Cytokines in cerebrospinal fluid as a prognostic predictor after treatment of nusinersen in SMA patients

Objectives: Recent studies have suggested that neuroinflammation may play a role in the progression of spinal muscular atrophy (SMA), and this may influence the efficacy of antisense oligonucleotide treatment. This study explored the biomarkers associated with SMA and the efficacy of nusinersen therapy. Methods: Fifteen patients with SMA were enrolled and their motor function (World Health Organization motor milestone, Hammersmith Functional Motor Scale Expanded (HFMSE), and Revised Upper Limb Module [RULM] scores, and 6-minute walking test) was evaluated before, during (63 days), and after (6 months) nusinersen treatment. The concentrations of monocyte chemoactive protein 1 (MCP1), tumour necrosis factor-alpha (TNF-α ), and interleukin (IL)- 10 in the cerebrospinal fluid were measured at the indicated time points, and their correlations with motor function were analysed. Results: A significant increase in MCP1 was observed after 6 month ’ s treatment compared with that before treatment, while TNF-α gradually decreased over the course of treatment. IL-10 levels were negatively correlated with HFMSE scores before treatment, and reductions in IL-10 levels were correlated with improvements in RULM scores. Conclusions: This study suggests that neuroinflammation may be associated with the severity of SMA and with the therapeutic effects of nusinersen, which could have clinical implications in the treatment of SMA.


Introduction
Classic chromosome 5q spinal muscular atrophy (SMA) is a rare genetic neurodegenerative disorder caused by mutations in the survival of motor neuron 1 (SMN1) gene [1].Over the past decade, the advent of genetic molecular therapies (onasemnogene abeparvovec, nusinersen and risdiplam) has greatly improved patient outcomes [2][3][4].Nusinersen is an antisense oligonucleotide (ASO) that specifically binds SMN2 and modulates its transcription to produce full-length, functional SMN proteins [5].Nusinersen is delivered into the cerebrospinal fluid (CSF) via intrathecal injection.Since the approval of nusinersen in China in 2019 for the treatment of 5q SMA, it has improved the prognosis of Chinese patients with SMA [6].Real-world data from different countries and disease registries have also shown overall improvement in motor function and quality of life in patients with SMA treated with nusinersen [7].However, for individual patients, the degree of improvement varied greatly [7,8].Some patients did not respond to nusinersen treatment, suggesting that there may be mechanisms underlying differential responses to nusinersen.
SMA is a multisystem disease characterized by neuroinflammation.The involvement of neuroinflammation in the pathogenesis of SMA has been demonstrated in animal and patient studies [9,10].However, the role of neuroinflammation in the response of patients with SMA to gene modification therapy remains unclear.Ando  SMN-ASO injection inhibited the activation of spinal microglia and reduces the expression of oxidative stress markers in the microglia of SMA model mice [11].
Microglia are the immune cells of the central nervous system.They closely interact with inflammatory T lymphocytes and astrocytes to regulate inflammatory responses.Microglia secrete different combinations of pro-inflammatory and anti-inflammatory cytokines and chemokines depending on the stimulation of the surrounding microenvironment and other cells [12].Several studies have shown that an increase in microglia leads to spinal cord inflammation in SMA model mice [13].
A recent study found that the levels of some pro-inflammatory cytokines in the CSF of patients with SMA1 were significantly reduced after nusinersen treatment [14].In another prospective multicenter observational study, chitotriosidase 1, which is expressed by polymorphonuclear neutrophils and activated macrophages, was elevated in the CSF of patients with SMA during treatment with nusinersen [15].
Based on existing evidence, we hypothesized that neuroinflammatory cytokines could serve as biomarkers to predict the progression of SMA and the efficacy of gene modification treatment.The aim of this study was to identify biomarkers in the CSF of patients with SMA associated with the efficacy of nusinersen therapy.

