α‐Synuclein implicated in Parkinson's disease is present in extracellular biological fluids, including human plasma

Abstract

SPECIFIC AIMS We investigated whether neuronal cells in culture normally secrete α-synuclein protein (α-syn) into their surrounding media. We also looked for the presence of α-syn in human cerebrospinal fluid (CSF) and blood plasma. The methods used were immunocapture with anti-α-syn antibodies coupled to magnetic beads, followed by detection on Western blots using five different antibodies to α-syn. PRINCIPAL FINDING 1. Neuronal cells normally secrete α-syn To determine whether cells secrete α-syn during normal metabolism, we used human neuroblastoma M17 cells stably transfected with cDNA encoding the wild-type form of α-syn (M17-wt), untransfected M17 cells, or cells transfected with vector alone (M17-vec). Serum-free OPTI-MEM medium conditioned for 6–48 h with each cell type was immunoprecipitated using magnetic dynabeads coupled to 211 mouse monoclonal antibody which recognizes amino acid residues 121-125 of α-syn. Bound proteins were eluted from the beads, after washing, and separated on SDS-gels before transfer to nitrocellulose membranes for Western blotting. The blots were probed with FL-140 rabbit polyclonal antibody raised against a recombinant full-length human α-syn (Fig. 1 ⤻ A), N-19 goat polyclonal antibody to an amino-terminal region of α-syn (5-19) or α-synuclein C-20 goat polyclonal antibody which recognizes the carboxyl-terminal region of α-syn (120-135) (data not shown). In all cases, a 15 kDa protein that comigrated with recombinant α-syn and comigrated with α-syn from human brain homogenates was identified in the conditioned medium (Fig. 1)⤻ . In accord with its identification as α-syn, the 15 kDa protein was much more abundant in medium conditioned by M17-wt cells than medium conditioned with M17-vec or untransfected M17 cells (Fig. 1A,B⤻ ). Prolonged exposure of the blotting membrane revealed detection of the 15 kDa band in immunoprecipitates prepared from medium conditioned with M17-vec or M17 cells for as little as 6 h (data not shown), with a progressive increase in the amount of the 15 kDa protein released into the medium over time. The 15 kDa band was not detected when the membrane was probed with control mouse IgG (data not shown) or when the anti-α-syn antibodies used to probe the blots were pre-incubated with recombinant α-syn (Fig. 1A⤻ ). A similar 15 kDa band was detected when Fl-140 antibody was used for immunocapture and the blots were probed with mouse monoclonal anti-α-syn antibodies 211 (Fig. 1B⤻ ) or LB509, which recognizes α-syn (115-122), or with goat polyclonal anti-α-syn antisera N-19 or α-synuclein C-20 (data not shown). In all these experiments, >96% of the cultured cells were found to be viable at the end of the 48 h conditioning period. Figure 1. Detection of α-syn in culture medium. Medium conditioned by M17-wt cells (expressing α-syn), M17-vec cells (expressing vector only) or non-transfected M17 cells were collected after 6, 24 and 48 h, then immunoprecipitated with 211 (A) or FL-140 (B). The captured proteins were separated on Bis-Tris 4–12% SDS-PAGE and Western blotting was carried out with FL-140 (A) and 211 (B). A) Lanes 1, 12, and 13 contain 3 ng recombinant human α-syn; lane 2 contains 1 μg of human brain lysate; lanes 3–5, M17 medium; lanes 6–8, M17-vec; lanes 9–10 and 14–16, M17-wt medium. Lanes 12–16 were probed with FL-140 antibody pre-absorbed with recombinant human α-syn (1 μg α-syn/mL antibody). B) Lanes 1–3, 14, and 15 contain M17-wt medium; lanes 4–6, M17-vec; lanes 7–9, M17 medium. Lanes 10 and 12 were loaded with 3 ng of recombinant human α-syn; lanes 11 and 13, 1 μg of human brain lysate; lanes 12–15 were probed with 211 antibody pre-absorbed with recombinant human α-syn (1 μg α-syn/mL antibody). Download figure Download PowerPoint 2. Detection of α-syn in human CSF Using the same experimental approach, we determined whether α-syn is present in human CSF from non-neurodegenerative disease controls and from patients with DLB and PD. We used FL-140 to capture the protein from the samples, with LB509, 211, N-19, or α-synuclein C-20 for detection on the Western blots. Recombinant α-syn and human brain lysates were run on the same gels as positive controls, with recombinant β-synuclein (β-syn) and γ-synuclein (γ-syn) as additional negative controls. A 15 kDa protein that comigrated with recombinant α-syn and with α-syn present in the brain lysate was identified in all CSF samples tested. Similar results were obtained when the immunoprecipitation step was carried out with 211 and the blots were probed with FL-140, N-19, or α-synuclein C-20. The 15 kDa band was absent in negative controls using mouse or rabbit IgG for capture and when membranes were probed with LB509, 211, N-19, or α-synuclein C-20 pre-absorbed with recombinant α-syn. 3. Detection of α-syn in human plasma A similar 15 kDa protein was readily detected in human plasma samples obtained from normal control subjects or from patients with PD and DLB. Again, this protein could be captured with magnetic beads derivatized with FL-140 or 211 and detected on Western blots using antibodies LB509, 211, N-19, FL-140, or α-synuclein C-20 (Fig. 2 ⤻ ). On Western blots, this protein was indistinguishable from recombinant α-syn and from α-syn present in human brain homogenates. There was no detection of recombinant β-syn or γ-syn (Fig. 2A,B⤻ ); the 15 kDa band was absent when rabbit or mouse IgG was used for capture (data not shown) or when membranes were probed with antibodies to α-syn pre-absorbed with the recombinant protein (Fig. 2)⤻ . Our results strongly suggest that the 15 kDa band detected in human plasma is also monomeric α-syn. Our initial survey indicates there is considerable overlap in the amount of α-syn observed in the CSF and plasma from the PD and non-PD groups. Neither β-syn nor γ-syn could be detected on Western blots using several specific antibodies to β-syn or γ-syn for immunoprecipitation of the plasma or CSF. Figure 2. Immunoprecipitation of α-syn from human plasma Human plasma from normal, DLB, and PD patients was incubated with beads cross-linked with primary antibodies FL-140 (A–C) or 211 (D, E) and the resulting immunoprecipitates were fractionated on SDS-PAGE and immunoblotted with LB509 (A), 211 (B), N-19 (C), FL-140 (D), and N-19 (E). A, B) 6 ng of recombinant human β-syn (lane 1); 6 ng of recombinant human γ-syn (lane 2); 6 ng of recombinant human α-syn (lane 3); 3 μg of normal human brain lysates (lane 4); plasma from PD patients (lanes 5, 6, and 13); plasma from DLB patients (lanes 7 and 8) and age-matched control plasma (lanes 9–12 and 14) were used. Lanes 13 and 14 were probed with the antibodies pre-absorbed with recombinant human α-syn (1 μg α-syn/mL antibody). C) Plasma from PD patients (lanes 1, 2, and 7); age-matched control plasma (lanes 3, 4, 8, and 9); 6 ng of recombinant human α-syn (lanes 6 and 10); 3 μg of normal human brain lysates (lane 5). Lanes 7–10 were probed with N-19 antibody pre-absorbed with recombinant human α-syn (1 μg α-syn/mL antibody). D) PD patients plasma (lanes 1, 5, and 6); age-matched control plasma (lanes 2, 3, 7, and 8); 6 ng of recombinant human α-syn (lanes 4 and 9); 3 μg of normal human brain lysates (lane 10). Lanes 1–4 were probed with FL-140 antibody pre-absorbed with recombinant human α-syn (1 μg α-syn/mL antibody). E) Plasma from PD patients (lanes 3, 4, and 7); age-matched control plasma (lanes 5, 6, 8, and 9); 6 ng of recombinant human α-syn (lanes 1 and 10); 3 μg of normal human brain lysates (lanes 2) were used. Lanes 7–10 were probed with N-19 antibody pre-absorbed with recombinant human α-syn (1 μg α-syn/mL antibody). Download figure Download PowerPoint CONCLUSIONS α-Syn lacks a signal sequence for targeting to the ER and so is generally thought to exist only as an intracellular protein. In contrast, our results provide convincing evidence for the existence of an extracellular form of α-syn, detected here in the culture medium of neuronal cells and in CSF and plasma. Although we do not know the mechanism by which α-syn is released, a point of possible significance is the fact that the carboxyl-terminal tail of α-syn contains two di-acidic motifs of the form Asp-X-Glu. This motif is known to be an ER-to-Golgi directing signal and to interact with the ER coat protein II (COPII) complex (Fig. 3 ⤻ ). Although α-syn does not have an ER directing signal, its amphipathic α-helical, lipid binding domain binds strongly to membranes and even permeabilizes them. These membrane binding properties might allow α-syn to associate with the ER membrane, allowing the di-acidic motifs on α-syn to interact with the COPII complex transporting proteins from the ER to Golgi. Once in the Golgi, the default pathway is secretion from the cell (Fig. 3)⤻ . β-Syn and γ-syn, not detected extracellularly in CSF and plasma, do not contain any di-acidic motifs. Figure 3. Schematic diagram showing possible mechanisms for the secretion of α-syn from neuronal cells. Classical ER/Golgi-dependent and nonclassical independent pathways are shown. Download figure Download PowerPoint It has recently been reported that platelets from PD and normal individuals contain α-syn and γ-syn. However, neither protein is secreted upon platelet activation. Thus, it is unlikely that the α-syn in human plasma originates from platelets. It has been reported that in a transgenic mouse model of Alzheimer’s disease, after peripheral administration of a monoclonal antibody to the β-amyloid (Aβ) peptide, a rapid increase in plasma Aβ is observed; the magnitude of this increase was highly correlated with amyloid burden in the hippocampus and cortex. These results demonstrate that Aβ can efflux from the brain to the plasma, so it is possible that a similar mechanism could operate for other neuronal proteins, including α-syn. This type of mechanism could explain the presence of the α-syn in plasma samples. In MSA brains, glial cytoplasmic inclusions in oligodendrocytes contain α-syn fibrils, but α-syn is not expressed in glial cells, suggesting that the α-syn deposits found in glial cells of MSA brains may be due to neuronal secretion. This evidence suggests that neuronal cells normally secrete α-syn into the surrounding media in the brain; this could circulate to the CSF and then to the blood. An interesting possibility is the potential use of α-syn and/or its derivatives in biological fluids as a biomarker for PD and related disorders. Further detailed studies of the levels of normal, nitrated, phosphorylated, glycosylated, or oligomeric forms of α-syn are required to determine whether this is a viable approach. The development of a reliable biomarker would dramatically accelerate research on the etiology, pathology, disease progression and therapeutics for synucleinopathies. ¹To read the full text of this article, go to http://www.fasebj.org/cgi/doi/10.1096/fj.03-0098fje; doi: 10.1096/fj.03-0098fje ²These authors contributed equally to this work.