Defining mutational signatures of lung cancer-associated carcinogens through in vitro exposure of human airway epithelial cells

This study establishes a physiologically relevant in vitro system using human airway epithelial cells and whole genome sequencing to characterize mutational signatures of lung carcinogens, successfully identifying a novel, distinct mutational pattern induced by NTCU exposure while finding no distinct signature for NNK.

The Gist

Imagine your DNA as a massive, intricate instruction manual for building and running a human body. Every time a cell divides, it makes a copy of this manual. Sometimes, the copying machine makes a typo. Usually, these typos are random and harmless, but sometimes, harmful chemicals (carcinogens) come along and smash the keyboard, forcing the machine to make very specific, predictable mistakes.

Scientists call these specific patterns of mistakes "mutational signatures." Think of a signature like a unique fingerprint left behind by a criminal. If you find a specific type of typo in a cancer cell, you can look at the "fingerprint" and say, "Ah, this cancer was caused by cigarette smoke," or "This one was caused by UV light from the sun."

However, there are many known cancer-causing chemicals where we don't know their fingerprint yet. This study set out to find the fingerprints of two specific culprits: NTCU and NNK, both of which are linked to lung cancer but have been mysterious in terms of exactly how they mess up our DNA.

The Experiment: A "Test Kitchen" for DNA Damage

Instead of waiting for people to get sick and studying their tumors (which is messy because tumors have accumulated many different mistakes over years), the researchers built a controlled test kitchen.

1. The Ingredients: They used human airway cells (the lining of your lungs) grown in a lab.
2. The Test: They exposed these cells to three different "ingredients":
   • BaP (Benzo(a)pyrene): A known villain found in cigarette smoke. This was the "control" to make sure their kitchen was working.
   • NTCU: A chemical that causes a specific type of lung cancer in mice, but whose human "fingerprint" was unknown.
   • NNK: Another tobacco-related chemical that causes a different type of lung cancer.
3. The Process: They let the cells grow, divide, and accumulate mistakes for four weeks. Then, they sequenced the entire DNA of the cells to read every single typo.

The Results: Who Left the Mess?

Here is what they found, translated into everyday terms:

1. The Known Villain (BaP): "The Master of Ceremonies"
When they exposed the cells to BaP, the DNA went wild with a very specific pattern of errors. The researchers found a "fingerprint" that matched perfectly with what we already know about cigarette smoke.

• The Takeaway: Their "test kitchen" works! They successfully recreated the known fingerprint, proving their method is accurate.

2. The Mystery Solved (NTCU): "The New Fingerprint"
When they exposed the cells to NTCU, the DNA didn't just get messy; it got messy in a brand new way.

• They found a distinct pattern of errors that had never been seen before in the global database of cancer fingerprints.
• The Analogy: Imagine you've only ever seen fingerprints with loops and whorls. Then, you find a fingerprint that looks like a perfect spiral. You know exactly who made it (NTCU), but you've never seen that spiral before.
• The Connection: This new "spiral" pattern looked almost identical to a pattern found in mice that were exposed to the same chemical. This is huge because it means we can finally say, "If we see this spiral in a human lung cancer, it was likely caused by NTCU."

3. The Disappointment (NNK): "The Ghost"
When they exposed the cells to NNK, nothing happened. The DNA looked exactly the same as the cells that weren't exposed to anything.

• The Analogy: It's like handing a suspect a weapon, but when you check the crime scene, there are no bullet holes.
• Why? The researchers suspect the "kitchen" wasn't set up right for this specific chemical. NNK is a "pro-carcinogen," meaning it's a sleeping giant. It needs a specific enzyme (a biological key) to wake up and become dangerous. The human cells in the lab didn't have enough of that specific key, so NNK stayed asleep and didn't leave a fingerprint.

Why Does This Matter?

This study is like upgrading the police database.

• Before: We knew some criminals (like cigarette smoke) left fingerprints, but for others (like NTCU), we had no idea what to look for.
• Now: We have a new, high-tech scanner that can identify the "NTCU spiral." If doctors find this pattern in a patient's tumor, they can trace the cancer back to its source, even if the patient doesn't remember being exposed to it.

In short: The researchers built a super-accurate lab model to catch DNA "fingerprint" makers. They confirmed their model works, discovered a brand-new fingerprint for a lung cancer chemical (NTCU), and learned that some chemicals (NNK) need a very specific setup to leave their mark. This helps us understand exactly how different environmental toxins cause cancer.

Technical Summary

1. Problem Statement

While the Catalogue of Somatic Mutations in Cancer (COSMIC) contains over 100 characterized mutational signatures, the specific causal patterns for many known carcinogens remain undefined, particularly in relevant human tissues.

• The Gap: Clinical history often lacks the resolution to pinpoint specific chemical exposures. While in vivo mouse models exist for certain carcinogens, they may not perfectly recapitulate human mutagenesis.
• Specific Targets: The study focuses on two potent lung carcinogens with undefined mutational footprints:
  • NTCU (N-nitrosotris-(2-chloroethyl) urea): Induces lung squamous cell carcinoma (SCC) in mice and mimics human SCC progression, but its mutational signature is unknown.
  • NNK (4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone): A tobacco-specific nitrosamine inducing lung adenocarcinoma in mice, yet it lacks a documented mutational signature.
• Goal: To develop a physiologically relevant in vitro system using human airway epithelial cells and whole-genome sequencing (WGS) to definitively characterize the mutational signatures induced by these agents.

