The relationship between Bayesian motor unit number estimation and histological measurements of motor neurons in wild-type and SOD1G93A mice
Introduction
Amyotrophic Lateral Sclerosis (ALS, motor neuron disease) is a neurodegenerative disorder, clinically characterized by inexorable loss of upper (corticospinal) neurons and lower (cranial/spinal) motor neurons, causing progressive weakness and death within 3–5 years of diagnosis (Rowland and Schneider, 2001, Magnus et al., 2002, Mitchell and Borasio, 2007, Dupuis et al., 2011). The fundamental processes that lead to the loss of corticospinal and motor neurons are not understood (Rothstein, 2009). Theories of disease pathogenesis include effects of abnormal proteins, such as Cu/Zn superoxide dismutase 1 (SOD1) (Orrell, 2009), fused in sarcoma protein (FUS) (Blair et al., 2009), tar-DNA binding protein of molecular weight 43 kDa (TDP43) (Kabashi et al., 2008), optineurin (Maruyama et al., 2010) and C9orf72 (DeJesus-Hernandez et al., 2011, Renton et al., 2011), mitochondrial dysfunction and glutamate excitotoxicity (Rothstein, 2009).
The need to quantify motor unit number in disorders of muscle innervation led to the development of motor unit number estimation (MUNE) techniques that estimate the number of motor units innervating a muscle or group of muscles (McComas et al., 1971, Arasaki et al., 1997, Shefner et al., 2002, Daube, 2006, Bromberg and Brownell, 2008). A number of MUNE methods exist. The Poisson method (Daube, 1995) has been used in clinical studies but there are limitations in applying this method through the course of ALS (Shefner et al., 2007). The incremental method uses gradual increases in strength of low intensity electrical stimulation but is limited by sampling issues to determine motor unit size and alternation (Galea et al., 1991). The multiple point stimulation method, which relies on the sampling of motor units that are activated at low strength of electrical stimulation has been widely applied and shown to be reproducible (Felice, 1995). The multi-point incremental method, that has recently been described, incorporates features of the incremental and multipoint methods and appears widely applicable (Shefner et al., 2011).
We have developed a Bayesian method of MUNE (Ridall et al., 2006, Ridall et al., 2007, Henderson et al., 2007) using Bayesian statistics to analyze stimulus–response curves obtained by graded incremental electrical stimulation (from baseline to the maximum compound muscle action potential, CMAP) of a motor nerve. This method is based on the physiology of motor unit activation and accounts for variability in motor unit threshold, variability in size between and within single motor unit action potentials, and alternation of motor unit firing (McCombe et al., 2009). Our method has the advantages of directly recording all motor units as they are progressively activated, of dealing with motor unit variability by allowing motor unit size to vary as part of the Bayesian model, and of successfully incorporating motor unit alternation by recognizing that the size of the applied stimulus will influence probabilistic motor unit activation. Hence, unlike current statistical methods, our Bayesian statistical model makes assumptions that better reflect the natural physiological mechanisms that are involved in the stimulus and response of motor units.
As there is no “gold-standard” for determining the number of motor units in a muscle, debate remains about which MUNE method gives the most accurate results. Hence rather than comparing Bayesian MUNE with an existing MUNE method, we chose to evaluate our Bayesian MUNE method by a direct comparison of the estimated motor unit number with the number of motor neurons in the region of the spinal cord innervating that muscle. Previously, attempts have been made to compare an incremental MUNE method with histological quantification; however this study produced poor correlations between MUNE and histological nerve fiber counts (Arasaki et al., 1997, David et al., 2010). In the present study, we have followed the loss of motor units over time in transgenic mice over-expressing human superoxide dismutase 1 with a glycine to alanine switch at the 93rd codon (SOD1G93A) mutation (Gurney et al., 1994), which is an animal model of ALS (Scott et al., 2008).
There have been previous MUNE studies in SOD1G93A mice (Hegedus et al., 2007, Hegedus et al., 2009b). The incremental twitch MUNE method detected approximately 60 motor units in the medial gastrocnemius at 40 days of age but required surgically invasive isolation of the gastrocnemius tendon (Hegedus et al., 2007, Hegedus et al., 2009b). A modified incremental MUNE method measured approximately 100 motor units in the distal hindlimb at 90 days of age (Shefner et al., 2006) while comparison of the incremental MUNE method with a multipoint stimulation method in 120 day old SOD1G93A mice found approximately 100 motor units for incremental MUNE compared to 110 motor units for multipoint stimulation MUNE (Shefner et al., 2002).
