Spinal Muscular Atrophy Patient iPSC-Derived Motor Neurons Have Reduced Expression of Proteins Important in Neuronal Development

Abstract

Spinal muscular atrophy (SMA) is an inherited neuromuscular disease primarily characterized by degeneration of spinal motor neurons, and caused by reduced levels of the SMN protein. Previous studies to understand the proteomic consequences of reduced SMN have mostly utilized patient fibroblasts and animal models. We have derived human motor neurons from type I SMA and healthy controls by creating their induced pluripotent stem cells (iPSCs). Quantitative mass spectrometry of these cells revealed increased expression of 63 proteins in control motor neurons compared to respective fibroblasts, whereas 30 proteins were increased in SMA motor neurons vs. their fibroblasts. Notably, UBA1 was significantly decreased in SMA motor neurons, supporting evidence for ubiquitin pathway defects. Subcellular distribution of UBA1 was predominantly cytoplasmic in SMA motor neurons in contrast to nuclear in control motor neurons; suggestive of neurodevelopmental abnormalities. Many of the proteins that were decreased in SMA motor neurons, including beta III-tubulin and UCHL1, were associated with neurodevelopment and differentiation. These neuron-specific consequences of SMN depletion were not evident in fibroblasts, highlighting the importance of iPSC technology. The proteomic profiles identified here provide a useful resource to explore the molecular consequences of reduced SMN in motor neurons, and for the identification of novel biomarker and therapeutic targets for SMA.

Keywords: Human iPSCs; Motor neuron; Neuronal Development; Proteomics; SMA; UBA1; UCHL1.

Figure 1

Characterization of iPSC line and neuronal cultures representative of a healthy control and SMA Type 1 patient iPSC line. (A) Representative positive immunostaining for nuclear and surface pluripotency antigens with normal G-band karyotype of the iPS cells shown at the right. (B) Gene-chip and bioinformatics based PluriTest characterization of control and SMA iPS cell lines used in this study. H9 human embryonic stem cells (hESCs) were used as positive control, while human dermal fibroblasts and primary human neural progenitor cells (hNPCs) were negative controls. (C) Upon neuronal induction and differentiation to the cultures analyzed contain: few Nestin progenitors (< 10%) and Map2 a/b neurons (dendritic marker), pan-neurons marker beta III-tubulin (>60%) with few astroglial (GFAP) cells, mostly SMI32- and ISL1 (Islet-1) positive motor neurons (~40%). Nkx6.1 and ChAT are spinal motor neuron markers that are expressed in both control and SMA-derived motor neurons. Scale bar for A is 75 μm. Scale bar for C is 200 μm. (D) Representative western blot showing SMN protein levels in three different control and SMA motor neuron cell lines, along with Coomassie stained gel as loading control. The graph represents mean integrated density of SMN bands from this blot/total protein (Coomassie gel), as determined by ImageJ software. Error bars represent standard error from the mean and statistical significance was calculated using an unpaired, one-tailed t-test with two-sample unequal variance. Please note that the Coomassie loading control shown here is the same that is shown for Figure 4A because they were both derived from the same blot. (E) Average gene expression levels of full-length (FL)-SMN in the control and SMA motor neurons [the same six cell lines shown in (D)], as determined by qRT-PCR (p = 0.002; unpaired, one-tailed t-test). Relative fold expression was normalized to H9 hESCs. Error bars represent standard error from the mean. ^**p ≤ 0.01.

Figure 2

SMN depletion has contrasting downstream effects on the proteome of motor neurons compared to genetically matched fibroblasts. While 99 proteins were differentially expressed with statistical significance when SMA motor neurons (n = 1) were quantitatively compared to control motor neurons (n = 1) (A), only 18 proteins were differentially expressed when SMA fibroblasts (n = 1) were quantitatively compared to control fibroblasts (n = 1) (B). Of one these 18 proteins, only one was also differentially expressed in the same direction with statistical significance when the SMA MNs were compared to control MNs, and six of the 18 proteins were differentially expressed in one direction (i.e., up- or down-regulated) when SMA fibroblasts were compared to control fibroblasts and then expressed in the opposite direction when SMA MNs were compared to control MNs (C) (arrows indicate proteins that were verified biochemically). Plot (C) was generated using Plotly software (Plotly Technologies Inc. (2015), https://plot.ly).

