Total tau in cerebrospinal fluid detects treatment responders among spinal muscular atrophy types 1–3 patients treated with nusinersen

Abstract Aims Considering the substantial variability in treatment response across patients with spinal muscular atrophy (SMA), reliable markers for monitoring response to therapy and predicting treatment responders need to be identified. The study aimed to determine if measured concentrations of disease biomarkers (total tau protein, neurofilament light chain, and S100B protein) correlate with the duration of nusinersen treatment and with scores obtained using functional scales for the assessment of motor abilities. Methods A total of 30 subjects with SMA treated with nusinersen between 2017 and 2021 at the Department of Pediatrics, University Hospital Centre Zagreb, Croatia, were included in this study. Cerebrospinal fluid (CSF) samples were collected by lumbar puncture prior to intrathecal application of nusinersen. Protein concentrations in CSF samples were determined by enzyme‐linked immunosorbent assay in 26 subjects. The motor functions were assessed using functional motor scales. Results The main finding was significantly decreased total tau correlating with the number of nusinersen doses and motor improvement in the first 18–24 months of treatment (in all SMA patients and SMA type 1 patients). Neurofilament light chain and S100B were not significantly changed after administration of nusinersen. Conclusions The measurement of total tau concentration in CSF is a reliable index for monitoring the biomarker and clinical response to nusinersen therapy in patients with SMA.


