Analysis of proteome-wide degradation dynamics in ALS SOD1 iPSC-derived patient neurons reveals disrupted VCP homeostasis

Abstract

Mutations in SOD1 cause amyotrophic lateral sclerosis (ALS) through gain-of-function effects, yet the mechanisms by which misfolded mutant SOD1 (mutSOD1) protein impairs human motor neurons (MNs) remain unclear. Here, we use induced-pluripotent-stem-cell-derived MNs coupled to metabolic stable isotope labeling and mass spectrometry to investigate proteome-wide degradation dynamics. We find several proteins, including the ALS-causal valosin-containing protein (VCP), which predominantly acts in proteasome degradation and autophagy, that degrade slower in mutSOD1 relative to isogenic control MNs. The interactome of VCP is altered in mutSOD1 MNs in vitro, while VCP selectively accumulates in the affected motor cortex of ALS-SOD1 patients. Overexpression of VCP rescues mutSOD1 toxicity in MNs in vitro and in a C. elegans model in vivo, in part due to its ability to modulate the degradation of insoluble mutSOD1. Our results demonstrate that VCP contributes to mutSOD1-dependent degeneration, link two distinct ALS-causal genes, and highlight selective protein degradation impairment in ALS pathophysiology.

Keywords: ALS; CP: Neuroscience; CP: Stem cell research; SILAC-based mass spectrometry; SOD1; VCP/p97; amyotrophic lateral sclerosis; iPSCs; motor neurons; protein degradation; ubiquitin.

Figure 1.. Patient neurons exhibit high levels of disordered soluble and insoluble SOD1 protein without impact on the major clearance pathways

(A) Experimental schematic of MN generation and biochemical fractionation for ELISA and western blot (WB) analysis of SOD1 protein. (B) Representative images of iPSC-differentiated MNs expressing ISL1/2 and TUJ1. Scale bar, 20 μm. (C) Quantification of soluble disordered SOD1 by ELISA (n = 5 independent biological replicates, two-way ANOVA across time and genotypes, p = 0.4254; Sidak’s multiple comparisons test per time point, day 16 **p = 0.0068, day 30 **p = 0.0029, day 35 *p = 0.0215, day 51 ***p = 0.0003). (D and E) WB analysis (D) and quantification (E) of detergent-insoluble SOD1 (n = 6 independent biological replicates, two-way ANOVA across time and genotypes, *p = 0.046; Sidak’s multiple comparisons test per time point, day 16 p = 0.8352, day 25 *p = 0.0120, day 30 ***p = 0.0005, day 35 ****p < 0.0001, day 40 ****p < 0.0001, day 51 **p = 0.0022). The detergent-insoluble SOD1 is expressed relative to the normalized soluble SOD1. (F) Assessment of the ubiquitination flux in mutSOD1 and isogenic control MNs by WB. The proteasome activity was blocked with MG132 (10 μM, 8 h), and polyubiquitinated proteins were isolated from total cell lysate extracts using TUBE magnetic beads. Alternatively, the autophagosome-lysosome fusion was blocked with the inhibitor bafilomycin A1 (20 nM, 24 h), and the polyubiquitinated proteins were isolated in the same way. Representative blots (top) and quantification (bottom). Unpaired t test, for MG132 treatment, n = 6 independent biological replicates, p = 0.236; for BAFA1 treatment, n = 3 independent biological replicates, p = 0.240; ns, not significant.

Figure 2.. Mutant SOD1 iPSC-derived patient neurons exhibit slower protein degradation dynamics of a small panel of proteins including VCP/p97

