DNMT3A is one of the principal enzymes responsible for establishing DNA methylation patterns, the chemical marks that help determine which genes are switched on or off. Mutations in the gene encoding DNMT3A are among the most frequent alterations in acute myeloid leukemia and other hematologic malignancies, and they have also been observed in solid tumors. Yet the clinical behavior of DNMT3A-mutant cancers has often appeared contradictory. In some settings, loss of DNMT3A seems to sensitize tumors to immune attack; in others, it appears to help malignant cells escape immune surveillance. The new screening effort was designed to resolve this paradox by asking, at whole-genome scale, what happens when DNMT3A and its interaction partners are perturbed inside tumors that are actively fighting off an immune response.

The researchers used CRISPR-based genetic screens conducted in vivo, a technique that allows scientists to disable thousands of genes simultaneously in tumor cells and then observe which perturbations change the tumors’ fate inside a living animal with an intact immune system. This approach is far more faithful to real tumor biology than conventional in vitro screens because it captures the full complexity of the tumor microenvironment, including cytotoxic T cells, natural killer cells, macrophages, and the inflammatory signals that coordinate an immune response. Tumor cells carrying a library of gene-targeting guide RNAs were implanted into immunocompetent hosts, and the guide RNAs that became enriched or depleted over time revealed which genes the tumors depended on to survive immune pressure.

The results pointed repeatedly to DNMT3A and the molecular machinery surrounding it. When the screens interrogated tumors under strong immune selection, loss of DNMT3A function emerged as a route through which tumor cells could reshape their epigenetic landscape in ways that dampen immune recognition. The data indicate that DNMT3A normally helps maintain the expression of molecules that the immune system uses to identify and eliminate cancerous cells. When that maintenance is lost, the resulting methylation changes can silence antigen presentation components and interferon-responsive genes, effectively lowering the tumor’s visibility to patrolling cytotoxic lymphocytes. In this role, DNMT3A functions as a guardian of immunogenicity, and its absence gives tumors a survival advantage in the face of immune attack.

At the same time, the screens revealed a second, opposing role. In other contexts, intact DNMT3A activity appeared to support immune evasion through a different mechanism: the enzyme’s methyltransferase function can reinforce transcriptional programs that suppress inflammatory signaling within the tumor. By methylating and repressing specific regulatory regions, DNMT3A can help keep certain immune-stimulatory genes quiet, allowing tumor cells to persist even when inflammatory cytokines such as interferon gamma are present in the microenvironment. The dual functions mean that the consequence of altering DNMT3A is not straightforward. Whether restoring or inhibiting the enzyme benefits the patient depends on which of its activities dominates in a given tumor and which immune pressures that tumor faces.

This duality has immediate implications for the growing number of clinical strategies that target DNA methylation. Hypomethylating agents such as azacitidine and decitabine are already used to treat myelodysplastic syndromes and certain leukemias, and there has been intense interest in combining these drugs with immune checkpoint inhibitors, on the theory that demethylation can unmask tumor antigens and make tumors more visible to T cells. The new findings refine that logic. If DNMT3A loss itself can drive immune evasion by silencing antigen presentation, then simply erasing methylation marks may not uniformly increase immunogenicity. Instead, the specific genes reactivated or repressed by demethylating treatment, and the baseline status of DNMT3A in the tumor, could determine whether combination immunotherapy helps or hurts.

The screening platform also identified genes that cooperate with DNMT3A in these immune-related functions, mapping a network of epigenetic regulators, interferon pathway components, and antigen presentation machinery that together control the tumor-immune interface. Perturbing members of this network produced effects that converged on the same phenotypes seen with DNMT3A disruption, strengthening the conclusion that DNA methylation homeostasis is a central determinant of immune recognition. Notably, the in vivo setting was essential for detecting many of these effects. Parallel analyses in culture conditions did not fully reproduce the immune-dependent phenotypes, underscoring how microenvironmental pressure sculpts which genetic dependencies actually matter in cancer.

For patients with DNMT3A-mutant malignancies, the work offers a framework for thinking about why responses to immunotherapy vary so widely. Clonal hematopoiesis, the age-related expansion of blood cells carrying DNMT3A mutations, affects a large fraction of older adults and is increasingly recognized as a factor that modifies cancer risk and treatment outcomes. The new results suggest that the same epigenetic rewiring that characterizes DNMT3A-mutant clones may influence their visibility to the immune system, with potential consequences for both cancer development and the efficacy of immune-based therapies. Measuring the immune-evasion programs identified in the screens could help stratify patients and predict who is most likely to benefit from checkpoint blockade or epigenetic drug combinations.

Technically, the study demonstrates the power of pooled CRISPR screening performed inside living hosts as a discovery engine for cancer immunology. Rather than cataloging gene expression differences after the fact, in vivo screens impose real selective pressure and let the organism reveal which genes determine survival. The enrichment and depletion patterns of guide RNAs, analyzed by sequencing, provide a genome-wide functional map of immune evasion that would be nearly impossible to assemble through observation alone. The strategy also highlights the value of testing epigenetic regulators under immune pressure, since their roles are inherently context-dependent and often invisible in simplified experimental systems.

The researchers emphasize that the dual functions of DNMT3A open multiple therapeutic avenues. In tumors where DNMT3A loss silences immune recognition, interventions that restore antigen presentation or reconstitute interferon signaling could resensitize cancer cells to immunotherapy. In tumors where DNMT3A activity actively suppresses inflammatory genes, inhibiting the enzyme’s methyltransferase function might unleash the immune-stimulatory programs the tumor has been holding in check. Distinguishing between these states will require biomarkers that report on the enzyme’s downstream methylation targets, an effort the screening data now make more concrete. As epigenetic therapies continue to merge with cancer immunotherapy, the lesson from this study is that the epigenetic state of a tumor is not merely background biology; it is an active, manipulable determinant of whether the immune system can see and destroy the cancer, and DNMT3A sits at the fulcrum of that decision.

Subject of Research: In vivo CRISPR screening of DNMT3A's dual roles in tumor immune evasion

Article Title: In vivo CRISPR screens identify dual functions of DNMT3A in mediating tumor immune evasion

Article References: In vivo CRISPR screens identify dual functions of DNMT3A in mediating tumor immune evasion. (n.d.). https://doi.org/10.1038/s41467-026-77495-z

Image Credits: AI Generated

DOI: 10.1038/s41467-026-77495-z

Keywords: DNMT3A, CRISPR screens, tumor immune evasion, DNA methylation, immunotherapy, epigenetics, Nature Communications, antigen presentation, interferon signaling, cancer genetics, in vivo screening, tumor microenvironment