| INTRODUC TI ON
Proximal spinal muscular atrophy (SMA) is a progressive autosomal recessive neurodegenerative disease with an incidence of 1 in 6-11,000 live births. 1 It is caused by homozygous deletion or mutation of the SMN1 gene (survival motor neuron 1, telomeric) on chromosome 5q11.2-q13.3. The leading sign is progressive muscle weakness and atrophy.Depending on the age of onset of the first signs, the disease is clinically classified into five phenotypes; the most common type of SMA that occurs in the first 6 months, SMA type 1 (Werdnig-Hoffmann disease), SMA type 2 (the first signs manifest at the age of 7-18 months), SMA type 3 (Kugelberg-Welander disease) that occurs after 18 months and is subdivided into 3a that manifests before 3 and 3b that manifests after 3 years of age, SMA type 4 with the onset in the second or the third decade of life, and SMA type 0 that manifests with fetal hypomotility or akinesia prenatally. 2 Patients with SMA have a mutation or deletion of the SMN1 gene, with all patients having at least one copy of the SMN2 (survival motor neuron 2, centromeric) gene as they would not survive in utero without it.It seems that the SMN2 gene is unique to humans. 3[5] The SMN gene on chromosome 5 is a duplicate gene that has two highly homologous copies, a telomeric copy called SMN1 and a centromeric copy known as the SMN2 gene.SMN2 differs from SMN1 in cytosine substitution by thymine in exon 7 (840C<T), which is sufficient to silence the inclusion of exon 7 in the SMN2 gene. 6e SMN1 gene provides a full-length functional SMN protein, and in healthy individuals, its activity is dominant and sufficient.
Homozygous disruption of the SMN1 gene due to deletion of exon 7 or gene conversion of SMN1 to SMN2 leads to the development of SMA in about 95% of cases. 7About 3% of SMA patients are compound heterozygotes for deletion of one SMN1 allele and small intragenic mutations.However, the type of SMA, i.e., the time of onset and severity of clinical signs, also depends on the involvement of surrounding genes in the 5q13 region as well as some other genes located on other chromosomes.Among the listed disease-modifying genes, SMN2 stands out as the most important.Although it is responsible for the production of only 10% of functional SMN protein in healthy people, in case of a deletion of the SMN1 gene, its activity is essential for survival. 8,9Unlike SMN1, the number of copies of the SMN2 gene is variable and there may be more than one copy of the gene.The greater the number of SMN2 copies, typically milder the SMA phenotype that is also less progressive and has a later onset.About 80% of patients with SMA type 1 have one or two SMN2 copies.About 82% of patients with SMA type 2 carry three SMN2 copies, while 96% of patients with SMA type 3 have three or four copies of this gene. 10The fact that other genes also influence SMA severity 11 may explain why other studies did not observe a correlation between SMN levels and SMA severity. 12SMN2 gene variant c.859G>C in exon 7 causes a milder phenotype as it increases the inclusion of exon 7.In addition to SMN2, other disease-modifying genes have been found as well, one of which is the NLR family apoptosis inhibitory protein (NAIP)/BIRC1 gene, encoding a protein from the NLR family of apoptosis inhibitors.Although the role of the NAIP gene in SMA has not yet been fully elucidated, it is associated with the severity of the clinical presentation.Patients with a homozygous SMN1 exon 7 deletion who do not have a copy of the NAIP gene have a more severe phenotype: in 75% of patients with SMA type 1, the NAIP gene is absent. 13Additional SMA-modifying genes are PLS3 (plastin 3), 14 CORO1C (coronin 1C), 15 and others. 16e discovery of the intronic splice silencing site (ISS-N1) in 2004 led to the first SMA disease-modifying treatment, nusinersen, and antisense oligonucleotide [2'-O-(2-methoxy)-ethyl-phosphorothioat e] for intrathecal administration, approved by the U.S. Food and Drug Administration (FDA) in 2016. 17Later, the FDA approved gene replacement therapy onasemnogene abeparvovec for the treatment of SMA in children under 2 years of age, designed as one dose for life, administered intravenously. 18,19Also, a small-molecule risdiplam, a pyridazine derivative administered orally, acts similar to nusinersen by increasing the inclusion of exon 7 in the mRNA of the patient's SMN2 gene. 16sinersen targets the ISS-N1 sequence in intron 7 of the SMN2 gene that prevents the inclusion of exon 7 in SMN2 mRNA.When nusinersen binds to the intronic splicing silencer (ISS)-N1 sequence, it allows the inclusion of exon 7 in SMN2 mRNA, and consequently, increases the production of full-length SMN protein. 16,17N protein is involved in the biogenesis of small nuclear ribonucleoproteins and mRNA splicing and probably has a major role in motoneuron cellular as well as neuromuscular junction functions.SMN protein levels are significantly reduced in cells of the central and peripheral nervous system in SMA patients. 19 or by replacement of fully functional recombinant SMN1 transgene (onasemnogene abeparvovec).
Neurofilaments (Nf), especially phosphorylated Nf heavy-chain (pNFH) and light-chain Nf (NfL) levels have emerged as potential biological markers in monitoring SMA progression and response to nusinersen therapy in infants. 21Elevated levels of NfL in the CSF reflect neuronal and axonal degeneration 22 ; therefore, it is not a disease-specific biomarker. 23In addition to Nf, serum creatinine and creatine kinase may also correlate well with motor function and disease severity, and their measurement in the blood might be useful to determine patients who will respond to therapy. 24her, potentially useful biological markers for monitoring response to nusinersen are total tau (t-tau) protein and S100 calciumbinding protein B (S100B). 25 Tau proteins maintain the stability of microtubules.The increase in the concentration of t-tau protein in the CSF occurs due to neuronal death and axonal degeneration. 26 increase in the concentration of t-tau protein has been observed in several neurodegenerative diseases, especially Alzheimer's disease 4,26-33 and Creutzfeldt-Jakob disease. 3400 proteins belong to the large family of small dimeric multigenic calcium-binding proteins. 35S100B is specific for astrocytes, especially mature astrocytes whose end-feet are crucially involved in the formation of the blood-brain barrier (glia limitans). 36An increase in S100B occurs in many diseases of the central nervous system such as Alzheimer's disease, Parkinson's disease, amyotrophic lateral sclerosis, and multiple sclerosis, while elevated S100B levels in the peripheral nervous system indicate inflammatory autoimmune diseases or axonal degeneration and regeneration. 37ile some studies showed that NfL levels decrease in SMA patients upon nusinersen treatment, [38][39][40][41][42] data on S100B and t-tau are quite variable.Only one study measured S100B levels in patients with SMA type 3, showing no difference in S100B after nusinersen treatment. 25Two studies, including mainly SMA type 1 patients, observed a significant decrease in t-tau levels upon nusinersen treatment, whereas others, including mainly SMA type 2 and type 3 patients, did not report changes in t-tau levels after nusinersen. 25,43,44Thus, we aimed to test the potential of t-tau, NfL, and S100B proteins as theragnostic biomarkers in SMA.We hypothesized that nusinersen therapy would lead to a decrease in the concentrations of t-tau, NfL, and S100B proteins in the CSF of SMA patients.7][48][49] Genomic DNA was isolated from a peripheral whole blood sample.The laboratory participates in the quality control scheme EMQN.

