Background Neurofilament (NF) proteins detection in biological fluids as a by-product of axonal loss is technically challenging and to date relies mostly on cerebrospinal fluid (CSF) measurements. Plasma antibodies against NF proteins and particularly to their soluble light chain (NF-L) could be a more practical surrogate marker for disease staging in amyotrophic lateral sclerosis (ALS), an invariably fatal and clinically heterogeneous neuromuscular disorder.
Methodology We have used a recombinant neurofilament light chain (NF-L) protein for the ELISA detection of antibodies against NF proteins in plasma samples from a well-characterised cohort of ALS individuals (n:73). The use of an established functional rating scale and of a recently proposed staging of disease progression allowed stratification of the ALS cohort based on disease stage, site of onset, survival and speed of disease progression.
Results Antibody levels to NF proteins in plasma were significantly higher in ALS individuals compared to healthy controls (p<0.001). Higher NF plasma immunoreactivity was seen in advanced ALS cases (stage IVA-B) compared to earlier phases of the disease (p<0.05). There was no difference in anti-NF plasma antibodies between ALS individuals treated with riluzole and untreated patients; although riluzole-treated ALS cases with an earlier age of onset and with a shorter diagnostic delay displayed higher anti-NFL antibody levels compared to untreated ALS patients with similar features.
Conclusions Immunoreactivity to plasma NF-L and homologous NF proteins is informative of the stage of disease progression in ALS. The determination of NF antibody levels in plasma could be added to the growing panel of disease-monitoring biomarkers in ALS targeting cytoskeletal antigens.
Statistics from Altmetric.com
The search for prognostic biomarkers tracking the natural disease progression in amyotrophic lateral sclerosis (ALS), a fatal and rapidly progressive neuromuscular condition, is a challenging process. Investigations into disease biomarkers are complicated by the significant clinical heterogeneity of ALS, a condition which entails different sites of disease initiation as well as a variable speed of neurological decline.1 The monitoring of the disease progression in ALS using the ALS functional rating scale revised (ALSFRS-r) is now complemented by a disease staging approach, taking into account easily identifiable clinical milestones.2 While the most robust indicator of a poor prognosis in ALS remains the diagnostic delay, the time-interval between symptoms onset and diagnosis,3 some investigations have tentatively shown that NF-L and NF heavy chain (NF-H) expression in biological fluids, particularly in CSF, holds a predictive power with regard to the rate of disease progression.4–7 Recent studies on superoxide dismutase-1 (SOD1) mutated animal models of ALS have provided a method for optimal detection of NF-H in plasma, which has also been shown to be a good surrogate biomarker of treatment response.8 ,9
Factors affecting the bioavailability of NF proteins hence their exploitability as markers of axonal loss, the unifying pathobiology of neurodegenerative disorders, include NF propensity to aggregate in fluids,8 ,9 their rapid turnover and the organism's immune reaction reducing the yield of antibody-based assays. Sensing the immune response to NF proteins release, aggregation and abnormal phosphorylation may be a better strategy for biomarkers discovery in ALS and may be a more reliable representation of the regional propagation of the disease beyond motor areas. The potential value of immunomonitoring ALS by studying anti-NF-L immunoreactivity draws on the intrinsic stability of antibodies and on their homogeneous distribution in blood and CSF.10 In a small cohort of ALS patients, serum antibodies against NF-L have already been shown to be up-regulated compared to healthy controls.11 Previous studies have shown that serum and CSF anti-NF-L antibodies are elevated in other neurological disorders such as multiple sclerosis.12 ,13
Here we present data suggesting that plasma anti-NF antibody levels are particularly elevated in ALS and vary with the disease stage.
Materials and methods
Sporadic ALS patients (n:73) with a diagnosis of clinically definite, probable, laboratory-supported probable and possible ALS (El Escorial criteria)14 were recruited from motor neuron disease (MND) clinics within the North-East London and Essex MND Regional Care Centre-UCLP. Participants signed an informed consent under the East London REC 1 ethical approval.
