Motoneuron afterhyperpolarisation duration in amyotrophic lateral sclerosis

Non-technical summary

Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disease of still unknown aetiology, although over 130 years have passed since its first description. Since it is not possible to directly record from motoneurons (MNs) in ALS patients, a significant proportion of the research on ALS takes place in animal models of the disease with specific genetic mutations. However, these results are received with scepticism by many of the clinical researchers, since these mutations are responsible for only about 2% of all human ALS cases. We developed a method to characterise indirectly the human MN afterhyperpolarisation (AHP) duration by analysis of muscular electrical activity. In early stages of ALS muscle impairment we observed substantial shortening of the AHP, which was consistent with the results from acute experiments in ALS animal models. Thus, our study lays a bridge between animal and clinical research that may be relevant for identification of mechanism(s) underlying neurodegeneration in ALS.

Abstract

Abstract

Motor unit (MU) potentials were registered from 20 ALS patients and 13 age-matched control individuals during isometric constant force contractions of brachial biceps (BB). The registered signals were decomposed into single MU potential trains. The estimates of duration of the afterhyperpolarisation (AHP) in MNs, derived from the interspike interval variability, was compared between ALS patients (124 MNs) and control subjects (111 MNs) and no significant differences were encountered. However, the relationship between TI and age for patients appeared to be qualitatively different from that of the control group. The dependence of patients’ AHPs on relative force deficit (RFD), which quantified muscle involvement, was more specific. For RFDs below 30%, the AHP estimate was significantly lower than control values and then increased thereafter with increasing RFDs. Moreover, firing rates of patients with the smallest RFDs were significantly higher while firing rates of patients with the greatest RFDs were significantly lower than control values. The AHP shortening in the early stages of muscle impairment is consistent with the decrease in firing threshold of ‘fast’ MNs found in spinal cord slices from neonatal SOD1 mice. The later elongation of the AHP may be caused by the higher vulnerability of ‘fast’ MNs to degeneration and by the influence of reinnervation. Our results are comparable to what has been observed in acute experiments in animal models, providing a bridge between animal and clinical research that may be relevant for identification of mechanism(s) underlying neurodegeneration in ALS.

Introduction

Amyotrophic lateral sclerosis (ALS) is a late-onset neurodegenerative disease of unknown aetiology. Since it was first described, several mechanisms, such as oxidative stress, glutamatergic excitotoxicity, mitochondrial degeneration, defective axonal transport systems and dysfunction of non-neuronal cells, have been suggested to be involved in its pathogenesis (for review, see Vucic & Kiernan, 2009). Environmental factors, such as stress or diet, could also influence the disease (Bradley & Mash, 2009; Chio et al. 2009; Vanacore et al. 2010; Carunchio et al. 2011). Nowadays, a multifactorial origin is widely accepted for the neurodegeneration in ALS, but the mechanism(s) underlying its initiation remain to be defined.

ALS is classically thought to be a disease that causes the progressive loss of upper and lower motoneurons (MNs) followed by axonal degeneration and muscle atrophy, resulting in the virtually complete disappearance of spinal and cortical MNs (Charcot, 1874; Strong & Rosenfeld, 2003). This point of view is still prevalent among clinical neurologists (de Carvalho et al. 2008), and ongoing debates focus mostly on the question of which MN subsystem (upper, UMN or lower, LMN) is affected first.

For obvious reasons, it is not possible to directly characterise the properties of MNs in ALS patients. Therefore, a significant proportion of the research on ALS takes place in animal models of the disease that carry mutations in the Cu/Zn superoxide dismutase gene (SOD1 mice and rats).

In recent years, research in SOD1 animal models has found MN pathology to begin at the distal axon terminals and to proceed in a ‘dying back’ pattern (Fischer et al. 2004; Xie et al. 2005; Parkhouse et al. 2008; Sotelo-Silveira et al. 2009; Carrasco et al. 2010). This differs from the classical view emerging from clinical studies, and thus these results have been received with scepticism by many clinical researchers. Their acceptance of these results is further hindered by the fact that SOD1 mutations are responsible for only about 2% of all human ALS cases (Vucic & Kiernan, 2009).