Patients and procedures
This single-center prospective observational study was conducted at Nanjing Medical University, Nanjing, China between September 2021 and May 2022.The inclusion criterion for patients to be enrolled in this study was genetically confirmed 5q SMA type 2 or 3 patients.None of the patients had received disease-modifying therapy before and were treated with nusinersen for the first time.A dose of 12 mg nusinersen was administered intrathecally on days 1 (baseline), 14, 28, and 63, with maintenance injections administered every 4 months thereafter.Demographics, clinical characteristics, and laboratory test results were also collected at baseline.Motor function and biomarkers in the CSF samples were evaluated at baseline (pretreatment), 63 days, and 6 months after nusinersen administration.
This study was approved by the Ethical Review Board of the First Affiliated Hospital of Nanjing Medical University.All patients provided informed consent before participation in the study.

Motor function
Motor function was assessed using the World Health Organization (WHO) motor milestones, Hammersmith Functional Rating Scale Expanded (HFMSE), Revised Upper Limb Module (RULM), the Children's Hospital of Philadelphia Infant Test of Neuromuscular Disorders (CHOP INTEND), and the 6-minute walk test (6MWT).
The WHO standard assessment of milestone achievement in gross motor development evaluates unsupported sitting, helpful standing, hand and knee crawling, helpful walking, standing alone, and walking alone [16].
The HFMSE consists of 33 items, with higher scores indicating better motor function.The maximum total score is 66 points, with a change of at least 3 points considered a clinically meaningful improvement.The RULM has 20 items with a maximum score of 37 points, with higher scores indicating better arm function.An improvement of at least 2 points in the RULM was considered clinically meaningful [17].The CHOP INTEND scale was used to assess infant motor function; it consists of 16 items ranging from 0 to 64 points with lower scores indicating reduced motor function.The 6MWT measures the distance that a patient can walk within six minutes.

Laboratory testing
CSF samples were collected at the lumbar puncture sites and immediately stored at − 80 • C.After thawing, the samples were centrifuged at 14000×g for 4 minutes at room temperature.The multiplex Ella™ platform and Ella reagent kit (Protein Simple, CA, USA) were used to measure the concentrations of the following cytokines: monocyte chemoactive protein 1 (MCP1), tumor necrosis factor-alpha (TNF-α) and interleukin-10 (IL-10).The Ella platform was used to conduct immunoassays using a microfluidic cartridge.The samples were inserted into the cartridge wells and passed through a microfluidic channel with three glass nanoreactors that were coated with a capture antibody.Relative fluorescence units were used to calculate concentrations based on manufacturer-generated and calibrated standard curves.The concentrations of cytokines were expressed as pg/mL.

Statistical analysis
Statistical analyses were performed using SPSS software (version 26.0;IBM Corporation, Armonk, NY, USA) and GraphPad Prism 9.0 software (San Diego, CA, USA).A two-sided p value of <0.05 was considered statistically significant.The Wilcoxon matched-pairs signedrank test was used to analyze changes in motor function outcomes from baseline to day 63 and month 6.Associations between CSF biomarker levels and motor function scale scores were analyzed using Spearman's rank correlation.

Baseline characteristics
The study enrolled 15 patients, including nine with SMA type 2 and six with SMA type 3. Demographic and clinical characteristics at baseline are summarised in Table 1.The median age of patients with SMA type 2 was 4 years (range:1-12years), whereas that of patients with SMA type 3 was 19.5 years (range: 7-32 years).Among them, 33.3 % of the patients with SMA type 3 were able to walk, whereas none of the patients with SMA type 2 were able to walk.

Effects of nusinersen on motor function
The changes in motor function test scores compared to baseline are summarized in Table 2.The median WHO motor milestone scores was sustained at 8 before and after nusinersen treatment.Five (33.3 %)

Inflammatory cytokines levels were correlated with motor functional in patients with SMA
In all patients, higher IL-10 levels were significantly correlated with lower HFMSE scores before nusinersen treatment (p = 0.033, r = − 0.571; Fig. 3A).A significant correlation was observed in patients with SMA type 3 (p = 0.045; Fig. 3D) but not in those with SMA type 2 (p = 0.223; Fig. 3G).Baseline IL-10 levels did not correlate with RULM scores (Figs.3B, 3E, and 3H).However, a decrease in IL-10 significantly correlated with an improvement in the RULM scores at 6 months (p = 0.007, r = − 0.788; Fig. 3C), and a significant correlation was observed in both the SMA type 3 and SMA type 2 subgroups (Figs.3F and 3I).

