Are eccDNAs Apoptotic Products? Innate Immunostimulatory Activity and Experimental Interpretation

If your innate immune readouts light up when you add purified circular DNA to cells, what exactly are you seeing—functional extrachromosomal circles or a signature of cell death? The claim "eccDNAs are apoptotic products with high innate immunostimulatory activity" has energized the field and, at times, muddied interpretation. In this guide, we take a balanced stance: some eccDNAs do arise during apoptosis and can strongly activate cGAS–STING, while others originate via active biogenesis (including larger ecDNAs) with different localization and regulatory behaviors. Getting the distinction right is essential for experimental design, artifact control, and any downstream mechanistic claims.

This article emphasizes primary experimental evidence, then translates it into practical workflows and analysis rules that immunology labs can implement. We also include schematics you can reproduce in lab meetings and methods parameters that make your results more auditable.

Apoptosis vs Biogenesis: What Exactly Are We Measuring?

Apoptosis produces a characteristic "DNA ladder" due to nucleosomal fragmentation at ~180–200 bp intervals. On a 1–2% agarose gel, you'll see evenly spaced bands (monomer, dimer, trimer, and so on). In contrast, eccDNA preparations that start from intact, high‑molecular‑weight (HMW) genomic DNA typically lack this periodic ladder; the input DNA remains near the well or as a diffuse HMW smear before exonuclease enrichment. This crude but powerful check helps you avoid starting from material already dominated by apoptotic debris.

• Apoptotic ladder assay references show nucleosome‑sized repeats in classic gels; an improved non‑enzymatic assay was described by Suman and colleagues in 2011, providing a visual baseline for laddering behaviors in diverse contexts.
• Size distributions of detected eccDNAs vary by source. Plasma often shows peaks around ~202 and ~338 bp with 10‑bp sub‑periodicity consistent with nucleosome phasing, while cancer cell lines and tissues contain broader distributions, including kilobase‑to‑megabase circles linked to gene amplification.

Localization adds another layer. Biogenesis linked to genome instability frequently involves micronuclei and nuclear hubs; cGAS can surveil ruptured micronuclei and cytosolic DNA, creating fertile ground for innate activation if circles escape to the cytosol. Large ecDNAs, however, tend to be nuclear and participate in transcriptional reprogramming rather than floating freely in the cytoplasm.

If you want a baseline for "healthy" eccDNA patterns beyond disease models, compare size profiles and complexity with eccDNAs in non‑malignant contexts. For additional context on profiles outside cancer, see our companion article on eccDNA in somatic systems and aging, which discusses distribution and potential functions across tissues: eccDNA in Somatic Cells and Aging Research: What We Know and How to Profile It.

Figure 1. Gel schematic comparing apoptotic ladder (left) and HMW input for eccDNA enrichment (right).

eccDNAs Are Apoptotic Products with High Innate Immunostimulatory Activity: When the Claim Holds—and When It Doesn't

The innate DNA sensor cGAS binds double‑stranded DNA without strict sequence requirements, synthesizes the second messenger cGAMP, and activates STING on the endoplasmic reticulum to trigger TBK1 and IRF3/7, culminating in type I interferon and ISG expression. Two conditions matter for circular DNA to engage this pathway: availability in the cytosol and physical form that supports cGAS oligomerization. Circularity can protect DNA ends and, depending on size and protein coating, potentially enhance persistence in the cytosol.

Primary experimental evidence supports the idea that apoptosis can be a rich source of immunostimulatory circles. Wang and colleagues reported that eccDNAs formed via apoptotic fragmentation and ligation are potent cGAS–STING stimulants; they increased upon apoptosis induction and required circularity for maximal activity. Mechanistic components included DNase γ and DNA ligase III in formation, while STING‑dependent responses (e.g., IFN‑β transcription, IRF3 phosphorylation) confirmed pathway engagement; see the original experimental evidence in the 2021 Nature paper by Wang et al. in the Nature study on apoptotic eccDNAs and innate activation.

Separately, replication stress and micronuclei rupture provide another route to cytosolic DNA exposure. Chan and colleagues showed that eccDNA associated with genome instability promotes inflammation through cGAS–STING in models of chronic tissue injury; their Science Advances work connects eccDNA abundance with immune infiltration signatures and highlights size diversity—from hundreds of base pairs to kilobase‑scale and beyond. For details, see Chan et al. 2025 on eccDNA-driven inflammation via cGAS–STING.

