iiinova — where your digital self keeps evolving

Finally, I'd like to thank our collaborators in the Yilmaz and Breault labs. As well as team PROSPECT, a Cancer Grand Challenges partnership between @cancergrand.bsky.social, NCI/NIH, the Bowelbabe Fund for Cancer Research UK, and the French National Cancer Institute! (18/18)

These persistent epigenetic alterations also create the potential for developing new diagnostic tools. Can we measure "memory" in patient populations and quantify cancer risk before a tumor forms? (17/18)

This work creates a new framework for thinking about how environmental exposures and experiences shape the health of our tissues over a lifetime. We've built concepts and tools to now expand into studying long-term effects of diet, metabolism and more. (16/18)

Altogether, this created a model where inflammation: → creates epigenetic memory → which is clonally propagated, even when the inflammation subsides and tissue looks normal to the eye → and once an oncogenic mutation occurs... → memory enhances tumor growth (15/18)

We found tumors in colitis recovered mice had higher expression of regenerative AP-1 associated genes. Consistently, inhibition of AP-1 blocked tumor growth specifically in colitis-recovered mice. This told us epigenetic memory is functionally required for the phenotype (14/18)

Finally, we went back to cancer. After APC loss, mice that had recovered from cancer had larger tumors. This was true even at the earliest stages of tumorigenesis when lesions were microscopic. This told us → memory increases initial tumor outgrowth (13/n)

This suggested cooperativity between the AP-1 and FOX families. Using a novel footprinting-based in vitro binding assay, we showed FOX factors stabilize AP-1 on DNA. This suggested a mechanisms for maintaining AP-1 memory at specific sites (12/n)

We asked how this general TF family (AP-1) was able to create a specific memory in our system. Using the seq2PRINT method developed by our lab, we found composite binding motifs between AP-1 and FOX factors. (11/n)

Hence, some forms of epigenetic memory could be propagated clonally and create hetergeneous stem cell lineages. Interestingly, we found some of this could be explained by changes in DNA methylation. But not all. So, we took a deeper dive into mechanisms of AP-1 memory. (10/n)

We then developed a novel computational framework to evaluate whether motif accessibility for a given TF was clonally inherited. This revealed that for certain TF motif families, cells within a clone looked significantly more like one another than chance (including AP-1). (9/n)

We took this one step further - is memory clonal? Passed from a mother cell to a daughter cell? To answer this, we developed SHARE-TRACE, allowing for simultaneous measurement of: - clonal lineage history - gene expression - chromatin accessibility All in a single cell. (8/n)

Is this memory intrinsic to stem cells? Organoids derived from colitis tissue demonstrate a hyperproliferative/regenerative phenotype. This told us memory is encoded within stem cells, not dependent on niche. (7/n)

100

The strongest signal was in a progressive gain in AP-1 motif accessibility. This accumulates throughout injury, with strongest signal after inflammation resolves. Critically, this memory is durable through proliferation (>100 days post-DSS) and without elevated Fos protein. (6/n)

Using SHARE-seq (joint RNA + ATAC in single cells), we find stem cells in mice recovered from colitis are epigenomically distinct from controls. This told us memory is encoded in chromatin accessibility. (5/n)

We found ~97% of gene expression changes during colitis injury returned to baseline following recovery from colitis. This told us there wasn't meaningful transcriptional memory in our model. However, this wasn't the case in the chromatin (4/n)

We modeled this using chronic DSS colitis and examined acute and chronic injury, as well as following recovery from colitis. By 21 days post-DSS: ✔️body weight restored ✔️crypt architecture normal ✔️ immune infiltration resolved → tissue appears fully recovered (3/n)

Many cancer risk factors are episodic (inflammation, infection, diet, etc), yet risk can persist. Mutations don’t fully explain this. Instead, we hypothesized that prior exposure induces a stable epigenetic state in stem cells that alters future tumorigenesis (2/n)

Epigenetic memory of colitis promotes tumour growth - Nature

A little late to Bluesky but my postdoc work w/ @jbuenrostro.bsky.social now out in @nature.com "Epigenetic memory of colitis promotes tumour growth" www.nature.com/articles/s41... We wanted to understand how transient inflammation can create an increase in cancer risk, even after full recovery 🧵

nature.com

Colonic stem cells retain a memory of inflammation following disease resolution and there is a mechanistic link between chronic inflammation and malignancy, suggesting potential strategies to mitigate...