Our group developed a method to compare human MNs with respect to the duration of their afterhyperpolarisation (AHP) by the analysis of their interspike interval (ISI) variability. The method is based on the fact that there is one-to-one relationship between the MU and the MN discharge (Kernell, 2006), and on the hypothesis proposed by Person and Kudina (Person & Kudina, 1972; Person, 1992) that the AHP influences the plot of the sd of ISIs vs. their mean value (MISI). On such a plot, introduced by (Tokizane & Shimazu, 1964), two ranges of ISI can be distinguished. In the short-interval range, MN discharges regularly with low ISI variability, whereas in the long-interval range, the MN discharge is irregular, and the variability increases rapidly with the mean ISI. This is due to the fact that the short-interval range corresponds to the rhythmic firing mode, whereas the long-interval range corresponds to the occasional spike mode (Calvin, 1974). In this last case, a large fraction of ISIs are longer than the AHP of a MN, and consecutive spikes are generated by random fluctuations of the membrane potential just below the threshold, after all post-spike ionic phenomena have been completed (Matthews, 1996). Obviously, the occasional spike mode produces substantially more ISI variability than does the rhythmic spike mode. Using computer simulations (Piotrkiewicz, 1999) and also in human experiments (Piotrkiewicz et al. 2001) we have previously shown that the ISI corresponding to the transition between both ranges (transition interval, TI) is correlated with the AHP duration. This result was also confirmed by recording directly from cat MNs (Powers & Binder, 2000). Although the TI is not equal to the AHP duration, the sd (MISI) plots with shorter TIs correspond to ‘faster’ MNs (with shorter AHPs). Thus, this method can be used to compare AHP durations of single MNs.

In this manner, we can obtain information on the physiology of human MNs that is comparable to what could be derived from acute experiments in animal models. In the present study, we applied this method to investigate live human ALS MNs.

Methods

Subjects and ethical approval

The data analysed in this study were collected from 20 patients, aged 31–75 years (mean 55.0 years), and 13 control subjects, aged 27–79 years (mean 54.8 years), with no known neurological disorders. For this investigation, we only chose patients who presented clinical signs of lower and upper MN degeneration and who had been diagnosed as having definite ALS according to El Escorial criteria (Brooks, 1994). Thus, patients differed only in age and in the degree of lesion of the muscle under investigation (Table 1). The control group included 11 subjects who had also participated in earlier studies (Piotrkiewicz et al. 2001, 2007). Repeated experiments were run on a few of the control subjects. Every subject gave written informed consent for the experimental procedures, which conformed to the standards set by the latest revision of the Declaration of Helsinki and had the approval of the Ethical Committee of the Medical Research Centre, Polish Academy of Sciences.

Table 1.

Summary of data from patients

				Measurements
Patient group	Patient	Age (years)	RFD* (%)	Rate/force	AHP estimation
1	P01	38	0.0	+	+
	P02	44	0.0	+	−
	P03	48	5.9	+	+
	P04	55	11.8	+	+
2	P05	55	41.2	+	+
	P06	58	41.7	+	+
	P07	49	47.1	+	+
	P08	41	47.1	+	−
	P09	58	50.0	+	−
	P10	61	54.2	+	+
	P11	67	64.7	+	+
	P12	34	66.7	+	+
	P13	63	70.6	+	+
3	P14	75	75.0	+	+
	P15	74	83.3	+	+
	P16	64	91.7	+	+
	P17	67	92.6	+	−
4	P18	31	NM	−	+
	P19	59	NM	−	+
	P20	59	NM	−	+

Experiment

The subject was comfortably seated in an armchair with his/her left forearm placed on a support, assuring isometric contractions of brachial biceps (BB). The arm was positioned horizontally and the angle of the elbow was 90 deg. The wrist was supported by a cuff suspended from the ceiling. During the experiment the subject exerted force by pressing a lever attached to tensometric strain gauges. The subject was provided with auditory and visual feedback on the MU discharges and was instructed to perform a series of constant force isometric contractions, keeping the MUs firing steadily. For the convenience of the subject, the signal of the force transducer and the force reference level were displayed on the screen together with the electromyogram. Between consecutive recordings, the subject was allowed to rest for 3–4 min.