Context still dictates outcome. Large ecDNAs tied to oncogene amplification often reside in the nucleus and contribute to transcriptional regulation rather than innate activation, while small apoptotic circles or micronucleus‑leaked fragments are more likely to engage cGAS. Compartmentalization, DNA‑binding proteins, and packaging into extracellular vesicles further modulate immunogenicity.

Figure 2. cGAS–STING activation by cytosolic circular DNA; cytosolic availability and circularity are central to signaling potency.

There is also conceptual overlap with tumor microenvironments, where cell death, replication stress, and micronuclei are common. For a broader discussion of circles in oncology and how they reshape gene regulation alongside potential immune signaling intersections, see: eccDNA in Cancer: Gene Amplification, Oncogene Regulation, and Research Applications.

Experimental Interpretation and Controls: Avoiding Apoptosis-Driven Misreads

You can reduce misinterpretation by baking apoptosis checks and innate pathway specificity into your workflow from sample preparation to readout.

Wet‑lab controls to separate origins

• Limit apoptosis during culture or handling. Where biologically acceptable, include apoptosis limiters or reduce stressors; quantify Annexin V/PI positivity and caspase activity before DNA extraction to benchmark cell health.
• Confirm DNA integrity with gels before enrichment. Prefer HMW inputs without laddering. If laddering is present, document and decide whether the experiment explicitly studies apoptotic eccDNAs.
• Use linear DNA depletion and validate it. ATP‑dependent exonucleases (e.g., Plasmid‑Safe) digest linear DNA. Refresh ATP and enzyme for dense samples; validate depletion with qPCR of linear genomic markers and recovery of circular spike‑ins.
• Amplify circles with RCA cautiously. Phi29‑based RCA boosts signal but can bias size distributions; record incubation time and conditions and consider parallel libraries with and without RCA when material allows.
• Validate circularity and junctions. Enzymatically linearize controls, perform restriction digests to show expected mobility shifts, and sequence across junctions. Long‑read platforms help resolve longer circles and confirm full‑length structures.

Innate immune assay specificity

• Confirm cGAS–STING dependence. Use cGAS‑ or STING‑deficient cells, or pathway inhibitors (e.g., RU.521 for cGAS), and read out TBK1/IRF3 phosphorylation and IFN‑β/CXCL10 expression.
• Rule out endotoxin/TLR artifacts. Test DNA preparations by LAL assay; treat with polymyxin B or use endotoxin‑removal columns. Include TLR4‑deficient or MyD88‑deficient cells where relevant.
• Include nuclease sensitivity controls. Treat DNA preparations with DNase and heat‑inactivated DNase controls to demonstrate DNA‑dependent stimulation.

Workflow and method resources

• For an end‑to‑end view of enrichment and library preparation steps, see our practical guide: Experimental Workflow for eccDNA Sequencing: Enrichment, Library Prep, and Common Pitfalls.

Bioinformatic filters matter as much as wet‑lab controls. Size‑based heuristics, mapping quality thresholds, and junction support cutoffs reduce the risk that an immune‑active small‑circle population is misread as evidence of active biogenesis. We outline recommended thresholds in the Bioinformatics section below, and a more detailed discussion appears in: Bioinformatics for eccDNA: Detection Algorithms, Filtering Artifacts, and Reporting Standards.

Several independent groups have adapted Circle‑seq–style workflows, which strengthens method portability but highlights that formal multi‑center benchmarking remains limited. For example, Chan et al. adapted eccDNA isolation for liver tissue and linked eccDNA abundance to inflammation in their Science Advances study (2025) (Chan et al. 2025). Likewise, Lin et al. applied eccDNA sequencing pipelines to placenta and plasma (Frontiers in Genetics, 2023), demonstrating cross‑tissue utility while formal multi‑lab validation efforts are still needed.

Worked Example A: Apoptosis‑Induced eccDNAs and Immune Readouts

Scenario: Primary cells exposed to a pro‑apoptotic stimulus show ~20–40% Annexin V positivity. Pre‑enrichment gels reveal laddering. After exonuclease digestion and RCA, short‑read libraries detect a dominant 150–400 bp circle population and robust IRF3 phosphorylation in stimulated dendritic cells.