Clinical and demographic information are summarised in table 1, including time interval between symptom onset and diagnosis, treatment with riluzole, site of ALS onset (bulbar vs limb), stage of the disease at inclusion (according to Roche et al 2011) and the ALSFRS-r score at sampling. ALS participants were at an early, intermediate or advanced phase of the disease, according to the above staging criteria2 (legend table 1) and defined ranges of ALSFRS-r scores (ALSFRS-r> 40, 39> ALSFRS-r> 26 and ALSFRS-r <25, respectively). The progression rate was calculated as the monthly ALSFRS-r score slope in the time interval between the time of sampling and the last follow-up visit (the average follow-up time was 18 months; range 12–32). Healthy controls (n:52) with no significant past medical history were included. Patients and controls had no history of spinal surgery or other known pathologies conditioning NF proteins release including peripheral neuropathy, compressive radiculopathy or chemotherapy/neurosurgical treatment and no clinical evidence of ongoing autoimmune disorders. All patients and controls had negative ANA determination. Two ALS patients with a possible family history of fronto-temporal dementia (FTD) had an expansion of the C9orf72 gene.
Recombinant mouse neurofilament light (rmNF-L) protein was produced as previously described.15 ,16 An aliquot of each rmNF-L fraction was applied on a SDS-PAGE gel and eluted fractions containing rmNF-L, as determined by western blotting, were pooled and stored at 4°C until use. Immunoreactivity of both monoclonal antibodies and of serum from rmNF-L immunised mice tested against spinal cord and brain tissues in a western blot experiment, revealed a single band at 68KDa, the expected molecular weight for NF-L. While this may suggest a degree of specificity of the detected antibodies to rmNF-L, it is possible that our target antibodies react against a number of homologous neurofilaments including NF heavy and medium chains, internexins and peripherins. For the purpose of this work, we define the detected immune response as anti-NF antibodies. Plasma samples from ALS patients and healthy controls were analysed to test immunoglobulin reactivity against recombinant NF-L by ELISA. Briefly, Nunclon plates (Nunc, Roskilde, Denmark) were coated with 10 μg/mL rmNF-L in carbonate buffer pH 9.6. Plates were blocked with 2% bovine serum albumin (BSA)/PBS and incubated with the plasma samples at 1 : 100 dilution. After washing with PBS-Tween 0.1%, the plates were incubated with horseradish-peroxidase (HRP)-conjugated goat anti-human IgG (Sigma, UK) and the reaction product was developed using TMB substrate (Thermo Fisher Scientific, UK). Results were normalised by subtracting the absorbance derived from uncoated wells and pooled serum samples. The samples were tested in triplicate and the differences among them were negligible. The absorbance was measured at 450 nm using a Synergy HT microplate reader (Bio-Tek instruments, Vermont).
Comparisons between ALS patients and controls were performed using the nonparametric Mann–Whitney U test. Receiver-operating characteristic (ROC) curves and area under the curve (AUC) were performed to calculate the optimal cut-off values for evaluation of reactivity of plasma samples from ALS patients and controls against NF-L recombinant protein. The ROC curve was adjusted for the relevant variables including age and gender. The Spearman rank-order correlation coefficient was used to evaluate the strength of the association between the anti-NF antibody levels and the ALSFRS-r scores. Univariate Kaplan-Meier analysis and multivariate Cox proportional hazard model adjusted for age and sex were used to assess the overall and the disease specific survival and to assess the association between the ALS progression rate and the risk of death. p Values <0.05 were considered as statistically significant. Statistical analysis was performed using SigmaPlot software 11 (Systat, San Jose, California).
Baseline anti-NF antibodies: ALS staging and site of onset
The mean intra-assay coefficient of variation of optical densities was 10.9% for all samples measured. Plasma anti-NF antibody levels were found to be significantly higher in ALS patients compared to healthy donors (p<0.001). The median OD values of anti-NF antibody levels in the ALS and healthy control groups were 0.81 (IQR 0.55 to 1.07) and 0.58 (IQR 0.44 to 0.81), respectively (figure 1A). To obtain the optimal cut-off of antibody reactivity which discriminates between patients and healthy controls, the ROC curve was determined by assessing sensitivity and specificity values. The ROC analysis, adjusted for age and gender, showed a high diagnostic accuracy, AUC=78%, 95% CI (69.3% to 86.7%) (figure 1B). We observed that a specificity of 77%, 95% CI (63.2% to 87.5%) and a sensitivity of 71%, 95% CI (59.4% to 81.2%) could be achieved for these samples, for a cut-off of 0.59 in NF antibody levels.