The force of maximum voluntary contraction (MVC) was measured at the beginning of each experiment, except in the first three experiments (Table 1, group 4). Mean control MVC was determined separately for men and women using data from control subjects. For patients, MVC was used to calculate the relative force deficit (RFD), expressing the reduction in patient's force as a percentage of the sex-respective mean control MVC. Different muscles in one patient may be lesioned to different degrees, and the extent of lesions may not necessarily correlate with disease severity (Emeryk-Szajewska et al. 2003). Thus, we used RFD to characterise the involvement of the MNs supplying the muscle under investigation.

To assess the dependence between MN firing rate and force, short (10–20 s) contractions were recorded at the levels of 100%, 75%, 50%, 25% and 10% of the MVC. For the analysis of ISI variability, longer fragments of MU firing (60–100 s) were recorded at contractions not exceeding 30% MVC.

MU potentials (MUPs) were recorded by intramuscular bipolar electrodes hand-made from a Teflon-coated tungsten wire (90 μm diameter) following the procedure described by (Basmajian & Stecko, 1962). Two perpendicularly cut fragments of wire were introduced into a hypodermic needle, and their ends were hooked. The recording surfaces thus were of 0.0064 mm² each. The electrodes were sterilised in an autoclave (130°C) before use and disposed of thereafter. The needle was inserted into a muscle and then withdrawn, whereupon the hooked ends fixed the wires in the muscle. The electrode position was not changed during the experiment.

MUPs were amplified by a dual channel isolated 1902 preamplifier (CED, Cambridge, UK) and bandpassed between 500 Hz and 10 kHz. Data were transferred for off-line analysis to a personal computer using the CED acquisition unit Micro 1401 and recorded on magnetic tape for safety. Sampling rates varied between 15 and 20 kHz, depending on the frequency content of the signal (5–7 kHz maximum), thus preventing aliasing.

Data analysis

In each experiment, we routinely recorded the potentials of several MUs simultaneously (Fig. 1). MU recordings have been decomposed into their constituent single MUP trains by an operator–computer interactive method described elsewhere (Mazurkiewicz & Piotrkiewicz, 2004). The results of the preliminary computer identification were verified by an experienced human operator who corrected the misclassifications. Only the sections of the recordings in which all the MU potentials were reliably identified were used for further analysis. It ought to be mentioned that the method of identification of single MUPs is based on their shapes. If the differences in MUP shapes happened to be small, the identification process became unreliable. Therefore for some patients the analysis was limited to calculation of MU firing rates, for which the identification of 10–50 consecutive MUPs was sufficient. For most subjects, this could be done even at the highest force levels (50–100% MVC). Longer recordings with proper MUP identification were subjected to the variability analysis, as described below.

Figure 1. Examples of MUP recordings made during moderate contractions.

The potentials of single MUs are marked with distinctive numerals. A, normal subject; B, ALS patient from group 1; C, ALS patient from group 3. Note the very high firing rates in B and the differences in firing rates between single MUs in B and C.

sd is a reliable measure of variability if it is calculated from long, stationary MUP trains (at least 50 consecutive potentials). However, maintaining steady MU firing is difficult for untrained subjects. Therefore, to get a complete sd–MISI plot for a single MU, one needs to join together shorter periods of stationary discharge with similar MISI and sd values (Piotrkiewicz et al. 2001). This procedure is therefore subjective, requiring that the experimenter impose his or her own ‘joining criteria’. We found the solution to this problem in the study of Holt et al.(1996), who developed a new technique for measuring variability that was less sensitive to changes in the firing rate and therefore was suitable for studying the variability of irregular neuronal spike trains. Thus, instead of sd, we calculated the absolute value of the difference between each two adjacent ISIs, CD2 = |ISI_i+1− ISI_i|. CD2 was plotted against the mean ISI duration, which was calculated from the same two intervals: MISI2 = (ISI_i+1+ ISI_i)/2.

An example of a plot of CD2 vs. MISI2 is shown in Fig. 2 (circles). Dealing with the cloud of data was inconvenient, so we compiled mean CD2 values (CD_m) for different ISIs. We averaged CD2 data in 10 ms bins (moved every 5 ms) and plotted these values vs. the central value of the bin (see inset to Fig. 2). The resulting CD_m–MISI plot (squares) was comparable to the sd–MISI plot (diamonds). Computer simulations based on our threshold-crossing model (Piotrkiewicz, 1999) have confirmed that both measures of variability are equivalent with respect to estimation of MN AHP duration (not illustrated).

