Potassium channel complex autoimmunity induced by inhaled brain tissue aerosol

Abstract

Objective: To test the hypothesis that autoimmunity induced by inhalation of aerosolized brain tissue caused outbreaks of sensory-predominant polyradiculoneuropathy among swine abattoir employees in the Midwestern United States.

Methods: Mice were exposed intranasally, 5 days per week, to liquefied brain tissue. Serum from exposed mice, patients, and unaffected abattoir employees were analyzed for clinically pertinent neural autoantibodies.

Results: Patients, coworkers, and mice exposed to liquefied brain tissue had an autoantibody profile dominated by neural cation channel immunoglobulin Gs (IgGs). The most compelling link between patients and exposed mice was magnetic resonance imaging (MRI) evidence of grossly swollen spinal nerve roots. Autoantibody responses in patients and mice were dose-dependent and declined after antigen exposure ceased. Autoantibodies detected most frequently, and at high levels, bound to detergent-solubilized macromolecular complexes containing neuronal voltage-gated potassium channels ligated with a high affinity Kv1 channel antagonist, 125I-α-dendrotoxin. Exposed mice exhibited a behavioral phenotype consistent with potassium channel dysfunction recognized in drosophila with mutant ("shaker") channels: reduced sensitivity to isoflurane-induced anesthesia. Pathological and electrophysiological findings in patients supported peripheral nerve hyperexcitability over destructive axonal loss. The pain-predominant symptoms were consistent with sensory nerve hyperexcitability.

Interpretation: Our observations establish that inhaled neural antigens readily induce neurological autoimmunity and identify voltage-gated potassium channel complexes as a major immunogen.

FIGURE 1

Serum IgGs reactive with antigens in central and peripheral neural tissues detected by indirect immunofluorescence on substrates of cerebellar cortex, midbrain and stomach; scale bar: 200 μm. IgG of a patient (A) and a mouse (D) exposed nasopharyngeally to liquefied porcine brain tissues. (A and D) Note prominent synaptic staining in molecular layer (ml) of cerebellar cortex. Somata and processes of neural cells also are immunoreactive in the molecular layer and granular layer (gl), and enteric neurons and nerve fibers in the myenteric plexus (my pl); gut smooth muscle (sm) and mucosa (muc) are not immunoreactive. (B) IgG in serum of a control patient with idiopathic VGKC complex-IgG autoimmunity yields similar synaptic staining in the cerebellar molecular layer (and midbrain) but, in contrast to the “Signature IgG” pattern yielded by serum of aerosol –exposed patients and mice, this IgG does not bind to neuronal somata and processes; (C) IgG titers in the aerosol-exposed patients' sera correlated significantly with levels of VGKC complex-IgG (patients, dark line, R = 0.67, p < 0.01; asymptomatic coworkers, gray line, R = 0.73, p < 0.001).

FIGURE 2

Autoantibody profiles in mice exposed 5 days per week to liquefied brain tissue by oronasopharyngeal exposure. (A) Median serum levels of signature IgG (), myelin basic protein (MBP)-IgG (–◇–) and VGKC complex-IgG (–X–) increase in parallel in 10 mice for which exposure frequency was doubled at day 57. (B) Dose-dependent production of VGKC complex-IgG in three groups of 10 mice. Median serum levels rose progressively 28 days after the dose of inhaled brain tissue was doubled (–X– day 0; –◇– day 29; day 57).

FIGURE 3

Patients' serum autoantibody levels fell after neural antigen exposure ceased. (A) Signature IgG and (B) VGKC complex-IgG levels. Patients' antibody levels fell, with (●—●) and without (◯—◯) immunotherapy. Horizontal dotted lines indicate detection thresholds. (C) Serum autoantibody levels in mice fell after neural antigen exposure ceased; MBP-IgG () and VGKC complex-IgG (–◇–).

FIGURE 4

Clinical, behavioral and imaging observations in patients and mice exposed to liquefied brain tissue. (A) Frequency of neural-specific autoantibodies in serum of patients with documented motor, sensory and imaging abnormalities. (B) Sensitivity to isoflurane-induced anesthesia was reduced in mice exposed to liquefied neural tissue (***p < 0.001, n = 30). (C, D) Representative magnetic resonance images of spinal cord revealing nerve root abnormalities (circled) in a patient and a mouse. Images are post-gadolinium T1-weighted coronal scans of sacral (patient) and thoracic (mouse) spinal roots. Nerve root brightness reflects enhanced blood-nerve barrier permeability.