Interpretation: The size distribution, laddering at input, and cGAS‑dependent readouts point to apoptosis as a major source of immunostimulatory circles. The correct path is to label this explicitly as apoptosis‑enriched eccDNA, include cGAS/STING genetic controls, and avoid conflating with active biogenesis.

Controls to add next time: Reduce apoptosis during handling; add synthetic circular spike‑ins to estimate absolute recovery; perform long‑read validation to test for larger circles; add LAL tests and polymyxin B treatment to exclude LPS artifacts.

Worked Example B: ecDNA Amplification Without Innate Activation

Scenario: A cancer cell line with known ecDNA‑driven oncogene amplification (identified by FISH and long‑read sequencing) yields large, kilobase‑to‑megabase circles located in nuclear hubs. Cytosolic DNA staining is minimal; immune cells exposed to purified nuclear eccDNA fraction do not show significant IFN‑β induction.

Interpretation: Large nuclear ecDNAs primarily act as regulatory scaffolds and replication units rather than innate agonists. Absence of cGAS–STING activation is consistent with nuclear confinement and possibly protein coating that prevents cytosolic sensing. This reinforces that not all eccDNAs are immunostimulatory, and origin/compartment matter.

Figure 3. Immunofluorescence schematic showing cGAS co‑localization with cytosolic DNA foci and micronuclei.

Bioinformatics Heuristics that Separate Debris from Data

Sequencing alone won't tell you origin, but disciplined filters raise confidence and make results comparable across labs.

• Junction support. Require at least 3 independent long‑read supports per predicted junction (or 5–10 for short‑read junctions) when coverage permits. Report per‑junction support counts in deliverables.
• Mapping quality. Enforce MAPQ ≥30–60 depending on aligner and platform to reduce spurious split reads.
• Size‑bin logic. Treat <200 bp as apoptotic‑enriched when interpreting innate activation unless the goal is explicitly to study apoptosis; prioritize >1 kb candidates as potential ecDNA when paired with nuclear localization data.
• Exclusions. Remove mitochondrial and plasmid reads; cross‑check with plasmid spike‑ins used for process control.
• Junction validation. Randomly sample junctions for PCR/Sanger confirmation, especially for high‑value findings.
• Reporting standards. Provide a per‑sample data dictionary: raw FASTQs, aligned BAM/CRAM, a junction table with genomic coordinates, size estimates, support counts, QC metrics, and a README with methods and software versions. Use community pipelines when possible (e.g., ECCsplorer, ecc_finder, nf‑core/circdna) and record exact parameters.

Reproducibility package (download)

We provide a reproducibility starter package with parameter checklists, an example data‑dictionary template, and an nf‑core/circdna profile to help audit and rerun analyses. Download or inspect:

• nf‑core/circdna pipeline (examples & docs): https://github.com/nf-core/circdna
• nf‑core/configs (institution profiles): https://github.com/nf-core/configs

These resources speed reproducibility, standardize deliverables, and provide an auditable nf‑core profile for local or cloud runs.

Methods Appendix: Practical Parameters for Enrichment and Sequencing

Pipeline reproducibility note — brief benchmarking and notebook pointers

Several pipelines have public benchmarks and reproducible executables but few ready‑to‑run notebooks. In comparative studies, ecc_finder showed high sensitivity and speed on short/long reads (Zhang et al., 2021, Frontiers), while ECCsplorer produces consensus sequences for non‑model species but is resource‑intensive (Mann et al., 2022, BMC Bioinformatics). For workflow reproducibility use nf‑core/circdna (Nextflow profile and release notes: https://nf-co.re/circdna/1.0.4/). If you require a runnable Jupyter/R notebook for inter‑pipeline comparison, we recommend exporting pipeline outputs to a standardized BED/CSV table and running a light notebook to compute size‑stratified concordance (junction support, false‑positive suppression) as described in the cited methods.

Below is a compact set of parameters and options that have worked well in published methods and lab practice. Adapt to your system and include proper controls.

