Article Text

Download PDFPDF

Plasma pNfH levels differentiate SBMA from ALS
  1. Vittoria Lombardi1,
  2. Alessandro Bombaci2,3,
  3. Luca Zampedri3,
  4. Ching-Hua Lu3,
  5. Bilal Malik3,
  6. Henrik Zetterberg4,5,
  7. Amanda J Heslegrave5,
  8. Carlo Rinaldi6,
  9. Linda Greensmith3,
  10. Michael G Hanna3,
  11. Andrea Malaspina1,
  12. Pietro Fratta3
  1. 1 Neuroscience and Trauma, Blizard Institute, London, UK
  2. 2 Department of Neuroscience, ALS Center, University of Truin, Torino, Italy
  3. 3 Department of Neuromuscular Diseases, UCL Queen Square Institute of Neurology, London, UK
  4. 4 Department of Psychiatry and Neurochemistry, Institute of Neuroscience and Physiology, The Sahlgrenska Academy, Goteborgs Universitet, Goteborg, Sweden
  5. 5 Dementia Research Institute, University College London, London, UK
  6. 6 Department of Paediatrics, University of Oxford, Oxford, UK
  1. Correspondence to Dr Pietro Fratta, Department of Neuromuscular Diseases, University College London Institute of Neurology, London WC1N 3BG, UK; p.fratta{at}; Dr Andrea Malaspina; a.malaspina{at}

Statistics from


Spinal and bulbar muscular atrophy (SBMA), known as Kennedy disease (KD), is a slowly progressive adult-onset X-linked neuromuscular disorder with no effective treatment. It is characterised by progressive limb and bulbar muscle weakness, associated with metabolic and endocrine alterations.1 2 SBMA is caused by the expansion of a CAG repeat in exon 1 of the androgen receptor (AR) gene; more than 37 repeats are pathogenic.1 While the genetic test is diagnostic, biomarkers would aid the initial differential diagnosis, and furthermore, there is a strong need for disease activity and progression markers to inform effective clinical trials design.

Neurofilaments (Nfs), both light and heavy chains, are now becoming a widely accepted marker of neuronal damage and a prognostic biomarker for amyotrophic lateral sclerosis (ALS) and other neurodegenerative disease.3–7 Recently, plasma neurofilament light chain (NfL) levels were unexpectedly found not to be raised in patients with SBMA.8 This finding supports other lines of evidence, including an increase in plasma muscle damage markers, myopathic changes in biopsies and a series of genetic experiments in mouse models, that point to a primary myopathic involvement in SBMA.2 9 10

We here used the highly sensitive single molecule array (SIMOA) platform to investigate plasma levels of phosphorylated neurofilament heavy chain (pNfH), another well-established marker of neuronal damage, in patients with SBMA and in a rodent model of disease.

Materials and methods

We have undertaken cross-sectional pNfH analysis using the SIMOA platform in plasma from 46 patients with SBMA, 50 patients with ALS (25 ALS-Fast and 25 ALS-Slow, as previously described)8 and 50 healthy controls (HCs) previously tested for NfL. Participant’s demographic and clinical data are summarised in figure 1A, and detailed methods and statistical analysis are listed in the online supplementary file 1.

Supplementary data

Figure 1

(A) Cohort demographic, genetic information, pNfH and clinical scales values. (B) pNfH concentration in HCs, SBMA, ALS-Slow and ALS-Fast. Using Kruskal-Wallis test and Dunn’s multiple comparison test, we observed reduced levels of both pNfH compared with ALS groups (all p value <0.0001). (C) pNfH levels from mice AR100 (SBMA) and littermate controls (WT) show a small, but statistically significant difference (Mann-Whitney test). (D) Correlation between pNfH and NfL (rs=0.77, p<0.0001). (E) ROC curve, created comparing pNfH and NfL levels, shows a high AUC (0.95; 95% CI 0.90 to 1.00). The value associated to the highest Youden-Index is 105 pg/mL (98% sensitivity, 86% specificity, PPV 0.88, NPV 0.98). ALS, amyotrophic lateral sclerosis; ALSFRSr, ALS Function Rating Revised Scale; AMAT, adult myopathy assessment tool; AUC, area under curve; F, female; Fa-ALS, fast-progressing ALS; HCs, healthy controls; M, male; n° CAG, number of CAG repeats in androgen receptor gene; n°, number; NfL, light chain neurofilaments; NPV, negative predictive value; ns, not statistically significant result; pNfH, phosphorylated heavy neurofilaments; PPV, positive predictive value: PRL, progression rate to last visit; ROC, receiver operating characteristic; SBMA, spinal bulbar muscular atrophy; SBMAFRS, SBMA Function Rating Scale; Sl-ALS, slow-progressing ALS; WT, wild type; Y, years. *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001.

Sera from SBMA (AR100) and wild-type littermate controls (n=10 for each group; 18 months) were also investigated.