In this study, we evaluate the Bayesian MUNE method by comparing the numbers of motor units estimated by MUNE with the numbers of motor neurons in the spinal cords of the same wild-type mice and SOD1G93A mice. Groups of these mice were studied before the onset of overt signs of motor dysfunction, at the onset of overt symptoms and at the terminal stages of disease, to follow the course of disease. This study is novel in using both neurophysiological and histological techniques in the same mice to estimate motor unit numbers at different stages of disease progression.
Section snippets
Animals
Wild-type and SOD1G93A mice (B6.Cg-Tg(SOD1-G93A)1Gur/J) were bred at the University of Queensland. B6.Cg-Tg(SOD1-G93A)1Gur/J carry a high copy number of the mutated allele of the human SOD1 gene. The University of Queensland Animal Ethics Committee approved all animal procedures, which were in accordance with national guidelines. SOD1G93A heterozygous males were mated to wild-type females and the offspring were genotyped for the SOD1G93A transgene. Experiments were conducted on age-matched
MUNE studies
To perform MUNE estimation, we record the stimulus response curve, which is a graphical display of the CMAPs evoked by progressive increases in the strength of stimulus delivered to the nerve. The MUNE program plots stimulus–response curves as the stimulus level against the area of the CMAP response. In all wild-type animals the stimulus–response curves were continuous (Fig. 1A, C and E). The stimulus–response curves from pre-symptomatic SOD1G93A mice were comparable to age matched wild-type
Discussion
The SOD1G93A mouse is currently the standard animal model of ALS (Turner and Talbot, 2008, Ludolph et al., 2010). There have been previous studies using MUNE to follow disease progression in SOD1G93A mice, but there have been no previous studies that used histological evaluations and neurophysiology studies in the same animals to determine whether the MUNE methods were accurate (Shefner et al., 2002, Shefner et al., 2006, Hegedus et al., 2009a, Hegedus et al., 2009b, Shefner, 2009). In this
Acknowledgements
We thank C. Pfluger for his technical assistance and Dr. F. Steyn for critical reading of the manuscript. This work was supported by the National Health and Medical Research Council of Australia (NHMRC 569698 to PAM, RDH, ANP, PGR and MCB). The authors have no conflicts of interest.
References (66)
- et al.
Motor unit number estimation in the assessment of performance and function in motor neuron disease
Phys Med Rehabil Clin N Am
(2008) Motor unit number estimates – from A to Z
J Neurol Sci
(2006)- et al.
Expanded GGGGCC hexanucleotide repeat in noncoding region of C9ORF72 causes chromosome 9p-linked FTD and ALS
Neuron
(2011) - et al.
Formation and maturation of subneural apparatuses at neuromuscular junctions in postnatal rats: a scanning and transmission electron microscopical study
Dev Biol
(1987) - et al.
Energy metabolism in amyotrophic lateral sclerosis
Lancet Neurol
(2011) - et al.
Time course of preferential motor unit loss in the SOD1 G93A mouse model of amyotrophic lateral sclerosis
Neurobiol Dis
(2007) - et al.
Development and use of the incremental twitch subtraction MUNE method in mice
Suppl Clin Neurophysiol
(2009) - et al.
Reproducibility of the CMAP scan
J Electromyogr Kinesiol
(2011) - et al.
Biological basis for motor unit number estimation through Bayesian statistical analysis of the stimulus-response curve
Suppl Clin Neurophysiol
(2009) - et al.
