Colibactin-associated mutations in the human colon appear to reflect anatomy and early exposure, not oncogenesis

This study demonstrates that colibactin-associated mutational signatures in the human colon are primarily driven by anatomical location (specifically the rectum) and early exposure rather than playing a causal or prognostic role in colorectal cancer oncogenesis.

The Gist

The Big Picture: A Mystery in the Colon

Imagine the human colon (large intestine) as a long, winding highway stretching from the right side of your belly to the left. Along this highway, there are tiny service stations called "crypts." These are the places where new cells are born to replace old ones.

For a long time, doctors and scientists have been worried about a specific type of cancer called Colorectal Cancer (CRC). Recently, this cancer has been showing up more often in younger people (under 50), and it seems to happen more often on the left side of the colon (the rectum and sigmoid colon) than the right.

Scientists suspected a specific culprit: a "bad actor" bacterium called E. coli that carries a toxic weapon called Colibactin. They thought this toxin was like a vandal that scratched up the DNA of the cells in the colon, causing them to turn into cancer. They believed that if you had this vandal, you were more likely to get cancer, especially at a young age.

The Experiment: Taking a Snapshot of the Highway

To test this theory, the researchers from the University of Utah decided to look at the "service stations" (crypts) before any cancer had even started.

• The Subjects: They studied 21 men. Some had perfectly healthy colons, and some had small, non-cancerous growths called polyps (which are like warning signs).
• The Method: Instead of looking at the whole highway at once, they used a high-tech laser (like a microscopic scalpel) to snip out individual service stations (crypts) from six different spots along the highway: from the start (cecum) to the end (rectum).
• The Goal: They sequenced the DNA of these tiny stations to see what kind of "scratches" (mutations) were already there. They wanted to see if the "vandal" (Colibactin) was more active in people who had polyps or were destined to get cancer.

The Findings: The Vandal is Everywhere, But Not Dangerous

Here is what they discovered, which turned the old theory upside down:

1. The Vandal is a Local Resident, Not a Criminal
They found that the "Colibactin scratches" (specific mutations in the DNA) were indeed present. However, they weren't more common in people with polyps or cancer. Instead, they were heavily concentrated in the rectum (the very end of the highway) for everyone, regardless of whether they were healthy or had polyps.

• The Analogy: Imagine that the rectum is a neighborhood where it rains a lot. Everyone living there gets wet (has these mutations) because of the rain (anatomy/environment), not because they are criminals. The fact that the rectum is "wet" doesn't mean the people there are more likely to commit crimes (get cancer).

2. The Damage Happened Early in Life
The researchers noticed that these specific scratches didn't pile up as people got older. They were there in young people and stayed the same in older people.

• The Analogy: It's like getting a sunburn when you were a child. You have that scar for the rest of your life, but it didn't happen because you stayed out in the sun yesterday. The "Colibactin vandalism" happens very early in life (perhaps in infancy) and then just sits there. It doesn't keep getting worse over time.

3. The "Cancer" Connection Was a Coincidence
Because the rectum is the place where early-onset cancer happens most often, and because the rectum is also the place where these "Colibactin scratches" are most common, scientists assumed the scratches caused the cancer.

• The Analogy: It's like noticing that all the people who get into car accidents on a specific curve of the road are wearing red hats. You might think the red hats cause the crashes. But in reality, everyone in that town wears red hats. The red hats (Colibactin) are just a common feature of that location, not the cause of the crash. The crash is caused by something else (like the sharp curve of the road or bad brakes).

The Conclusion: Rethinking the Risk

The study concludes that Colibactin is likely not the main driver of colon cancer.

• Old Theory: The bad bacteria attacks your cells, causing cancer.
• New Reality: The bad bacteria (or its toxin) is just a common guest in the rectum that leaves a harmless mark early in life. The fact that cancer happens there is due to other factors, not because of this specific bacterial mark.

Why does this matter?
For a long time, doctors have been trying to screen for this specific bacteria to predict who will get cancer. This study suggests that finding the bacteria doesn't mean you are at high risk.

Instead of looking for a "bad bacteria" to predict cancer, doctors should focus on the specific driver mutations (the actual "engine failures" in the cells) that turn a healthy cell into a cancer cell. The "Colibactin scratches" are just background noise—like static on a radio—that happens to be louder in one part of the colon, but they aren't the song playing.

Summary in One Sentence

The "Colibactin" bacteria leaves a specific mark on the DNA of the rectum in almost everyone early in life, but this mark is just a harmless souvenir of where you live in your body, not a warning sign that you are going to get cancer.

Technical Summary

1. Problem Statement

Colorectal cancer (CRC) is a leading cause of cancer death, with a rising incidence of early-onset CRC (diagnosed before age 50). A significant clinical mystery is the regional enrichment of early-onset CRC in the distal colon (left side/rectum).

• Current Hypothesis: The bacterial genotoxin colibactin (produced by pks+ Escherichia coli) has been proposed as a causal and prognostic factor for early-onset CRC. This hypothesis is supported by the observation that colibactin-associated mutational signatures (specifically SBS88 and ID18) are enriched in distal CRC tumors.
• The Gap: It remains unclear whether colibactin mutagenesis is a driver of carcinogenesis (accumulating over time or increasing in disease states) or if the observed enrichment in tumors is merely a reflection of regional anatomical differences or early-life exposure that persists without causing cancer. Previous studies lacked the resolution to distinguish between clonal (stem cell) and subclonal mutations in premalignant tissue across different anatomical regions.

