Neurofilament light proteins (NFL) are a structural element of the neuronal cytoskeleton and are released with neuronal damage. Its levels are increased in cerebrospinal fluid (CSF) in the setting of neurodegenerative diseases, including Alzheimer`s disease 1, frontotemporal dementia 2, 3 ,Huntington’s disease (HD) 4, 5 , and amyotrophic lateral sclerosis 6 , being considered a marker of neuronal damage. Recent studies have shown that NFL can also be reliably measured in blood and are a potential prognostic marker of neurodegeneration in patients with HD 7 , Alzheimer’s disease 8 . Moreover, NFL have been used as marker of neuronal injury in HIV- infection 9 , traumatic brain injury 10 and multiple sclerosis 11 . Also, NFL have revealed as a useful to monitor disease activity and response to treatment in multiple sclerosis12 .

HD is a paradigm of neurodegeneration in the need of appropriate biomarkers for monitoring disease progression. In previous studies various biomarkers of neuronal damage in blood and CSF were investigated, such as tau protein13, cytokines14, mitochondrial DNA15, mutant huntingtin protein16, low brain-derived neurotrophic factor (BDNF)17, different neurotransmitors18, 19, 20, 21, equilibrate nucleoside transporter ENT122 as well as different markers of neuroinflammation23. Nevertheless, a number of studies were limited to animal models. To this time point, NFL seem to be the most promising and accessible biomarker of HD´s progression as the changes of level have been confirmed both in blood and CSF4, 7.

Two studies exploring NFL as biomarker in HD have been published in 2009 and 2017. Constantinescu et al. examined the levels of NFL in CSF of 35 HD patients in the setting of a clinical trial4. The CSF-NFL levels were significantly higher in HD subjects compared with age and gender matched controls, and were correlated with scores on the Unified Huntington’s Disease Rating Scale Total Functional Capacity (TFC), suggesting that NFL could be used as a potential biomarker. Recently, a study by Byrne et al. gave another, more practical perspective on this issue when levels of NFL were studied in blood7. They included subjects enrolled in TRACK-HD study, both pre-manifest and manifest carriers and matched healthy controls and correlated levels of NFL with different clinical parameters such as MRI neuroimaging findings, cognitive and motor evaluation and brain volume (global and regional). They were able to obtain blood sample at baseline and follow-up from 97 controls and 201 individuals with positive genetic testing and additionally compare it with NFL levels in CSF of 37 participants (23 mutation carriers and 14 healthy controls). All in all, they found that blood levels of NFL were higher in mutation carriers and correlated with disease staging as well as changes in MRI, cognitive decline and brain atrophy. What is even more interesting, elevated NFL in blood of pre-manifest carriers were associated with disease onset during the following 3 years. Concentrations of NFL in blood and CSF were correlated in mutation carriers. These results may result key in the setting of future trials.

In order to replicate previous published results we examined the CSF-NFL levels of HD patients (participating in a clinical trial SAT-HD), compared the results with a sample of healthy controls and correlated CSF-NFL levels with demographic and clinical variables (baseline demographic characteristics and HD measures of disease severity).


The ELISA for NFL was performed according to previously published methods24. NFL-CSF levels were determined using commercial sandwich ELISA kits according to the manufacturers’ recommendations (UmanDiagnostics AB, Umeå, Sweden). As specified in the trial protocol, a standardized lumbar puncture (LP) procedure was performed at the L4-L5 level in supine position in all cases. LP were performed at the end of period one and period two of the trial. Treatment assignment did not have impact on the NFL levels and no differences were found between CSF levels after period one and two. The samples analyzed correspond to the end of period one. Plasticware throughout the processing were of polypropylene material to minimize protein absorption. All the CSF samples were aliquoted and stored at -80º until assayed. In the HD group all the patients underwent the LP between 8:00 and 9:00am in a fasting state. The control samples were obtained from the Hospital Ramón y Cajal CSF Biobank collection. This samples were collected and stored according to the local ethical and legal requirements. All participants signed an informed consent form (ICF) for storing the samples at the Biobank for future research. All the controls met clinical criteria for “Symptomatic controls” (SCs) (patients with neurological symptoms, but have no objective clinical or paraclinical findings to define a specific neurological disease at the time of sampling). The time of the day, fasting status and collection vessels was not registered in any of the samples from the control group, however, all the samples were aliquoted and stored at -80º without delay after the LP. Also the storage time was longer (data not available) in the control group25.

Twenty-four EH participants (11 women; mean age 47.3 (SD 12.3)), two pre-manifest carriers and 34 controls (23 women; mean age 34.5 (SD 9.5)) were included. We compared levels of NFL-CSF levels with a regression model adjusted for age and gender. We examined the relationships (Pearson correlations) between NFL-CSF levels and demographic (age, gender) and clinical (CAG repeat number, disease burden, UHDRS motor, behavioural, functional total scores and NPI scale) variables. All the participants gave their signed consent for the procedure.


Demographic, clinical data and NFL-CSF levels of all participants are shown in Table 1. NFL-CSF results were significantly higher in all HD subjects [5014.4 (1557.3) ng/l] as compared to controls [331.4 (200.2) ng/l] (p<0.00) and were correlated with age (correlation coefficient -0.37, p<0.01) and CAG triplet number (0, 51, p<0.05) in the subset of HD patients. NFL levels were not correlated with age in the control group. We did not find any correlation with the remaining variables.

