Extrachromosomal circular DNA is exactly what the name suggests: small rings of DNA that exist outside the chromosomes, looping free of the linear genome. Unlike chromosomal DNA, these circles can carry intact gene fragments and regulatory elements, and because they replicate independently of the normal cell cycle, they can be amplified rapidly in numbers. That property makes eccDNA a potent engine of oncogene overexpression, a mechanism that has been implicated in drug resistance in several cancers, including notoriously amplified circles carrying the MYC oncogene in neuroblastoma and other tumors. Yet in ovarian cancer, the interplay between eccDNA and the epigenetic machinery that controls gene activity has remained poorly charted territory.

The epigenetic machinery in question centers on histone deacetylases, or HDACs, enzymes that remove acetyl chemical groups from histone proteins around which DNA is wound. When histones are highly acetylated, DNA is more accessible and genes are more active; HDACs strip those acetyl groups away, tightening chromatin and silencing genes. In many cancers, HDAC activity is dysregulated, contributing to the abnormal silencing of tumor suppressors and the miswiring of gene-control networks. HDAC inhibitors, including the drug panobinostat, also known by its code name LBH589, were designed to reverse that silencing and restore normal gene expression patterns in tumor cells.

Panobinostat is a broad-spectrum HDAC inhibitor that has been tested in various hematologic and solid malignancies. The new study set out to ask a question that is often overlooked in drug development: what happens to the eccDNA landscape when a tumor is treated with an epigenetic drug? The answer, according to the team’s experiments, is that the drug does far more than simply reactivate silenced genes. It induces a significant remodeling of the eccDNA population within ovarian cancer cells, with a notable enrichment of circles derived from messenger RNA genes and from long noncoding RNAs, a class of regulatory RNA molecules with their own roles in cancer biology.

To connect this eccDNA remodeling to gene activity, the researchers integrated two powerful genome-wide datasets. The first came from RNA sequencing, which measures the complete set of RNA transcripts produced by the cells and therefore reveals which genes are switched up or down under treatment. The second came from chromatin immunoprecipitation sequencing, or ChIP-seq, targeted at HDAC3, one of the key histone deacetylase enzymes. ChIP-seq identifies the precise genomic locations where HDAC3 binds, marking the genes it directly regulates. By overlaying these maps with the eccDNA data, the team identified a set of 52 oncogenes that are linked to eccDNA, bound by HDAC3, and consistently upregulated by panobinostat treatment.

That convergence of three independent molecular signals, eccDNA presence, HDAC3 binding, and treatment-induced upregulation, is what gives the finding its weight. It suggests that the drug’s inhibition of HDAC3 does not uniformly suppress oncogenic potential. Instead, in at least a subset of genes, blocking HDAC3 appears to coincide with an adaptive activation program in which oncogenes carried or associated with circular DNA become more highly expressed. The authors describe this as an adaptive oncogenic response, a kind of molecular counterpunch thrown by the tumor as the epigenetic pressure of the drug takes hold.

The clinical relevance of these 52 genes was tested using survival analysis across multiple cancer cohorts. The team found that alterations in these oncogenes correlated with worse disease-free survival, meaning that patients whose tumors carried changes in these genes tended to remain cancer-free for shorter periods after treatment. This kind of population-level correlation does not prove causation on its own, but it flags the gene set as a meaningful marker of aggressive disease and provides a rationale for digging deeper into individual members of the group.

One member stood out. LMO2, a gene encoding a transcriptional regulator best known for its roles in blood cell development and, historically, as a gene disrupted in certain leukemias, emerged from the analysis as a novel oncogene in ovarian cancer. The mechanistic work behind this claim is technically detailed but conceptually clear. Supplementary analyses showed that panobinostat increases the levels of H3K27ac, an acetyl mark on histone H3 that signals active regulatory regions, at the promoter-proximal regions of the LMO2 gene, without altering HDAC3 binding at those same sites. In other words, the drug does not move HDAC3 off the DNA; it changes the chemical balance of chromatin in a way that favors LMO2 activation.

Functional experiments reinforced the picture. In TOV-112D, an ovarian cancer cell line used in the study, inhibiting HDAC3 with drugs or knocking the enzyme down with interfering RNAs both led to increased LMO2 expression. Conversely, when the researchers knocked LMO2 down, the cells became more sensitive to panobinostat. That last result is the most consequential: it implies that LMO2 upregulation is not merely a byproduct of treatment but a functional contributor to reduced drug sensitivity. A tumor that can ramp up LMO2 in response to HDAC inhibition may thereby cushion itself against the drug’s effects, and removing that cushion restores vulnerability.

The broader significance of the study lies in its demonstration that epigenetic therapy and the eccDNA layer of cancer genetics are deeply intertwined. HDAC-associated epigenetic regulation, the authors conclude, is connected to eccDNA-linked oncogene activation, and this connection may modulate how ovarian cancer patients respond to HDAC inhibitors. For clinicians, the work suggests a potential combination strategy: pairing HDAC inhibitors with approaches that suppress LMO2 or its downstream effects could prevent the adaptive escape that otherwise undermines treatment. For researchers, the study provides a template for integrating Circle-seq, RNA-seq, and ChIP-seq data to find similar vulnerabilities in other tumor types. The work, supported by funding from the National Natural Science Foundation of China and provincial programs, was conducted with approval from the Animal Welfare Committee of Fudan University Shanghai Cancer Center, and the authors report no competing interests. As the field moves toward therapies that target the nonchromosomal genome, LMO2 now stands as a concrete, testable target in the fight against ovarian cancer’s most stubborn defense mechanisms.

Subject of Research: The interplay between extrachromosomal circular DNA, HDAC3-mediated epigenetic regulation, and oncogene activation in ovarian cancer

Article Title: Integration of eccDNA and transcriptomic landscapes uncovers LMO2 as a therapeutic target in ovarian cancer

Article References: Han, T., Yan, Q., Bu, H., Gan, Y., Shen, X., Li, K., Guan, L., Dong, W., Li, P., Zhao, M., Zhou, X., Gao, B., & Hao, Q. (2026). Integration of eccDNA and transcriptomic landscapes uncovers LMO2 as a therapeutic target in ovarian cancer. Journal of Ovarian Research. https://doi.org/10.1186/s13048-026-02296-3

Image Credits: AI Generated

DOI: 10.1186/s13048-026-02296-3

Keywords: eccDNA, ovarian cancer, LMO2, HDAC3, panobinostat, epigenetics, oncogene, drug resistance, transcriptomics, ChIP-seq, therapeutic target, Journal of Ovarian Research