| ME THODS
CSF samples were recovered from 13 SMA type 1, 4 SMA type 2, and 9 SMA type 3 patients.Each subject underwent a thorough neurological examination.Healthy subjects were not included in the study.
CSF samples were taken upon the informed consent of a parent or legal guardian.CSF samples (5 ml) were collected by spinal tap (lumbar puncture) from the L3/L4 intervertebral space just before nusinersen administration.A small portion of CSF (1 ml) was aliquoted and stored at −80°C in polypropylene tubes for later analysis (Table 2).However, we were not able to collect CSF for analysis of biomarkers after the administration of every nusinersen dose.The time points at which CSF was collected for analysis of biomarkers are indicated in Table 2.

| Assessment of motor functions
The motor abilities of SMA patients were measured using the Children's Hospital of Philadelphia Infant Test of Neuromuscular Disorders (CHOP INTEND) functional scale (score 0-64) 50 and the Hammersmith Functional Motor Scale Expanded (HFMSE) scale (score 0-66). 51Patients with SMA type 1 and type 2 were tested using both the CHOP INTEND and HFMSE, as recommended for children up to the age of 2 years and older/>2 years, non-ambulant were assessed by CHOP INTEND, whereas children who scored 50 or more CHOP INTEND points were additionally tested using HFMSE.SMA patients older than 2 years of age with SMA types 2 and 3 were tested using the HFMSE only (a small number of the patients were sometimes also tested using the HFMS (Hammersmith Functional Motor Scale)). 52It is recommended that all patients with SMA be neurologically examined at baseline prior to the first initial four doses and every 4 months afterward before each successive dose (5th, 6th, etc.), which means four times in the first year and three times per each successive year, and the examination should include an assessment of motor functions measurements based on functional scales as well as breathing functions. 53available kits for t-tau (Innotest hTau Ag, Fujirebio, Tokyo, Japan), NfL (Human Nf-L/NEfL ELISA Kit, LifeSpan BioSciences Inc.), and S100B (Human S100B, R&D Systems) were used.

| Statistical analysis
Statistical analysis was performed with SPSS, version 19.0.1 (SPSS).
The statistical significance level was set at α = 0.05 for all tests.
The normality of the distribution of individual variables within the groups was tested using the Kolmogorov-Smirnov test.Additionally, due to the small number of subjects, non-parametric Kruskal-Wallis and Mann-Whitney tests were used.The relationship between the tested variables (dependent variables) and the received number of nusinersen doses (independent variable) was tested using linear regression.Outliers were not included in the statistical analysis.A comparison of genetic data among patients with different types of SMA was performed using the χ 2 test.Given the exploratory nature and design of the study, no correction for multiple testing was performed.

| Comparison of t-tau protein levels in patients with SMA of all types
Linear regression showed a statistically significant negative association between t-tau protein concentrations (dependent variable) and nusinersen dose values (independent variable, β = −0.23;SE = 3.64; p = 0.04, Figure 1A,B).T-tau protein levels were significantly reduced in all SMA patients (SMA 1, SMA 2, and SMA 3) who received the 4th dose of nusinersen compared to baseline values.The concentrations of t-tau protein were significantly reduced in patients who received the 5th, 6th, 7th, 9th, 10th, and 11th dose of the drug compared to baseline values (Figure 1A).T-tau protein levels were significantly reduced in patients who received the 7th and 9th dose of the drug compared to those who received the 3rd dose of the drug (Figure 1A).
As means of grouped data are more accurate than means of ungrouped data, we analyzed patient data in four groups: before Note: CSF for analysis of biomarkers was not collected during the administration of every nusinersen dose.

