Longitudinal cohort study of muscle function and metabolic biomarkers in disease-modifying treatment of spinal muscular atrophy

Abstract

Disease-modifying treatments (DMT) for spinal muscular atrophy (SMA) developed in the past decade have improved the prognosis of patients with this severe condition. However, treatment effects vary, highlighting the need for sensitive predictive biomarkers to identify individuals more likely to respond to DMT. A previous study detected changes in mitochondrial oxidative capacity in SMA arm muscles, using 31-phosphorus magnetic resonance spectroscopy (³¹P MRS). Here, we investigated whether ³¹P MRS can distinguish responders from non-responders to DMT. In this longitudinal observational cohort study, patients with SMA were assessed at baseline, two months, and ten months after initiation of DMT treatment. We used a 7 Tesla MR scanner equipped with a custom-built ergometer that allowed for individual mechanical loading and metabolic analysis of the upper arm muscles. Maximal voluntary contraction (MVC) force and motor functional scores were measured before each MR examination. Responders to DMT were defined as those who gained ≥20% in MVC after ten months of treatment. Seven non-ambulatory and five ambulatory patients with SMA starting DMT were included. Four patients fulfilled the MVC response criterion. While significant correlations were observed between MVC force and in vivo metabolic readouts of muscle phenotype, the latter could not longitudinally discriminate responders from non-responders to DMT. Our MR platform for single-arm muscle exercise provides a non-invasive tool for monitoring treatment response in both ambulatory and non-ambulatory patients. Further studies involving larger cohorts will be necessary to establish the diagnostic value of ³¹P MRS in evaluating DMT efficacy in SMA and other neuromuscular diseases.

Keywords

spinal muscle atrophy magnetic resonance spectroscopy biomarker disease-modifying treatment muscle function

Introduction

Spinal Muscular Atrophy (SMA) is an autosomal recessive neurodegenerative disease caused by a homozygous loss of function of the survival motor neuron 1 (SMN1) gene.¹ The incidence is approximately one in 6000–10000 live births. SMA is characterized by the degeneration of spinal cord alpha-motor neurons. This leads to progressive, generalized weakness, as well as muscle fatigability, ultimately resulting in severe functional limitations.² It has a wide range of severity and is classified into four clinical types. The most severe phenotype is characterized by an onset of weakness in the first six months of life with rapid progression to respiratory insufficiency and death (SMA type 1).³ The more chronic variants with childhood onset are characterized by the ability to sit but not to walk independently (SMA type 2), or deteriorating ambulation skills (SMA type 3).⁴ Adult-onset SMA with mild impairments is rare (SMA type 4).⁵

New disease-modifying treatments (DMT) that increase intracellular SMN protein levels have been developed in the past decade and have changed the prognosis of SMA. Therapy outcomes vary, in particular with longer disease duration, with responder rates in adults ranging from 25% to 70%.^6–9 The treatment efficacy of DMT in SMA is currently objectified using motor functional tests such as the Hammersmith Functional Motor Scale Extended (HFMSE) and Revised Upper Limb Module (RULM).^10,11 However, objectification is hampered both by ceiling and floor effects in very strong or very weak patients.¹² In addition, HFMSE has shown an unequal distribution of the constituent items in the total score, explained by multidimensionality and dependency of some of the motor function items.^10,13

Against this background, there is a pressing need for more sensitive biomarkers for patients with SMA to identify individuals who are more likely to benefit from DMTs. Given the limitations of current functional outcome measures, imaging-based biomarkers offer a promising alternative. In vivo phosphorus MR Spectroscopy (³¹P MRS) has the potential to evaluate the effects of interventions such as physical training or pharmacological treatment of myopathies.^14–17 ³¹P MRS can detect changes to muscle at the microscopic, cellular level.¹⁸ Specifically, increased mitochondrial biosynthesis in individual myofibers in response to a training intervention has been detected within eight weeks by ³¹P MRS, resulting in improved energy and pH balance during exercise and faster metabolic recovery post-exercise.^18–20

The expected effect of DMTs for SMA is stabilization or rescue of motor neurons, in particular high-threshold motor neurons associated with fast-twitch (FT) myofibers that produce high force but at a higher adenosine triphosphate (ATP) cost per contraction than slow-twitch (ST) fibers.²¹ The higher vulnerability of this particular population of motor neurons in SMA is evidenced by the common finding of fast-to-slow phenotypic transformation of skeletal muscles in SMA patients that contributes to the symptom of muscle weakness.^22–25 On a macroscopic scale, any effect of DMT would hence manifest as a gain in muscle strength and physical function, while, on a microscopic scale, any availability of rescued FT motor units for daily use may be expected to stimulate mitochondrial biosynthesis in the associated myofibers. Therefore, in theory, a positive response to DMT in SMA may be identifiable both by longitudinal measurements of maximal voluntary contraction (MVC) force and motor functional scores as well as in vivo ³¹P MRS evaluation of, on the one hand, muscular ATP turnover and acidification during contractile work and, on the other hand, the speed of energetic recovery following muscular work.

Here, we tested this hypothesis in a longitudinal study of mechanical and metabolic biomarkers of upper arm muscle function in a cohort of 12 patients with SMA treated with the DMTs Nusinersen (Spinraza^®, Biogen) or Risdiplam (Evrysdi^®, Roche). We previously demonstrated that dynamic in vivo ³¹P MRS interrogation of upper arm muscles of patients with SMA in combination with arm-cycling exercise inside a wide-bore 3 Tesla (T) MR scanner is feasible.²² For logistical reasons, the present study employed a narrow-bore 7 T MR scanner that we fitted with a novel MR-compatible ergometer affording individual mechanical loading of the m. biceps brachii (BB) and m. triceps brachii (TB) muscles. A minimal gain in MVC of ≥20% at 10 months after the start of DMT was used as a criterion for a positive response (‘responder’) to DMT.