Figure 2. Variability measures plotted against mean ISI.

Circles, CD2 vs. MISI2, both calculated for two adjacent intervals; squares, CD_mvs. MISI, both calculated over 10 ms bins; diamonds, sdvs. mean ISI, both calculated from stationary fragments of at least 50 ISIs. Arrow, transition interval. The inset explains the calculation of CD_m: point no. 1 for 75 ms is an average of the points falling into the MISI range delimited by dashed lines at 70 and 80 ms; point no. 2 for 80 ms is an average of the points falling into the MISI range delimited by dotted lines at 75 and 85 ms, and so on

In most plots, linear sections in the short- and long-interval ranges could be distinguished by visual inspection. For each of these plots, the TI delimiting short- and long-interval sections were estimated according to the procedure described in detail elsewhere (Piotrkiewicz et al. 2007).

Statistical analyses included Student's t test and regression analysis, and significance was determined by ANOVA (all built-in in Microsoft Excel 2003).

Results

The CD_m–MISI relationships were obtained for 124 MUs from 16 ALS patients. The results were compared with 111 MUs from control subjects (Fig. 3). The patients’ data overlapped control data, but were slightly shifted towards higher variability. To assess the statistical significance of this shift, the variability was compared between patients and controls in the short-interval range, meaning that from every MU of each subject we only took the CD_m values corresponding to MISIs shorter than its TI. The mean values for patients and controls were 11.2 and 9.4 ms, respectively, and the difference was statistically significant (P < 0.0001).

Figure 3. Pooled CD_m–MISI relationships.

Diamonds, ALS patients; dots, control subjects.

The average of the transition intervals of ALS patients was not significantly different from that of control subjects. However, the mean TI (calculated for each subject) was significantly correlated with the degree of muscle lesion, expressed as RFD (linear regression coefficient R = 0.87, P < 0.0002). For patients with minor BB lesions, TIs were shorter than control ones and were independent of RFD for values up to 30% (Fig. 4A). Above this value, TIs increased with RFD but did not exceed the limits of control subjects’ TIs (linear regression coefficient R = 0.81, P < 0.01). A comparison of TIs of single MNs for patients with low RFD (RFD < 30%: Table 1, group 1) with those of age-matched control subjects revealed that TIs of patients were significantly shorter (P < 0.001) than those of controls. For remaining patients (RFD > 30%: Table 1, groups 2 and 3), TIs did not differ from control values.

Figure 4. Correlations between mean TIs, age and RFD.

Triangles, ALS patients; circles, control subjects. A, mean TIs plotted vs. RFD; dashed lines represent upper and lower limits of control data. B, mean TIs plotted vs. age; dotted line, regression line for control data. C, RFD plotted vs. age; regression line fitted by the least squares method.

ALS is known as an age-dependent disease (Eisen et al. 1993; Czaplinski et al. 2006). In our study the RFD was significantly correlated with the patient's age (R = 0.71, P < 0.002, Fig. 4B). The age factor influences also the AHP duration in healthy subjects. It has been shown (Piotrkiewicz et al. 2007) that mean TI linearly increases with age over a range of 5.5–79 years (regression coefficient R = 0.766, P < 0.001). It seemed therefore reasonable to compare ALS patients with control subjects in this respect. The relationship between TI and age for patients appeared to be qualitatively different from that of the control group (Fig. 4C). However, the substantial scatter of RFD values for younger patients (Fig. 4B) suggests that the factor of age has a lesser impact on AHP duration than the disease per se.

Firing rates were analysed in those 17 ALS patients for whom MVCs were measured. The range of firing rates in control subjects was 8.4–20.4 spikes s⁻¹. The relationship between MU firing rates and contraction levels varied between individual ALS patients. Firing rates of patients from group 2 did not differ from control values, in contrast to the other groups, representing extremes of the RFD range. Firing rates of patients from group 1 were higher than those of controls (range: 10.9–29.5 spikes s⁻¹), whereas those of patients from group 3 (range: 4.7–18.0 spikes s⁻¹) were lower than those of controls. These differences were statistically significant (P < 0.01, Fig. 5).