| Comparison of t-tau protein levels in SMA patients in relation to the duration of treatment and SMA type
Similar to pooling the number of doses, we grouped the patients who were receiving nusinersen into six groups: 0-6 months, 6-12 months, 12-18 months, 18-24 months, 24-30 months, and 30-36 months.The rationale for such division into subgroups was the assumption that half-year periods were sufficiently long to detect significant changes if they occur.In comparison to baseline values, the results for all 30 SMA patients taken together were highly significant for all subgroups up to the 18-24 months group (Figure 2A; Table 3), whereas the results for SMA type 1 subjects only, probably due to a lower number of cases analyzed, were significant for the 18-24 months group only (Figure 2B).

| Comparison of NfL and S100B protein levels in SMA patients
The NfL protein values were overall not significantly decreased be- In patients with SMA of all three types who received a different number of nusinersen doses, the S100B protein levels did not differ significantly from baseline values (H test = 13.33;df = 13; p = 0.42; Figure 3) in all SMA patients.This was confirmed by linear regression

| CHOP INTEND and HFMSE scores in patients with SMA of all three types
A statistically significant difference in the number of points on the CHOP INTEND scale was seen between SMA type 1 patients who received a different number of nusinersen doses (H test = 17.9; df = 8; p = 0.02; Figure 4).The most significant differences were seen between SMA type 1 patients who received the 6th, 7th, 8th, and 9th dose of nusinersen compared to baseline values (Table 4).
Patient data were also analyzed in four groups (before treatment, Figure 5).The CHOP INTEND value was significantly higher in patients in the 7th to 9th dose group (second year of therapy or 12-24 months after therapy onset) compared to baseline values, in patients in the group from 1st to 3rd dose, and in the group from 4th to 6th dose (Figure 4 and Table 4).
A statistically significant difference was also observed in HFMSE scores between patients with SMA types 2 and 3 who were  from baseline levels in all three SMA types, although in the longer term, this relationship was no longer maintained.The NfL protein showed no significant reduction in CSF independent of the number of nusinersen doses, and the S100B protein CSF levels did not correlate with either the nusinersen doses or the assessed motor function measurement scores.
To date, only five studies analyzed CSF t-tau concentrations after nusinersen treatment in SMA, 25,38,39,43,44 but none of these studies followed SMA patients as long as we did.following treatment, which also correlated well with improvements in motor abilities. 38,39Other studies did not detect significant changes in t-tau concentrations as a result of treatment. 25,43,44These studies mainly included SMA type 2 and 3 patients, while our study and those of Olsson et al. 38  the median age at the beginning of the therapy was 39 years, 25 and SMA 3, the median age at the beginning of the therapy was 12 years, respectively. 43Moreover, in our study, we did not observe a change in t-tau levels upon nusinersen treatment in patients with SMA type 2 and type 3. Therefore, it is likely that an earlier start of the treatment with nusinersen results in a significant decrease in CSF t-tau.
While the role of tau protein changes in mechanisms underlying motoneuron degeneration in SMA remains largely unknown 54 based on results presented in this and other previously mentioned reports, 38,39 it appears that the measurement of t-tau in CSF is a valuable tool for monitoring the response to nusinersen in SMA patients.
9][40][41][42] These outcomes reflect the large variance among patients regarding this potential marker, making it a weak candidate for monitoring response to nusinersen treatment.
The concentration of S100B (together with NfH and NSE) was analyzed in a single study in CSF of 11 patients with SMA type 3 after the administration of nusinersen. 25No changes in the S100B value in the patients' CSF were detected after the administration of nusinersen in that study as well.
A shortcoming of this study is the lack of a control group of patients/respondents.However, lumbar puncture is considered an invasive procedure and is performed on healthy subjects only in exceptional circumstances.Despite this limitation, the conclusions derived from this study retain significance in monitoring nusinersen treatment in patients with SMA.

| CON CLUS ION
Our study confirms that nusinersen therapy is associated with an increase in the functional motor scales' scores.[58][59][60] Even though NfL reduction in CSF had been reported in a few studies of SMA patients on nusinersen, NfL, in general, does not seem to represent a reliable biomarker to monitor the response of SMA patients to nusinersen treatment.Finally, our study shows that t-tau concentration is emerging as a reliable biomarker for monitoring the response of SMA patients to nusinersen treatment in the first 2-year period.