Materials and methods

Participants

We recruited 12 patients with SMA types 2, 3, and 4 from the Dutch SMA registry, before initiation of Nusinersen and Risdiplam. Inclusion criteria were: 1) confirmed homozygous deletion of the SMN1 gene, 2) age ≥12 years, 3) ability to perform active arm exercise movements in supine position, 4) ability to follow test instructions, and 5) sufficient understanding of the Dutch language. Exclusion criteria were: 1) pregnancy, 2) comorbidities affecting exercise tolerance, 3) being under examination for a non-diagnosed disease at the time of investigation, 4) contra-indications to 7 T MR examination (University Medical Center Utrecht (UMCU) screening guideline) 5) risk factors for exercise testing registered by a Dutch version of the Preparticipation Questionnaire (American College of Sports Medicine and American Heart Association). Medications that influence metabolic processes were not used by the participants. We included participants between November 2020 and July 2021.

Sample size calculation

Due to the exploratory design and unknown effect size, no a priori power analysis was performed. In vivo ³¹P MRS accurately detects metabolic dysfunction by combining both static and dynamic indices (phosphocreatine (PCr), inorganic phosphate (Pi), and pH during rest, exercise, and recovery). Therefore, ³¹P MRS is highly suitable for metabolic assessment in orphan diseases with small sample sizes, such as SMA.^26–28 We aimed to include ≥10 participants, consistent with or exceeding previous UMCU studies.²²

Study design

This longitudinal prospective clinical investigation consisted of five visits during approximately 300 days (Figure 1). At the first visit (test), the participants performed: 1) muscle function testing, 2) an incremental supine exercise test of the dominant upper arm, and 3) an endurance test at a personalized 80% of peak workload. The tests were repeated at the second visit (retest), either after a break or within two weeks. We established the test-retest reliability of the MR-compatible ergometer on the resistance achieved by the participants measured by the steep-ramp test and the time to fatigue during the endurance test (eTable 1). The third, fourth, and fifth visits (MR visits 1 to 3) were mostly coinciding with regular hospital visits of Nusinersen or Risdiplam treatment on day 0, day 60, and day 300. During MR visits 1 to 3, the participants performed: 1) muscle function testing and 2) an endurance test at a personalized 80% of peak workload inside the MR scanner.

Figure 1.

A) Schematic overview of the study visits, where during the test and retest, the participants performed muscle function testing, an incremental supine exercise test of the dominant upper arm, and an endurance test at a personalized 80% of peak workload. During Magnetic Resonance (MR) visits 1, 2, and 3, the patients performed muscle function testing and an endurance test at a personalized 80% of peak workload inside the 7 Tesla (7 T) MR scanner. (B and C) The setup of the dynamic supine MR-compatible upper arm exercise test with the 7 T MR-compatible supine arm ergometer. Permission was granted by the person depicted in the figure.

Ergometer

The MR-compatible ergometer was custom-made in collaboration with medical technicians and clinical physicians (Figure 1). The ergometer can be adjusted to be used on the 3 T and 7 T Philips Ingenia MR scanner and on a treatment table. With this ergometer, a load can be imposed on the upper arm muscles of participants unilaterally. The force can be varied by the use of weights, with a maximum of 5000 g. The ergometer was calibrated each time as friction of the system was present.

Demographics and anthropometry

Data concerning medical history were obtained from the Dutch SMA database. Height and weight were measured, and body fat percentage was obtained using a Bodystat scan (Quadscan 4000 Euromedix, Belgium).

Motor function

The gross motor function was assessed with the SMA-specific HFMSE at MR visits 1 to 3. The HFMSE has 33 items with a maximum total score of 66.²⁹ The upper limb motor function was assessed with the SMA-specific RULM at MR visits 1 to 3. The RULM has 20 items with a maximum score of 37.³⁰ We assessed the participants treated with Risdiplam only with the RULM. A clinically meaningful improvement is defined as a ≥2-point increase on the RULM total score.³⁰

Isometric muscle strength

We measured MVC of the BB and TB with a handheld dynamometer (MicroFET2, Hoggan Health Industries, USA), using the break test with the elbow in 90-degree flexion and supination of the lower arm. MVC was collected before the in-magnet supine exercise test at MR visits 1 to 3. A responder to DMT was defined as an increase of ≥20% in MVC over the approximately 10 months of treatment (difference in MVC between MR visit 1 and MR visit 3). This ≥20% threshold was chosen based on the minimal detectable change for this outcome measure.³¹ Since MVC of the BB and TB muscles was measured in isolation, we defined responders separately for each muscle.

Dynamic supine MR-compatible upper arm exercise test

At the test and retest visits, participants were positioned supine with their dominant arm resting alongside, in a 90-degree elbow flexion at the neutral starting position. The lower arm was fitted in a hand-made orthosis to fixate the wrist and hand and to minimize involvement of the forearm muscles during the exercise test (Figure 1). To determine individualized peak workload, participants performed a Steep Ramp Test for elbow flexion and elbow extension. The lower arm moved between 60–90 degrees of elbow flexion or extension at 60 rpm, paced by an audio cue. After 10 s of unloaded exercise, we increased resistance with either 200 g (MVC <40 Newton (N)) or 500 g (MVC >40 N) every 10 s until exhaustion or a maximal test duration of 110 s. After ≥10 min (min) of rest and recovery to baseline levels of perceived fatigue, the participant performed an endurance exercise test at 80% of peak workload or the maximal resistance of 5000 grams. The audio cue was set at 60 rpm, with exercise continued until exhaustion or a maximal duration of 6 min. At MR visits 1 to 3, the endurance tests were repeated within the MR scanner with equal workloads (80% of the highest peak workload determined at the test or retest visit). After 30 s of rest, the participants started with dynamic elbow flexion against an 80% peak workload until exhaustion. After a recovery period of 10 min, the same protocol was executed for elbow extension. Perceived fatigue and muscle pain were rated on the OMNI 0–10 scale after the exercise tests.

