A method to solubilise protein aggregates for immunoassay quantification which overcomes the neurofilament “hook” effect

https://doi.org/10.1016/j.jneumeth.2010.11.026Get rights and content

Abstract

Introduction

Neurofilament (Nf) aggregates are a common pathological feature of neurodegenerative disorders. Although Nf levels have been investigated as a potential disease biomarker, Nf aggregates may mask Nf epitopes, preventing accurate quantification by immunoassay. Using the SOD1G93A mouse model of amyotrophic lateral sclerosis, we developed a method to disrupt Nf aggregates, allowing optimal immunoassay performance.

Methods

Phosphorylated (NfHSMI35) and hyperphosphorylated (NfHSMI34) Nf levels in plasma from 120-day SOD1G93A mice were quantified using an in-house ELISA modified for use with small volumes. Different pre-analytical methods were tested for their ability to solubilize Nf aggregates and immunoblotting was used for qualitative analysis.

Results

A ‘hook effect’ was observed for serially diluted plasma samples quantified using an ELISA originally developed for CSF samples. Immunoblotting confirmed the existence of high molecular-weight NfH aggregates in plasma and the resolving effect of timed urea on these aggregates. Thermostatic (pre-thawing) and chemical (calcium chelators, urea) pre-analytical processing of samples had variable success in disrupting NfH aggregates. Timed urea-calcium chelator incubation yielded the most consistent plasma NfH levels. A one hour sample pre-incubation with 0.5 M urea in Barbitone-EDTA buffer at room temperature resolved the “hook effect” without compromising the ELISA. In SOD1G93A mice, median levels of NfHSMI34 were over 10-fold and NfHSMI35 levels 5-fold greater than controls.

Conclusion

NfH aggregates can be solubilised and the “hook effect” abolished by a one-hour sample pre-incubation in a urea-calcium chelator-enriched buffer. This method is applicable for quantification of NfH phosphoforms in experimental and disease settings where Nf aggregate formation occurs.

Research highlights

▶ Plasma neurofilament (Nf) levels may be a biomarker of neurodegenerative disease. ▶ Nf aggregates cause a ‘hook effect’ that prevents accurate quantification by ELISA. ▶ A new method to dissolve Nf aggregates and abolish the hook effect was developed. ▶ The method is sensitive, accurate and reproducible. ▶ Detects pathological increases in Nf levels in plasma of mSOD1 mice that model ALS.

Introduction

Neurodegeneration is the final pathological pathway for a number of neurological conditions in which the presence of protein aggregates is a common feature (Troy et al., 1997, David et al., 2010). Current strategies to treat these fatal conditions are largely focused on neuroprotection as neuronal death and axonal loss are still irreversible in the human central nervous system (CNS). For the patient this means an inexorable loss of function, progressive disability and premature death. An important limitation for the assessment of disease progression and the effectiveness of neuroprotective treatments is that accurate quantification of loss of function on a clinical scale remains the primary outcome measure (Winhammar et al., 2005, Tuner et al., 2009). Because of functional redundancy in the CNS, a large number of neurons need to be lost before disease progression becomes detectable clinically. Consequently clinical scales are not very sensitive for picking up early or small levels of neurodegeneration. Therefore it would be desirable to quantify and monitor the loss of neurons and axons in vivo before deterioration becomes detectable with current clinical scales (Shefner, 2001, Petzold, 2005, Winhammar et al., 2005, Tuner et al., 2009). One approach is to measure proteins which are released from the degenerating neurons and axons into the adjacent body fluid compartment. Neurofilaments are specific protein biomarkers that have been previously investigated for neurodegenerative disorders (Petzold, 2005, Shaw et al., 2005).

Neurofilaments (Nf) are heteropolymers consisting of at least four subunits, a heavy (NfH), medium (NfM), light (NfL) chain and alpha-internexin (Petzold, 2005, Perrot et al., 2008). Typically these heteropolymers can readily be dissolved with the aid of Ca2+-chelators (Yabe et al., 2001a, Yabe et al., 2001b, Kushkuley et al., 2009, Kushkuley et al., 2010). For this reason the sample incubation buffer of ELISA systems used to quantify Nf levels typically contain EDTA or EGTA (Yabe et al., 2001a, Yabe et al., 2001b). We have previously shown that formation of NfH protein aggregates may occur in vitro and that this aggregation can be overcome mechanically by sonication or by time (Petzold et al., 2003). Nf aggregate formation is known to occur in vivo in a number of neurodegenerative diseases and animal models (see Table 1) and in many cases Nf aggregates increase as disease progresses. Accurate measurement of Nf levels may therefore be a useful biomarker of disease progression in such disorders. However, precise quantification of NfH levels in biological samples by immunoassay may be corrupted in at least four ways: (i) the immunoassay relevant NfH epitope may be masked by the aggregate, (ii) there may be reduced solubility of NfH aggregates, (iii) the protein stability of NfH monomers in solution may differ from the stability of NfH in aggregates, (iv) aggregates may bind soluble NfH. All of these mechanisms will cause a doubling dilution of biological samples to behave differently from a doubling dilution of a soluble, stable NfH protein standard. Typically in biological samples such as blood, cerebrospinal fluid (CSF) or other body fluids, one observes a “hook effect” diverging from the linear range of the NfH standard curve. If this is the case, then parallelism is not achieved (Plikaytis et al., 1994). Consequently the results are inconclusive because after correcting for the dilution factor, the protein concentrations calculated from a 1:1 dilution will still be very different to the protein concentrations calculated from other sample dilutions.

