Troponin T is elevated in a relevant proportion of patients with 5q-associated spinal muscular atrophy

Troponin T concentration (TNT) is commonly considered a marker of myocardial damage. However, elevated concentrations have been demonstrated in numerous neuromuscular disorders, pointing to the skeletal muscle as a possible extracardiac origin. The aim of this study was to determine disease-related changes of TNT in 5q-associated spinal muscular atrophy (SMA) and to screen for its biomarker potential in SMA. We therefore included 48 pediatric and 45 adult SMA patients in this retrospective cross-sequential observational study. Fluid muscle integrity and cardiac markers were analyzed in the serum of treatment-naïve patients and subsequently under disease-modifying therapies. We found a TNT elevation in 61% of SMA patients but no elevation of the cardiospecific isoform Troponin I (TNI). TNT elevation was more pronounced in children and particularly infants with aggressive phenotypes. In adults, TNT correlated to muscle destruction and decreased under therapy only in the subgroup with elevated TNT at baseline. In conclusion, TNT was elevated in a relevant proportion of patients with SMA with emphasis in infants and more aggressive phenotypes. Normal TNI levels support a likely extracardiac origin. Although its stand-alone biomarker potential seems to be limited, exploring TNT in SMA underlines the investigation of skeletal muscle integrity markers.

5q-associated spinal muscular atrophy (SMA) is a rare neuromuscular disorder that is caused by loss-of-function mutations of the ubiquitously expressed survival of motor neuron 1 gene (SMN1, MIM *600,354) and is clinically characterized by progressive, proximally pronounced muscle weakness and atrophy.Neuropathological studies revealed underlying degeneration of the motoneurons in the ventral horn of the spinal cord with secondary skeletal muscle atrophy and myopathy.However, developmental alterations of the postsynaptic motor endplate and neurodegeneration of other extra-motor tissues have been reported.SMA is classically divided into five subtypes according to the best ever reached motor milestone and the age at disease onset.Type zero, the congenital form, shows the most severe form of progression, while type four is relatively mild.While there were no treatment options until recently, the first disease-modifying therapy with nusinersen has been approved in 2016 by the FDA 1,2 .The stunning story of beneficial disease-modifying treatments for SMA has continued with the introduction of the SMN2-RNA splice modulator risdiplam 3,4 and the SMN1 gene replacement therapy onasemnogene abeparvovec-xioi 5 .While these disease-modifying therapies are targeting SMN depletion as the main contributor of the disease, more symptom-oriented approaches aiming to increase muscle strength and functional performance are under investigation.Fast skeletal muscle troponin activators lead to increased muscle strength relative to the neuronal input and are currently assessed in clinical trials 6,7 .Other approaches include myostatin inhibition 8 or modulators of muscle metabolism 9 .
As there has no biomarker been validated of these new therapeutic options for the clinical routine yet, clinical scales remain the most relevant outcome parameter to date 10 .Major efforts for identifying suitable biomarkers have been undertaken but none have been implemented into clinical routine yet.Although muscle plays a key role in SMA pathology, research on fluid muscle biomarkers has been scarce.However, creatinine and creatine kinase have been discussed as candidate biomarkers in SMA, indicating that markers of muscle integrity are potentially useful for monitoring disease progression 11,12 .
Troponin I (TNI) and Troponin T concentration (TNT) are highly sensitive markers for myocardial ischemia 13 with evidence for age-and sex-dependent upper reference limits of TNT 14,15 .However, elevated TNT but not TNI levels have recently been found in numerous systemic diseases, such as paraneoplastic systemic sclerosis 16 , inclusion body myositis 17 , polymyositis 18 and other diseases involving skeletal myopathies 19,20 .An extracardiac origin of TNT 21 and laboratory cross reactions with skeletal isoforms have been discussed 22,23 .Chronically elevated TNT levels were reported in amyotrophic lateral sclerosis (ALS) 24 , where TNT levels increased with disease duration and progression rate 25 .Furthermore, TNT levels correlated with motor function in patients with ALS.Of note, patients with the pure upper motoneuron variant primary lateral sclerosis (PLS) did not exhibit elevated TNT levels.Therefore, TNT was proposed as a biomarker of lower motoneuron or neuromuscular involvement 24 .Stevens et al. reported an altered expression of troponin isoforms and myosin heavy chains in infants with SMA type 1-3 (aged 1-5 months) 26 , while Djordjevic et al. did not find elevated TNT levels in children with SMA type 2 (mean age 5.4 years, range 2.5-11.7 years) and 3 (mean age 10.2 years, range 5.2-13.7 years) 27 .
This study aims to investigate disease-related changes of TNT and to screen for its biomarker potential in SMA.

