome; MS: multiple sclerosis; NfH: neurofilament protein type heavy; NfL: neurofilament protein type light; PPMS: primary progressive MS; ROC: receiver operating characteristic; RRMS: relapsing remitting MS; SPMS: secondary progressive MS. neurodegenerative processes. 25 In CIS, RRMS and SPMS, the levels of NfL correlated with the extent of blood-csf barrier damage, relapses, and for CIS and RRMS, with inflammatory cell counts in the CSF. This further supported the concept that in MS, neurofilament release reflects two parallel neurodegenerative processes: first as a result of chronic brain-diffuse neuroinflammation and second due to acute focal inflammatory activity in the course of plaque formation. Replicating our findings on NfH SMI35, the NfL levels correlated with disability in earlier (CIS and RRMS), but not in progressive (SPMS and PPMS) stages. 14 Likewise, NfL and NfH SMI35 concentrations in CIS and RRMS (but not in controls) showed a robust correlation, whereas it was not seen in progressive disease. It remains speculative whether the relatively small sample number of progressive MS stages, the difficulty in quantifying neurological deficits by the EDSS, or a dissociation of the liberation of different Nf subunits in progressive MS contributed to these findings. NfL levels in CSF were reported to be higher in different stages of MS, compared to healthy controls and in relapses versus remissions. 26,27 Previous studies also report relatively weak correlations of NfL levels with the EDSS 28 3 and, in fewer studies, with age in the controls. 31 In a study by Teunissen and colleagues, NfL is determined by the Uman-Diagnostics NF-light ELISA and NfH by a modified conventional ELISA assay. 21,32 In comparison to our findings, their differences for NfH using this assay system were less pronounced, information towards correlation of NfL and NfH and/or age in controls and separate stages of MS are not given and performance of the NfL versus the NfH assay is not described. 14,32 Taken together, we confirmed and expanded on previous findings describing Nf as quantitative markers of neurodegeneration in CSF. We found NfL and NfH SMI35 are both stable proteins, an important prerequisite for biomarkers. It is important to note that based on our findings, we could not conclude that NfL is a superior analyte over NfH, in general. Some of our results are likely to reflect the properties of the assays used, and not wholly the properties of these proteins. Rather, in comparison to the ECL-NfH SMI35 assay, we found that the NF-light ELISA differentiated better between health and disease, especially in the CIS stage. All our analyses were performed in CSF; conversely, serum Nf measurements would be the most relevant for clinical practice, an aim so far reached more frequently by analyses of NfH as compared to NfL. 37 Based on this study, further development of a NfL assay that would include the benefits of ECL technology seemed a promising approach in aiming towards NfL measurement in serum/plasma samples. Our findings support the role of Nf as a useful measure of neurodegeneration and their potential usefulness as a surrogate measure in MS treatment studies. Acknowledgments We thank M Limberg and H Wariwoda for excellent technical assistance. JK s work is supported by an ECTRIMS Research Fellowship Programme and by the Forschungsfonds of the University of Basel, Switzerland. Conflict of interest K Plattner, JP Bestwick, S Ramagopalan, A Nissim and A Malaspina report no potential conflict of interest. J Kuhle receives research support from Novartis Pharma, Protagen AG, Roche and he served on scientific advisory boards for Genzyme/Sanofi-Aventis, Merck

