Human ALS Seeds Transmit Two Distinct SOD1 Aggregation Strains in Mice

In a finding that strengthens one of the most provocative ideas in modern neurodegeneration research, scientists in Sweden have shown that microscopic protein aggregates extracted from the spinal cords of patients with a specific inherited form of amyotrophic lateral sclerosis can trigger the disease when introduced into laboratory mice. The study, published in Acta Neuropathologica, demonstrates for the first time that tissue from patients homozygous for the D90A mutation in the SOD1 gene contains seeding-competent material capable of transmitting motor neuron disease, and that this material can carry two structurally distinct strains of misfolded superoxide dismutase-1 aggregates. The work lends powerful new support to the hypothesis that ALS, at least in its SOD1-linked forms, propagates through the body by a prion-like mechanism in which misfolded proteins impose their abnormal shape on their normal counterparts.

Amyotrophic lateral sclerosis is characterized by the adult-onset degeneration of the upper and lower motor neurons, the nerve cells that command voluntary movement. The disease typically begins in a focal region of the nervous system and then spreads contiguously, producing progressive paralysis and, ultimately, death from respiratory failure. Mutations in the gene encoding the free radical scavenging enzyme superoxide dismutase-1 are a well-established cause of the disease and are found in roughly one to nine percent of all patients. Since the SOD1 gene was first linked to familial ALS in 1993, more than 235 coding mutations have been catalogued. Most of these mutations are inherited as dominant traits, but the most prevalent of them all, D90A, behaves differently: disease develops primarily in individuals who carry two copies of the mutation, one inherited from each parent.

The central mystery that has driven this field for years concerns how the disease spreads. Cytosolic inclusions containing aggregated SOD1 are a pathological hallmark of ALS, both in patients and in transgenic animal models expressing mutant human SOD1. Using a technique called binary epitope mapping, the Umeå University team, led by researchers including Caitlin Henne, Isabelle Sigfridsson, Thomas Brännström, Stefan Marklund, Per Zetterström and Peter Andersen, previously discovered that two structurally different strains of human SOD1 aggregates, designated A and B, can arise in mice. Strain A forms in most mutant models, whereas homozygous D90A mice characteristically produce the distinct strain B, alongside strain A. Critically, when seed preparations of either strain are injected into the spinal cords of recipient mice expressing a human SOD1 transgene, the aggregates propagate in a templated fashion, spreading through the nervous system and precipitating premature, fatal motor neuron disease that closely resembles human ALS.

Earlier experiments had already demonstrated that seeds prepared from the central nervous systems of patients carrying the aggressive G127X truncation mutation could transmit strain A aggregation and disease to mice. But those patients carry a destabilized, inactive protein present only in minute quantities in the nervous system. The D90A mutation presents a very different challenge: the D90A protein is molecularly stable, retains wild-type-like enzymatic activity, and accumulates at high concentrations in the central nervous system. Patients homozygous for D90A typically survive more than a decade after onset, with a median of fourteen years, and their spinal ventral horns become profoundly degenerated with massive motor neuron loss. Seeds prepared from such tissue were therefore expected to contain only vanishingly small amounts of seeding-competent material, raising real doubt about whether transmission would be detectable at all.

The researchers addressed this question by preparing seeds from the ventral horns, including the entire lamina IX region, of six patients homozygous for D90A who had died of ALS. The preparation protocol involved homogenization in buffer containing detergent and guanidinium chloride, followed by ultracentrifugation through dense iohexol cushions that pelleted very large proteinaceous complexes. Quantitative analysis revealed that the seeds contained only picogram amounts of aggregated SOD1 per microliter, diluted within roughly fifty thousand times more protein from other ventral horn components. One microliter of each seed was then inoculated stereotactically into the lumbar ventral horn of the left side of the spinal cord in one-hundred-day-old, still asymptomatic mice carrying the human SOD1 G85R transgene, a slow model of disease in which aggregation arises spontaneously in late life.