Discussion
In this study, half or more of the patients with SMA type 2 or 3 achieved clinically meaningful improvements in motor function after nusinersen treatment.In our cohort study, improvements in HFMSE and RULM scores appear to be more pronounced following nusinersen treatment when compared to certain clinical studies [18,19].
A multicenter observational study found that the HFMSE scores of patients with SMA type 2 improved by an average of 0.6 points from baseline, while those of patients with SMA type 3 improved by an average of 2.4 points at 6 months of nusinersen treatment; the RULM scores improved by an average of 1.1 and 0.4 points in patients with SMA type 2 and SMA type 3, respectively, at 6 months [20].Nonetheless, our results align with a number of recent studies that have indicated a 3-point rise in HFMSE score after just 2 months from the baseline [21], along with another study reporting a mean score increase of 5.3 points over a 38-month observation period [22].It is speculated that the reason for this could be due to the promotion of Multi-Disciplinary Treatment (MDT) leading to most patients incorporating rehabilitation functional exercises alongside drug treatment.Besides, patients with types 2 and 3 in our cohort had a higher degree of motor dysfunction than adult-type SMA, and the ceiling effect was absent in most patients, which may also be one of the reasons for the more significant treatment effect.Certainly, we would incorporate a greater number of patients with extended follow-up in order to evaluate the lasting impacts of enhanced motor function.
Neuroinflammation is involved in the development of many neurodegenerative diseases [23,24]; however, its role in the SMA has not been confirmed.We analysed the association of inflammatory cytokines in the cerebrospinal fluid of patients with SMA at baseline with SMA type and SMN2 copy number.We grouped patients by SMN2 copy number (2, 3, and 4), and found no significant differences in IL-10 and TNF levels at baseline between the groups.There were significant differences in MCP1 concentrations between the groups.The concentration of MCP1 was significantly higher in patients with 3 copies of SMN2 than in those with 2 copies of SMN2.However, the concentration of MCP1 in patients with 4 copies of SMN2 decreased slightly, which may be due to the small number of patients in this category.
In addition to motor neurons, the pathogenesis of SMA involves glia.Reactive gliosis has been observed in the spinal cords of patients with SMA type 1, and reactive astrocyte proliferation was detected in a SMA mouse model; restoration of SMN protein levels in astrocytes decreased the release of proinflammatory cytokines and alleviated disease progression [25] [26].We used MCP1, a neuroprotective chemokine secreted by astrocytes, as a candidate marker of spinal cord astrocyte activation in patients and found that, compared with baseline, the level of MCP1 was significantly upregulated in the CSF of patients with SMA after 6 months of treatment.Consistent with the results of our study, Scheijmans et al. showed that the MCP1 was increased in the CSF of patients with SMA during treatment with nusinersen [9].Martin et al. showed that the addition of MCP1 increased neuron length in SMA-patient-derived induced pluripotent stem cells, suggesting that restoration of MCP1 might improve motor neuron growth in patients with SMA [27].Consistent with this hypothesis, we found that MCP1 levels increased after nusinersen treatment, indicating that this chemokine may be involved in the neuroprotective mechanism of nusinersen.
We found a significant decrease in TNF-α in the CSF of patients with SMA at both 63 days and 6 months of nusinersen treatment relative to baseline.Signaling with the proinflammatory cytokine TNF-α signaling was significantly enhanced in SMN-deleted mice [11].Nuzzo et al. found that the concentrations of TNF-α in the CSF of patients with SMA type 1 were higher than in patients with SMA types 2 and 3, and this cytokine was significantly decreased at 10 months of nusinersen treatment relative to baseline in patients with SMA type 1 [14].Yang et al. conducted a comprehensive transcriptome analysis and found that, in the CSF of mice with severe SMA, the expression of TNF-α was upregulated, but this was not upregulated in mice with pre-symptom and mild-symptom SMA [28].Furthermore, in a mouse model of amyotrophic lateral sclerosis, motor neuron degeneration could be prevented by intraspinal injection of TNF-α-neutralizing antibodies, which also highlights the important role of TNF-α in the development of neurodegenerative disease [29].
To further determine whether cytokine levels were associated with clinical responses to treatment, we conducted a regression analysis of baseline and post-treatment CSF cytokine levels and motor function scores.We found a negative correlation between CSF IL-10 concentration and motor function before treatment.A significant negative correlation was also found between the reduction in IL-10 levels and the improvement in motor function after nusinersen treatment.In consistent with our finding, Bonanno et al. showed that patients with higher posttreatment serum IL-10 concentrations had better HFMSE scores after 6 months of nusinersen treatment, but no such correlation was detected for IL-10 in the CSF [30].A possible explanation for this inconsistency is that there is a weak, or no, relationship between the concentrations of cytokines in the CSF and blood.Indeed, it was found that the concentrations of various cytokines in the CSF were correlated with each other, but there was little correlation between CSF and serum values [31].Under neuroinflammatory conditions, various CSF cytokines are likely produced by glial cells in the central nervous system.IL-10, as an anti-inflammatory cytokine, is produced by different cell types such as T regulatory cells (Tregs) and T helper 2 cells.T helper 1 cells limit the proinflammatory activity by producing IL-10, while Tregs release cytokines such as IL-10 and exert immunosuppressive effects on other cells [32].In addition, we revealed a trend of an increase in IL-10 levels in the CSF, although the difference was not signifucant, indicating that IL-10 is a potential pharmacodynamic biomarker worthy of further investigation.
The main limitations of our study include the small sample size as well as the absence of the lack of an untreated control group.Similar to some studies on cerebrospinal fluid markers in SMA that lack a control group [33,34], it can be hard to collect CSF samples from SMA patients who have not been treated with nusinersen.Considering the lack of an untreated control group, our analysis of CSF data in comparison to the cytokine reference range reported by Pranzatelli et al. [35] for children with non-inflammatory neurological diseases indicates a decrease in MCP1 and IL-10 levels, and an increase in TNF-α levels in our patients with SMA (Supplementary Table 1).It remains to be discussed whether the changes in cytokines are caused by the progression of the disease itself in our study.Comparison of subgroups of patients showed an increasing trend of IL-10 concentrations in the CSF between 2,3 and 4copies of SMN2, but this was not significant.Thus, studies on the correlation between cytokines in cerebrospinal fluid and improved motor function are purely exploratory.Further studies with larger sample sizes and longer observation periods are required to identify the interaction between cytokines and prognosis in patients with SMA during nusinersen treatment.
The highlight of the present investigation is that it was a real-world study of a rare diseases.The levels of pro-inflammatory cytokines decreased, while those of protective factors that inhibit inflammation increased during nusinersen treatment in the CSF of patients with SMA.The cytokines in the CSF identified in this study may serve as potential biomarkers for future investigations to monitor disease progression and treatment efficacy.

