Micro-Cancers in the Thyroid: When Autoimmunity Becomes Somatic Evolution
> Paper: Polyclonal selection of immune checkpoint mutations in thyroid autoimmunity > Authors: Pantelis A. Nicola, Andrew R. J. Lawson, Iñigo Martincorena, et al. > Institutions: Wellcome Sanger Institute, University of Cambridge, University of Edinburgh, and others > Published: Nature, April 14, 2026 > DOI: 10.1038/s41586-026-10493-9
---
A Long-Ignored Hypothesis
How do autoimmune diseases arise? The textbook answer: the immune system "mistakes" the body's own tissue for an enemy and attacks. But over sixty years ago, immunologist Frank Macfarlane Burnet proposed a more radical idea—perhaps the problem lies not in immune recognition but in somatic mutations within lymphocytes.
Burnet suggested that self-reactive lymphocytes that should have been eliminated might escape checkpoint surveillance by acquiring somatic mutations. The hypothesis was never tested because the technical challenge was enormous: you need to find the very few lymphocytes carrying key mutations among a vast population of normal cells.
That hypothesis has now been confirmed.
---
What Was Found: A Mutinous Legion in the Thyroid
Using the ultra-accurate single-molecule DNA sequencing technology NanoSeq, the team deeply scanned thyroid tissue from patients with Hashimoto's thyroiditis and Graves' disease.
They found hundreds of independent B-cell clones, each carrying different somatic mutations, all doing the same thing: disrupting immune checkpoint genes.
Convergent Evolution: Different Roads, Same Destination
The most striking finding is convergent evolution. These B-cell clones independently acquired mutations, yet all targeted the same genes:
| Target gene | Alias | Function | Mutation type | |--------|------|------|----------| | TNFRSF14 | HVEM | Key checkpoint inhibiting T-cell activation | Loss-of-function | | CD274 | PD-L1 | Sends "don't attack" signals to T cells | Loss-of-function |
TNFRSF14 mutations were especially prevalent. In some biopsy samples, the team detected dozens to hundreds of independent TNFRSF14-mutant clones. Each clone accounted for only a small fraction of total cells (usually <1%), but together the mutant clones made up a substantial proportion of B cells.
Even more strikingly, some clones carried biallelic loss—both copies of the checkpoint gene were disrupted. Others accumulated 4–6 driver mutations—advancing step by step, like cancer cells.
---
Not a Runaway Immune System, but "Micro-Cancers"
The paper's core argument: autoimmune disease is not simply immune overreaction, but the product of somatic evolution occurring locally within tissue.
How does cancer arise? A cell acquires a mutation, escapes growth control, and proliferates into a clone. The clone keeps accumulating mutations, gains an ever-stronger competitive advantage, and eventually forms a tumor.
And autoimmune disease? A B cell acquires a checkpoint mutation, escapes tolerance control, and forms a positive feedback loop with self-reactive T cells. The B cell expands into a clone, keeps accumulating mutations, and becomes ever more capable of attacking the body's own tissue.
Both are Darwinian evolution replayed locally within the body. The only difference is the goal:
- Cancer: clones aim for unlimited proliferation
- Autoimmunity: clones aim to attack self tissue
- Whole-exome sequencing: identifying candidate driver mutations
- Laser microdissection: pinpointing the spatial location of mutant cells
- Methylation sequencing: confirming the identity of mutant cells
- Spatial transcriptomics: mapping mutant cell distribution in tissue
- Single-nucleus DNA sequencing: confirming clonal independence
- Antibody synthesis: validating the self-reactivity of clones
- Different autoimmune diseases may share common mutational lineages
- Disease progression may result from clonal competition (like cancer's evolutionary tree)
- Therapeutic targets may shift from "suppressing the whole immune system" to "eliminating specific mutant clones"
- Develop therapies targeting checkpoint-mutant clones (analogous to targeted cancer therapy)
- Use immune checkpoint inhibitors? Wait—autoimmunity is already checkpoint failure; further inhibiting checkpoints would pour fuel on the fire
- In fact, the goal may be to restore checkpoint function—for example, gene therapy to supplement PD-L1, or cell therapies that clear mutant B cells
- Some cancer patients develop autoimmune side effects after immune checkpoint inhibitor therapy
- Some autoimmune diseases raise or lower cancer risk
- Original paper: https://www.nature.com/articles/s41586-026-10493-9
- Companion commentary: https://www.nature.com/articles/d41586-026-01415-w
- Burnet, F. M. The Clonal Selection Theory of Acquired Immunity (1959)
Co-author and cancer genomicist Iñigo Martincorena is known for his earlier work on mutations in apparently normal tissue and early cancer evolution. This time, the team turned a cancer-research lens on autoimmunity—and found the two share the same underlying logic.
---
The Technical Breakthrough: NanoSeq
Why was this hypothesis only testable now? Because detection is so hard.
Conventional bulk sequencing can only detect clones at >5% frequency in tissue. But B-cell clones in the immune system are extremely diverse, each making up a tiny fraction of total cells. You need to detect mutations at <1% frequency.
NanoSeq is an ultra-accurate single-molecule sequencing protocol that can detect mutations in individual DNA molecules with very high accuracy. The team used it together with:
Cross-validation across these methods shows these mutations are not sequencing noise: they are real, functionally important, located in B cells, and self-reactive.
---
What It Means
Rethinking Disease
Autoimmune diseases may need to be redefined as "benign cancers"—not cells that proliferate without limit, but cells that acquire mutations enabling escape from immune control.
This means:
Implications for Treatment
Current autoimmune treatments (glucocorticoids, immunosuppressants) are "carpet bombing"—suppressing the entire immune system, at the cost of increased infection risk and severe side effects.
If disease is driven by specific mutant clones, then in principle one could:
Explaining the Cancer–Autoimmunity Connection
Clinically, links between autoimmunity and cancer have long been observed:
This study offers a unifying framework: both are products of somatic mutations escaping immune control—only the mutations occur in different cell types and lead to different phenotypes.
---
Limitations and Open Questions
The paper leaves several puzzles:
1. Source of mutations: How do these somatic mutations arise? Are they normal byproducts of AID (activation-induced cytidine deaminase) during B-cell activation, or accelerated by some environmental factor?
2. Initial trigger: Why do mutant clones first appear in the thyroid? Is thyroid tissue special in some way, or is it just the random wandering of B cells?
3. The role of T cells: The paper focuses on B-cell mutations, but do T cells carry similar mutations? TNFRSF14 is a B–T cell interaction checkpoint—are T-cell-side mutations also important?
4. Other autoimmune diseases: Do similar patterns exist in multiple sclerosis, rheumatoid arthritis, or type 1 diabetes?
5. Therapeutic feasibility: How can mutant clones be specifically eliminated without damaging normal immune function?
---
One-Sentence Takeaway
Autoimmune disease may not be the immune system "going haywire," but thousands of micro-cancers unfolding in the body—each one a B-cell clone that has escaped the immune brakes. This shift in perspective may redefine how we understand and treat these diseases.
---