Big thanks to our funders Stichting Long COVID and ZonMw for making this research possible! Hats off to the team: @oliverch77.bsky.social, Brent Appelman, Hanneke Willemen, the incredible Niels Eijkelkamp, and all others. And a special thanks to all involved patients! 12/12
Big picture:
We establish a patient-to-mouse model of Long COVID driven by antibodies.
This gives us a way to:
🔬 study mechanisms
💊 test treatments 11/12
This opens the door to new treatment strategies:
If antibodies are causal, therapies that target them might help patients:
• immunoadsorption → which we are already testing @amsterdamumc.bsky.social : www.youtube.com/watch?v=fE7A...
• FcRn targeting
• Plasma cell–targeting therapies 10/12
So what does this mean?
Our results provide functional evidence that autoantibodies can drive symptoms in Long COVID, not just correlate with them. 9/12
Another cool thing: our findings have been independently validated by 3 different groups. First of all by the great @virusesimmunity.bsky.social: www.medrxiv.org/content/10.1.... But also in Belgium (www.biorxiv.org/content/10.1...) and the UK (www.researchsquare.com/article/rs-7...). 8/12
Another key finding:
Antibodies collected 2 years later from the same patients still induced symptoms in mice.
This suggests:
- The immune system remains persistently altered
- Autoantibodies are long-lasting and functional 7/12
We also found that not all patients are the same.
Using biomarkers, we identified distinct subgroups, each with different antibody profiles and effects.
👉 Long COVID is not one disease, but multiple biological subtypes. 6/12
Even more remarkable:
This effect was not short-lived.
The induced pain persisted over time, mirroring chronic symptoms seen in patients. 5/12
Strikingly, mice that received Long COVID antibodies developed clear symptoms.
Most prominent:
⚡ Increased pain (mechanical + heat hypersensitivity)
Measured using standard behavioral assays like the Von Frey test. 4/12
We purified IgG antibodies from Long COVID patients and transferred them into mice.
If these antibodies are harmless → nothing should happen.
If they are pathogenic → mice should develop symptoms.
Here’s what we found 👇 3/12
Long COVID has long been linked to autoimmunity.
But a key question remained:
👉 Are these antibodies causing disease, or just a byproduct?
We set out to test this directly. 2/12
Our paper is (finally) out in @cp-cellrepmed.bsky.social: antibodies from #LongCOVID patients transfer symptoms to mice, pointing to a potential causal role for autoimmunity: www.cell.com/cell-reports... @amsterdamumc.bsky.social @umcutrecht.bsky.social 1/12
Transferring IgG antibodies from patients with #LongCovid to mice recapitulates many symptoms
www.sciencedirect.com/science/arti...
I had almost forgotten about this #MECFS interview last year, but suddenly there it is on youtube! Still holds true 🙂: www.youtube.com/watch?v=GVvG...
I'm humbled by all reactions. But alongside the excitement, I’d like to also share a more "real life" personal perspective on what the VICI trajectory meant for my professional and private life over the past year: www.linkedin.com/posts/jdendu...
For the first time, a prestigious Dutch #VICI grant (€ 1.5M) has been awarded to research on post-acute infection syndromes (PAIS). I’m deeply grateful for this opportunity to further indentify the underlying mechanisms: www.zonmw.nl/en/news/39-r...
Dank! We voegen het toe aan onze lijst met ideeën. In januari maken we keuze voor welke behandelingen we een aanvraag gaan indienen (liefst zo veel mogelijk, maar tijd/funding is beperkende factor).
We already knew that severe acute #COVID19 is induced by pathogenic IgG. But now Lynn Mes & @oliverch77.bsky.social show in @eurjimmunol.bsky.social that IgA2 adds even more fuel to the fire 🔥: onlinelibrary.wiley.com/doi/10.1002/...
Taken together, both “ends” of the antibody response appear affected in Long COVID:
– the antigen binding part (Fab)
– the tail (Fc)
This may not only impede SARS-CoV-2 clearance, but also hint at broader immune dysfunction.
🔹 Aberration 2: Impaired class switching.
Normally, B cells switch from IgM to potent, long-lasting IgG.
In Long COVID, this switch is blunted — reducing the production of high-affinity IgG needed for viral clearance.
🔹 Aberration 1: Immunological imprinting (“original antigenic sin”).
Instead of mounting fresh responses to SARS-CoV-2, B cells fall back on memory from seasonal coronaviruses. The result: antibodies less effective against the new virus.
The results:
– Reduced SARS-CoV-2 spike-specific IgG and IgA
– Elevated IgG against spike of seasonal coronaviruses
This points to two distinct immunological abnormalities 👇
Aberrant immune responses have long been suspected to play a role in #LongCOVID.
Here, we examined antibody profiles against both SARS-CoV-2 and the “everyday” seasonal coronaviruses. 2/6
Long COVID patients show immunological imprinting and impaired antibody class switching: fewer IgG/IgA antibodies against SARS-CoV-2 spike, but more against seasonal coronaviruses.
Interesting finding by my PhD student Ashwin Mak in J Med Virol: onlinelibrary.wiley.com/doi/10.1002/... 1/6
Also pretty proud of my team Oliver Chen, Amélie Bos, and Ashwin Mak (2 oral presentations, 3 posters) that resp. showed that Long COVID autoantibodies remain pathogenic for years, can directly affect skeletal muscle cells, and that patients show impaired antibody isotype switching. 7/
@virusesimmunity.bsky.social showed that IgG of Long COVID patients that is transferred to mice partially ends up in the brain, indicating that these IgGs may disrupt the blood-brain barrier. Could be an important step in the development of neurological symptoms. 6/
Maureen Hansen presented her recently published work on identification of circulating cell-free RNA signatures for the characterization and diagnosis of ME/CFS: www.pnas.org/doi/10.1073/... 5/
Esther Melamed showed that 25% of the patients in their (large) cohort display anti-ganglioside antibodies (not included in most autoantibody arrays): www.medrxiv.org/content/10.1... Let’s all check this in our cohorts! 4/
In vivo models: more and more mouse models are being developed to study pathophysiology and therapies. Nadia Roan showed that machine learning is a promising strategy to boost the quality of Long COVID models, and (successfully) test therapies. 3/
First, a call to action from @virusesimmunity.bsky.social to combine all our data internationally. Maureen Hansen mentioned that such a repository website already exists for #MECFS MAPMECFS.org. Could we make this for Long COVID as well? 2/