2. Methodology

The authors developed a controlled in vitro platform combining human immortalized tracheobronchial epithelial cells (AALEs) with WGS.

• Cell Model: SV40 and hTERT-immortalized human airway epithelial cells (AALEs), which express cytochrome P450 enzymes (CYP1A1, CYP1B1) necessary for bioactivating carcinogens.
• Experimental Design:
  • Treatments: Cells were exposed for four weekly cycles to:
    • BaP (Benzo(a)pyrene): 5 μM (positive control, known to induce SBS4).
    • NTCU: 0.5 μM.
    • NNK: Two concentrations (0.01 μM and 1.0 μM).
    • Controls: Media, DMSO, and Acetone.
  • Metabolic Activation: BaP and NTCU treatments included liver microsomes and NADPH where required to simulate metabolic activation.
  • Clonal Expansion: Post-exposure, live single cells were sorted via flow cytometry, expanded into clones, and harvested for DNA extraction.
• Sequencing & Analysis:
  • WGS: Performed on the DNBseq platform (100bp paired-end, >30x depth) for 25 samples (22 treated + 9 controls).
  • Variant Calling: GATK Best Practices workflow (Mutect2) with a comprehensive "Panel of Normals" to filter out subclonal background mutations inherent to the cell line.
  • Signature Extraction:
    • SBS, DBS, INDEL: Mutational counts were extracted using the musicatk R package.
    • Decomposition: Non-negative Matrix Factorization (NMF) via SigProfilerExtractor and musicatk was used to discover de novo signatures.
    • Validation: Signatures were compared against COSMIC v3.0 and a recent in vivo mouse study (Gómez-López et al., 2025). A Bayesian workflow (BayesPowerNMF) was used for orthogonal validation.

3. Key Results

A. Validation of the Platform (BaP)

• The system successfully recapitulated the known mutational signature of Benzo(a)pyrene.
• SBS: Dominated by C>A mutations (specifically C[C>A]C context), matching COSMIC Signature SBS4 (cosine similarity = 0.92).
• DBS & INDEL: Identified specific double-base substitutions (CC>AA) and indels matching COSMIC DBS2 and ID3, respectively.
• Conclusion: The platform is robust and capable of detecting known carcinogen-specific signatures.

B. Discovery of a Novel NTCU Signature

• Mutation Burden: NTCU-exposed cells showed a significant increase in somatic SNVs compared to controls (Average ~9,735 mutations vs. ~2,925 in controls; adjusted p = 0.000097).
• Signature Profile:
  • A novel Signature 3 was identified, accounting for 82.3% of mutations in NTCU samples.
  • Characteristics: Present across all six SBS types but with an elevated probability of C>T mutations.
  • COSMIC Comparison: No strong match to existing COSMIC signatures (max cosine = 0.81), indicating a previously uncharacterized mutational pattern.
  • Cross-Species Validation: The in vitro human NTCU signature showed a remarkably high correlation (cosine = 0.95) with the aggregate mutational profile from the in vivo mouse model (Gómez-López et al.).
• DBS/INDEL: No distinct DBS or INDEL signatures were found for NTCU; activity in these categories was driven by background noise.

C. Negative Result for NNK

• Mutation Burden: NNK-exposed samples (both low and high dose) showed no significant increase in mutation counts compared to controls.
• Signature Profile: The mutational patterns of NNK samples were highly correlated with controls (cosine > 0.9) and lacked a distinct signature.
• Interpretation: The authors hypothesize this is due to metabolic limitations. NNK requires specific enzymes (CYP2A6/CYP2A13) for $\alpha$-hydroxylation to become mutagenic. The AALE cell line and liver microsomes used likely lacked sufficient CYP2A activity to bioactivate NNK effectively in this in vitro setting.

D. Computational Validation

• Application of BayesPowerNMF confirmed the three primary signatures (BaP, NTCU, Background) and identified a fourth, low-abundance signature (SBS18-like) present in controls, likely representing in vitro oxidative damage artifacts.

4. Key Contributions

1. Novel System: Established a robust, human-relevant in vitro platform using AALEs and WGS to define causal links between environmental exposures and mutational signatures.
2. New Signature Discovery: Defined a distinct, novel mutational signature (Signature 3) specifically associated with NTCU, a major inducer of lung squamous cell carcinoma. This signature was validated against in vivo mouse data.
3. Methodological Rigor: Demonstrated the necessity of using a comprehensive "Panel of Normals" to filter subclonal background mutations in cell-line-based studies and validated findings using orthogonal Bayesian modeling.
4. Clarification of NNK: Provided evidence that the lack of a detectable NNK signature in this study is likely due to metabolic activation constraints rather than an inherent lack of mutagenicity, highlighting the need for tailored metabolic systems for specific nitrosamines.

5. Significance

• Etiology of Lung Cancer: This study provides a definitive mutational "fingerprint" for NTCU, aiding in the understanding of lung squamous cell carcinoma etiology and potentially helping to identify NTCU-like exposures in human tumors.
• Toxicology & Risk Assessment: The platform offers a high-throughput, controlled alternative to animal testing for characterizing the mutagenic potential of unknown or poorly characterized carcinogens.
• Precision Medicine: By linking specific environmental exposures to distinct genomic footprints, this work supports the development of better biomarkers for cancer prevention and the interpretation of tumor genomes in clinical settings.
• Limitation Insight: The study underscores the critical importance of matching metabolic activation capabilities (specific CYP enzymes) to the specific carcinogen being tested in in vitro models.