6 6 Multiple Sclerosis Journal () Serono and Novartis Pharma that were used exclusively for research. His work is supported by an ECTRIMS Research Fellowship Programme and by the Forschungsfonds of the University of Basel, Switzerland. RL Lindberg received research support from the Swiss MS Society, Swiss National Science Foundation, European FP6 and IMI JU programs, Roche Postdoc Fellowship Program, and unrestricted research grants from Novartis and Biogen. N Norgren is employed by UmanDiagnostics AB, Sweden; D Leppert is employed by Hoffmann-La Roche AG, Basel, Switzerland. G Giovannoni serves on scientific advisory boards for Merck Serono, Biogen Idec and Vertex Pharmaceuticals; served on the editorial board of Multiple Sclerosis; received speaker honoraria from Bayer Schering Pharma, Merck Serono, Biogen Idec, Pfizer, Teva Pharmaceutical Industries, Sanofi-Aventis, Vertex Pharmaceuticals, Genzyme Corporation, Ironwood and Novartis Pharma; served as a consultant for Bayer Schering Pharma, Biogen Idec, GlaxoSmithKline, Merck Serono, Protein Discovery Laboratories, Teva Pharmaceutical Industries, Sanofi-Aventis, UCB, Vertex Pharmaceuticals, GW Pharma, Novartis Pharma and FivePrime; serves on the speakers bureau for Merck Serono; and received research support from Bayer Schering Pharma, Biogen Idec, Merck Serono, Novartis Pharma, UCB, Merz Pharmaceuticals, Teva Pharmaceutical Industries, Sanofi-Aventis, GW Pharma and Ironwood. L Kappos participated in the last 24 months as a principal investigator, member or chair of planning and steering committees or advisory boards in corporate-sponsored clinical trials in MS and other neurological diseases, sponsored by: Abbott, Actelion, Advancell, Allozyne, BaroFold, Bayer Health Care Pharmaceuticals, Bayer Schering Pharma, Bayhill, Biogen Idec, BioMarin, CSL Behring, Elan, Genmab, GeNeuro SA, Genmark, GlaxoSmithKline, Lilly, Merck Serono, Novartis Pharma, Novonordisk, Peptimmune, Sanofi- Aventis, Santhera, Roche, Teva, UCB and Wyeth. He also lectured at medical conferences or in public on various aspects of the diagnosis and management of MS: in many cases, these talks were sponsored by non-restricted educational grants to his institution from the above companies. Honoraria and other payments were used exclusively to fund departmental research. Both research and clinical operations (nursing, patient care services) of the MS Centre in Basel were supported by non-restricted grants from one or more of these companies. References 1. Bjartmar C and Trapp BD. Axonal and neuronal degeneration in multiple sclerosis: Mechanisms and functional consequences. Curr Opin Neurol 21; 14: Trapp BD, Peterson J, Ransohoff RM, et al. Axonal transection in the lesions of multiple sclerosis. N Engl J Med 1998; 338: Deisenhammer F, Egg R, Giovannoni G, et al. EFNS guidelines on disease-specific CSF investigations. Eur J Neurol 29; 16: Teunissen CE and Khalil M. Neurofilaments as biomarkers in multiple sclerosis. Mult Scler 212; 18: Lee MK and Cleveland DW. Neuronal intermediate filaments. Ann Rev Neurosci 1996; 19: De Waegh SM, Lee VM and Brady ST. Local modulation of neurofilament phosphorylation, axonal caliber, and slow axonal transport by myelinating Schwann cells. Cell 1992; 68: Petzold A. Neurofilament phosphoforms: Surrogate markers for axonal injury, degeneration and loss. J Neurol Sci 25; 233: Teunissen CE, Dijkstra C and Polman C. Biological markers in CSF and blood for axonal degeneration in multiple sclerosis. Lancet Neurol 25; 4: Heins S, Wong PC, Muller S, et al. The rod domain of NF-L determines neurofilament architecture, whereas the end domains specify filament assembly and network formation. J Cell Biol 1993; 123: Geisler N and Weber K. Self-assembly in vitro of the 68, molecular weight component of the mammalian neurofilament triplet proteins into intermediate-sized filaments. J Mol Biol 1981; 151: Lee MK, Xu Z, Wong PC, et al. Neurofilaments are obligate heteropolymers in vivo. J Cell Biol 1993; 122: Norgren N, Karlsson JE, Rosengren L, et al. Monoclonal antibodies selective for low molecular weight neurofilaments. Hybrid Hybridomics 22; 21: Kuhle