Spike Protein Amyloid Fibrils: How Spike685 Makes Clots Resist Lysis

This explainer is for anyone who's heard that the spike protein "forms amyloid" and wants to know what that means for blood clots. By the end you'll know what the 2025 Linköping study tested, which two spike segments mattered, why it's a test-tube result rather than a clinical one, and how amyloid in clots gets detected.

What are spike protein amyloid fibrils?

Amyloid fibrils are long, stacked strands of protein held together by a cross-beta-sheet core. They're unusually stable and hard for the body to clear. Many proteins, or short pieces of them, can fold this way under the right conditions.

In 2022, Sofie Nyström and Per Hammarström at Linköping University scanned the spike protein and found seven amyloid-prone segments. All seven aggregated at body temperature, and three met every strict amyloid criterion. Intact folded spike didn't form fibrils on its own, so something had to cut it first. Our walkthrough of how neutrophil elastase cuts spike into amyloid-forming pieces covers that earlier paper step by step.

The 2025 follow-up asked a narrower question: if you make fibrils from each of those seven segments, what do they do to clotting and clot breakdown?

Why does fibrinolysis matter for microclots?

When you clot, thrombin turns soluble fibrinogen into a fibrin mesh. Later, plasmin dissolves that mesh, a process called fibrinolysis. If fibrin forms but won't dissolve, small clots can linger in circulation.

Groups led by Etheresia Pretorius and Douglas Kell have reported fibrin amyloid microclots, or fibrinaloids, in Long COVID plasma that stain with amyloid dyes and resist breakdown (Pretorius et al., 2021; Kell et al., 2022). Earlier, the same group showed that adding recombinant spike S1 to healthy plasma produced fibrin(ogen) resistant to fibrinolysis. The molecular step in between stayed unclear.

What did the 2025 study find?

Hanna Westman, Per Hammarström and Sofie Nyström published their results in Biochemistry in late 2025 (Westman et al., 2025). They used purified fibrinogen, thrombin and plasmin, then added amyloid fibrils made from each of the seven spike peptides. Most did very little. Two behaved differently from the rest.

Spike601: a pre-clotting effect

Spike601 fibrils (residues 601 to 620) bound fibrinogen and pulled it out of solution. With less fibrinogen available, thrombin produced less fibrin. Once a clot had formed, Spike601 didn't stop plasmin from dissolving it.

Spike685: a post-clotting effect

Spike685 fibrils (residues 685 to 701, immediately after the furin cleavage site) did the opposite. Clots still formed, but they showed marked resistance to plasmin. The authors tested final fibril concentrations of 1.4, 2.8, 5.6, 10 and 20 micrograms per millilitre, and resistance grew with dose.

Under fluorescence and electron microscopy, Spike685 produced dense fibrin networks and drew fibrin into larger amyloid-positive aggregates. Fibrin left over after lysis sat alongside the spike amyloid.

Next step

Curious whether fibrin amyloid microclots show up in your blood?

Our at-home amyloid fibrin test uses Thioflavin T fluorescence microscopy and returns a videographic report you can share with your doctor.

See the at-home test

How do normal and Spike685 clots compare?

The paper's background notes that dense clots made of thinner fibers are already known to slow the diffusion of tPA and plasminogen into a clot. That's a plausible reason a denser Spike685 clot dissolves slowly, though the study didn't measure enzyme diffusion directly.

Does this explain microclots in Long COVID?

It offers one plausible route, not proof. The chain the authors sketch runs like this: spike is cleaved, amyloidogenic pieces such as Spike685 are exposed, those pieces form fibrils, and the fibrils seed fibrin clots that plasmin struggles to clear.

Every link still needs testing in people. Nobody knows what concentration of spike amyloid, if any, circulates in patients. Antibody-based estimates of soluble spike in plasma land far below the micrograms per millilitre used in these assays, which the authors acknowledge in their discussion. Our guide to amyloid microclots in Long COVID sets this against the patient studies.

What the study could not show

The authors are explicit about the limits, and so are we.

• The system used purified fibrinogen, thrombin and plasmin. Real plasma holds hundreds of other proteins, cells and regulators.
• They used 20-amino-acid synthetic peptides, not full-length spike. Behaviour in isolation can differ from behaviour inside the intact protein.
• The fibril concentrations were chosen so the assay would work, and sit above published estimates of soluble spike in human plasma.
• Five of seven spike fibrils did little, so broad statements that "spike causes clotting problems" skip past the specificity.
• No patients were studied, so no claim about symptoms or treatment follows from this paper alone.

For what's been reported in symptomatic people after vaccination, including a small cohort that found amyloid-positive microclot material, see our review of spike-driven microclots after mRNA vaccination. The wider claims made in the 2025 JPandS review, and which rest on which kind of evidence, are covered in our walkthrough of that review.

How are amyloid fibrils actually detected?

The detection toolkit here is consistent across labs. Thioflavin T binds the beta-sheet core of amyloid and fluoresces. Congo red gives apple-green birefringence under polarized light. Electron microscopy shows fibril shape. ThT kinetics track how quickly fibrils assemble.

That same ThT principle is what fibrin amyloid microscopy applies to plasma. Our pillar guide to amyloid fibril testing and ThT detection explains those methods, while the amyloid fibrin test guide focuses on microclots in blood. To see how commercial antibody and antigen panels compare with downstream clot detection, see our guide on spike protein test options and microclot analysis. If you're working out which test answers which question, the comparison of types of amyloid tests lines up Alzheimer's blood tests, PET, CSF, amyloidosis biopsy and fibrin amyloid microscopy.

Where our test fits

Our at-home amyloid test is a finger-prick dried blood sample mailed to a partner laboratory in Mexico City. The lab stains it with Thioflavin T and records 6K 12-bit RAW video microscopy at 20x magnification under UV, reviewed with AI-assisted fibril detection plus trained human oversight. The result is a Videographic Microscopy Report in a secure portal.

It's laboratory observational data for informational and research purposes, not a diagnosis and not a medical device. It doesn't measure circulating spike protein, which needs a separate spike-specific assay, and it doesn't quantify D-dimer or fibrinogen. Our comparison of D-dimer versus microclot testing explains why a normal D-dimer doesn't settle the microclot question.

What to read next

If clotting is your main concern, the amyloid fibrin test guide goes deeper, and the research on nattokinase and lumbrokinase for microclots covers what's been studied for breaking fibrinaloids down. Bring any result to a clinician who can read it alongside your history and standard labs.