No significant correlations between age and anti-NF antibody levels were observed in the group of healthy controls (r=0.21, p=0.09) or in the group of ALS patients (r=0.26, p=0.06). Anti-NF antibodies were higher in the intermediate and advanced phases of the disease compared with the early phase. Significant differences were found between advanced and early phases of the disease (p<0.05); the median OD value was 0.61 (IQR 0.43 to 0.97) in the early phase and 0.94 (IQR 0.76 to 1.19) in the advanced stage of the disease (figure 1C). The Cox proportional hazard model adjusted for age and sex shows that a change of one unit in ALSFRS-R slope gives an average of threefold increase in the risk of death (95% CI (1.9 to 4.6), p<0.001). There were no statistically significant differences in plasma anti-NF antibodies between bulbar and spinal onset ALS individuals.
Plasma NF antibody levels, disease progression and treatment
Anti-NF plasma antibodies displayed a negative correlation with the ALSFRS-r scores at the time of sampling (figure 1D). Anti-NF antibodies immunoreactivity did not correlate with the rate of disease progression calculated as the ALSFRS-r slope over the follow-up period. Survival analysis using Kaplan-Meier curves was performed for the overall ALS cohort and for each disease phase group. The Kaplan-Meier curve is a method that takes into account censored data (withdraws) and deaths. No data were missing in this study. The median overall survival of our ALS cohort was 12 months (95% CI (10 to 17.3)) and the median survival of early, intermediate and advanced disease stage ALS subgroups were 26 (95% CI (12 to 40)), 12 (95% CI (8 to 23)) and 4.5 (95% CI (2.6 to 10)), respectively (figure 1E). We observed that the anti-NF antibody levels affected the survival curves. The Cox proportional hazard model adjusted for age and sex shows an association between NF antibody levels and likelihood of death, where a change of one unit in NF antibody levels increases for an average of twofold and half the risk of death (HR: 2.6, 95% CI (1.15 to 6.10), p=0.022). The comparison of anti-NF antibody plasma levels between riluzole-treated and untreated ALS patients showed no significant difference. However, the analysis of subgroups of riluzole-treated and untreated ALS patients with similar disease characteristics such as diagnostic delay (1.5 to 15 months) and age of ALS onset (34 to 59 years) showed higher levels of anti-NF antibodies in the riluzole-treated groups (figure 1F). No correlation between anti-NF antibody levels and the ALSFRS-r monthly slope or disease duration was observed in riluzole-treated and untreated ALS patients. There was no correlation between anti-NF antibodies and diagnostic delay.
It is proposed that the levels of NF proteins in biological fluids, mainly in CSF, correlate with the rate of neurological decline observed in ALS.4–9 ,17 Although technically not impossible, serial lumbar punctures may become a limiting factor if NF analysis in CSF remains the standard approach in biomarkers studies where longitudinal samples are required. Here we show that the humoral immune response to NF proteins whose bioavailability increases with axonal loss in ALS is a good discriminant of the disease stage. We used a recombinant NF-L protein in our assay, which is likely to capture also the antibody response to NF proteins sharing partial sequence homology with NF-L. In our ALS cases, the anti-NF antibody signal relates to the disease phase, but it does not correlate with the rate of disease progression. This may depend on the timeline in which a humoral response against NF proteins is established during neurodegeneration, which may differ from the time of NF proteins release as a result of axonal loss. Ultimately, structural antigens like NF-H or NF-L may be better surrogate markers to rate disease progression since their release in biological fluids may accompany neuronal destruction. Previous studies in animal models of neurodegeneration have highlighted the dynamic interplay between antigen and related antibody by showing that increased levels of NF proteins in blood during active disease and remission are mirrored by a loss of NF-specific antibodies.18 Anti-NF antibodies may reduce NF proteins ELISA detection by a process of antigen masking. We therefore advocate that plasma NF immunoreactivity is used in conjunction with the measurements of NF protein as the combined analysis may be more informative of disease staging and rate of progression.