Study design
The study was performed as a multicenter, retrospective, observational analysis.Inclusion criteria were the presence of 5q-associated SMA with molecular genetic proof of homozygous deletion or other mutation in the SMN1 gene and the absence of a relevant cardiac disease based on the patients' disease history.Data were collected from n = 45 children and n = 48 adults with SMA from the neurological and pediatric departments of the university hospitals in Bonn and Dresden (Germany).Informed consent was obtained from all subjects or their legal guardians.The conduct of the study was compliant with the World Medical Association Declaration of Helsinki and the guidelines and regulations of the respective local ethic committees, which approved the study (Bonn: Ethics Board decision letter 324/20, Dresden: EK 393,122,012 and EK 183,042,019).

Laboratory analysis and functional scores
We retrospectively collected laboratory data prior to (treatment-naïve, baseline) and during treatment with disease-modifying treatments (nusinersen, risdiplam or onasemnogene abeparvovec-xioi) up to 38 months.According to the standard of care, nusinersen was administered intrathecally every four months after four initial loading doses.Risdiplam is a daily administered oral drug and onasemnogene abeparvovec-xioi is administered intravenously once, with the application being recommended before two years of age in Germany.The laboratory assays were performed at the in-house laboratory of the respective university hospital with standardized protocols used for clinical routine.Serum concentrations of TNT were analyzed in all samples using an electrochemiluminescence immunoassay (ECLIA) with an upper limit of normal of < 14 ng/L for all age and sex categories (Elecsys® Troponin T hs, pharmaceutical, Roche Diagnostics).Cardiac troponin I (TNI) was analyzed, whenever data were available, in order to screen for cardiac involvement (upper limit of normal: 10 ng/L).As TNI is more cardiospecific, TNT elevation without corresponding TNI elevation was postulated to originate from extracardiac tissue.Creatine kinase activity (CK, as a marker of muscle destruction, upper limit of normal: male 3.2 µkat/L, female 2.85 µkat/L 28 ), myoglobin concentration (Myo, as a second marker of muscle destruction, upper limits of normal age-and gender-dependent, see 29 ) and creatinine concentration (Crn, a surrogate marker of muscle mass, upper limits of normal age-and gender-dependent, see 30 ) were included when available.Disease severity was assessed with the Revised Upper Limb Module (RULM) 31 , Hammersmith Functional Motor Scale Expanded (HFMSE) 32 or the revised ALS-Functional Rating Scale (ALSFRS-R) 33 , which are commonly used scales to assess the motor skills and disease-related functional properties of patients with SMA.

Statistical analysis
Statistical analysis and data visualization were performed using IBM SPSS Statistics Version 21.0.Unless otherwise stated, data are presented as mean ± standard deviation (SD).Shapiro-Wilk test was performed to test data on normal distribution and rank-based, non-parametric tests (Wilcoxon test, Mann-Whitney U test) were applied for non-normally distributed data.ANOVA was used additionally as it is insensitive of violations of normality in appropriate sample sizes 34 .The association between variables was analyzed rank-based using Spearman correlations and Eta 2 for nominal variables.A correlation coefficient (ρ) of ρ < 0.3 was considered as a weak, ρ = 0.3-0.59 as a moderate, and ρ ≥ 0.6 as a strong correlation.An Eta 2 (η 2 ) of η 2 = 0.01-0.059was considered as weak, η 2 0.06-1.39 as moderate and ≥ 1.4 as strong effect 35 .Regressions were analyzed after adjustment for multicollinearity.TNT was entered as the dependent variable and those parameters that differed significantly between the groups with and without TNT elevation were entered as independent variables.Based on these results, further analyses were performed separately for adults and children.Longitudinal TNT data were examined using rank-based, non-parametric tests (Wilcoxon test).Data sets with missing values were excluded pairwise for cross-sequential analysis.Statistical significance was set as p < 0.05 two-sided.