Figure 3

Perturbation of developmental and differentiation pathways in SMA motor neurons. (A) A Venn diagram and bar chart illustrates the number of differentially expressed proteins seen in control motor neurons (n = 1) compared to control fibroblasts (n = 1) (blue circle and blue bars) and SMN motor neurons (n = 1) compared SMA fibroblasts (n = 1) (green circle and green bars). (B) Bioinformatics analysis of the 46 proteins that were only increased in the control motor neurons vs. control fibs was conducted using the Database for Annotation, Visualization and Integrated Discovery (DAVID). CTR, control; MNs, motor neurons.

Figure 4

Reduction of beta III-tubulin levels in SMA motor neurons. (A) Representative western blot showing beta III-tubulin protein levels in three different control and SMA motor neuron cell lines, along with Coomassie stained gel as loading control. The graph above it illustrates the average integrated density of the beta III-tubulin bands from this blot/total protein (Coomassie gel), as determined by ImageJ software. Error bars represent standard error from the mean and statistical significance was calculated using an unpaired, ced t-test with two-sample unequal variance. Please note that the Coomassie loading control shown here is the same that is shown for Figure 1D because they were both derived from the same blot. (B) Representative confocal images indicate a reduction of beta III-tubulin levels in SMI-32 positive 32i SMA motor neurons. CTR, control; MNs, motor neurons. ^*p ≤ 0.05.

Figure 5

Decreased UCHL1 levels in SMA motor neurons. (A) Representative western blot showing UCHL1 protein levels in three different control and SMA motor neuron cell lines. The graph above it shows the integrated density of the UCHL1 bands from this blot/total protein (Coomassie gel), as determined by ImageJ software. The graph to the right shows the average integrated density of the UCHL1 bands from this blot/total protein (Coomassie gel). Error bars represent standard error from the mean and statistical significance was calculated using an unpaired, one-tailed t-test with two-sample unequal variance. (B) Representative confocal images indicate a reduction of UCHL1 levels in SMI32 positive cells. (C) Average gene expression levels of UCHL1 in the control and SMA motor neuron cell lines [the same six cell lines shown in (A)], as determined by qRT-PCR (p = 0.04; unpaired, one-tailed t-test). Relative fold expression was normalized to H9 hESCs. Error bars represent standard error from the mean. CTR, control, MNs, motor neurons. ^*p ≤ 0.05.

Figure 6

Reduction and differential localization of UBA1 in SMA motor neurons. (A) Representative western blot showing UBA1 protein levels in three different control and SMA motor neuron cell lines, along with Coomassie stained gel as loading control. The graph illustrates the average integrated density of the UBA1 bands from this blot/total protein (Coomassie gel), as determined by ImageJ software. Error bars represent standard error from the mean and statistical significance was calculated using an unpaired, one-tailed t-test with two-sample unequal variance. (B) Representative confocal images indicate a reduction of UBA1 levels in SMI32 positive 32i SMA motor neurons and mostly cytoplasmic distribution (in comparison to the mostly nuclear distribution seen in the 14i control (CTR) motor neurons). (C) Average gene expression levels of UBA1 transcript variants 1 (TV1) (p = 0.14; unpaired, one-tailed t-test) and 2 (TV2) (p = 0.08; unpaired, one-tailed t-test) in the control and SMA motor neuron cell lines [the same six patient and control cell lines shown in (A)], as determined by qRT-PCR. Relative fold expression was normalized to H9 hESCs. Error bars represent standard error from the mean. CTR, control; MNs, motor neurons. ^**p ≤ 0.01.