Muscle ³¹P magnetic resonance spectroscopy

³¹P MRS data acquisition

A whole-body 7 T MR system (Philips Healthcare, Best, The Netherlands) equipped with an in-house designed double-tuned ²H/³¹P whole-body birdcage transmit coil embedded on the outside of the patient tube (Futura, Heerhugowaard, the Netherlands) was used for ³¹P MRS acquisition.³² Two ³¹P surface receive coils (WaveTronica, the Netherlands) were fastened over the dominant frontal and dorsal upper arm, respectively. A series of scout MR images was acquired using a ¹H dipole antenna (WaveTronica, the Netherlands) positioned at the lateral side of the upper arm to evaluate the correct positioning of the participant and the ³¹P coils. First- and second-order shim settings were optimized over the dominant upper arm with an MR image-based method using the MR Code software (TeslaDC, Zaltbommel, the Netherlands). ³¹P free induction decay signals were acquired following a 28° radio frequency block pulse (calibrated for each scan session) (repetition time 500 ms; number of signal averages 4; 512 data points; bandwidth 5000 Hz) with a time resolution of 2 s during 30 s of rest, exercise (variable time, with a maximum of 360 s), and recovery (variable time, with a minimum of 410 s) after completion of the exercise test.

³¹P MRS data analysis

Datasets were analyzed blindly by one observer (L.P.). Free induction decays were denoised using the principal component analysis-based denoising tool using Wolfram Research (Mathematica, version 14.2, IL).³³ Thereafter, ³¹P MRS data were processed in MATLAB (version 9.14.0 R2023a, The MathWorks Inc., USA), and analyzed using AMARES time domain fitting (www.jMRUI.org) with customized starting values and prior knowledge files. PCr (0.0 ppm = reference peak), alpha-ATP (−7.3 ppm), gamma-ATP (−2.3 ppm), and Pi (5.0, 4.4, and 4.0 ppm, respectively) were fitted. Myofiber pH was derived from the chemical shift difference between the Pi and PCr resonances.³⁴ Here, the mean chemical shift of the three Pi peaks was calculated, weighted by their amplitudes. Post-exercise metabolic recovery kinetics were determined by non-linear curve fitting of mono- or double-exponential functions using MATLAB.³⁵ Pi and PCr 95% recovery times (PCr 95RT; in s) were calculated similarly to previous methods.²²

Statistical analysis

We presented continuous variables as mean and standard deviation (SD) or median and interquartile range (IQR), whichever was more appropriate. The data of the patients treated with Nusinersen and Risdiplam were tested as one group. Missing data were addressed by performing the statistical analyses using all available (non-missing) data without imputation. Significant differences between responders and non-responders to DMT in isometric muscle strength were tested using a non-parametric Wilcoxon rank sum exact test. Due to the small and uneven group sizes between responders and non-responders, no statistical comparisons were made between responders and non-responders to DMT (based on isometric muscle strength or RULM total score) on ³¹P MRS metabolic measures. To test the linear correlation between variables, we used Spearman's rank correlation and used the data from MR visits 1 to 3 altogether or longitudinal changes between MR visit 1 and MR visit 3. All statistical analyses were performed using RStudio (version 2024.09.0 + 375, RStudio, Boston, USA), and the significance level was set at 0.05.

Results

Demographics

We included a total of 12 participants, of whom eight men and four women, with an age range of 19 to 55 years (Table 1). A total of three participants had SMA type 2, eight SMA type 3, and one SMA type 4. Nine of the participants started treatment with Nusinersen, and three with Risdiplam, following the regular dosing schedule.^36,37

Table 1.

Cohort demographic characteristics at baseline.

	All (n = 12)
Sex, m/f	8/4
Age, y	19–55
SMA sub-type 2/3/4, n	3/8/1
SMN copy no. 3/4, n	3/9
Disease-modifying treatment - Nusinersen/Risdiplam	9/3
Ambulant/non-ambulant	5/7
Body fat percentage, % (IQR)	32.8 (23.6–40.4)

Median + interquartile range (IQR) when data is not normally distributed. Abbreviations: m = male; f = female; y = years; SMA = Spinal Muscular Atrophy; SMN = survival motor neuron.

Study execution

The study visits were typically planned simultaneously with the planned DMT visits to reduce the burden on the participants (Figure 2). Ten participants completed the BB MVC tests, and ten participants completed the TB MVC tests. Eight patients completed both MVC tests. The reason for not completing the MVC test was not being able to perform the test due to muscle weakness. Ten participants performed the BB MR exercise test, and eight participants performed the TB MR exercise test. Six participants performed both MR exercise tests. Missing data points were caused by muscle weakness. The increased number of dropouts in the MR exercise test is caused by the higher muscular demand required to complete the test compared to the MVC test. After the MR exercise tests, patients experienced pain scores of 0.8 out of 10 in the BB and 0 out of 10 in the TB, respectively, averaged over the three MR exercise tests.

Figure 2.

Overview of the timeline of MR visits 1 to 3, combined with the regular hospital visits of Nusinersen or Risdiplam treatment for all participants.

Response to DMT

At MR visit 1, median BB MVC and TB MVC were 39 N (IQR 24–255) and 13 N (IQR 7–31), respectively (Table 2). Using the ≥20% MVC criterion, comparison of absolute TB MVC at baseline and 10 months after DMT identified four responders and six non-responders. For BB, no patient reached the response threshold. Figure 3 shows the absolute and relative changes in MVC after approximately 10 months of DMT for the responders and non-responders to DMT. On a group level, absolute and relative TB MVC in responders was increased after approximately 10 months of DMT compared to non-responders (p = 0.0381 and p = 0.0095, respectively). No group comparisons were possible for BB (ten non-responders vs. zero responders).

Figure 3.

A) Absolute [N] and relative (%) change in maximal voluntary contraction (MVC) in the m. biceps brachii between magnetic resonance visit 1 (MR V1) and MR V3 for responders (R) and non-responders (NR). B) Absolute [N] and relative (%) change in MVC in the m. triceps brachii between MR visit 1 and MR visit 3 for responders and non-responders. The absolute and relative MVC of the TB muscle in responders was significantly increased after approximately 10 months of DMT in comparison to non-responders. *p < 0.05, **p < 0.01.