(A) Experimental schematic of SILAC-MS pulse-chase strategy for iPSC-MNs. (B) Assessment of labeling efficiency on day-30 MNs based on the quantification of heavy labeled peptides (lysine/arginine amino acids) over total peptides. The number of detected peptides is shown on the y axis and the percentage of heavy labeled peptides over total peptides (heavy and light) on the x axis. MN cultures with mutSOD1 are shown in red and isogenic controls in green. (C) Quantification of the total number of labeled and unlabeled peptides in both mutSOD1 and isogenic control MN cultures shown over time in culture. (D) Normalized mean peptide intensity of all labeled proteins detected in both mutSOD1 and isogenic control MNs within each one of the three time points interrogated. All values are normalized to the respective values on day 30. Each dot represents a single protein, and average and standard deviation values are shown for each time point. On day 31 (24 h after “chase”) there were 39 common proteins with a mean intensity of 0.62 and 0.42 in mutSOD1 and isogenic control MNs, respectively. On day 35 (5 days after “chase”) there were 87 common proteins with a mean intensity of 0.44 and 0.38 in mutSOD1 and isogenic control MNs, respectively. On day 51 (21 days after “chase”) there were 129 common proteins with a mean intensity of 0.20 and 0.15 in mutSOD1 and isogenic control MNs, respectively. Paired t test (two-tailed), day 31 ****p < 0.0001, day 35 **p = 0.0053, day 51 **p = 0.0012. (E) Venn diagram of the number of proteins that are more labeled (i.e., persist) in SOD1^+/A4V MN cultures relative to isogenic controls across all three time points (days 31, 35, and 51). The eight proteins that are more labeled across all the three time points examined are highlighted. (F) Over-representation analysis (PANTHER database) of all persisting proteins shown in (E). Based on their categorization, enriched proteins are represented in black (Biological Process) or blue (Protein Class) circles. For all the over-represented proteins, false discovery rate (FDR) < 0.05. (G) Analysis of the labeled VCP protein at the level of labeled peptides. Each dot represents a single labeled peptide on day 35, with a value normalized to the respective value of the same peptide on day 30. Unpaired t test (two-tailed), ***p = 0.0001. All data from (A) to (G) represent n = 2 independent differentiation and labeling experiments for both genotypes. (H) Schematic of the SILAC-IP LC-MS/MS experimental approach. (I) Immunoprecipitation (IP) of VPC from whole-cell lysate extracts (600 μg/reaction) derived from mutSOD1 and isogenic control MNs. IgG of the same isotype with the anti-VCP antibody was used as a negative control of the reaction. Input: 3.3%. (J) Average intensity of heavy VCP peptides that are enriched in both mutSOD1 and isogenic control MN cultures on day 35 upon immunoprecipitation. The comparison is made between identical (common) VCP peptides in both genotypes. Paired t test, Wilcoxon correction, n = 3 independent differentiations; experiment 1, **p = 0.0024; experiment 2, ****p < 0.0001; experiment 3, ****p < 0.0001. (K) Total VCP protein levels under baseline conditions on day 35 in patient and isogenic control MNs. Unpaired t test (two-tailed), n = 3 independent differentiations, p = 0.228; ns, not significant.

Figure 3.. The SOD1 A4V mutation is sufficient to alter VCP turnover

(A) Experimental schematic of the HUES3 stem cell editing and MN-directed differentiation. (B) Representative images of stem-cell-differentiated MNs expressing ISL1/2 and TUJ1 on day 25. Scale bar, 10 μm. (C and D) WB analysis (C) and quantification (D) of the detergent-insoluble SOD1 levels in HUES3-SOD1^+/+ or HUES3-SOD1^+/A4V MNs across time. The detergent-insoluble SOD1 (top blot) is expressed relative to the normalized soluble SOD1 (bottom blot). Two-way ANOVA across time and genotypes, ****p < 0.0001; Sidak’s multiple comparisons test per time point, day 16 p = 0.9571, day 25 ***p = 0.0008, day 30 **p = 0.0033, day 35 **p = 0.0037, day 40 ****p < 0.0001, day 51 ****p < 0.0001; ns, not significant; n = 3 independent differentiations. (E) Schematic representation of the SILAC-IP LC-MS/MS approach. (F) Average intensity of heavy VCP peptides that are enriched in both HUES3-SOD1^+/+ and HUES3-SOD1^+/A4V MN cultures on day 51 upon immunoprecipitation. The comparison is done between identical, common VCP peptides in both genotypes. Paired t test, Wilcoxon correction; n = 2 independent differentiations; experiment 1, ****p < 0.0001; experiment 2, **p = 0.009.

Figure 4.. Accumulated VCP in postmortem tissue of an SOD1^+/A4V ALS patient

(A and B) Immunohistochemistry of VCP and SOD1 in postmortem motor (affected) (A) and occipital (unaffected) (B) cortex from an ALS patient (patient #1-JHU74) carrying the A4V mutation in the SOD1 gene. The pictures on the right column represent the magnified regions within the yellow dashed squares. Scale bar, 20 μm. (C) Quantification of SOD1 and VCP intensities in MAP2⁺ neurons within the motor or occipital cortex. Motor cortex, n = 17 neurons; occipital cortex, n = 16 neurons. Unpaired t test (two-tailed), SOD1 p < 0.0001; VCP p < 0.0001.

Figure 5.. The interactome of VCP alters in patient MNs

(A) Schematic representation of the crosslink/IP/MS experimental workflow. The MNs on day 35 were crosslinked with the reversible crosslinker DSP (1 mM) for 20 min before they were collected. The cell extracts were subjected to immunoprecipitation (IP) using a specific antibody against VCP, and the precipitates were further analyzed by LC-MS/MS. (B and C) Venn diagrams showing the distribution of VCP co-precipitated proteins between the patient or isogenic control MNs (B) and their overlap with the reported (known from the literature or predicted) VCP interactors in the BioGrid database (C). (D) Volcano plot of the 399 shared VCP interactors between the genotypes. The red dots on the right and the green dots on the left represent VCP interactors that are enriched in either of the two genotypes. The red and green columns represent some of the exclusive VCP interactors in patient or isogenic control MNs, respectively. (E) Subcellular localization of some of the exclusive or enriched VCP interactors in either SOD1^+/A4V (red) or SOD1^+/+ (green) MNs. (F) Validation by IP/WB of HSPB1 as an enriched interactor of VCP in isogenic control MNs. The levels of co-precipitated HSPB1 were normalized to the respective level of precipitated VCP. (G) GO analysis (WebGestalt web tool) of the over-represented groups of the VCP interactome in either SOD1^+/A4V (red) or SOD1^+/+ (green) MNs. FDR < 0.05. (H) Schematic representation of our working model. Mutant SOD1 protein progressively becomes more insoluble than WT protein (red and green panels, respectively) and is associated with the slower turnover of VCP, which exhibits both gain- and loss-of-function interactions in mutSOD1 MNs.

