Minimal Changes in Excitation-Contraction Coupling in Spastic Human Muscle

Abstract

Spasticity results from upper motor neuron lesions and can create a deforming force, pain, and is often accompanied by contracture. While the origin of spasticity is neural, there is ample evidence of secondary muscle changes. Here we use direct measurement of the force-frequency relationship (FFR) to characterize human muscle’s physiological properties. This study directly quantified the FFR of both healthy and spastic human skeletal muscles. Muscle force was measured intraoperatively in healthy gracilis (n=13; aged 39.4 ± 10.6 years; surgery due to brachial plexus injury) and spastic biceps brachii muscle (n=8; aged 53.3 ± 10.3 years; surgery due to stroke or traumatic brain injury). Nerve stimulation was applied at frequencies ranging from 1 to 70 Hz. Twitch contraction parameters, including time to peak tension (TPT) and half-relaxation time (HRT), were also compared. The FFR of the two muscles were modelled with sigmoid functions, and differences between muscles were assessed with an extra sum-of-squares F-test. Time to peak tension (TPT) did not significantly differ between groups (p=0.12), whereas half relaxation time (HRT) was prolonged in the spastic biceps (p<0.05). Despite small differences in twitch kinetics, both muscles exhibited nearly identical FFR profiles. This study represents the first direct in-vivo report of spastic human muscle kinetic properties and shows that these contractile kinetics are similar in healthy and spastic muscles. This may suggest that there are not dramatic calcium handling or myosin heavy chain changes in the biceps muscle secondary to spasticity.

Graphical Abstract

New & noteworthy

This study presents the first in-vivo intraoperative measurement of kinetic properties of spastic human muscle. Despite slower relaxation in spastic biceps, the force-frequency relationship was similar to that of healthy gracilis muscle. This suggests that spasticity does not substantially alter frequency-dependent force summation, possibly due to similar fiber-type compositions, and limited changes in calcium handling or myosin isoforms in human spastic muscle.

Introduction

The classic definition of skeletal muscle spasticity is, “…a velocity dependent increase in the stretch reflex” (1, 2). This definition clearly implies a neural etiology which, presumably, would require a neural approach to treatment. Indeed, this idea forms the mechanistic basis for the use of neurotoxins (3) and hyperselective neurectomy (4) in the treatment of spasticity. However, recent reports from upper motor neuron (UMN) lesion patients have revealed changes in skeletal muscle structure (5–7) and biology (8–10) that occur secondary to an UMN lesion. Unfortunately, it is quite difficult to determine the properties of spastic skeletal muscle and the extent to which its structural and physiological properties are secondarily altered by spasticity.

In this short report, we take advantage of our significant experience treating the spastic human biceps muscle by hyperselective neurectomy (4, 11). Direct exposure of the biceps brachii affords the opportunity to isolate the insertion tendon and measure force directly using a buckle transducer (12) by stimulating the musculocutaneous nerve to generate a force-frequency relationship (FFR). The skeletal muscle FFR is a sigmoidal correlation between activation frequency and isometric force output (13). With increasing activation frequency, muscle force increases due to temporal summation of muscle twitches until force ultimately reaches a fused tetanic contraction. This phenomenon, known as “excitation-contraction coupling,” (ECC) is mechanistically based on two phenomena: (a) the cyclic release of intracellular calcium ion (Ca²⁺) (14) that is faster than the rate at which the sarcoplasmic reticulum (SR) can pump it back from the myofilaments, resulting in increased force (15) and, (b) the rate at which cross-bridges cycle, that depends on muscle fiber type (myosin heavy chain isoform) (16). Thus, simple FFR measurements of spastic and healthy skeletal muscles can provide insight into the physiological and structural consequences of spasticity.

Although the muscle FFR has been widely studied, most studies were performed on animal muscles, which have a significantly higher percentage of fast-twitch fibers compared to humans due to their small size (16–18). Additionally, in the limited human studies that were performed, only muscle peak force production was estimated indirectly by measuring external torque without any report of muscle kinetics. Such torque measurements are also confounded by the activation of synergistic muscles with different properties, leading to potential inaccuracies in muscle force and speed estimation (19, 20). Finally, most human studies rely on transcutaneous electrical stimulation, which can result in unintended activation of adjacent muscles and reverse recruitment of motor units (21, 22) resulting in uncertainties regarding recruitment efficacy.