2. 1 |
Subjects, genetic analyses, administration of nusinersen, and CSF sampling All procedures were performed with the approval of the Ethics Committee of the Clinical Hospital Center Zagreb (class: 8.1-20/185-2, no.02/21AG) and by the standards of the Helsinki Declaration and International Standards on Good Clinical Practice (International Conference on Harmonization, 2013). 45This longitudinal study included 30 SMA subjects who were treated at the Department of Pediatrics (Clinic of Pediatrics), Division of Neurology at the University Hospital Centre Zagreb (Table 1; 16 SMA type 1, 4 SMA type 2, and 10 SMA type 3 patients), from 2017 to 2021.Genetic analyses were performed at the Department of Molecular Laboratory Diagnostics of the Clinical Institute for Laboratory Diagnostics, Clinical Hospital Center Zagreb, by multiplex ligation-dependent probe amplification (MLPA).SALSA MLPA kit MRC -P021-B1 SMA, Applied Biosystems capillary sequencer, and the publicly available Coffalyser.Net (MRC-Holland, Amsterdam,
ence in the concentrations of t-tau protein (H test = 9 0.55; df = 3; p = 0.02).T-tau protein levels were significantly lower in the group of all SMA patients (SMA 1, SMA 2, and SMA 3) who received 1-5 doses compared to baseline values, as well as in patients who received 6-10 doses and 11-15 doses compared to baseline values (Figure 1B; Table

F I G U R E 2 T
-tau protein concentrations in (A) all SMA patients, (B) SMA 1 patients, and (C) SMA 2/3 patients grouped based on nusinersen treatment duration.Each patient is represented by a dot on the graph.The red horizontal line represents the mean value, while the black horizontal lines represent the standard deviation.*p < 0.05 F I G U R E 3 (A) NfL and (B) S100B levels in all SMA patients (with SMA 1, SMA 2, and SMA 3) who received different doses of nusinersen.Each patient is represented by a dot on the graph.The red horizontal line represents the mean value, while the black horizontal lines represent the standard deviation.*p < 0.05 administered a different number of nusinersen doses (Figure4).The reason for this was the fact that nusinersen could not be administered intrathecally in some patients due to scoliosis.Compared to the scores attained before the treatment, the most significant differences were seen in patients who received the 8th dose (U = 1; Z = −2.37;p = 0.02), the 6th to 10th dose (U = 32.5;Z = −2.02;p = 0.04), and after 12 months of total treatment duration (U = 1; Z = −2.37;p = 0.02).

E Y W O R D S biomarker, nusinersen, spinal muscular atrophy, tau protein | 3 of 11 ŠIMIĆ et al.
K Demographics, genetics, and treatment characteristics of SMA subjects Concentrations of specific proteins in CSF were determined by enzyme-linked immunosorbent assay (ELISA).CommerciallyTA B L E 1 Time points of CSF samples collection (indicated by +) and the total number of doses received Olsson et al. and Johannsen et al. reported results similar to ours and detected the most signifi- and Johannsen et al. 39 included patients with SMA type 1 as well.Johanssen et al. observed a significant decrease in t-tau levels in patients with SMA type 1 with an average age of 16 months and SMA type 2 with an average age of 78.8 months.
38Olsson et al. observed a significant decrease in t-tau levels after administration of nusinersen in SMA type 1 patients with an average age of 2.2 months.38Studies th not observe a change in t-tau levels mainly included SMA type 2 and type 3 patients who started nusinersen treatment at an older age: SMA type 2 and type 3, the median age at the beginning of the therapy was 41 years,44 SMA 3,