The results were striking despite the technical odds. Seeds from two of the six patients, designated A1 and A2, significantly shortened the survival of the recipient mice compared with non-inoculated controls. Mice receiving the A2 seed developed fatal paralysis and their spinal cords showed strain A aggregation patterns by binary epitope mapping, indicating that the patient’s ventral horn had contained strain A aggregates. The A1 seed told a subtler story. The two shortest-lived mice in that group displayed unmistakable strain B patterns, whereas the longer-lived mice in the same group showed strain A patterns, which arise spontaneously in the G85R model. This suggests the A1 patient harbored strain B aggregates, which propagate roughly thirty percent more slowly than strain A, and that the spontaneous strain A aggregation eventually overwhelmed any later-seeded B aggregation in the surviving animals. Confocal immunohistochemistry confirmed the biochemical findings, revealing both strain A and strain B aggregates in the tissue of the two most short-lived A1-inoculated mice, while all other inoculated mice showed strain A alone.

The specificity of these effects was rigorously controlled. Nine different seed preparations from four neurologically normal individuals, prepared with three distinct protocols, produced lifespans indistinguishable from non-inoculated mice, as did seeds from control C57BL/6 mice. Notably, the postmortem interval was significantly shorter in the control group than in the ALS group, a factor that should have favored, rather than undermined, seeding activity in the controls, since seeding-competent material is sensitive to proteolytic degradation. The pattern of disease onset also told a coherent story: mice receiving the active seeds overwhelmingly developed hindleg symptoms first, consistent with aggregation initiating at the lumbar inoculation site, and the aggregates were found to have spread along the neuraxis in terminally ill animals, exactly as expected from a prion-like propagation process.

An intriguing aspect of the findings is that the researchers detected no obvious difference in the total quantity of detergent-insoluble SOD1 aggregates between the two active seeds and the four inactive ones. This implies that the prion-active aggregates represent only a subfraction of the total insoluble SOD1 in the tissue, meaning that total aggregate burden alone is a poor predictor of biological activity. What matters is the structural nature of the aggregates present. This has practical implications: understanding which aggregate structures are seeding-competent could inform the selection of epitopes for antibody-based therapies and other structure-dependent interventions. It may also help explain why ALS phenotypes vary so dramatically between different SOD1 mutations, since different aggregate strains could propagate at different rates and produce different disease courses.

Perhaps the most compelling comparison in the study is between the two patient mutations whose seeds have now been shown to transmit disease. The truncated G127X protein is inactive, disordered, and present in only minute amounts, yet its seeds transmitted strain A aggregation and rapidly progressive disease. The D90A protein is stable, active, and abundant, and its seeds transmitted both strain A and strain B aggregation in patients with a uniform, slowly progressive clinical phenotype. That aggregates from two mutations with such radically different biochemical properties and clinical courses both prove capable of templating their misfolding in recipient animals provides the strongest evidence yet that prion-like propagation of misfolded SOD1 is the primary pathogenic mechanism in SOD1-linked ALS. While the current sample of two transmitting patients is too small to draw firm conclusions about which clinical phenotypes map to which strains, the researchers note that the strain B-transmitting seed came from the patient with the earliest onset and shortest symptomatic disease, a pattern consistent with prior observations in transgenic mice. The work opens a path toward diagnosing and, ultimately, intercepting these pathological protein strains before they march through the nervous system.

Subject of Research: Prion-like transmission of SOD1 aggregate strains from D90A ALS patient tissue to transgenic mice

Article Title: Seeds from ALS patients homozygous for the SOD1 D90A mutation transmit two types of SOD1 aggregation and motor neuron disease

Article References: Henne, C., Sigfridsson, I., Brännström, T., Forsberg, K. M. E., Marklund, S. L., Zetterström, P., & Andersen, P. M. (2026). Seeds from ALS patients homozygous for the SOD1 D90A mutation transmit two types of SOD1 aggregation and motor neuron disease. Acta Neuropathologica, 152(1), Article 27. https://doi.org/10.1007/s00401-026-03078-3

Image Credits: AI Generated

DOI: 10.1007/s00401-026-03078-3

Keywords: ALS, SOD1, D90A mutation, prion-like propagation, protein aggregation, motor neuron disease, binary epitope mapping, aggregate strains, transgenic mice, amyotrophic lateral sclerosis, Seeds, patients