J, Regeniter A, Leppert D, et al. A highly sensitive electrochemiluminescence immunoassay for the neurofilament heavy chain protein. J Neuroimmunol 21; 22: Kuhle J, Leppert D, Petzold A, et al. Neurofilament heavy chain in CSF correlates with relapses and disability in multiple sclerosis. Neurology 211; 76: Koel-Simmelink MJ, Teunissen CE, Behradkia P, et al. The neurofilament light chain is not stable in vitro. Ann Neurol 211; 69: Teunissen CE, Petzold A, Bennett JL, et al. A consensus protocol for the standardization of cerebrospinal fluid collection and biobanking. Neurology 29; 73: Polman CH, Reingold SC, Edan G, et al. Diagnostic criteria for multiple sclerosis: 25 revisions to the McDonald Criteria. Ann Neurol 25; 58: DeLong ER, DeLong DM and Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach. Biometrics 1988; 44: Goldstein ME, Sternberger NH and Sternberger LA. Phosphorylation protects neurofilaments against proteolysis. J Neuroimmunol 1987; 14: Pant HC. Dephosphorylation of neurofilament proteins enhances their susceptibility to degradation by calpain. Biochem J 1988; 256: Petzold A, Keir G, Green AJ, et al. A specific ELISA for measuring neurofilament heavy chain phosphoforms. J Immunol Methods 23; 278: Sternberger LA and Sternberger NH. Monoclonal antibodies distinguish phosphorylated and nonphosphorylated forms of neurofilaments in situ. Proc Natl Acad Sci USA 1983; 8: Van Geel WJ, Rosengren LE and Verbeek MM. An enzyme immunoassay to quantify neurofilament light chain in cerebrospinal fluid. J Immunol Methods 25; 296: Tortelli R, Ruggieri M, Cortese R, et al. Elevated cerebrospinal fluid neurofilament light levels in patients with amyotrophic lateral sclerosis: A possible marker of disease severity and progression. Eur J Neurol 212; 19:

7 Kuhle et al Khalil M, Enzinger C, Langkammer C, et al. CSF neurofilament and N-acetylaspartate related brain changes in clinically isolated syndrome. Mult Scler 212. Aug Malmestrom C, Haghighi S, Rosengren L, et al. Neurofilament light protein and glial fibrillary acidic protein as biological markers in MS. Neurology 23; 61: Norgren N, Rosengren L and Stigbrand T. Elevated neurofilament levels in neurological diseases. Brain Res 23; 987: Lycke JN, Karlsson JE, Andersen O, et al. Neurofilament protein in cerebrospinal fluid: A potential marker of activity in multiple sclerosis. J Neurol Neurosurg Psychiatry 1998; 64: Semra YK, Seidi OA and Sharief MK. Heightened intrathecal release of axonal cytoskeletal proteins in multiple sclerosis is associated with progressive disease and clinical disability. J Neuroimmunol 22; 122: Norgren N, Sundstrom P, Svenningsson A, et al. Neurofilament and glial fibrillary acidic protein in multiple sclerosis. Neurology 24; 63: Rosengren LE, Karlsson JE, Karlsson JO, et al. Patients with amyotrophic lateral sclerosis and other neurodegenerative diseases have increased levels of neurofilament protein in CSF. J Neurochem 1996; 67: Teunissen CE, Iacobaeus E, Khademi M, et al. Combination of CSF N-acetylaspartate and neurofilaments in multiple sclerosis. Neurology 29; 72: Boylan KB, Glass JD, Crook JE, et al. Phosphorylated neurofilament heavy subunit (pnf-h) in peripheral blood and CSF as a potential prognostic biomarker in amyotrophic lateral sclerosis. J Neurol Neurosurg Psychiatry (4): Petzold A, Mondria T, Kuhle J, et al. Evidence for acute neurotoxicity after chemotherapy. Ann Neurol 21; 68: Lu CH, Kalmar B, Malaspina A, et al. A method to solubilise protein aggregates for immunoassay quantification which overcomes the neurofilament hook effect. J Neurosci Methods 211; 195: Petzold A, Rejdak K and Plant GT. Axonal degeneration and inflammation in acute optic neuritis. J Neurol Neurosurg Psychiatry 24; 75: Rana OR, Schroder JW, Baukloh JK, et al. Neurofilament light chain as an early and sensitive predictor of long-term neurological outcome in patients after cardiac arrest. Int J Cardiol 212.