3429

Surya Nagaraja @snaga13.bsky.social · 26/02/2025

We do see AP-1 go up and nuclear receptors like HNF and PPAR go down in colitis-associated adenomas though. So their oppositional nature seems somewhat similar to the program described and to be amplified following colitis.

010

Surya Nagaraja @snaga13.bsky.social · 26/02/2025

The answer is a bit mixed. We see a lot of these fetal/embryonic programs get activated during injury (particularly in the healing phases) but they don’t seem to be remembered following tissue recovery. Rather, the role of AP-1 in non-neoplastic memory seems to be maintaining a more mature program

Surya Nagaraja @snaga13.bsky.social · 25/02/2025

That’s right. For generally measuring the clonality of different features, we did it across all clones. There’s some analysis in the Extended Data showing differences between the conditions. Thanks!

040

We’d like to thank our collaborators in the Yilmaz and Breault labs, and PROSPECT @cancerresearchuk.org / NCI for their support. Find out about binding partners, drugging AP-1, the new tools we made and more here at our pre-print: tinyurl.com/ColitisNagar... (14/14)

041

With rates of early onset colon cancer and IBD rising, our findings also create opportunities for novel diagnostics to track cancer risk (tracking epigenetic memory in stool?) and nominate therapeutic targets development to prevent malignancy (AP-1 inhibition?). (13/14)

We find a central role for the AP-1 in oncogenic memory of inflammation. Given the diversity of processes regulated by this family, we suspect that diverse environmental factors (diet, infection, aging) may all accumulate through this pathway to influence tissue health. (12/14)

100

Our findings suggest a new model, where: - Chronic inflammation creates heritable and clonal epigenetic memory - This primes stem cells for increased expression of pro-proliferative programs - Therefore, lowering the threshold for oncogenesis once a mutation occurs. (11/14)

131

Spatial RNA-seq on >200 individual tumors revealed that epigenetic memory of colitis primes tumors for higher AP-1 associated gene expression. Again, we find a remarkably similar proportion of tumors with high activation and that also show additional oncogenic programs. (10/14)

100

Looking at the earliest stages of microscopic tumor formation, we found that epigenetic memory promoted initial tumor outgrowth. This suggested that grossly larger tumors occurred due to the clustering of many microscopic tumors in close proximity to one another. (9/14)

100

What about cancer? We induced adenoma formation via APC loss after recovery from colitis. 🚨Colitis-recovered mice developed significantly larger tumors than controls🚨. However, we didn’t see any differences in proliferation or initiation. So why are they bigger? (8/14)

We found a subset of clones to display exceptionally strong memory of inflammation, resembling what we saw in vivo with stem cells. This raises an intriguing possibility that colitis selectively expands certain stem cell populations with altered epigenetic states. (7/14)

100

To study this further, we developed SHARE-TRACE, a method that enables simultaneous profiling of gene expression, chromatin accessibility, and clonal history in single cells. This enabled us to find that epigenetic memory is propagated through clonal lineages. (6/14)

210

We next asked whether this memory was stored in the stem cells themselves. Using an ex vivo organoid culture system, we found this epigenetic memory to be cell-intrinsic and promote growth. Colitis-derived organoids were both larger and hyperproliferative. (5/14)

100

Once inflammation resolved and the epithelium healed, gene expression returned to normal (>97% of genes). However, we found that stem cells stored an “epigenetic memory” of damage in their chromatin! Highlighted by a cumulative increase in AP-1 motif accessibility. (4/14)

220

We focused on one of the most well-established connections - ulcerative colitis and colorectal cancer. Using a mouse model of colitis, we performed SHARE-seq to track chromatin accessibility and gene expression in single cells over the course of disease and recovery. (3/14)

130

How does inflammation fuel cancer? Chronic inflammation has long been connected to cancer, but the molecular mechanisms are unclear. We hypothesized that inflammation and damage create lasting epigenetic alterations in stem cells, reducing the threshold for oncogenesis. (2/14)