Methods Appendix — Registered protocols and recommended citations

For reproducibility, follow and cite canonical, peer‑reviewed protocols where possible. Representative citable methods: JoVE's Circle‑seq purification protocol (JoVE, 2016; DOI: 10.3791/54239) for cell‑ and tissue‑derived eccDNA prep, and CIDER‑Seq long‑read processing (Mehta et al., Nat Biotechnol 2020; DOI: 10.1038/s41587-020-0528-7) for deconcatemerization and full‑length validation. For the ATP‑dependent linear‑DNA depletion step, cite the Circle‑seq/Nature Protocols workflow (see the 2022 Circle‑seq methods PDF) and consult community entries such as the scCircle‑seq protocol on protocols.io (https://www.protocols.io/view/sccircle-seq-unveils-the-diversity-and-complexity-q26g7m2w1gwz) for implementation notes. Suggested citation formats: "JoVE. Genome‑wide purification of eccDNA. JoVE (2016). doi:10.3791/54239"; "Mehta D. et al. CIDER‑Seq. Nat Biotechnol (2020). doi:10.1038/s41587-020-0528-7"; and protocol.io links as web‑citable methods (include version/date). Adding these protocol DOIs/links improves authority by pointing readers to stepwise, versioned procedures and facilitates method adoption and cross‑lab reproducibility.

Enrichment and validation

• Linear DNA depletion: ATP‑dependent exonuclease digestion at 37°C; refresh ATP and enzyme every 12–24 h for complex samples; validate with qPCR reduction of linear markers (>10× reduction) and recovery of circular spike‑ins.
• RCA amplification: Phi29 polymerase at 30°C for 16–72 h; record time and input mass; consider parallel libraries without RCA to assess bias when input allows.
• Circularity checks: Restriction digest to linearize known circles, mobility shift on gels, and junction PCR across predicted breakpoints.
• Long‑read confirmation: ONT or PacBio for >1 kb circles; apply deconcatemerization (e.g., CIDER‑Seq) to reconstruct full‑length units.

Library prep and platforms

• Short‑read (Illumina): Useful for junction discovery and size distributions in 150–600 bp ranges; aim for ≥30–50 million reads per sample when surveying low‑abundance circles.
• Long‑read (ONT/PacBio): Critical for resolving longer circles and complex repeats; target N50 sufficient to span >1–5 kb; basecalling and polishing parameters should be reported.

Neutral, methods‑focused example of outsourcing specific steps (disclosure)

• Disclosure: The following example references CD Genomics as a methods resource. We do not make clinical claims, and all details should be validated in your lab context.
• Example: For teams that prefer to outsource linear‑DNA depletion and library construction while retaining control of immune assays, some labs combine in‑house sample prep and QC with vendor‑executed Circle‑seq style enrichment and sequencing. When selecting platforms, review the Genomics services overview to match short‑ and long‑read options to your size targets. If longer circles are expected, long‑read validation (e.g., ONT/PacBio) after RCA can be requested so full‑length sequences are available for junction confirmation and size‑bin analyses.

Extended Practical Guidance: Quantification, Controls, and Study Design Notes

Absolute quantification and normalization

• Spike‑in circles: Introduce synthetic circular DNAs of known copy number during extraction to estimate absolute recovery and to normalize across batches. Report recovery as copies per million cells and discuss RCA bias when applicable.
• Digital PCR: Use ddPCR for quantifying specific eccDNA junctions or ecDNA loci; report absolute copies per cell alongside sequencing‑derived size distributions.
• Coverage and depth: For short‑read surveys of low‑abundance circles, ≥50 million reads per sample improves junction support in noisy backgrounds; for long‑read validation, target sufficient N50 and per‑circle read supports to reconstruct full units.

Contamination and specificity controls

• Endotoxin/LPS: Run LAL assays on DNA preparations and include polymyxin B treatment where needed. Pair this with pathway specificity controls (e.g., STING‑KO) to confirm cGAS–STING dependence rather than TLR‑mediated artifacts.
• Mitochondrial and plasmid DNA: Linearize mtDNA before exonuclease steps if contamination is expected; exclude plasmid reads bioinformatically and confirm with known spike‑in plasmids.
• Nuclease controls: Benzonase or DNase treatments can demonstrate DNA dependence of the immune readout; include heat‑inactivated enzyme controls and buffer‑only controls.

Study design considerations

• Compartment and context: If your biological question concerns innate activation, enrich and verify cytosolic availability (e.g., fractionation and imaging); if your question concerns regulation, prioritize nuclear ecDNA profiles and transcriptional impacts.
• Replication stress models: Use hydroxyurea or other cell‑cycle manipulations cautiously; monitor micronuclei and cGAS localization to interpret co‑occurring innate signals.
• Reporting: Pre‑register QC thresholds (size bins, MAPQ, junction support) and acceptance criteria for deliverables to ensure reproducibility across collaborators.