Table 2.

Longitudinal data of muscle function and motor function measures.

	MR visit 1 (0 d)	MR visit 2 (60 d)	MR visit 3 (300 d)
MVC BB, N (IQR)	39 (24–255) (n = 11)	34 (24–214) (n = 12)	35 (22–210) (n = 10)
MVC TB, N (IQR)	13 (7–31) (n = 11)	9 (6–26) (n = 11)	10 (7–33) (n = 10)
TTF BB, s (IQR)	97 (65–217) (n = 10)	99 (77–163) (n = 10)	112 (52–177) (n = 11)
TTF TB, s (IQR)	55 (42–135) (n = 8)	84 (42–160) (n = 8)	80 (55–143) (n = 8)
RULM, total score (IQR)	27 (17–37) (n = 12)	28 (17–37) (n = 12)	29 (19–37) (n = 12)
HFMSE, total score (IQR)	52 (11–55) (n = 9)	53 (22–56) (n = 9)	54 (20–56) (n = 9)

Median + interquartile range (IQR) when data is not normally distributed. Abbreviations: MR = magnetic resonance; d = days; MVC = maximal voluntary contraction force; BB = m. biceps brachii; TB = m. triceps brachii; N = Newton; TTF = time to fatigue; s = seconds; RULM = Revised Upper Limb Module for SMA; HFMSE = Hammersmith Functional Motor Scale Extended.

The participants showed a median HFMSE total score of 52 (IQR 11–55) and a median RULM total score of 27 (IQR 17–37) at the start of this study. Using the ≥2-point increase criterion on the RULM, we identified three responders and nine non-responders to DMT.

During the dynamic supine upper arm exercise test at MR visit, the median time to fatigue was 97 s (IQR 65–217) and 55 s (IQR 42–135). This measure did not show a significant change at MR visit 3 compared with MR visit 1 for both the BB and TB (Table 2).

In vivo ³¹P MRS observations

Figure 4A shows representative in vivo ³¹P MR spectra of the BB of a patient in the studied cohort recorded at rest, during exercise, and subsequent recovery. The corresponding quantitative time courses of concentration dynamics for PCr, Pi, and intramuscular pH are shown in Figure 4B. A total of 54 ³¹P MRS datasets were recorded from the BB and TB. Of these, 48 (89%) met the signal-to-noise ratio (SNR) of ≥3, resulting in ten and five complete longitudinal datasets for the BB and TB, respectively. The median PCr depletion of MR visits 1 and 3 is 77.3% (IQR 60.3–92.7) and 46.9% (IQR 27.6–71.7) for the BB and TB, respectively. The median pH end exercise of MR visits 1 and 3 are 6.9 (IQR 6.5–7.0) and 6.8 (IQR 6.5–6.9) for the BB and TB, respectively. The median PCr 95RT end-exercise of MR visits 1 and 3 are 180.6 s (IQR 71.3–359.6) and 171.6 s (IQR 83.5–289.9) for the BB and TB, respectively.

Figure 4.

Dynamic in vivo ³¹P MR spectra of the proximal arm musculature of patients with SMA during rest, exercise, and recovery. (A) Time-resolved series of ³¹P-MR spectra of the m. biceps brachii (BB) of a patient with SMA acquired at rest, during elbow flexion exercise, and subsequent recovery at a temporal resolution of 2 s. Note the exercise-induced depletion of phosphocreatine (PCr) and concomitant rise of inorganic phosphate (Pi) that recovered upon cessation of exercise. α, γ-ATP = α-, γ-phosphate groups in adenosine 5’-triphosphate. (B) Time course of PCr (black), Pi (5.0 ppm in blue, 4.4 ppm in red, and 4.0 ppm in yellow), and muscle pH (dashed, light grey) of the BB muscle of an individual patient with SMA performing the upper arm exercise test until exhaustion.

Correlations between functional and metabolic biomarkers of muscle quality

MVC positively correlated with the motor function tests RULM and HFMSE for the BB (r = 0.838, p = 3.9e⁻⁶ and r = 0.780, p = 5.1e⁻⁵, respectively) and the TB (r = 0.622, p = 0.0058 and r = 0.901, p = 3.4e⁻⁷, respectively). Correlations between MVC and ³¹P MRS outcome measures of MR visit 1–3 are shown in Figure 5A-C. For BB, MVC correlated positively with ³¹P MRS outcomes PCr depletion (r = 0.840, p = 3.6e⁻⁶) and PCr 95RT (r = 0.635, p = 0.0026), and negatively with muscle pH end-exercise (r = −0.754, p = 0.00012). For the TB, MVC correlated positively with PCr depletion (r = 0.739, p = 0.00046) and PCr 95RT (r = 0.516, p = 0.028), and negatively with muscle pH end-exercise (r = −0.520, p = 0.027). No significant longitudinal correlations between MVC and ³¹P MRS measures across MR visit 1 and MR visit 3 were observed (Figure 5D-F). Therefore, the subjects who improve more in MVC do not show greater metabolic changes over time.

Figure 5.