Table 1
Demographics and clinical characteristics of 15 patients with SMA.
SMA, spinal muscular atrophy; SMN, survival motor neuron.X.Cheng et al.patients gained a new motor milestone, whereas only one (6.7 %) patient had reduced motor milestones after 6 months of nusinersen treatment.Among the 14 patients who completed the HFMSE test, HFMSE scores significantly improved by a mean (standard deviation) of 4.43 (4.01) points at 63 days (p = 0.003) and 5.21 (4.59) points at 6 months from baseline (p = 0.004).Clinically meaningful improvements in HFMSE scores were observed in nine (64.3 %) patients at 63 days and ten (71.4 %) patients at 6 months compared with baseline.Among the 13 patients who completed the RULM, the median RULM scores showed an increasing trend during nusinersen treatment.RULM scores were also significantly increased compared with baseline at 63 days (mean change, 3.77±2.98;p = 0.005) and 6 months (2.85 ± 2.79; p = 0.007) of nusinersen treatment.Clinically meaningful improvements in RULM scores were observed in 11 (84.6 %) patients at 63 days and in eight (61.5 %) patients at 6 months.The 6MWT was performed in two ambulant patients with SMA type 3.One patient showed a decline in

Table 2
Changes in motor function test scores vs. baseline.Hospital of Philadelphia Infant Test of Neuromuscular Disorders; HFMSE, Hammersmith Functional Motor Scale Expanded; RULM, Revised Upper Limb Module; SD, standard deviation; WHO, World Health Organization.a CHOP INTEND score for one patient and 6WMT data for two patients are shown.