Results
In total, we included data from 93 patients with SMA of which 48 patients were adults (51.6%) and 45 were children below the age of 18 years (48.4%, see Table 1 and S1 2 and Fig. S1).
A regression analysis with TNT as dependent variable and age, SMA type, CK, Crn and BMI as independent variables revealed that age has a significant impact on TNT (standardized coefficient B = -0.61;p = 0.003).We therefore separated the cohort into adults and children for further analyses.
Regression analysis in adult SMA including the above-mentioned variables revealed that sex had a significant impact on TNT (p = 0.003).
Risdiplam.19 adult patients with SMA (9 female; mean age 31.9 years, ranging from 18 to 58 years) received risdiplam.Baseline TNT data were available for 17 patients (TNT: 13.8 ± 7.9 ng/L) with a TNT elevation in 6/17 patients (35.3%, TNT: 22.1 ± 7.6 ng/L).Only a short observation period is available, which leads to strong limitations to draw robust conclusions on the following results.In this limited data set, TNT did not change under risdiplam, neither after two months (TNT: 13.4 ± 7.9 ng/L, p = 0.17) nor after six months (TNT: 12.1 ± 4.6 ng/L, p = 0.60, see Fig. S4).Subgroup analyses revealed decreased TNT concentrations in the group with initial TNT elevation after two months (5 patients, TNT: 18.4 ± 8.6, p = 0.042).Later analyses were hindered by data  S6).Subgroup analyses of the cohort with initial TNT elevation revealed no changes up to 6 months of treatment (TNT: 59.7 ± 25.4 ng/L, p = 0.17).

Discussion
We found elevated concentrations of TNT in more than half of treatment-naïve patients with SMA in our cohort, a marker primarily used for myocardial infarction.In contrast, none of the samples studied had concentrations of the myocardial specific TNI above the cut-off values.This strongly argues against myocardial damage and we therefore speculate that TNT originates from degenerating or regenerating skeletal muscle fibers in patients with SMA.Animal studies have demonstrated that cardiac TNT is physiologically expressed in the rat skeletal muscle throughout embryonic and fetal development, disappears during the first weeks after birth and is re-expressed in regenerating muscle fibers after an injury or after degeneration 37 .Similarly, TNT but not TNI were expressed in muscles of patients with other neuromuscular disorders such as Duchenne muscular dystrophy 21,38 , and TNI was not expressed in fetal and healthy or diseased adult skeletal muscle 39 .Existing literature indicates that TNT levels are higher in healthy neonates and gradually decrease with age, largely during the first 3 months of life 40 .TNT values above the cut-off of 14 ng/L in healthy children were previously reported, particularly up to the age of 6 months 14 .The highest TNT levels measured up to 37 ng/L (97.5th percentile) and declined with age.At one year of age, the 97.5th percentile dropped to the adult cut-off of 14 ng/L.As a result, the upper limit of normal of the assay we used does not seem to be suitable for young children.We therefore compared our data with the reference values developed by Kiess et al.: In 86% of the treatment-naïve infants with SMA and in 90% of children in the second year of life in our cohort, we found TNT concentrations above this 97.5thpercentile of healthy (compare Fig. S1).TNT elevation was more than twice as prominent and three times higher in children compared to adults.Furthermore, treatment-naïve TNT concentrations in children correlated inversely with the SMA type, SMN2 copy number and age, suggesting an association of TNT concentration with more aggressive subtypes in combination with early disease stages 41 .During aging, restoration of muscle maturation and nerve function following injuries are delayed 42,43 and non-neurodegenerative processes such as inflammation and secondary myopathy might outweigh motoneuron loss in later stages 44 .Thus, more active denervation and regeneration processes in muscles of the aggressive phenotype SMA type 1 might result in a stronger re-expression of TNT in the skeletal muscles compared to milder phenotypes.Still, treatment-naïve  11,36 .Possible explanations include secondary myopathic processes in patients with more remaining vulnerable muscle mass, leading to increased absolute CK levels in the peripheral blood.As most of our SMA patients show strong muscle atrophy, the standard absolute CK reference values might not be suitable for these patients.Whether or not patients with severe muscle atrophy show relatively increased CK must be studied under consideration of the lean body mass in future studies.
A second striking result was a clear sex difference in TNT elevation in adult patients with SMA, which seems to be much more pronounced than recent cardiologic studies reported for sex differences between healthy males and females [45][46][47] .Although we did not observe differences in motor function, male patients showed significantly higher baseline TNT levels than females.We suspected an effect of muscle mass, however, TNT correlated only moderately to Crn but not to BMI in our adult cohort and there were no significant differences of Crn or BMI between adult male and female patients.However, further analyses including quantitative analyses like dual x-ray absorptiometry or other assessments of lean body mass are warranted to exclude an impact of muscle mass on sex-dependent TNT concentrations in patients with SMA.A second hypothesis was an association with sex hormones.Several studies reported a TNT decrease under testosterone therapy 48,49 , making it rather unlikely that the sex hormone testosterone is responsible for the TNT difference between male and female patients.At present, we do not have a conclusive explanation for the sex difference in our cohort.
Longitudinal analysis of TNT concentrations during disease-modifying treatment with nusinersen in adult patients with SMA revealed a significant decrease after 34 and 38 months of treatment in the subcohort of patients with elevated TNT concentrations at baseline.Quantitative magnetic resonance imaging showed stabilized muscle mass under nusinersen treatment, indicating a slowed or reduced muscle wasting.Such reduced muscle wasting under nusinersen might result in decreasing CK and TNT levels (compare 11,50 ).Compared to the dynamics of the muscle integrity markers CK and myoglobin during treatment in our study, the aforementioned dynamic of TNT seems to be much weaker.
No sufficient data for adults treated with risdiplam or children and infants treated with risdiplam and nusinersen were available.In the group of onasemnogene abeparvovec-xioi-treated infants and children with SMA, no significant decrease of TNT concentration was observed in the very short observation period of 6 months.We did not observe a TNT or TNI increase shortly after administration of onasemnogene abeparvovec-xioi, suggesting that none of our patients experienced a drug-related myocarditis, a feared side effect of the treatment.Longer observation periods would be needed to draw proper conclusions on TNT dynamics in infants and children during disease-modifying treatment.
In conclusion, we observed an extracardiac TNT elevation in a relevant proportion of patients with SMA, supporting the potential usefulness of fluid molecular skeletal muscle markers as biomarkers for SMA.Interestingly.TNT was more pronounced in more aggressive subtypes in early disease stages.We hypothesized that possible mechanisms might include differential effects in children and adults: While TNT elevation in children might result from more active denervating and regeneration processes with altered TNT expression in the muscle, it might stem from muscle destruction processes in progressed disease stages.Nonetheless, this hypothesis is highly speculative and further research is warranted.Taking into account the limited longitudinal dataset, the dynamic of TNT during the disease-modifying treatment nusinersen showed a slight decline in the subset of patients with initially elevated TNT levels in adults with SMA.However, in comparison to CK or Myoglobin, the added biomarker value of TNT for monitoring therapeutic effects under disease-modifying therapies seems to be low.Nevertheless, our data argue for further investigations on molecular biomarkers of the skeletal muscle in patients with SMA.
Our study has limitations.It is a retrospective cross-sequential observational study and as such, there were numerous missing data points.The cohorts for subgroup analyses were relatively small and the observation periods relatively short, which further limited data analysis, particularly in infants and children.Further, as lined out above, CK analyses and Crn analyses with the Jaffe method showed a relevant floor effect that limited our data analysis.Investigating lean body mass could have added valuable information but was not collected.Data were collected in two different centers in Germany and reference values and units differed for some variables.Finally, we are aware that we operated with cut-off values for TNT and TNI that were optimized to detect cardiac events.Although TNI levels below the threshold of detection have a high negative predictive value for cardiac diseases 51 , it must be noted that TNI alone cannot rule out underlying cardiac disease.However, given the low pretest probability in our cohort due to application of the exclusion criteria, relevant cardiac diseases seem unlikely.Future research will show whether monitoring of neuromuscular processes requires different cut-off values than cardiac events.