Figure 1. Demonstrates NFL-CSF concentration in different study subgroups, while Figure 2. shows correlations between NFL-CSF and age (Supplementary materials)

Table 1: Demographic and clinical characteristics of participants.
group n male/female age CAG repeat length Disease Burden Total Functional Capacity mUHDRS mUHDRS
controls 34 11/23 34.5 (9.5) 331.4 (200.2)
Premanifest (early) 1 0/1 42 40 189.0 13 0 900
Premanifest (late) 1 1/0 45 41 247.5 13 0 1200
early 10 5/5 47.1 (12.9) 45.4 (2.6) 445.3 (102.3) 11.0 (0.6) 16 (6.7) 5416.0 (1734.4)
moderate 10 8/2 44.7 (12.8) 46.9 (4.6) 463.1 (125.4) 5.0 (1.0) 26 (9.1) 5416.0 (1734.4)
advanced 4 1/3 53.7 (8.8) 44.7 (2.9) 478.3 (93.3) 2.5 (0.5) 38 (10.2) 5416.0 (1734.4)

Fig 1. NFL-CSF concentration in different study subgroups


Fig 2. Correlations between NFL-CSF and age.


These results indicate, as in previous studies, that CSF-NFL levels are a marker of neuronal damage in HD. It seems to be a highly sensitive, but non-specific marker of axonal damage. One of the remaining questions is whether NFL shows superiority over other biomarkers. The biggest discussion considers total tau and NFL. Niemelä et al. published recently study aiming to compare those two biomarkers, but only in CSF23. They included in the analysis both 11 pre-manifest and 12 manifest carriers. While both biomarkers were correlated with each other and showed correlations with clinical parameters, NFL showed significant superiority over tau (TFC, r = -0.70 p < 0.01 for NFL vs r = -0.59 p < 0.01 for tau and Total Motor Score (TMS) r = 0.83p < 0.01 for NFL vs r = 0.67 p < 0.01 for tau). NFL was also significantly correlated with 5-year probability of disease onset, whereas tau was not. This strengths the thesis that rather NFL should be used in clinical trials although authors suggest that both could be used.

CSF-NFL levels were inversely correlated with age at CSF collection and directly correlated with CAG repeat number. We did not find any significant correlations between clinical assessment scores and NFL levels in HD subjects. Increased levels of NFL in younger patients may be an indicator of a more active neurodegenerative process in early stages of the disease. The transversal nature of our study is a limitation to our understanding of the progression of this biomarker.

All the HD patients included in the study were recruited for a clinical trial (SAT-HD with Sativex(®)) and only symptomatic, ambulatory HD patients were included. We were able to obtain samples from two more pre-manifest carriers, but this group was too small to carry out a valid comparative analysis. This limits our results to a subset of HD patients, excluding pre-symptomatic and advanced patients. Furthermore, control samples were collected in not homogenous conditions as we do not dispone with exact information about hour of collection and fasting status. That is why, there are differences in two study samples, the controls are younger than the HD population, and they had a narrower age range. This could be one an explanation why we did not find the widely replicated association between age and NFL.

Further investigations should be done in order to confirm preliminary results provided by us and previously published studies. In particular, CSF-NFL levels should be evaluated in pre-symptomatic, early symptomatic and advanced patients. Furthermore, CSF-NFL levels should be tracked prospectively and correlated with clinical tests, especially in perspective of usefulness to monitor disease progression. Although one study already confirmed practical application of NFL in blood, we are in strong need for validate quantification of NFL levels in accessible bio fluids such as blood, which would serve for longitudinal studies. Finally, it is necessary to evaluate NFL reactivity for therapeutic interventions and its potential to become an outcome for clinical trials in HD.

Data availibility statement

The datasets generated during the current study are available in Dryad repository [, DOI: doi:10.5061/dryad.j089n4j ]. The data is associated to the title of the article.

Authors roles

1. Research project: A. Conception: JLLS, B. Organization: JLS, C. Execution; JLLS, CP, JGC, JGY, JCAC

2. Statistical Analysis: A. Design: JLLS, B. Execution: JLLS, C. Review and Critique; JLLS

3. Manuscript Preparation: A. Writing of the first draft: JLLS, NSZ, B. Review and Critique; NSZ, JLLS

Corresponding Author

Correspondence to:

José Luis López-Sendón Moreno

Servicio de Neurología. Planta 7D. Hospital Ramón y Cajal. 28034 Madrid. Spain.


Telephone number: 0034-913368821

Conflict of interests

The authors have declared that no competing interests exist.

Ethical Approval

The study was approved by the “Comité Etico de Investigación Clínica (CEIC) del Hospital Ramón y Cajal” (hospital’s local Research Ethical Committe). SAT-HD trial was approved by the “Comité Etico de Investigación Clínica (CEIC) del Hospital Ramón y Cajal” (hospital’s local Research Ethical Committe) and the Spanish Agency of Medicines. Comité Etico de Investigación Clínica (CEIC) del Hospital Ramón y Cajal” (Hospital’s local Research Ethical Committe) and the Spanish Agency of Medicines approved the collection and storage of samples at the Hospital Ramón y Cajal CSF Biobank. All participants provided written informed consent for this study.