The immune response to protein misfolding and/or aggregation is increasingly recognised as one of the main drivers of disease progression in ALS, along with mitochondrial dysfunction, oxidative damage, defective axonal transport and excitotoxicity.1 In ALS, there is ample evidence of T-cell and monocyte/macrophage infiltration and activation, in affected areas and in the peripheral circulation. This occurs along with a florid humoral response to aberrantly oxidized or modified proteins and a potential switch from a benign TH2 to a toxic TH1 inflammatory milieu in the affected tissue.19–21 In this context, the build-up of NF immunoreactivity from an early to a late stage of the disease may reflect the propagation of the pathology in ALS, with the involvement of the bulbar or fronto-temporal regions and a crescendo of reactivity to antigens like NF-L.22 Our results also show that increased levels of anti-NF antibodies correlate with low scores on the ALSRS-r scale, which points towards an association of these antibodies with disease severity. This observation is in line with a previous study that shows a significant relationship between ALSFRS-r and serum levels of anti-NF antibodies in ALS patients.11 Furthermore, the association between NF antibody levels and likelihood of death observed in this study could indicate an important role of these antibodies in the pathology of ALS. The treatment of ALS with riluzole has been associated with a marginal effect in prolonging survival of a few months in patients with a predominantly bulbar form of the disease.23 Riluzole treatment does not seem to be effective enough to change the homeostasis of plasma NF antibodies in our ALS cases. The higher levels of antibodies against NF proteins in younger ALS patients with a shorter diagnostic delay compared to a similar cohort of untreated patients may simply reflect a sampling bias and the selection of individuals with a more severe form of the disease, who are known to have a shorter diagnostic latency in virtue of the rapidity of disease progression. Higher anti-NF levels in this group of patients may reinforce the perception of a lack of efficacy of riluzole treatment. Nevertheless, larger, longitudinal and placebo-controlled studies may be required to ascertain whether anti-NF antibodies are a good readout of the marginal treatment response described in riluzole-treated ALS patients.
The homogeneous distribution in different biological fluids of antibodies against axonal cytoskeletal proteins in patients with different neurological diseases is a good platform for the discovery of biomarkers and may provide the basis for the development of disease-modifying treatments with an immunomodulatory effect.12 Our experimental observations in ALS may be extrapolated to other neurological conditions, where the proposed measurement may help to stratify the associated neuropathology.
This study supports the role of anti-NF antibodies as biomarkers for the clinical staging of ALS. Our findings also suggest that other components of the immune response to antigen pools created by degenerating neurons and axons may be relevant targets in the search for biomarkers of disease progression in neurodegenerative disorders.
We thank Bert van het Hof for technical assistance with the preparation of the recombinant NF-L protein. Dr Abdel Douiri acknowledges financial support from the National Institute for Health Research (NIHR) Biomedical Research Centre (BRC) based at Guy’s and St Thomas’ NHS Foundation Trust. The views expressed in this publication are those of the authors and not necessarily those of the NHS, the National Institute for Health Research or the Department of Health.
Contributors AM and FP contributed to the conception and design of the study. AM has led the recruitment of patients and controls and the data collection process. AD carried out the statistical analysis. FP and AM carried out the analysis of data and FP, JT, JK, BS, GG, DB, SA and AM contributed to the interpretation of the statistical analyses, writing and revision of the manuscript.
Funding This work was funded by the Motor Neuron Disease Association UK, grant number Malaspina/Apr08/RF/6039 and by the Barts and the London Charity, grant number 468/17/14. This work was also supported by Stichting MS Research, the Netherlands, grant number 07–627 and by the MS Society of Great Britain and Northern Ireland, grant number NSCG-1F7R.
Competing interests None.
Ethics approval East London REC 1 ethical approval.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement Anonymised data about clinical samples are available upon request to the corresponding author.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.