Conclusion: Ensuring Your eccDNA Signal Isn't Just a Marker of Cell Death

Here's the deal: eccDNAs encompass a spectrum—from small, apoptosis‑derived circles that readily trigger cGAS–STING to large, nuclear ecDNAs that remodel gene regulation without necessarily activating innate immunity. The balanced stance is to interpret your data through origin, size, and compartment. Build apoptosis checks into sample handling, remove linear DNA with validated exonuclease protocols, verify circularity and junctions, and apply disciplined bioinformatic thresholds. Confirm innate pathway specificity and rule out endotoxin artifacts before attributing immune activation to circular DNA.

If you're planning an enrichment and sequencing campaign, align platform choices to your size targets and include orthogonal validation. For method background and options, review the Genomics services overview and the Circle‑seq method explainer linked above. You can maintain complete scientific control by defining acceptance criteria (size‑bin thresholds, junction support, QC metrics) up front and auditing vendor deliverables against them.

References:

1. Wang Y. et al. eccDNAs are apoptotic products with high innate immunostimulatory activity. Nature. 2021. https://doi.org/10.1038/s41586-021-04009-w
2. Chan L.K. et al. Extrachromosomal circular DNA promotes inflammation through the cGAS–STING pathway. Science Advances. 2025. https://doi.org/10.1126/sciadv.adw0272
3. Yang H. et al. cGAS is essential for cellular senescence. PNAS. 2017. https://doi.org/10.1073/pnas.1705499114
4. Kumar P. et al. ATAC‑seq identifies thousands of extrachromosomal circular DNA in cancer and cell lines. Science Advances. 2020. https://doi.org/10.1126/sciadv.aba2489
5. Sin S.T.K. et al. Identification and characterization of extrachromosomal circular DNA in human plasma. PNAS. 2020. https://doi.org/10.1073/pnas.1914949117
6. Mann L. et al. ECCsplorer: a pipeline to detect extrachromosomal circular DNA. BMC Bioinformatics. 2022. https://doi.org/10.1186/s12859-022-04589-9
7. Zhang P. et al. ecc_finder: A robust and accurate tool for detecting eccDNA. Frontiers in Plant Science. 2021. https://doi.org/10.3389/fpls.2021.743742
8. Fang M. et al. eccDNA‑pipe: an integrated pipeline for eccDNA analysis. Briefings in Bioinformatics. 2024. https://doi.org/10.1093/bib/bbae034
9. Ablasser A. cGAS in action: expanding roles in immunity and beyond. Science. 2019. https://doi.org/10.1126/science.aat8657
10. JoVE Protocol. Genome‑wide purification of extrachromosomal circular DNA. 2016. https://doi.org/10.3791/54239
11. Zhao Y. et al. Extrachromosomal circular DNA: Current status and future prospects. eLife. 2022. https://doi.org/10.7554/eLife.81412
12. Turner K.M. et al. Extrachromosomal oncogene amplification drives tumour evolution. Nature. 2017. https://doi.org/10.1038/nature21356
13. Kim H. et al. Extrachromosomal DNA is associated with oncogene amplification in cancer. Nature Genetics. 2020. https://doi.org/10.1038/s41588-020-0678-2
14. Bafna V. et al. Extrachromosomal DNA in Cancer. Annual Review of Genomics and Human Genetics. 2022. https://doi.org/10.1146/annurev-genom-120821-100535
15. Wu S. et al. Extrachromosomal DNA: An Emerging Hallmark in Human Cancer. Annual Review of Pathology. 2022. https://doi.org/10.1146/annurev-pathmechdis-051821-114223
16. Schwarz H. et al. Residual endotoxin in recombinant proteins. PLoS ONE. 2014. https://doi.org/10.1371/journal.pone.0113840
17. Tani T. Anti‑endotoxin properties of polymyxin B‑immobilized fibers. Adv Exp Med Biol. 2019;1145:321–341. doi: 10.1007/978-3-030-16373-0_19. (Full text: https://pmc.ncbi.nlm.nih.gov/articles/PMC7123644/)