2. Methodology

The study employed a high-resolution, region-specific approach to map mutational landscapes in the premalignant colon.

• Cohort: 21 male donors (ages ~30–70) undergoing colonoscopy.
  • Group 1: 11 donors with no polyps ("healthy").
  • Group 2: 10 donors with at least one adenomatous polyp ("at-risk").
  • Note: The study excluded CRC patients from the primary cohort but integrated public data from CRC patients (Lee-Six et al. and Cornish et al.) for comparison.
• Sample Collection: Biopsies were taken from six standardized anatomical regions: cecum, ascending, transverse, descending, sigmoid, and rectum.
• Experimental Workflow:
  1. Laser Capture Microdissection (LCM): Individual colon crypts (monoclonal, ~2,000 cells) were isolated to ensure genetic homogeneity.
  2. Whole-Genome Sequencing (WGS): 125 individual crypts were sequenced using Element Biosciences AVITI platforms with high-fidelity library preparation optimized for low DNA input.
  3. Bioinformatics:
     • Alignment to GRCh38.
     • Variant calling using FreeBayes and DeepSomatic (deep learning).
     • Strict filtering to remove germline variants, artifacts, and low-confidence calls.
     • Mutation Classification: Mutations were categorized as clonal (MAF > 0.2) or subclonal (MAF < 0.2).
  4. Signature Analysis: Extraction of mutational signatures (SBS and ID) using SigProfilerExtractor and comparison against COSMIC signatures.
  5. Motif Analysis: Direct identification of colibactin-associated SNVs and indels using sequence motifs derived from organoid studies (Huber et al.).
  6. Statistical Modeling: Generalized Linear Mixed Models (GLMM) with negative binomial distributions were used to model mutation burdens, adjusting for age, clinical group, and donor identity.

3. Key Contributions

• High-Resolution Premalignant Mapping: First study to perform WGS on individual crypts across six distinct anatomical regions in both healthy and polyp-bearing individuals, enabling the detection of subclonal mutations often missed in bulk tissue studies.
• Decoupling Anatomy from Disease: Demonstrated that mutational signatures associated with colibactin are anatomically specific (enriched in the rectum) but independent of disease state (present equally in healthy, polyp, and CRC-associated crypts).
• Refutation of Causality: Provided evidence that colibactin mutations are acquired early in life and do not accumulate progressively with age or disease severity, challenging the view that colibactin is a primary driver of early-onset CRC.

4. Key Results

A. General Mutational Landscape

• Accumulation Rate: Somatic mutations accumulated at ~23.7 SNVs/year, consistent with previous organ-level studies.
• Regional Differences: While total mutation counts were similar between the right and left colon, the mutation spectra differed significantly. The rectum showed a distinct enrichment of T>[A,C,G] mutations compared to other regions.
• Microsatellite Stability: All premalignant samples were microsatellite stable (MSS).

B. Colibactin Signatures (SBS88 and ID18)

• Regional Enrichment: Signatures SBS88 (single base substitution) and ID18 (indel) were significantly enriched in the rectum across all clinical groups (healthy, at-risk, and tumor data).
• No Disease-Specific Increase: Crucially, the burden of ID18 and SBS88 was statistically indistinguishable between:
  • Healthy crypts vs. Polyp-bearing crypts.
  • Premalignant crypts vs. CRC tumors.
  • Young vs. Old patients (when adjusting for total mutation burden).
• Clonality: ID18 and SBS88 were found primarily in clonal mutations (stem cell level) and did not accumulate in subclonal populations, suggesting an early-life exposure event rather than ongoing mutagenesis.

C. Motif-Based Analysis

• Direct counting of colibactin-associated sequence motifs confirmed that the absolute number of these mutations did not differ between clinical groups.
• Age Trajectory: Unlike clock-like signatures (e.g., SBS5), colibactin signatures did not correlate with age. Instead, they tracked with SBS89, a signature associated with early-life exposure.
• Proportion vs. Count: In younger patients, colibactin mutations appear as a higher proportion of the total mutational burden simply because the total burden is lower. This creates a statistical artifact where they appear "enriched" in early-onset cases, despite having the same absolute exposure as older adults.

5. Significance and Conclusion

• Paradigm Shift: The study argues against the hypothesis that colibactin is a causal driver of CRC or a prognostic marker for early-onset disease. Instead, the presence of colibactin mutations appears to be a common physiological event resulting from early-life colonization by pks+ E. coli, which persists in the distal colon (particularly the rectum) regardless of cancer status.
• Anatomical Specificity: The enrichment of early-onset CRC in the distal colon is likely driven by anatomical factors (e.g., higher baseline colibactin exposure in the rectum) rather than a unique susceptibility to colibactin-induced carcinogenesis.
• Clinical Implications:
  • Screening: Relying on aggregate colibactin signature burden to stratify risk may be misleading.
  • Future Directions: Clinical risk assessment should focus on identifying specific driver mutations within crypts rather than aggregate mutational signatures. Distinguishing between the carriage of colibactin-producing bacteria and the mutational footprint of past exposure is critical.
• Limitations: The study had a small sample size per group and did not sequence the polyps themselves, only the surrounding crypts. However, the integration of external datasets strengthens the generalizability of the findings regarding the lack of disease-state correlation.

In summary: Colibactin-associated mutations in the human colon are a marker of anatomical location and early-life exposure, not a specific driver of oncogenesis or a predictor of early-onset colorectal cancer.