Figure 1 .
Figure 1.TNT treatment-naïve concentrations.A: TNT treatment-naïve concentrations per SMN2 copy numbers, B: TNT treatment-naïve concentration between SMA types, C: TNT treatment-naïve concentrations between adults and children, D: TNT treatment-naïve concentration between male and female adult patients.Adult TNT cut-off (14 ng/L) displayed as dotted line.Whiskers indicate minimum and maximum, circles indicate extreme outliers with an interquartile range > 1.5 -3 and diamond shapes indicate outliers with an interquartile range > 3. **p < 0.001.Patients identified by newborn screening are not displayed.

Figure 2 .
Figure 2. Longitudinal dynamic of TNT and skeletal muscle markers (CK, Crn, Myoglobin) in adult patients with SMA under nusinersen.Mean and standard deviation are displayed.TNT cut-off (14 ng/L) displayed as dotted line.Number of patients included in the analyses are indicated in gray.Significant changes from baseline are indicated by *p < 0.05 or **p < 0.001.

Figure 3 .
Figure 3. Longitudinal display of TNT in adult patients with SMA under nusinersen in sex subgroups.Mean and standard deviation are displayed.TNT cut-off (14 ng/L) displayed as dotted line.Significant changes from baseline are indexed in the respective color as follows: *p < 0.05.

Table 1 .
Demographics of the study cohort.NBS, diagnosed pre-symptomatically by newborn screening.

Table 2 .
Comparison between the group with and without baseline TNT elevation.NBS Newborn screening, TNT Troponin T, BMI Body mass index, CK = Creatine kinase activity, Crn Creatinine concentration, * just informational display, no comparison due to age-and sex-specific reference values (see below), LLN Lower limit of normal.