Despite the previous effort of measuring human muscle contraction dynamics intraoperatively (23, 24), the complete FFR of whole human muscle has not been quantitatively established. In this study, our goal was to measure the FFR of both healthy and spastic human muscles to provide insight into the nature of muscle adaptation due to spasticity and, secondarily, to establish the FFR of normal human muscle.

Methods

Participants and recruitment

Healthy gracilis muscles were tested in patients (n=13) with brachial plexus injury who received a free functioning muscle transfer to restore the elbow flexion function (25) (Aged 39.4 ± 10.6 yrs [mean±SD]; brachial plexus injuries due to motor vehicle accidents). Spastic biceps muscles were tested in patients (n=8) with central nervous system (CNS) injury prior to hyperselective neurectomy for management of elbow flexor spasticity (11, 26) (Aged: 53.3 ± 10.3 yrs; 2/8 traumatic brain injury; 6/8 stroke; time since onset: 6.1 ± 6.1 yrs). Mild elbow flexion contracture (< 20°) was seen in only two patients and all others exhibited full elbow passive extension range. All patients had limited voluntary use of their affected arm with muscle strength graded from 2/5 to 4/5. These studies were approved by the Mayo Clinic Institutional Review Board (15–008865 and 21–006151). Each subject provided written informed consent prior to study inclusion.

Intraoperative nerve stimulation and force measurement

For gracilis muscle testing, the detailed protocol was previously reported (27). Briefly, the gracilis muscle was freed from surrounding tissues, and the obturator nerve was identified for stimulation. A buckle force transducer (BFT) was placed on the distal tendon to record tendon force (Fig. 1A), and a bipolar nerve stimulator was placed on the obturator nerve (Fig. 1B). The tested leg was positioned with the hip flexed to 60°, abducted to 45° and the knee flexed of 90° which placed the gracilis muscle near its optimal length (28, 29). A series of incremental square wave pulses of 0.1 ms duration were delivered, starting from 1 mA for each patient, to identify the minimal electrical current required to elicit the maximal compound muscle action potential (CMAP) resulting in a twitch contraction. Because it was not feasible to prevent movement artifact during tetanic contraction at the maximal CMAP current, 50% of the CMAP current amplitude was used for the FFR data reported here. A train of square-wave pulses was delivered at frequencies of 1, 2, 5, 10, 20, 50 and 70 Hz for a train duration of two seconds to induce tetanic contraction. Twitch contractions were recorded in all 13 patients, but only 5 were stimulated at all frequencies due to surgical time constraints and surgeons need to prioritize patients’ ultimate surgical goals.

Figure 1: Intraoperative setup of the buckle force transducer (BFT) and nerve stimulator.

(A) BFT (arrow) placed on the distal tendon of the gracilis muscle after identification and isolation from surrounding tissue. (B) The bipolar nerve stimulator was applied to the obturator nerve for gracilis activation. (C) BFT placed on the distal biceps brachii tendon after identification and isolation from surrounding tissue. (D) The bipolar nerve stimulator was applied to the musculocutaneous nerve (MCN) for biceps brachii activation.

Spastic biceps testing was performed during hyper-selective neurectomy (HSN) surgery, prior to neurectomy, using a similar approach. After the biceps and other elbow flexor motor nerves were identified, the BFT was placed on the distal biceps tendon (Fig. 1C). The arm was placed at 90° of elbow flexion and 90° of shoulder abduction, which is reported to be the optimal angle for elbow flexion torque (30). The stimulator was placed around the musculocutaneous nerve. The same activation procedure described above for the gracilis was used to define the biceps CMAP as well as its FFR.

Data processing and analysis

All force data were sampled at 1000 Hz and filtered using a fourth-order low-pass Butterworth filter with a cutoff frequency of 30 Hz. The BFT calibration and data processing methods were previously described in detail (12, 27). Twitch contraction measurements included time to peak tension (TPT) and half-relaxation time (HRT) (shown graphically in Fig. 2). All data are presented as mean±SEM in the text unless otherwise noted.

