High EBV DNA in Children’s Blood Linked to Immune Imbalance and Secondary Infections

Most people carry the Epstein–Barr virus quietly for life. The herpesvirus, first visualized in 1964 in cultured lymphoblasts from Burkitt lymphoma patients, establishes a lifelong latent infection in memory B cells and is normally held in check by a vigilant immune system. In some children, however, plasma levels of viral DNA climb far above baseline, a laboratory finding that has long been associated with more severe clinical courses but whose immunological underpinnings have remained murky. A new study published in 3 Biotech by Shasha Dai, Peng Wang, and Liming Cao now offers a detailed transcriptomic portrait of what distinguishes children with high plasma EBV-DNA loads from those with low loads, and it points to a striking pattern: innate immune hyperactivation running in parallel with weakened adaptive cellular immunity.

The research team analyzed peripheral blood samples from 34 children with high plasma EBV-DNA load and 31 children with low plasma EBV-DNA load, all recruited at the Children’s Hospital of Nanjing Medical University with ethical approval and written informed parental consent. Rather than relying on a single measurement, the investigators deployed a multi-layered analytical pipeline. Transcriptomic profiling of whole blood identified 459 differentially expressed genes between the two groups. Weighted gene co-expression network analysis, or WGCNA, was then used to group these genes into modules whose expression patterns correlate with clinical traits, a technique that helps separate coordinated biological programs from noise in high-dimensional gene expression data.

The pathway-level results were unambiguous. In the high-load group, gene sets involved in antiviral defense and inflammatory signaling were significantly upregulated, while pathways governing adaptive immunity were suppressed. This combination, sometimes described as a scorch-and-starve immune signature, suggests that the innate arm of the immune system was firing at full intensity, flooding the circulation with interferon-driven and pro-inflammatory signals, even as the T-cell compartment that normally delivers targeted, virus-specific killing appeared functionally diminished. Such a dissociation between innate and adaptive responses is a recognized feature of several chronic viral infections and has been linked to immune exhaustion in the EBV literature.

To move from bulk gene expression toward cellular composition, the team applied CIBERSORT-based deconvolution, a computational method that estimates the proportions of different immune cell types within a mixed sample by matching expression signatures to reference profiles. The estimates indicated higher proportions of neutrophils and monocytes and lower proportions of CD8-positive T cells and natural killer cells in the high-load group. The authors are careful to note that these deconvolution estimates should not be interpreted as direct cell counts; they are inferences drawn from gene expression, and their accuracy depends on the quality of the reference signatures. Nevertheless, the direction of the shift is biologically coherent. Prior work has shown that EBV can infect and induce apoptosis in human neutrophils and can impair dendritic cell development by promoting apoptosis of monocyte precursors, mechanisms that could plausibly distort myeloid-lymphoid balance during active viral replication.

The most clinically consequential part of the study concerns secondary infections. Children with high viral loads are suspected to be more vulnerable to additional pathogens, but identifying who is at risk has been difficult. The researchers therefore integrated multiple feature-selection strategies, combining LASSO regression, which shrinks coefficients to eliminate weak predictors, with Random Forest modeling, an ensemble method that ranks variables by their contribution to classification accuracy. This convergent approach prioritized three immune-related genes: STAT1, CXCL10, and IL6. All three sit squarely within the interferon and inflammatory axes. STAT1 is the canonical transcription factor downstream of interferon receptors, CXCL10 is an interferon-inducible chemokine that recruits activated T cells and monocytes, and IL6 is a pleiotropic pro-inflammatory cytokine central to acute-phase responses.

Expression of all three candidate markers was positively associated with plasma EBV-DNA load, and each was expressed at higher levels in children who developed secondary infections than in those who did not. When the three genes were combined into a single model, the three-gene signature discriminated secondary-infection status within the cohort with an area under the receiver operating characteristic curve, or AUC, of 0.85. An AUC of 0.85 indicates good but not exceptional discriminative performance, and the authors explicitly frame the model as exploratory pending independent validation. This caution is warranted. The cohort was small, the model was developed and evaluated within the same patient group, and biomarker signatures of this kind frequently lose performance when tested on external populations. Still, the finding suggests a practical path toward a blood-based test that could flag children at heightened risk of secondary infection before complications arise.

The bioinformatic predictions were then tested experimentally. Quantitative real-time PCR confirmed the elevated expression of the candidate genes, enzyme-linked immunosorbent assays measured the corresponding protein-level changes, and flow cytometry provided direct immunophenotyping of circulating lymphocyte subsets. The validation results were directionally consistent with the computational findings, strengthening the overall picture, although the authors emphasize that the study demonstrates association rather than causation. Whether immune dysregulation permits viral reactivation and secondary infection, or whether high viral burden itself drives the immune perturbation, cannot be resolved from a cross-sectional design. Longitudinal cohorts that track children from initial infection through viral-load dynamics would be needed to disentangle cause from consequence.

The findings sit within a broader and rapidly evolving literature on EBV immunopathology. Beyond infectious mononucleosis, EBV has been implicated in a growing list of conditions, including multiple sclerosis, systemic autoimmune diseases, and several malignancies ranging from Burkitt lymphoma to nasopharyngeal carcinoma and NK/T-cell lymphomas. Research into the JAK/STAT signaling pathway, which STAT1 anchors, has shown that the virus both activates and manipulates this axis to its advantage, and recent reviews have explored how EBV-specific humoral and cellular responses shape disease outcomes. The new study adds a pediatric dimension to this work, highlighting that the balance between interferon-driven inflammation and adaptive immune competence may be a decisive variable in how children handle the virus.

From a translational standpoint, the three-gene signature of STAT1, CXCL10, and IL6 is attractive because all three markers are measurable with routine laboratory techniques, and the study’s combination of transcriptomics, computational deconvolution, machine learning, and orthogonal wet-lab validation offers a template for biomarker development in other infection settings. The authors report no conflicts of interest and note that the study received no specific external funding. The data supporting the findings are available from the corresponding author upon request. For clinicians managing children with high plasma EBV-DNA loads, the study does not yet change practice, but it sharpens the questions worth asking: which patients will progress to secondary infection, and can early immune profiling identify them in time to intervene.

What remains to be established is equally clear. Independent cohorts, ideally from multiple geographic regions, must confirm the performance of the three-gene model before it can be considered a clinical decision-support tool. The cellular composition estimates need confirmation by direct flow cytometric quantification in larger samples, and the mechanistic link between innate hyperactivation, adaptive immune suppression, and susceptibility to secondary pathogens requires experimental dissection. If those steps succeed, the study could mark an early milestone in moving EBV management beyond viral-load monitoring toward genuinely immune-informed risk stratification, turning a longstanding laboratory observation into actionable clinical insight for the youngest and most vulnerable patients.

Subject of Research: Host immune dysregulation associated with high plasma Epstein–Barr virus DNA load and secondary infection in children

Article Title: Host immune dysregulation associated with high plasma EBV-DNA load in children and its potential association with secondary infection

Article References: Dai, S., Wang, P., & Cao, L. (2026). Host immune dysregulation associated with high plasma EBV-DNA load in children and its potential association with secondary infection. 3 Biotech, 16(10), Article 407. https://doi.org/10.1007/s13205-026-05039-9

Image Credits: AI Generated

DOI: 10.1007/s13205-026-05039-9

Keywords: Epstein-Barr virus, EBV-DNA load, children, immune dysregulation, secondary infection, transcriptomics, STAT1, CXCL10, IL6, CIBERSORT, biomarker, innate immunity