Linear correlation plots of the functional measure maximal voluntary contraction (MVC), measured with hand-held dynamometry, and magnetic resonance spectroscopy (³¹P MRS) outcomes of MR visits 1-3 data. The data of the m. biceps brachii (BB) are visualized in black and a closed circle, and the data of the m. triceps brachii (TB) with grey and an open circle. A) Correlation plot between MVC and phosphocreatine (PCr) depletion for the BB (r = 0.840, p = 3.6e⁻⁶) and TB (r = 0.739, p = 0.00046). B) Correlation plot between MVC and PCr 95% recovery time (PCr 95RT) for the BB (r = 0.635, p = 0.0026) and TB (r = 0.516, p = 0.028). C) Correlation plot between MVC and muscle pH end-exercise for the BB (r = −0.754, p = 0.00012) and TB (r = −0.520, p = 0.027). D) Longitudinal correlation plot between delta MVC and delta PCr across MR visit 1 and MR visit 3 for the BB (r = 0.212, p = 0.560) and TB (r = 0.200, p = 0.783). E) Longitudinal correlation plot between delta MVC and delta PCr 95RT across MR visit 1 and MR visit 3 for the BB (r = −0.333, p = 0.349) and TB (r = −0.300, p = 0.683). F) Longitudinal correlation plot between delta MVC and delta muscle pH across MR visit 1 and MR visit 3 for the BB (r = 0.188, p = 0.608) and TB (r = −0.400, p = 0.517). *p < 0.05, **p < 0.01, ***p < 0.001.

Identifiability of responders versus non-responders to DMT by metabolic biomarkers

We tested whether metabolic biomarkers of muscle quality assessed by ³¹P MRS may identify responders and non-responders to DMT. Figure 6 shows the 10-month changes in each of the three ³¹P MRS outcomes measured in the BB and TB muscles between participants classified as responders and non-responders, based on changes in MVC and RULM scores, respectively. Overall, PCr depletion and muscle acidification increased between MR visit 1 and MR visit 3, but differences were not statistically significant and showed high variability, regardless of treatment response.

Figure 6.

Comparison of the change in ³¹P MRS outcomes between MR visits 1 and 3 for the m. biceps brachii (BB) and m. triceps brachii (TB) for the responders (R) and non-responders (NR) to disease-modifying treatment (DMT). The R and NR distribution is based on a minimal gain of 20% in MVC for the BB or TB (left white panel) and a minimal gain of 2 points on the RULM total score (equal for BB and TB, right grey panel). The dashed grey line indicates no change between MR visit 1 and MR visit 3. A) Difference in phosphocreatine (PCr) depletion during exercise between MR visits 1 and 3 between R and NR to DMT for BB and TB muscles. B) Difference in PCr 95% recovery time (95RT) following exercise between MR visits 1 and 3 between responders and non-responders to DMT for BB and TB muscles. C) Difference in pH end exercise between MR visits 1 and 3 between responders and non-responders to DMT for BB and TB muscles.

Discussion

Three disease-modifying treatments for SMA introduced in the past decade have transformed the prognosis of children with this severe condition. Treatment effects vary, in particular in patients with longer disease duration. Therefore, there is an increasing need for sensitive predictive biomarkers of muscle function to identify individuals more likely to experience the effects of DMT. We previously demonstrated that non-invasive dynamic in vivo ³¹P MRS interrogation of working upper-arm muscles can inform on muscle quality regarding motor unit profile and mitochondrial oxidative capacity in patients with SMA.²² In this longitudinal prospective study, we investigated whether ³¹P MRS metabolic biomarkers correlate with muscle function and, if so, whether these metabolic parameters can identify responders versus non-responders to DMT. Below, we discuss our findings and how they may impact the design and follow-up of DMT in patients with SMA.

The findings of responders versus non-responders to DMT, based on muscle strength

In this study, a responder to DMT was defined as a patient exhibiting an increase of ≥20% of MVC after 10 months of DMT.³¹ MVC measured with hand-held dynamometry is known to be a highly reliable measure in patients with SMA.^38,39 We observed a 40% and 0% responder rate to DMT for the TB and the BB, respectively. This discrepancy in responder rate may relate to the pattern of affected and spared muscles in SMA, e.g., the BB is more preserved in comparison to the TB.^22,40 The observation of selective muscle sparing in SMA remains poorly understood.^41–43 The higher responder rate for the TB may indicate greater DMT efficacy in more affected muscles. In a previous study, the improvement of muscle strength measured with the Medical Research Council (MRC), was not different for the TB and BB after 14 months of treatment with Nusinersen in adult patients with SMA (n = 15).⁴⁰ Therefore, further research is needed to conclude the muscle-specific efficacy of DMT in SMA.

Other studies in SMA have used the clinically meaningful improvement of ≥2 RULM or ≥3 HFMSE points on the total score as a responder criterion.^44,45 Using this criterion, 3/12 (25%) and 4/9 (44%) of our patients qualified as responders to treatment by the RULM and HFMSE scores, respectively (eFigure 1). In a European cohort of 237 adults with SMA types 1–4, 28.7% showed clinically meaningful improvement on both HFMSE and RULM after 14 months of treatment with Nusinersen.⁴⁶ Interpretation of these particular motor function scores is, however, complicated by floor and ceiling effects.⁴⁷ Additionally, comparing only HFMSE scores may be insufficient to identify (non-)responders to DMT. MVC is a continuous variable and a muscle-specific outcome measure and, therefore, better related to the ³¹P MRS outcomes that we determined in each patient. We found that MVC correlates strongly with HFMSE and RULM (eFigure2). However, when observing the distribution of responders based on MVC over the change in HFMSE or RULM over the 10 months, this distribution is not homogeneous (eFigure1). Whether or not an increase of ≥20% of MVC, or even an improvement of ≥3 HFMSE points or ≥2 RULM points are suitable criterion to identify responders to DMT remains a subject of debate. Considering the natural decline of 0.5 HFMSE points and 0.8 RULM points per year (1.52 points over 2 years) in SMA, it may be argued that mere stabilization of muscle function over 10 months constitutes a positive treatment response.^41,48 Nevertheless, since the timeframe is likely too short to reliably capture the expected natural decline in motor scores, we decided on functional improvement as a criterion.

We did not study the influence of age or treatment regimen on the treatment effects as of the limited cohort size. No clear associations have previously been found between disease duration and the treatment effects of Nusinersen in adolescents and adults with SMA.⁴⁹ Moreover, a limitation of this study is that we did not monitor the physical therapy of the individuals during the study, as this may affect muscle or motor function over a longer period.⁵⁰ In future studies, it is recommended to take this into account.

