Extracellular Vesicles in Serum and Central Nervous System Tissues Contain microRNA Signatures in Sporadic Amyotrophic Lateral Sclerosis

Abstract

Amyotrophic lateral sclerosis (ALS) is a terminalneurodegenerative disease. Clinical and molecular observations suggest that ALS pathology originates at a single site and spreads in an organized and prion-like manner, possibly driven by extracellular vesicles. Extracellular vesicles (EVs) transfer cargo molecules associated with ALS pathogenesis, such as misfolded and aggregated proteins and dysregulated microRNAs (miRNAs). However, it is poorly understood whether altered levels of circulating extracellular vesicles or their cargo components reflect pathological signatures of the disease. In this study, we used immuno-affinity-based microfluidic technology, electron microscopy, and NanoString miRNA profiling to isolate and characterize extracellular vesicles and their miRNA cargo from frontal cortex, spinal cord, and serum of sporadic ALS (n = 15) and healthy control (n = 16) participants. We found larger extracellular vesicles in ALS spinal cord versus controls and smaller sized vesicles in ALS serum. However, there were no changes in the number of extracellular vesicles between cases and controls across any tissues. Characterization of extracellular vesicle-derived miRNA cargo in ALS compared to controls identified significantly altered miRNA levels in all tissues; miRNAs were reduced in ALS frontal cortex and spinal cord and increased in serum. Two miRNAs were dysregulated in all three tissues: miR-342-3p was increased in ALS, and miR-1254 was reduced in ALS. Additional miRNAs overlapping across two tissues included miR-587, miR-298, miR-4443, and miR-450a-2-3p. Predicted targets and pathways associated with the dysregulated miRNAs across the ALS tissues were associated with common biological pathways altered in neurodegeneration, including axon guidance and long-term potentiation. A predicted target of one identified miRNA (N-deacetylase and N-sulfotransferase 4; NDST4) was likewise dysregulated in an in vitro model of ALS, verifying potential biological relevance. Together, these findings demonstrate that circulating extracellular vesicle miRNA cargo mirror those of the central nervous system disease state in ALS, and thereby offer insight into possible pathogenic factors and diagnostic opportunities.

Keywords: amyotrophic lateral sclerosis; biomarker; central nervous system; extracellular vesicle; microRNA; neurodegeneration; pathway analysis; serum.

FIGURE 1

Immuno-affinity-based microfluidic EV isolation from ALS tissue. (A) Workflow for immobilizing and characterizing EVs. (1) EVs were captured from frontal cortex, spinal cord, and serum from ALS or control participants by a (2) CD63-antibody-coated ExoChip. (3) To quantify the captured EVs, the chip was processed with lipophilic (DiO) staining, and measured fluorescence intensity was normalized to background and correlated to the amount of fluorescently stained EVs. (4) Immobilized EVs were lysed for Western blotting and cargo profiling by NanoString. (B) Verification of the saturation capacity of the ExoChip device for serum and tissue homogenates. For frontal cortex and spinal cord, testing of 300 μL of homogenates with various total protein concentrations revealed a saturation concentration of 0.5 μg/μL for frontal cortex and 1 μg/μL for spinal cord tissue. For serum, the device reaches maximum fluorescence intensity when the volume of serum equals 400 μL. Data are presented as mean ± standard error of the mean (s.e.m.) (n = 3).

FIGURE 2

EV characterization from frontal cortex, spinal cord, and serum of ALS and control participants. EVs were captured by ExoChip. (A) Protein characterization from frontal cortex and serum EVs captured from anti-CD63 coated or non-coated devices. Immunoblotting was performed for β-actin and CD9. (B) SEM micrographs (bars = 1 μm). (C) On-chip immobilized and purified EVs were labeled with a fluorescent lipophilic dye (DiO). The fluorescence intensity values were normalized against the background. Fold-change is shown for minimum and maximum values with mean in a whisker plot. Frontal cortex (ALS = 12, Control = 8), spinal cord (ALS = 12, Control = 6), and serum (ALS = 7, Control = 9). (D) Size profile and frequency (%) by EV size measured from SEM micrographs in (B). Data are presented as mean ± s.e.m.; ALS (black), healthy controls (gray). Frontal cortex (ALS = 191 EVs from 2 ALS participants, Control = 446 EVs from 3 healthy controls), spinal cord (ALS = 325 EVs from 3 ALS participants, Control = 441 EVs from 3 healthy controls), and serum (ALS = 601 EVs from 3 ALS participants, Control = 804 EVs from 3 healthy controls). (E) Cumulative plots for each tissue show the number of EVs and their corresponding diameter for ALS (black) and control (gray) participants.

FIGURE 3

EV miRNAs are dysregulated in ALS frontal cortex, spinal cord, and serum. (A) Principal component analysis plot of unsupervised clustering of EV miRNA signal. (B) Fold-change of significant (P < 0.05) mature EV cargo DEmiRNAs was determined in ALS versus control groups. Increased fold-changes are represented in gray, while decreased fold-changes are represented in black. (C) Venn diagram showing common dysregulated EV-derived miRNAs from different ALS tissues. Frontal cortex (ALS = 12, Control = 8), spinal cord (ALS = 12, Control = 6), and serum (ALS = 14, Control = 8).

FIGURE 4

Association network of pathways in human ALS from frontal cortex, spinal cord, and serum. Significantly enriched KEGG pathways were combined and visualized in a network. KEGG pathways are represented by nodes, and shared gene content between pathways are represented by edges. Node shape indicates the tissue source of the enriched pathways: octagon (frontal cortex, spinal cord, serum), hexagon (frontal cortex, spinal cord), square (frontal cortex, serum), circle (frontal cortex), diamond (spinal cord), and triangle (serum). The size of each node denotes the number of miRNAs in that pathway, while the color gradient represents –log₁₀(P-value) (see color scale, yellow to red). All networks were organized by the Edge-weighted Spring Embedded layout in Cytoscape with minimal manual node rearrangement for visibility. Background colors delimit highly inter-connected subnetworks identified by the Cytoscape MCODE app (Bader and Hogue, 2003): green (brain function), gray (neurotransmitter systems), purple (intracellular signaling cascades), blue (cancer related), and pink (stem cell renewal). Non-clustered pathways are arranged at the bottom. FC, frontal cortex; SC, spinal cord.

FIGURE 5

Evaluation of predicted gene targets associated with EV DEmiRNAs. (A) Top 10 genes represented among pathways associated with the DEmiRNAs from frontal cortex (FC), spinal cord (SC), and serum. Node size represents the number of associated pathways containing that gene target (see Supplementary Tables 3–5). (B) qPCR for NDST4 in ALS and control iNeurons (one cell line per condition), or in control iNeurons incubated with ALS spinal cord EVs. Results are normalized to an internal reference (YWHAZ) and presented as fold change calculated by the 2^−ΔΔCT method. Data represent mean ± standard error of the mean from 6 independent experiments. **P < 0.01 compared to the control group. (C) Size distribution of ALS spinal cord EV particles used in (B) determined by NanoSight. (D) EV DEmiRNAs overlapping across published studies (see Table 3), where common DEmiRNAs are indicated in ovals, and predicted or verified targets are indicated in squares. Dark gray circles correspond to EV DEmiRNAs unique to each data set; connections between circles correspond to DEmiRNAs overlapping across two or more data sets.