Nusinersen Modulates Proteomics Profiles of Cerebrospinal Fluid in Spinal Muscular Atrophy Type 1 Patients

Abstract

Spinal muscular atrophy (SMA) type 1 is a severe infantile autosomal-recessive neuromuscular disorder caused by a survival motor neuron 1 gene (SMN1) mutation and characterized by progressive muscle weakness. Without supportive care, SMA type 1 is rapidly fatal. The antisense oligonucleotide nusinersen has recently improved the natural course of this disease. Here, we investigated, with a functional proteomic approach, cerebrospinal fluid (CSF) protein profiles from SMA type 1 patients who underwent nusinersen administration to clarify the biochemical response to the treatment and to monitor disease progression based on therapy. Six months after starting treatment (12 mg/5 mL × four doses of loading regimen administered at days 0, 14, 28, and 63), we observed a generalized reversion trend of the CSF protein pattern from our patient cohort to that of control donors. Notably, a marked up-regulation of apolipoprotein A1 and apolipoprotein E and a consistent variation in transthyretin proteoform occurrence were detected. Since these multifunctional proteins are critically active in biomolecular processes aberrant in SMA, i.e., synaptogenesis and neurite growth, neuronal survival and plasticity, inflammation, and oxidative stress control, their nusinersen induced modulation may support SMN improved-expression effects. Hence, these lipoproteins and transthyretin could represent valuable biomarkers to assess patient responsiveness and disease progression.

Keywords: ASO; apolipoprotein A1; apolipoprotein E; carbonyl groups; haptoglobin; neuromuscular disease; nusinersen; oxidized proteins; spinal muscular atrophy type 1; survival motor neuron (SMN); transthyretin.

Figure 1

CSF protein patterns of SMA type 1 patients at baseline and after six months of nusinersen therapy and control subjects. Silver stained reference 2D gels of CSF samples from: SMA type 1 pt. 5 before nusinersen treatment (A), the same SMA type 1 patient after six months of nusinersen therapy (B), and control subject Ctrl 4, who underwent a lumbar puncture for diagnostic reasons (C). Red circles and numbers point out differentially abundant protein spots detected by comparing patients’ and control groups: T0 vs. T1, Ctrl vs. T0, and Ctrl vs. T1. Numbers correspond to those listed in Table 2 and in Supplementary Table S1. The Venn diagram (D) shows the number of the differentially abundant proteins we detected and it allowed an overall evaluation on their distribution among the three different performed comparisons.

Figure 2

Principal Component Analysis. Principal Component Analysis (PCA) performed on relative % Vols of spots matched among T0, T1, and control samples. Graphs highlight spatial distribution of the 21 CSF analyzed samples [14 from SMA type 1 pts 1–7: 7 before (T0) and 7 after (T1) six months of nusinersen treatment; and 7 from control subjects] along the PC1 and PC2 (A) and along the PC2 and PC3 (B). T0 samples are in green, T1 in blue, and Ctrl in red.

Figure 3

Heatmap Analysis. Heatmap analysis performed on matching spots among the three conditions (T0, T1, and Ctrl). Color change blue to red indicating less or higher protein abundance, respectively. Each row corresponds to a protein spot while each column corresponds to individual sample gel. The dendrogram of spot % Vol values (on the left of the heatmap) highlights protein clustering according to protein abundance similarity (three main protein groups were delineated: Ia, Ib and II). The dendrogram of protein profiles (on the top of the heatmap) evidences three main gel clusters: they correspond to T0 (cluster A1), T1 (cluster A2), and Ctrl (cluster B) samples.

Figure 4

REVIGO scatterplot of Biological Process GO-terms annotating the significantly differing proteins identified between T0, T1, and Control samples. REVIGO scatterplot showing semantic similarities among Biological Process GO-terms, from identified differences, clustered in a 2D space by multidimensional scaling. Bubble radius indicates the frequency of individual GO terms in the GO annotation (GOA) referring database (SwissProt): bubbles corresponding to more general terms (higher in hierarchy) are larger than those representing more specific terms (lower in hierarchy). Bubble color indicates the GO annotation uniqueness in the GO list: from brilliant red, for common shared terms, to dark blue for unique individual terms (with lower p-values). Semantically similar GO terms are clustered and outlined by dashed colored lines. Summarized GO terms are listed in Supplementary Table S2.

Figure 5

APOE monodimensional and APOA1 two-dimensional western blot. (A). Immunostained APOE signals on CSF samples from pts not investigated by 2D-PAGE/MS (pts 8–10 = HT0/1, IT0/1, and JT0/1 samples), pts investigated by 2D-PAGE/MS (pts 1, 3, and 4 = AT0/1, CT0/1, and DT0/1 samples), and Ctrl subjects investigated and not investigated by 2D-PAGE/MS (Ctrl 3, 5, and 6; Ctrl 8 and 9, respectively). The histogram shows the mean of normalized relative-integrated-density ± SD values of T0 (baseline, n = 6), T1 (after six months of nusinersen therapy = nusinersen+, n = 6) and Ctrl (Ctrl, n = 5). A significant increase was detected among T0 and T1 samples and among T0 and Ctrl samples. # and § indicate significant abundance changes (p = 0.002) occurring between controls and T0 patients, and between T0 and T1 probands, respectively. Analysed patients and controls are reported on the top of the corresponding wb band. (B). on the left: CSF samples from pt. 8, before (T0) and after (T1) nusinersen therapy, and from Ctrl 8 were resolved by 2D SDS-PAGE and analysed by western blotting using an anti-apolipoprotein A1 monoclonal antibody. The 2DE wb evidenced a substantial increase in APOA1 signals induced by the treatment. Wb signals corresponding to identified APOA1 spot-differences are highlighted by red circles. Numbers corresponded to those reported in Figure 1 and Table 2, as well as in the enlargements of silver stained gels, from ET0, ET1, and Ctrl 4 samples, here shown: (B). on the right. (C). Immunostained APOA1 signals on CSF samples from three SMA type 1 patients (pts 5–7), previously investigated by 2D-PAGE/MS, at T0 (ET0, FT0, and GT0; baseline) and T1 (ET1, FT1, and GT1; after six months of nusinersen therapy = nusinersen+) and from previously investigated control subjects (Ctrls 1 and 2). The histogram shows the mean of normalized relative-integrated-density ± SD values of T0 (n = 3), T1 (n = 3) and Ctrl (n = 2). A significant increase was detected among T0 and T1 samples and among T0 and Ctrl samples. # and § indicate significant abundance changes (p = 0.006) occurring between controls and T0 patients, and between T0 and T1 probands, respectively.

Figure 6

Transthyretin western blot. CSF from pt. 9, before (A) and after (B) nusinersen therapy, and from Cntl9 (C) were resolved by 2D SDS-PAGE and analysed by western blotting using an anti-transthyretin monoclonal antibody. Immunoblot analysis highlighted the general restoring trend of transthyretin isoform pattern to that of the control sample.

Figure 7

Oxidized protein profile of CSF samples from patients 9 and 10, before and after nusinersen therapy, and from two control subjects. Western blot analysis with an anti-DNP antibody on CSF samples from pts 9 and 10, before (IT0 and JT0) and after (IT1 and JT1) nusinersen therapy, and from Ctrls 3 and 10 (Ctrl 3 and Ctrl 10). 2DE immunoblot analysis showed an evident reduction of carbonylated protein species after six months of nusinersen treatment. This signal decrease is mainly evident for protein species having a low-medium Mw (gel area marked by the blue arrows).