The potential of metabolic biomarkers assayed by ³¹P MRS to identify responders versus non-responders

We observed a strong correlation between ³¹P MRS outcomes and functional measures of the upper-arm muscles of patients with SMA (Figure 5A-C and eFigure 2). This supports our hypothesis that in responders, based on ≥20% MVC improvement at 10 months, changes in ³¹P MRS outcomes should be detectable. However, in this limited cohort, no significant correlations between within-subject changes in MVC and corresponding changes in ³¹P MRS measures were observed (Figure 5D-F). Additionally, we compared the outcome using the more clinically established responder criteria of a 2-point increase on the RULM score. Here, we selected RULM rather than HFMSE, as RULM has a better correlation with upper-arm functioning than HFMSE. Specifically, we reasoned that the improvement in MVC or RULM score in responders would reflect enhanced recruitment of FT myofibers, which produce more force and power than ST myofibers.⁵⁰ Given the differences in ATP cost per twitch, mitochondrial density, and capillary contact density among these myofiber phenotypes, this altered recruitment pattern was expected to influence muscle energetics during exercise. Responders were hypothesized to deplete more PCr, acidify more, and/or recover metabolically slower during the execution of the same physical task at visit 3 compared to visit 1.¹⁸ This was anticipated despite the fixed submaximal workload, as the increased MVC would allow responders to sustain the task for a longer duration before reaching exhaustion, thereby increasing the total energetic demand and enhancing FT fiber involvement. Unfortunately, we were unable to test this hypothesis because of an insufficient number of complete ³¹P MRS datasets (Figure 6). The enrollment of patients coincided with the COVID-19 pandemic, which severely limited the number of patients who eventually enrolled. Additionally, there was a substantial, almost bimodal variability in muscle weakness among the patients who did enroll, also visible in Figure 5. Consequently, on the one hand, the strong patients were limited to a maximum of 5000 g of weight during the MR exercise protocol, which may be below their 80% peak workload. On the other hand, some patients were limited in their strength to perform the MR exercise protocol (Table 2). Moreover, pronounced muscle atrophy in weaker patients likely reduced ³¹P MRS SNR, increasing variability in the PCr quantification.

Potential of in vivo ³¹P MRS as a generic non-invasive method to guide DMT trials

In vivo ³¹P MRS has the potential to guide DMT trials, as we observed strong correlations between MVC and the three metabolic biomarkers. There are two reasons why we could not objectify our hypothesis that ³¹P MRS has merit in this matter. Firstly, the narrow-bore MR scanner we used (forced by circumstance) was not suited for these patients with physical disabilities. If available, we would recommend using a wider bore (70 or 80 cm) MR scanner rather than the 60 cm bore MR scanner used in this study to better accommodate patients with SMA with arm contractures. Secondly, the in-magnet exercise protocol was, in retrospect, too demanding; only six patients completed the protocol for both muscles, leaving insufficient power for statistical testing. Future studies should consider lower contraction frequencies or mechanical loads below 80% MVC with a single-arm ergometer. This is based on our observation that in some patients, this combination of contraction frequency and load caused a progressive shift from a dynamic to a more isometric nature of the contraction as muscle fatigue set in. The latter was evidenced by a general upfield shift of the Pi resonance rather than Pi ‘peak-splitting’ dynamics previously observed in working upper arm muscles as the exercise progressed.²² The former indicates the onset of generic muscle acidification as a result of compromised blood flow, typically associated with isometric contractions at high frequencies and loads.¹⁸ However, this is not in line with the low post-exercise muscle pain scores reported by the patients (0.8/10 BB, 0/10 TB). With acidification, higher muscle pain scores are expected.

Despite these limitations, we observed a high test-retest reliability for resistance and time to fatigue of the MR exercise test using our MR-compatible ergometer, to provide a quantitative assessment of ATP metabolism in both affected (TB) and relatively spared (BB) muscles of the upper arm (eTable 1). Compared to ergometers in earlier studies,^22,52 this novel ergometer enabled the selective mechanical loading of individual BB and TB muscles, allowing targeted investigation of affected versus spared muscles in neuromuscular diseases.

Conclusion

This study longitudinally assessed disease-modifying treatment in patients with SMA using ³¹P MRS. Our MR platform for single arm-muscle exercise offers a non-invasive tool to monitor genetic treatment response, both in ambulatory as well as non-ambulatory patients with SMA. MVC correlated significantly with ³¹P MRS metabolic measures of PCr depletion, pH end-exercise, and 95% PCr recovery time post-exercise, in both the m. biceps brachii and m. triceps brachii. Further studies involving larger patient populations will be necessary to establish the diagnostic value of ³¹P MRS in evaluating DMT efficacy in SMA and other neuromuscular diseases.

Footnotes

Abbreviations

Acknowledgements

We thank all participants in this study for their willingness and commitment.

ORCID iDs

Lisa Pomp

Jeroen A.L Jeneson

Jeanine J Prompers

Mark Gosselink

Gabriela B.Q Davoli

Aidin Haghnejad

Fay-Lynn Asselman

W Ludo van der Pol

Bart Bartels

Ethical considerations

The study was conducted according to the principles of the Declaration of Helsinki, the code of conduct, and in accordance with the Medical Research Involving Human Subjects Act (WMO). The Medical Research Ethics Committee Utrecht (METC Utrecht) approved this study based on the requirements of the WMO (NL69612.041.19).

Consent to participate

Written informed consent was obtained from the participants before data collection.

Funding

This study was funded by Health∼ Holland and Stichting Spieren voor Spieren. The funding parties were not involved in the design of the study, the collection, analysis, and interpretation of the data, nor in writing the manuscript.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability

The data that support the findings of this study are available from the corresponding author (B.B.) on formal request.

References

Lefebvre

Bürglen

Reboullet

, et al. Identification and characterization of a spinal muscular atrophy-determining gene. Cell 1995; 80: 155–165.