Plasma pNfH levels were unchanged in SBMA compared with HCs (figure 1B). These results were also confirmed in SBMA mice, where serum pNfH levels were even lower in AR 100 compared with WT mice (p=0.009) (figure 1C). Conversely, in both fast-progressing and slow-progressing ALS groups, there was a statistically significant increase of plasma pNfH levels compared with HCs (ALS-Slow p<0.001; ALS-Fast p<0.0001), conforming with previous reports.3 5 Plasma pNfH levels differed between ALS-Fast and ALS-Slow patients (Mann-Whitney test: p=0.012, significance was not retained after Dunn’s multiple comparison correction). A weak but significant correlation between pNfH plasma levels and disease progression rate to last visit11 was observed (rs=0.36, p=0.01), supporting pNfH as a possible marker of disease progression rate in ALS. No correlation was found between pNfH and ALS Functional Rating Scale revised (ALSFRSr), SBMA Functional Rating Scale and Adult Myositis Assessment tool scales in patients with SBMA (rs=−0.03, p=0.86; rs=0.03, p=0.89; rs=0.02, p=0.89, respectively) or between pNfH and ALSFRSr in patients with ALS (rs=−0.19, p=0.18). pNfH plasma levels did not correlate with age.

Importantly, pNfH plasma levels were significantly different between SBMA and both ALS subgroups (p<0.0001). A receiver operating characteristic curve was highly significant (area under the curve (AUC) 0.95, p<0.0001; figure 1E) and identified the cut-off point of plasma pNfH that most effectively distinguishes SBMA from ALS as 105 pg/mL (highest Youden Index, 98% sensitivity, 86% specificity).

Lastly, we compared our pNfH results to measurements of NfL previously obtained on the same samples.8 Cox regression analysis showed a significant correlation between pNfH and NfL levels (rs=0.77, p<0.0001; figure 1D).


Contrary to what would be expected in a motor cell disorder,3 7 8 11 this study does not show a disease-related increase of pNfH in patients with SBMA. This finding was also confirmed in a well-established SBMA mouse model. Conversely, and in addition to what previously reported, pNfH levels increase in ALS and correlate with disease progression rate.11

We were able to identify a robust cut-off level (AUC=0.95) of pNfH that could distinguish SBMA from ALS. Although the diagnosis of SBMA has a firm genetic ground and it is compounded by robust clinical and neurophysiological observations, measurement of Nfs, which rise in the prodromal phase and in early stages of ALS,12 may help orientate the diagnostic approach at weakness onset, particularly when only signs of lower motor neuron involvement are present. In this context, the absence of a rise in Nf levels would make consideration for genetic testing mandatory. Our data also suggest that, differently form ALS, Nfs may be of limited use in SBMA clinical trials.

Limits of this study are represented by the cross-sectional design that does not allow us to infer about the variation of pNfH during the disease course. With regard to the analytical aspect, in this study, we employed one of the most sensitive platforms for neurofilament analysis; this approach was the same used for the NfL study.8 Although previous work has highlighted the inherent difficulties encountered in pNfH measuring in biological fluids,11 we show here a good correlation of pNfH and NfL expression levels in the same samples, possibly as a result of the sensitive immunodetection method employed.

The finding that pNfH is not increased in SBMA supports the recent discovery of normal level of NfL in SBMA,. pNfH and NfL are two proteins of the Nf core: the first involved in the cell structure homeostasis and axonal transport and the second is the most abundant and essential component of the core. We believe that there is no basis to prefer pNfH over NfL as a biomarker of axonal damage due to concerns of sample stability.

Lack of increase of Nf, both NfL and pNfH, in SBMA, traditionally considered a lower motor neuron disease, is surprising. Although the difference with ALS could be accounted for by the different progression and aggressiveness of the disorder, the finding of increased Nfs in peripheral neuropathies3 5 suggests progression rate cannot fully explain this discrepancy. Recent work on patients’ muscle biopsies have identified primary myopathic changes in SBMA, and different experiments on disease models have also suggested that the primary muscle changes could be the main driver of the neuromuscular phenotype.2 8 10 Our results support this view, although they do not exclude a role for neuronal loss.

In conclusion, plasma pNfH concentrations are not increased in patients with SBMA and in a mouse model of disease, as opposed to ALS. These results suggest pNfH could be a useful biomarker in the differential diagnosis between SBMA and ALS.


We thank the patients involved in the study and their families for the participation and support to KD research. All samples were obtained from the ALS Biomarker study. PF is supported by an MRC/MNDA LEW Clinician Scientist Fellowship and the NIHR UCLH BRC. This study was supported by the NIHR UCLH BRC, the UCL Kennedy’s Disease Fund and KDUK. CR is funded by a Welcome Trust Clinical Research Career Development Fellowship and the Muscular Dystrophy Association (MDA). HZ is funded by the UK Dementia Research Institute at UCL, the European Research Council, the Swedish Research Council and the Knut and Alice Wallenberg Foundation. The Simoa instrument was bought using a Welcome Trust multi-user equipment grant (PI: HZ).


View Abstract


  • Twitter @NA

  • VL and AB contributed equally.

  • Contributors VL and AB contributed to the writing of the manuscript, statistical analysis, data acquisition and data analysis. LZ, BM, HZ, AJH, CR, LG and MGH contributed to the critical revision of the manuscript. AM and PF contributed to the design and conceptualisation of the study and the critical revision of the manuscript.

  • Funding Study Funded by the NIHR UCLH Biomedical Research Centre Grant #BRC279566.

  • Competing interests None declared.

  • Patient consent for publication Not required.

  • Ethics approval Ethical approval was obtained from the East London and the City Research Ethics Committee (09/H0703/27) and the Ethical Review Panel of UCL Institute of Neurology (PPL PE83401B1).

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

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.