Protein aggregate formation and lack of parallelism is likely to be a fundamental problem for the accurate measurement of NfH levels in a number of neurodegenerative conditions and models. In this study, using plasma samples from 120-day SOD1G93A mice that model ALS, we tested a number of protocols designed to overcome protein aggregate formation and achieve parallelism without compromising protein stability or damaging the relevant epitope or NfH phosphorylation status.

Section snippets

Experimental animals

Mice expressing the human SOD1G93A mutant protein were bred and maintained within the biological services facility of the UCL Institute of Neurology under licence from the UK Home Office and following ethical approval from the Institute. The presence of the SOD1G93A mutation was confirmed by PCR reaction from ear biopsies of 3 weeks old mice. Male SOD1 mice were examined at 120 days of age, near to end stage in our colony of SOD1G93A mice (Kalmar et al., 2008). Wild type (WT), age- and

Plasma NfH aggregates cause a “hook effect”

A “hook effect” was observed for the dilution of plasma samples from 120-day-old SOD1G93A mice (Fig. 1A), while the plasma levels from their age-matched WT littermates were undetectable (data not shown). The resulting lack of parallelism between SOD1G93A plasma samples and standards introduced an analytical error averaging at 300% for a 1:16 dilution (Fig. 1B). Subsequent immunoblotting showed that this lack of parallelism was likely to be due to a combination of NfH aggregate formation (see

Discussion

Here we describe a method which overcomes a crucial analytical problem for the accurate quantification of NfH levels, a protein biomarker for neurodegenerative disorders such as amyotrophic lateral sclerosis, Alzheimer's and Parkinson's diseases. We and others (Personal communication Robert Bowser, Pittsburg, USA; Jens Kuhle, Basel, CH) have previously experienced analytical difficulties with reliable and reproducible quantification of NfH levels in blood samples of mutant SOD1 mice and

Conclusion

Nf aggregates are a hallmark of ALS and other neurodegenerative disorders and pose an important pre-analytical problem for quantitative immunoassay of Nf levels. In this study we showed that the presence of these aggregates is a source of endogenous binding for NfH in plasma which leads to a ‘hook effect’ during serial dilutions. To overcome this problem we developed a method which allows us to gently break up Nf aggregates, improving parallelism and thus allowing for quantification of plasma

Acknowledgements

This project is funded by Motor Neuron Disease Association. LG is the Graham Watts Senior Research Fellow, funded by The Brain Research Trust.

References (46)

  • A. Petzold

    Neurofilament phosphoforms: surrogate markers for axonal injury, degeneration and loss

    J Neurol Sci

    (2005)
  • G. Shaw et al.

    Hyperphosphorylated neurofilament NF-H is a serum biomarker of axonal injury

    Biochem Biophys Res Commun

    (2005)
  • T.B. Shea

    Triton-soluble phosphovariants of the high molecular weight neurofilament subunit from NB2a/d1 cells are assembly-competent. Implications for normal and abnormal neurofilament assembly

    FEBS Lett

    (1994)
  • J.M. Shefner

    Motor unit number estimation in human neurological diseases and animal models

    Clinical Neurophysol

    (2001)
  • J.H. Su et al.

    Plaque biogenesis in brain aging and Alzheimer's disease. I. Progressive changes in phosphorylation states of paired helical filaments and neurofilaments

    Brain Res

    (1996)
  • J.M.C. Winhammar et al.

    Assessment of disease progression in motor neuron disease

    Lancet Neurol

    (2005)
  • A.K. Ashbury et al.

    Giant axonal neuropathy: a unique case with segmental neurofilamentous masses

    Acta Neuropathol

    (1972)
  • K. Blennow et al.

    Cerebrospinal fluid and plasma biomarkers in Alzheimer disease

    Nat Rev Neurol

    (2010)
  • K. Boylan et al.

    Immunoreactivity of the phosphorylated axonal neurofilament H subunit (pNF-H) in blood of ALS model rodents and ALS patients: evaluation of blood pNF-H as a potential ALS biomarker

    J Neurochem

    (2009)
  • A. Charrié et al.

    Hook effect in immunometirc assays for prostate-specific antigen

    Clin Chem

    (1995)
  • D.C. David et al.

    Widespread protein aggregation as an inherent part of aging in C. elegans

    PLos Biol

    (2010)
  • G.M. Fabrizi et al.

    Giant axon and neurofilament accumulation in Charcot-Marie-Tooth disease type 2E

    Neurology

    (2004)
  • T. Ishii et al.

    Presence of neurofilament protein in Alzhermer's neurofibrillary tangles (ANT). An immunofluorescent study

    Acta Neuropathol

    (1979)
  • Cited by (54)

    • The proteome of neurofilament-containing protein aggregates in blood

      2018, Biochemistry and Biophysics Reports
      Citation Excerpt :

      Overall, it could be speculated that these hetero-complexes contain biological factors that are essential in maintaining homeostasis, like those regulating inflammation and metabolism, while they may also act as sinks for the sequestration of other proteins whose concentration increases with the development of pathological conditions. In this study, we have confirmed by Western blot that Nf are included in high molecular weight complexes in plasma as previously anticipated [21,22]. Our data also strengthen the hypothesis that NfH is more likely to be sequestered into metastable, circulating aggregates while smaller and perhaps less complex Nf isoforms like NfL are not, in line with the reported NfL linearity when tested in dilution curves compared to the hook effect seen with NfH [24].

    View all citing articles on Scopus
    View full text