Figure 5. MN firing rates of control subjects compared with those of patients from groups 1 and 3.

A, 10%; B, 50%; C, 100% MVC. Abscissa: 1, minimum; 2, mean; 3, maximum values.

It was already evident during MUP recognition that the ranges of firing rates of simultaneously active MUs were much broader in patients than in controls (Fig. 6, upper plots). Only the data for extreme patient groups 1 and 3 were analysed; it should, however, be mentioned that this finding concerns all patients. To assess these differences statistically, the mean firing rate values for the fastest and slowest MUs were evaluated for each force level at which more than one MU was identified. Next, ratios of maximum to minimum firing rates were calculated and were compared between subject groups (Table 2). The differences between the control group and patient groups 1 and 3 were significant (P < 0.002), but the patient groups were not significantly different from each other. Extreme values of firing rates were inversely related to the AHP durations of the respective MNs (Fig. 6, lower plots). In all patients two distinct populations of MNs supplying BB could usually be distinguished according to AHP duration. In group 1 fewer MNs exhibited longer AHPs. In patient group 3, MNs with longer AHPs prevailed, but MNs with shorter AHPs were also present.

Figure 6. Illustration of the presence of two MN populations in ALS patients.

Upper panels, moving average, calculated from five consecutive values of instantaneous firing rates of simultaneously active MNs, plotted vs. time; lower panels, CD_m–MISI relationships (TIs indicated by arrows). A, control subject; B, patient P04 from group 1; C, patient P13 from group 3. For each plot only data of MUs firing with maximum (dark diamonds) and minimum (light squares) rates are shown.

Table 2.

Statistical analysis of differences in firing rate ranges of simultaneously active MUs

				P value when compared with:
	Mean	sd	Range	Control	ALS 1	ALS 2
Control	1.30	0.13	1.12–1.66	—	0.0008	0.0019
ALS 1	1.69	0.37	1.22–2.58	0.0008	—	1.69
ALS 2	1.58	0.24	1.28–2.09	0.0019	1.69	—

Discussion

The analysis performed in this study has confirmed the higher variability of ISIs in the MNs of ALS patients (Schmied & Attarian, 2008). This finding is consistent with the increased excitability of ALS MNs, documented both in human ALS (Kohara, 1999; Kostera-Pruszczyk et al. 2002) and in an animal ALS model (Kuo et al. 2004, 2005). Enhanced excitability should result in an increase in the amplitude of the single EPSP in MN; this factor together with decreased number of excitatory inputs to the MN (Schutz, 2005; Avossa et al. 2006) would increase the variability of the synaptic inflow, which is directly related to ISI variability.

When pooled data on TIs from ALS and control MNs were compared, we found no significant differences. Since TI is correlated with AHP duration (Piotrkiewicz, 1999; Powers & Binder, 2000), this result is consistent with measurements from spinal cord slices from neonatal SOD1 mice (Kuo et al. 2004; Bories et al. 2007). However, there was a clear dependence of AHP duration on the degree of the BB lesion in patients (Fig. 4). Compared with MNs in controls, MNs supplying less lesioned muscles (patient group 1) had shorter AHPs, with AHP duration gradually increasing for MNs of patients with RFD values above 30%. This result indicates that the MN pools supplying muscles of ALS patients initially exhibit an excess of ‘faster’ MNs; as the muscle degeneration progresses, MNs change their activity patterns to become ‘slower.’

ALS is the disease, in which both LMNs and UMNs are affected. According to many clinical researchers, the disease begins in UMNs and proceeds in a ‘dying forward’ pattern. Therefore we initially considered the possibility that the shortening of AHP of patients’ LMNs supplying strong muscles could be attributed to the dysfunction of upper MNs. UMN dysfunction is likely to result in the decrease in excitatory input to MNs, which has been documented in ALS patients (Eisen et al. 1996) and in SOD1 mice (Schutz, 2005; Avossa et al. 2006). Indeed, a decrease in AHP duration has been observed following spinal cord transection in MNs supplying the cat soleus muscle (Cope et al. 1986) and the conversion of MNs to the faster phenotype has been documented after muscle immobilisation (Cormery et al. 2005). Human muscles have also been reported to become faster after spinal cord injury (SCI) (Hager-Ross et al. 2006). Thus, since AHP duration is closely matched to twitch duration of the muscle unit innervated by a given MN (Kernell et al. 1999; Kernell, 2006), decreased excitatory input could also lead to shortening of AHP duration. However, some animal muscles have been reported to become slower after SCI (Harris et al. 2006; Mrowczynski et al. 2006), which suggests that SCI impact on MN properties is generally muscle and species dependent. Moreover, neuromuscular system pathology in ALS can hardly be compared to the pathology observed in response to muscle immobilisation or SCI. Therefore, we sought an alternative hypothesis.