Finkel

Mercuri

Meyer

, et al. Diagnosis and management of spinal muscular atrophy: part 2: pulmonary and acute care; medications, supplements and immunizations; other organ systems; and ethics. Neuromuscul Disord 2018; 28: 197–207.

Wijngaarde

Stam

Otto

LAM

, et al. Muscle strength and motor function in adolescents and adults with spinal muscular atrophy. Neurology 2020; 95: E1988–E1998.

Wadman

Stam

Gijzen

, et al. Association of motor milestones, SMN2 copy and outcome in spinal muscular atrophy types 0-4. J Neurol Neurosurg Psychiatry 2017; 88: 364–367.

Piepers

Van Den Berg

Brugman

, et al. A natural history study of late onset spinal muscular atrophy types 3b and 4. J Neurol 2008; 255: 1400–1404.

Hagenacker

Wurster

Günther

, et al. Nusinersen in adults with 5q spinal muscular atrophy: a non-interventional, multicentre, observational cohort study. Lancet Neurol 2020; 19: 317–325.

Walter

Wenninger

Thiele

, et al. Safety and treatment effects of Nusinersen in longstanding adult 5q-SMA type 3 – A prospective observational study. J Neuromuscul Dis 2019; 6: 453–465.

Coratti

Cutrona

Pera

, et al. Motor function in type 2 and 3 SMA patients treated with Nusinersen: a critical review and meta-analysis. Orphanet J Rare Dis 2021; 16: 1–12.

Maggi

Bello

Bonanno

, et al. Nusinersen safety and effects on motor function in adult spinal muscular atrophy type 2 and 3. J Neurol Neurosurg Psychiatry 2020; 91: 1166–1174.

10.

van der Woude

Wadman

Asselman

, et al. Exploring functional strength changes during nusinersen treatment in symptomatic children with SMA types 2 and 3. Neuromuscul Disord 2024; 41: 1–7.

11.

Scheijmans

FEV

Cuppen

Van Eijk

RPA

, et al. Population-based assessment of nusinersen efficacy in children with spinal muscular atrophy: a 3-year follow-up study. Brain Commun 2022; 4: fcac269.

12.

Vuillerot

. State of the art for motor function assessment tools in spinal muscular atrophy (SMA). Arch Pediatr 2020; 27: 7S40–7S44.

13.

Cano

Mayhew

Glanzman

, et al. Rasch analysis of clinical outcome measures in spinal muscular atrophy. Muscle Nerve 2014; 49: 422–430.

14.

Kemp

Ahmad

Nicolay

, et al. Quantification of skeletal muscle mitochondrial function by ³¹P magnetic resonance spectroscopy techniques: a quantitative review. Acta Physiol 2015; 213: 107–144.

15.

Prompers

Jeneson

JAL

Drost

, et al. Dynamics MRS and MRI of skeletal muscle function and biomechanics. NMR Biomed 2006; 19: 927–953.

16.

Pomp

Jeneson

JAL

van der Pol

, et al. Electrophysiological and imaging biomarkers to evaluate exercise training in patients with neuromuscular disease: A systematic review. J Clin Med. 2023; 12: 6834.

17.

Pino

Rich

Kolb

. Update on biomarkers in spinal muscular atrophy. Biomark Insights 2021; 16: 11772719211035644.

18.

Meyerspeer

Boesch

Cameron

, et al. ³¹P Magnetic resonance spectroscopy in skeletal muscle: experts’ consensus recommendations. NMR Biomed 2021; 34: e4246.

19.

Adamopoulos

Coats

AJS

Brunotte

, et al. Physical training improves skeletal muscle metabolism in patients with chronic heart failure. J Am Coll Cardiol 1993; 21: 1101–1106.

20.

Handschin

. Peroxisome proliferator-activated receptor-γ coactivator-1α in muscle links metabolism to inflammation. Clin Exp Pharmacol Physiol 2009; 36: 1139–1143.

21.

Foley

Harkema

Meyer

. Decreased ATP cost of isometric contractions in ATP-depleted rat fast-twitch muscle. Am J Physiol Cell Physiol 1991; 261: 30–35.

22.

Habets

Bartels

Asselman

, et al. Magnetic resonance reveals mitochondrial dysfunction and muscle remodelling in spinal muscular atrophy. Brain 2022; 145: 1422–1435.

23.

Dubowitz

. Enzyme histochemistry of skeletal muscle. 3. Neurogenic muscular atrophies. J Neurol Neurosurg Psychiatry 1966; 29: 23–28.

24.

Arnold

Kassar

Kissel

. Spinal muscular atrophy: diagnosis and management in a new therapeutic era. Muscle Nerve 2014; 51: 57.

25.

Fenichel

Engel

. Histochemistry of muscle in infantile spinal muscular atrophy. Neurology 1963; 13: 1059–1066.

26.

Werkman

Jeneson

Helders

, et al. Exercise oxidative skeletal muscle metabolism in adolescents with cystic fibrosis. Exp Physiol 2016; 101: 421–431.

27.

Chabanon

Seferian

Daron

, et al. Prospective and longitudinal natural history study of patients with Type 2 and 3 spinal muscular atrophy: Baseline data NatHis-SMA study. PLoS One 2018; 13. doi:10.1371/JOURNAL.PONE.0201004

28.

Diekman

Visser

Schmitz

JPJ

, et al. Altered energetics of exercise explain risk of rhabdomyolysis in very long-chain acyl-coa dehydrogenase deficiency. PLoS One 2016; 11: e0147818.DOI: 10.1371/JOURNAL.PONE.0147818

29.

Hammersmith Functional Motor Scale-Expanded (HFMSE) Manual of Procedures and Score Sheets. Accessed September 12, 2024. www.biogen-pharma.si.

30.

Mazzone

Mayhew

Montes

, et al. Revised upper limb module for spinal muscular atrophy: development of a new module. Muscle Nerve 2017; 55: 869–874.

31.

Bohannon

. Reference values for extremity muscle strength obtained by hand-held dynamometry from adults aged 20 to 79 years. Arch Phys Med Rehabil 1997; 78: 26–32.