Amyotrophic lateral sclerosis
Lancet
(2007)
Selective loss of alpha motoneurons innervating the medial gastrocnemius muscle in a mouse model of amyotrophic lateral sclerosis
Exp Neurol
A hexanucleotide repeat expansion in C9ORF72 is the cause of chromosome 9p21-linked ALS-FTD
Neuron
Recent MUNE studies in animal models of motor neuron disease
Suppl Clin Neurophysiol
Postnatal maturation of nerve-muscle junctions in hindlimb muscles of the mouse
Dev Biol
Estimators of the precision of stereological estimates: an example based on the CA1 pyramidal cell layer of rats
Neuroscience
Time-course and characterization of orolingual motor deficits in B6SJL-Tg(SOD1-G93A)1Gur/J mice
Neuroscience
The pathology of the human muscle spindle: effect of denervation
J Neurol Sci
Transgenics, toxicity and therapeutics in rodent models of mutant SOD1-mediated familial ALS
Prog Neurobiol
Detection of preclinical motor neurone loss in SOD1 mutation carriers using motor unit number estimation
J Neurol Neurosurg Psychiatry
Validity of electromyograms and tension as a means of motor unit number estimation
Muscle Nerve
Glycinergic and GABAergic synaptic activity differentially regulate motoneuron survival and skeletal muscle innervation
J Neurosci
Postnatal development and cell death in the sciatic motor nucleus of the mouse
Exp Brain Res
FUS mutations in amyotrophic lateral sclerosis: clinical, pathological, neurophysiological and genetic analysis
J Neurol Neurosurg Psychiatry
Specific proteolytic cleavage of agrin regulates maturation of the neuromuscular junction
J Cell Sci
Consensus
Neuron death in vertebrate development: in vivo methods
Estimating the number of motor units in a muscle
J Clin Neurophysiol
Validation of an incremental motor unit number estimation technique in rabbits
Muscle Nerve
Thenar motor unit number estimates using the multiple point stimulation technique: reproducibility studies in ALS patients and normal subjects
Muscle Nerve
Localization of acetylcholine receptor by 125I-labeled alpha-bungarotoxin binding at mouse motor endplates
Proc Natl Acad Sci USA
Early and selective loss of neuromuscular synapse subtypes with low sprouting competence in motoneuron diseases
J Neurosci
The numbers and relative sizes of motor units estimated by computer
Muscle Nerve
Motor neuron degeneration in mice that express a human Cu, Zn superoxide dismutase mutation
Science
Cited by (33)
Loss of larger hypoglossal motor neurons in aged Fischer 344 rats
2023, Respiratory Physiology and NeurobiologyChemogenetic inhibition of TrkB signalling reduces phrenic motor neuron survival and size
2023, Molecular and Cellular NeuroscienceSerial measurements of phosphorylated neurofilament-heavy in the serum of subjects with amyotrophic lateral sclerosis
2015, Journal of the Neurological SciencesCitation Excerpt :In a study in SOD1 mutant mice, it has been shown that pNfH levels increase as disease progresses up to day 120 [31]. This study did not extend to the end stage of disease (150–180 days) when, as we have shown, few axons remain [32], and we suggest that it is possible that the levels would decline if mice were samples at this later stage. Our study included some subjects with ALS with prolonged survival, as has been reported in 14% of ALS subjects [21].
Use of biomarkers in ALS drug development and clinical trials
2015, Brain ResearchCitation Excerpt :Motor Unit Number Estimation (MUNE) has also been used to predict disease onset and progression in SOD1G93A mice (Shefner et al., 2006). Ngo et al. (2012) used MUNE to show differences in the numbers of motor units in SOD1G93A mice at all tested time points, including before symptom onset, and further correlated MUNE results to the number of motor units as determined by histological analysis. The best current biomarkers to stratify the ALS patient population are genetic markers for familial causes of ALS, with the most common genetic factor for ALS being the C9ORF72 repeat expansion (Majounie et al., 2012; Renton et al., 2011).
Targeting the motor end plates in the mouse hindlimb gives access to a greater number of spinal cord motor neurons: An approach to maximize retrograde transport
2014, NeuroscienceCitation Excerpt :We have recently detailed the organization of the MEP zone for several forelimb muscles in the rat and mouse (Tosolini and Morris, 2012; Tosolini et al., 2013). In the diverse mouse models of neuromuscular dysfunction mentioned above, the hindlimb is often the target of investigation (Flood et al., 1999; Turner et al., 2009; Wegorzewska et al., 2009; Kimura et al., 2010; Henriques et al., 2011; Ngo et al., 2012; Pratt et al., 2013). The aim of the present investigation was therefore to characterize the muscle–motor neuron organization in the mouse hindlimb.
NAP (davunetide) modifies disease progression in a mouse model of severe neurodegeneration: Protection against impairments in axonal transport
2013, Neurobiology of DiseaseCitation Excerpt :Deparaffinized lumbar spinal cord sections were Nissl-stained with cresyl violet (Sigma, Saint Louis, MO, USA). Motor neurons were counted in the ventral horns of the L4–L5 region of the spinal cord (Nishitoh et al., 2008; Wegorzewska et al., 2009), identified according to standard anatomical guidelines (Beaulieu et al., 1999; Ngo et al., 2012). It should be borne in mind that neuronal counting could be erroneous hence, additional quantitative methods were used.
- 1
These authors contributed equally to the paper.