On the other hand, after the UMN lesions low firing rates were often observed (Rosenfalck & Andreassen, 1980; Davey et al. 1990; Shahani et al. 1991; Gemperline et al. 1995; Frontera et al. 1997; Liang et al. 2010). Thus, UMN lesions could be rather responsible for low firing rates and their poor modulation observed in weak muscles.

Recently, researchers working with SOD1 mice found a marked decrease in firing thresholds of high input conductance (presumably ‘fast’) MNs cultured from neonatal spinal cord (Kuo et al. 2005). If this was also the case in human ALS, the levels of muscle contraction at which ‘fast’ MNs are recruited would be lower than normal. Consequently, in less affected MN pools, the proportion of ‘fast’ MNs recruited at low force levels would increase as compared to the case of a healthy neuromuscular system. In fact, this is what we have observed. Motoneurons with short AHPs (‘fast’ MNs) prevailed in a sample of MNs collected from patients with minor muscle lesions.

To our knowledge, disease- and age-dependent changes in AHP duration were not analysed in MNs of SOD1 mice, although it has been shown that ‘fast’ MNs are more vulnerable to degeneration (Mohajeri et al. 1998; Frey et al. 2000; Hegedus et al. 2008). Nevertheless, in our study we observed MNs with comparatively short AHPs (presumably ‘fast’ MNs) even in patients with severe muscle lesion. Since AHP duration was estimated from data collected at weak or moderate muscle contractions, these results may indicate that the firing thresholds of remaining ‘fast’ MNs are lower than normal even in advanced stages of the disease. It should, however, be noted that ‘fast’ MNs of patients from group 2 had longer AHPs as compared with those of patients from group 1.

Several factors may be responsible for the elongation of AHP duration as the disease progresses, including the following possibilities: (i) normal ageing, (ii) the vulnerability of ‘fast’ MNs to degeneration, which should be enhanced by their excessive use, and (iii) the process of reinnervation (Tötösy de Zepetnek et al. 1992). However, reinnervation should be a less important factor because it is impaired in ALS (Gordon et al. 2004).

The reliability of our method for estimation of AHP duration has been well documented both theoretically and experimentally (Piotrkiewicz, 1999; Powers & Binder, 2000; Piotrkiewicz et al. 2001). Our results indicating changes in AHP duration in ALS patients are further strengthened by observations on firing rates, which were inversely related to TI and thus to AHP duration. Increased firing rates were characteristic of MNs supplying strong muscles of ALS patients. Similar results were recently reported from the presymptomatic SOD1 mutant mouse model of ALS (Meehan et al. 2010).

In summary, the results of this study are compatible with results of studies conducted on ALS animal models. These results may contribute to further integration of clinical and animal research, leading to the eventual unravelling of ALS aetiology.

Glossary

Abbreviations

AHP

afterhyperpolarisation

ALS

amyotrophic lateral sclerosis

BB

brachial biceps

CD

consecutive interval difference

CD2

consecutive difference between two adjacent ISIs

CD_m

mean consecutive difference

CV2

coefficient of variation calculated from two adjacent ISIs

ISI

interspike interval

LMN

lower motoneuron

MISI

mean interspike interval

MN

motoneuron

MU

motor unit

MUP

motor unit potential

MVC

maximum voluntary contraction

RFD

relative force deficit

SCI

spinal cord injury

TI

transition interval

UMN

upper motoneuron

Author contributions

The experimental work was performed in the EMG Laboratory at the Neurological Clinic, Medical University of Warsaw, Warsaw, Poland. Both authors contributed to the conception and design of the experiments, as well as to the collection, analysis and interpretation of the data. M.P. drafted the manuscript and both authors critically revised the manuscript and approved the final version for publication.