Figure 2: Graphical illustration of the measurement of twitch time to peak tension (TPT) and half relaxation time (HRT).

Dashed lines from top to bottom represent peak force, half of peak force and baseline passive muscle tension respectively.

Peak force at each stimulation frequency was recorded and normalized to the force at 70 Hz, which was maximum. These data thus correct for intersubject variation in maximum tetanic tension.

Data were screened for normality to justify the use of the parametric unpaired t-test to compare differences in TPT and HRT between muscles. A sigmoid function was used to model the normalized force-frequency data for each muscle:

$$f\left(x\right)=2\ast \left(\frac{1}{1+e^{-ax}}-\frac{1}{2}\right)$$

where $f\left(x\right)$ is normalized force, $a$ is the slope coefficient, and $x$ is stimulation frequency. This function ensures that normalized force is zero at zero frequency and asymptotically approaches one as frequency increases.

To assess whether the FFR differed significantly between the healthy gracilis and spastic biceps muscles, an extra sum-of-squares F-test was performed. This test compared a reduced model, in which a single sigmoid curve was fit to all data (assuming shared parameters for both groups), against a full model, in which separate sigmoid curves were fit to each muscle group (allowing independent slope and midpoint parameters) (31, 32). (Matlab coded for extra sum-of-squares test located at https://doi.org/10.6084/m9.figshare.29963402.v1.) Significance level (α) was set to 0.05 for all statistical tests.

Results

TPT was not significantly different (p=0.12) between the healthy gracilis (66.9±2.9 ms) and spastic biceps (80.6±3.9 ms) (Fig. 3A). However, HRT was significantly slower (p<0.05) in spastic biceps (86.5± 9.3 ms) compared to healthy gracilis (66.7±4.5 ms; Fig. 3B). It is worth noting that the power of the statistical test was low (0.44) due to the small sample size but both effects point to the spastic biceps muscles being slightly “slower” compared to healthy gracilis in terms of twitch kinetics.

Figure 3: Comparison of muscle twitch characteristics between healthy gracilis (n=13) and spastic biceps brachii(n=8).

(A) Time to peak tension (TPT) of twitch contractions were not statistically different (p=0.12). (B) Half relaxation time (HRT) was significantly longer in spastic biceps compared to healthy gracilis (p=0.01), indicated by an *.

The healthy gracilis and spastic biceps had similar FFR relationships (Fig. 4A). Unlike data from animal models, the human FFR relationship is not flat at low frequencies but increases immediately as frequency increases from a twitch. A shared sigmoid model was fit to the pooled normalized force-frequency data from the healthy gracilis and spastic biceps muscles. Separate fits for each group produced slope parameters of $a=0.127$ for gracilis and $a=0.134$ for biceps, with corresponding midpoints at 8.65 Hz and 8.17 Hz, respectively. The extra sum-of-squares F-test yielded a p value of 0.7419, demonstrating no significant difference between the groups (Fig. 4). When data were pooled, the slope parameter of $a=0.1309$, and the normalized force reached 50% of its maximum at 8.41 Hz which, together, provide a quantitative estimate of the human muscle FFR.

Figure 4: Force-Frequency Relationship (FFR) of the healthy gracilis (orange, n=5) and spastic biceps (blue, n=8).

Absolute force data were normalized to force at 70Hz and plotted as mean ± SEM from 1Hz to 70Hz. A sigmoid function was fit to the mean normalized force values for each group using equation 1. Normalized muscle force reached half maximal force at 8.65 Hz for healthy gracilis and 8.17Hz for spastic biceps with no significant different differences between muscles identified (p=0.74). (Data from the two groups are slightly offset horizontally for visual clarity.)

Discussion

This study reports the first in-vivo human muscle kinetic data collected from direct intraoperative force measurements to quantify the force-frequency relationship (FFR) in both healthy and spastic human muscles. The results demonstrate that the FFR is similar between the normal gracilis and spastic biceps muscles despite slight differences in twitch kinetics.