32.

Gursan

Kahraman-Agir

Gosselink

, et al. Development of a double tuned 2H/³¹P whole-body birdcage transmit coil for 2H and ³¹P MR applications from head to toe at 7T. NMR Biomed 2025; 38: e5325.

33.

Froeling

Prompers

Klomp

DWJ

, et al. PCA Denoising and wiener deconvolution of 31 P 3D CSI data to enhance effective SNR and improve point spread function. Magn Reson Med 2021; 85: 2992–3009.

34.

Jeneson

JAL

Schmitz

JPJ

Hilbers

PAJ

, et al. An MR-compatible bicycle ergometer for in-magnet whole-body human exercise testing. Magn Reson Med 2010; 63: 257–261.

35.

Rzanny

Stutzig

Hiepe

, et al. Untersuchungen zur Reproduzierbarkeit verschiedener metabolischer Marker für die muskuläre Fasertypenverteilung durch funktionelle ³¹P-MRS während einer dynamischen Belastungsübung. Z Med Phys 2016; 26: 323–338.

36.

Intrathecal administration with SPINRAZA ®. (nusinersen).

37.

Evrysdi® (risdiplam) Prescribing Information. Genentech, Inc.

38.

Merlini

Mazzone

Solari

, et al. Reliability of hand-held dynamometry in spinal muscular atrophy. Muscle Nerve 2002; 26: 64–70.

39.

Kolber

Cleland

. Strength testing using hand-held dynamometry. Phys Ther Rev 2005; 10: 99–112.

40.

Schreiber-Katz

Siegler

Wieselmann

, et al. Improvement of muscle strength in specific muscular regions in nusinersen-treated adult patients with 5q-spinal muscular atrophy. Sci Rep 2023; 13: 6240.DOI: 10.1038/S41598-023-31617-5

41.

Wadman

Wijngaarde

Stam

, et al. Muscle strength and motor function throughout life in a cross-sectional cohort of 180 patients with spinal muscular atrophy types 1c–4. Eur J Neurol 2018; 25: 512–518.

42.

Otto

LAM

van der Pol

Schlaffke

, et al. Quantitative MRI of skeletal muscle in a cross-sectional cohort of patients with spinal muscular atrophy types 2 and 3. NMR Biomed 2020; 33: e4357.DOI: 10.1002/NBM.4357

43.

Brogna

Cristiano

Verdolotti

, et al. MRI Patterns of muscle involvement in type 2 and 3 spinal muscular atrophy patients. J Neurol 2020; 267: 898–912.

44.

Mercuri

Darras

Chiriboga

, et al. Nusinersen versus sham control in later-onset spinal muscular atrophy. N Engl J Med 2018; 378: 625–635.

45.

Pechmann

Behrens

Dörnbrack

, et al. Improved upper limb function in non-ambulant children with SMA type 2 and 3 during nusinersen treatment: a prospective 3-years SMArtCARE registry study. Orphanet J Rare Dis 2022; 17: 1–10.

46.

Günther

Wurster

Brakemeier

, et al. Long-term efficacy and safety of nusinersen in adults with 5q spinal muscular atrophy: a prospective European multinational observational study. Lancet Reg Health Eur 2024; 39. DOI: 10.1016/j.lanepe.2024.100862

47.

Coratti

Pera

Montes

, et al. Different trajectories in upper limb and gross motor function in spinal muscular atrophy. Muscle Nerve 2021; 64: 552–559.

48.

Coratti

Carmela Pera

Montes

, et al. Revised upper limb module in type II and III spinal muscular atrophy: 24-month changes. Neuromuscul Disord 2022; 32: 36–42.

49.

Hagenacker T, Maggi L, Coratti G, et al. Effectiveness of nusinersen in adolescents and adults with spinal muscular atrophy: systematic review and meta-analysis. Neurol Ther 2024;13(5):1483-1504.

50.

Mirea A, Leanca MC, Onose G, et al. Physical therapy and nusinersen impact on spinal muscular atrophy rehabilitative outcome. Front Biosci (Landmark Ed). 2022;27(6):179.

51.

Bottinelli

Canepari

Pellegrino

, et al. Force-velocity properties of human skeletal muscle fibres: myosin heavy chain isoform and temperature dependence. Journal of Physiology 1996; 495: 573–586.

52.

Vegter

Brink

Mouton

, et al. Magnetic Resonance-Compatible Arm-Crank Ergometry. Front Physiol 2021; 12. DOI: 10.3389/fphys.2021.599514

Longitudinal cohort study of muscle function and metabolic biomarkers in disease-modifying treatment of spinal muscular atrophy

Abstract

Keywords

Introduction

Materials and methods

Participants

Sample size calculation

Study design

Ergometer

Demographics and anthropometry

Motor function

Isometric muscle strength

Dynamic supine MR-compatible upper arm exercise test

Muscle 31P magnetic resonance spectroscopy

31P MRS data acquisition

31P MRS data analysis

Statistical analysis

Results

Demographics

Study execution

Response to DMT

In vivo 31P MRS observations

Correlations between functional and metabolic biomarkers of muscle quality

Identifiability of responders versus non-responders to DMT by metabolic biomarkers

Discussion

The findings of responders versus non-responders to DMT, based on muscle strength

The potential of metabolic biomarkers assayed by 31P MRS to identify responders versus non-responders

Potential of in vivo 31P MRS as a generic non-invasive method to guide DMT trials

Conclusion

Footnotes

Abbreviations

Acknowledgements

ORCID iDs

Ethical considerations

Consent to participate

Funding

Declaration of conflicting interests

Data availability

References

Muscle ³¹P magnetic resonance spectroscopy

³¹P MRS data acquisition

³¹P MRS data analysis

In vivo ³¹P MRS observations

The potential of metabolic biomarkers assayed by ³¹P MRS to identify responders versus non-responders

Potential of in vivo ³¹P MRS as a generic non-invasive method to guide DMT trials