Serum biomarkers of neuroinflammation and blood-brain barrier leakage in amyotrophic lateral sclerosis

Abstract

Amyotrophic lateral sclerosis (ALS) is an incurable and rapidly progressive neurological disorder. Biomarkers are critical to understanding disease causation, monitoring disease progression and assessing the efficacy of treatments. However, robust peripheral biomarkers are yet to be identified. Neuroinflammation and breakdown of the blood-brain barrier (BBB) are common to familial and sporadic ALS and may produce a unique biomarker signature in peripheral blood. Using cytometric bead array (n = 15 participants per group (ALS or control)) and proteome profiling (n = 6 participants per group (ALS or control)), we assessed a total of 106 serum cytokines, growth factors, and BBB breakdown markers in the serum of control and ALS participants. Further, primary human brain pericytes, which maintain the BBB, were used as a biosensor of inflammation following pre-treatment with ALS serum. Principal components analysis of all proteome profile data showed no clustering of control or ALS sera, and no individual serum proteins met the threshold for statistical difference between ALS and controls (adjusted P values). However, the 20 most changed proteins between control and ALS sera showed a medium effect size (Cohen's d = 0.67) and cluster analysis of their levels together identified three sample subsets; control-only, mixed control-ALS, and ALS-only. These 20 proteins were predominantly pro-angiogenic and growth factors, including fractalkine, BDNF, EGF, PDGF, Dkk-1, MIF and angiopoietin-2. S100β, a protein highly concentrated in glial cells and therefore a marker of BBB leakage when found in blood, was unchanged in ALS serum, suggesting that serum protein profiles were reflective of peripheral rather than CNS biofluids. Finally, primary human brain pericytes remained proliferative and their secretome was unchanged by chronic exposure to ALS serum. Our exploratory study suggests that individual serum cytokine levels may not be robust biomarkers in small studies of ALS, but that larger studies using multiplexed analysis of pro-angiogenic and growth factors may identify a peripheral signature of ALS pathogenesis.

Keywords: Amyotrophic lateral sclerosis; Blood-brain barrier; Cytokine; Neuroinflammation; Serum.

Fig. 1

Clustered proteomic analysis of sera partially segregates ALS and control. A Cytometric bead array analysis of participant blood sera. IL-6, IL-8, RANTES, MCP-1, IP-10, fractalkine in control (n = 15) and ALS (n = 15) sera and sICAM-1 and sVCAM-1 in control (n = 8) and ALS (n = 8) sera. Boxes, 25th -75th percentiles; whiskers, minima and maxima. Statistical significance determined by Mann Whitney test. B Proteome profiler analysis of an extended panel of 105 cytokines in control (n = 6) and ALS (n = 6) participant blood sera. Raw intensity normalized to reference spot intensity of each membrane. Proteins (n = 20) with the greatest difference between ALS/controls shown. Boxes, 25th -75th percentiles; whiskers, minima and maxima. Statistical analysis determined by Mann-Whitney test. C Heat map shows unsupervised hierarchical clustering of the listed cytokines. D Unsupervised hierarchical clustering of growth factors BDNF, EGF and PDGF-AB/BB/AA

Fig. 2

ALS serum pre-treatment did not influence the pericyte secretome in response to pro-inflammatory compounds. A Cytometric bead array analysis of selected cytokines secreted by primary human brain pericytes pre-treated with control (black, n = 15) or ALS (grey, n = 15) participant sera, then exposed to pro-inflammatory stimuli for 24 hours. Boxes, 25th -75th percentiles; whiskers, minima and maxima. Statistical analysis between ALS and control determined by multiple Mann-Whitney tests. B As per (A) but exposed to pro-inflammatory stimuli for 8 hours

Fig. 3

ALS serum pre-treatment did not influence pericyte intracellular expression of MCP-1 or VCAM-1 in response to pro-inflammatory compounds. A Representative images of ALS or control serum pre-treated pericytes 8 hours after vehicle or different pro-inflammatory stimuli were applied (IL-1β, IFN-γ or TNF-α). Scale bar = 50 μm. Inflammatory response is represented by an upregulation of MCP-1 and VCAM-1. B Quantification of percentage cells positive for MCP-1, mean +/− SEM. C Quantification of mean integrated intensity of VCAM-1 per positive cell, mean +/− SEM

Fig. 4

ALS serum pre-treatment did not significantly alter pericyte cell number. A There was a consistent but non-significant increase in ALS serum pre-treated cell nuclei observed across all pro-inflammatory stimuli and vehicle. Values are average cell counts per 10X magnification image site (equivalent to 1393.2 × 1393.2 μm). Two sites were imaged per well for each of the 15 ALS and 15 control sera-treated cell samples across all pro-inflammatory stimuli conditions and vehicle. Results are plotted as mean +/− SEM. B Overall combined cell nuclei count (from all treatments). Statistical analysis determined by Mann-Whitney test

Fig. 5

S100β levels in ALS sera are not different from control sera. A S100β levels detected in sera showed no difference between ALS and control groups. B ALS sera samples separated on the basis of disease duration. Results are plotted as mean +/− SEM and each dot represents the average concentration from one serum sample