Figure 6.. VCP impacts the solubility of SOD1

(A) Experimental workflow of HEK293T transfection and treatment. HEK293T cells were transfected with SOD1WT-MYC or SOD1A4V-MYC plasmids in combination with VCP-RFP or RFP plasmids. The cells were incubated up to 48 h and treated with the allosteric VCP inhibitor NMS873 (10 μM, 8 h). The cells were lysed and subjected to fractionation for biochemical analysis. (B) WB analysis of HEK293T-transfected cells. Lanes 1–5 correspond to DMSO-treated cells and lanes 6–10 to NMS873-treated cells. The bar graph corresponds to the levels of the detergent-insoluble SOD1-MYC (insoluble fraction, bottom blot). The detergent-insoluble SOD1-MYC is normalized to the soluble SOD1-MYC levels (soluble fraction, top blot); n = 2 independent transfections. (C) Schematic representation of patient or isogenic control MNs treated with NMS873 (10 μM, 8 h) on day 35 and lysed for biochemical analysis. (D and E) WB analysis and quantification of whole-cell extracts from DMSO- or NMS873-treated MNs and quantification of poly-Ub (D) or SOD1 (E) protein levels. For poly-Ub, two-way ANOVA (treatment × genotype), *p = 0.0152; Sidak’s multiple comparisons test per treatment, SOD1^+/A4V MN ****p < 0.0001, SOD1^+/+ MN p = 0.396. For SOD1 two-way ANOVA (treatment × genotype), p = 0.2170; Sidak’s multiple comparisons test per treatment, SOD1^+/A4V MN **p = 0.0099, SOD1^+/+ MN p = 0.4156; ns, not significant; n = 3 biological replicates. (F) WB analysis of detergent-insoluble SOD1 levels in SOD1^+/A4V or SOD1^+/+ MNs upon treatment with NMS873. Unpaired t test (two-tailed), SOD1^+/A4V MN *p = 0.0105, SOD1^+/+ MN p = 0.8696; ns, not significant; n = 3 biological replicates. (G) Impact of VCP on SOD1 accumulation. When VCP is overexpressed (left) in the context of mutSOD1, the levels of detergent-insoluble SOD1 protein are decreased. In contrast, upon chemical inhibition of VCP with NMS873 (right), there is significant accumulation of SOD1 protein within the detergent-insoluble fraction.

Figure 7.. VCP ameliorates mutSOD1 toxicity in iPSC-MNs in vitro and C. elegans models in vivo

(A) Experimental schematic of MN viability assay. MNs were differentiated from mutSOD1 and isogenic control iPSCs and infected with either LV-VCP-T2A-RFP or LV-RFP to monitor progressive degeneration by longitudinal time-lapse imaging microscopy. (B) Probability of survival of mutSOD1 and isogenic control iMNs upon VCP-RFP or RFP expression across 3 weeks in culture; n = 1 differentiation and 3 technical replicates; for SOD1^+/A4V iMN LV-VCP-T2A-RFP n = 220, LV-RFP n = 197; for SOD1^+/+ iMN LV-VCP-T2A-RFP n = 200, LV-RFP n = 198; Gehan-Breslow-Wilcoxon test. (C) Experimental schematic of genetic interaction experiments with C. elegans strains expressing human mutSOD1 protein (IW8), with overexpression (o-e) and knockout (ko) of the VCP ortholog cdc48.1. (D) Quantification of the average speed per second of C. elegans strains examined. One-way ANOVA for genotype, p < 0.0001; Unpaired t test to compare individual genotypes: (1) WT vs. mutSOD1 *p = 0.0206, (2) WT vs. ko VCP **p = 0.0042, (3) WT vs. o-e VCP p = 0.2096, (4) WT vs. mutSOD1; ko VCP **p = 0.0013, (5) WT vs. mutSOD1; o-e VCP p = 0.7650, (6) mutSOD1; o-e VCP vs. mutSOD1; ko VCP **p = 0.0023, (7) mutSOD1; o-e VCP vs. o-e VCP p = 0.1325, (8) mutSOD1; o-e VCP vs. ko VCP **p = 0.0044, (9) mutSOD1; o-e VCP vs. mutSOD1 *p = 0.0154; n = 3–5 independent experiments, n = 5–10 worms per experiment. (E) Quantification of SOD1 protein within the insoluble and soluble fractions in mutSOD1 worms and mutSOD1 overexpressing VCP. The amount of detergent-insoluble SOD1 is expressed relative to the amount of normalized soluble SOD1. Unpaired t test (two-tailed), *p = 0.0123; n = 4 independent preparations of worm cultures.