The TPT did not differ significantly in spastic biceps compared to healthy gracilis, but our statistical power was low (Power = 1-β =0.44) due to our small sample size. However, HRT was significantly longer in the spastic biceps. The most likely explanation for this result is slower calcium reuptake by the SR. Currently there are no direct studies defining the underlying mechanism, but it may be due to reduced sarcoplasmic reticulum calcium ATPase pump efficiency (33), reduced muscle extracellular matrix compliance (34, 35), and/or the decline of the neural inhibition mechanism (36).

Despite the expectation of a leftward shift due to prolonged HRT with spasticity, no significant difference was observed in the FFR between healthy and spastic muscle. This may be explained by the fact that the gracilis and biceps brachii both exhibit a roughly equal mix of fast- and slow-twitch fibers, which represents a fundamental determinant of the FFR (37, 38). There is no agreement to date on fiber type distribution change after CNS injury, as some reported increases in type 1 fiber (39) while others report the opposite (40) or no change (41). In the present study, any fiber type percentage associated with spasticity was likely not sufficient to produce measurable difference in FFR characteristics.

Both healthy gracilis and spastic biceps displayed limited force increase beyond 20 Hz. This is markedly different from most rodent models that continue to increase force even above 100 Hz (42). This difference can be explained by the faster twitch contractions in various animal models compared to humans (18). In contrast to human fast-twitch fibers that require approximately 50 ms to reach peak tension (43), feline slow-twitch fibers achieve peak tension in just 25 ms (44). Thus, a “slow” cat fiber contracts more rapidly than a “fast” human fiber. Along with a lower composition of fast fibers (45), human muscles reach fused tetanic contraction at dramatically lower frequencies compared to animal muscles. This result suggests caution in comparing so-called “fast” and “slow” muscle fibers among species.

While the overall FFR is similar between healthy and spastic muscles, the underlying subcellular mechanisms may differ and involve complex, potentially opposing changes that ultimately offset each other. There is a dearth of reports on how Ca²⁺ kinetics change in human muscles with CNS injuries. Animal studies have reported that CNS pathology is often accompanied by elevated intracellular Ca²⁺ (46). Chronically altered motor neuron activity also impacts SR Ca²⁺pump isoform profile (47, 48). Such changes could affect Ca²⁺ kinetics following CNS injury, but how and to what extent these findings can be translated to human muscle remains unknown.

We recognize the limitations of this study. Comparing healthy muscles to spastic ones was one major aim of the study, but it was practically and ethically difficult to measure the FFR on same muscle with and without spasticity simply based on which muscles are exposed surgically for which indication. We believe the comparison remains valid, given that the two groups had comparable fiber type compositions and both muscles are ultimately functioning as elbow flexors. Second, Although both muscles had a similar rest period of about one minute between trials, it is possible that spastic muscles are more susceptible to fatigue, which would shift their FFR rightward (49, 50). Longer resting periods were impractical due to time constraints in the operating room. Additionally, FFR can be affected by muscle length (51, 52). We performed the muscle tests near optimal length to better determine true maximal force, but have not accounted for the length factor, if any. As a result of these uncertainties, we only cautiously discuss the mechanistic underpinnings of our observations. Future studies that include direct measurement of calcium handling proteins, myosin heavy chain composition and calcium concentrations are required to disentangle potential subcellular effects of spasticity on muscle.

Conclusions

Despite slower twitch relaxation time in spastic biceps, the FFR was not significantly different from the healthy gracilis. This similarity likely reflects shared fiber-type composition and indicates that spasticity does not substantially alter the frequency-dependent force summation in human muscle. This may suggest that no dramatic calcium handling and/or myosin heavy chain changes occur secondary to muscle spasticity.

Supplementary Material

Matlab code for sigmoidal analysis is located at: https://doi.org/10.6084/m9.figshare.29963402.v1

Abbreviations:

BFT

Buckle Force Transducer

CMAP

Compound Muscle Action Potential

CNS

Central Nervous System

FFR

Force-Frequency Relationship

HRT

Half-Relaxation Time

SR

Sarcoplasmic Reticulum