When the Zika virus swept through the Americas in 2015 and 2016, it revealed a devastating capacity to damage the developing human brain, causing microcephaly and the constellation of features now known as congenital Zika syndrome. A new study published in Genomics & Informatics takes a fresh look at how different cell populations in fetal neural tissue respond transcriptionally to Zika virus exposure, and it arrives at a carefully hedged but informative conclusion: the progenitor cells that build the cortex mount the most prominent interferon-associated responses, and the two viral strains examined left distinguishable descriptive patterns in the data.
The study, led by Zahra Abedi and colleagues, is not a new experiment. It is a secondary, descriptive reanalysis of publicly available single-cell RNA sequencing data, principally a dataset generated by Stokes and coworkers and deposited under accession GSE238140. That dataset contains human fetal brain-derived neural tissue exposed ex vivo to one of four conditions: a mock treatment, the Brazilian Zika virus strain ZIKV-BR, the Cambodian 2010 Asian-lineage FSS13025 isolate referred to as ZIKV-FSS, or recombinant interferon-beta. Because the original dataset included only one biological sample per condition, the authors repeatedly and explicitly frame their findings as exploratory strain-associated patterns rather than definitive evidence of strain-dependent mechanisms.
After quality control filtering, the analysis retained 27,267 fetal neural cells out of 30,045 analyzed, distributed across the mock condition (5,222 cells), ZIKV-BR exposure (7,814 cells), ZIKV-FSS exposure (7,966 cells), and interferon-beta stimulation (6,265 cells). Unsupervised clustering resolved seven major neural and glial populations: progenitor-enriched radial glia-like cells, cycling neural progenitor cells, excitatory neurons, inhibitory neurons, astrocytes, oligodendrocyte progenitor cells, and microglia. The radial glia-like population was the largest, with 10,501 cells, followed by cycling neural progenitor cells at 5,459 and excitatory neurons at 5,670. A small oligodendrocyte progenitor population contained just 56 cells.
The central descriptive finding concerns where the antiviral response was strongest. Progenitor-enriched radial glia-like cells and cycling neural progenitor cells showed strong transcriptional interferon and interferon-stimulated gene responsiveness and a higher number of genes meeting the study’s exploratory differential expression criteria compared with several other cell types. Neuronal populations, by contrast, showed fewer transcriptional changes. The authors used cell-level Wilcoxon rank-sum tests to rank genes, applying thresholds of an absolute log2 fold change of at least 0.25, a false discovery rate below 0.05, and detection in at least 10 percent of cells in one of the two groups being compared. They stress that these P values and adjusted values should not be read as formal condition-level statistical inference, because cells from the same sample are not independent biological replicates and cannot estimate biological variance.
A crucial interpretive caveat runs through the entire paper. The analysis did not separate virus-positive cells from virus-negative bystander cells, and it did not directly measure interferon-beta production, viral RNA burden, infection frequency, or interferon receptor-mediated paracrine signaling. Consequently, the authors interpret interferon-stimulated gene induction in progenitor populations as transcriptional interferon responsiveness rather than evidence of cell-intrinsic innate immune initiation. This framing is consistent with the mechanistic picture from the original dataset’s creators, in which astrocytes were highlighted as interferon-beta-producing sentinel cells while interferon-stimulated gene expression in progenitors may reflect responsiveness to interferon made by other cells rather than autonomous antiviral signaling.
When the two viral strains were compared, exploratory analyses suggested strain-associated differences in response structure. In selected progenitor-focused and interferon-stimulated gene summaries, the patterns linked to ZIKV-FSS showed more descriptive similarity to interferon-beta-associated profiles, while ZIKV-BR exposure was linked to broader and more variable gene expression changes across lineages. The canonical interferon-stimulated gene heatmap, built from a curated panel of eleven classical antiviral genes including IFITM1, MX1, OAS1, OAS2, OAS3, IFIT1, IFI6, ISG15, STAT1, IRF7, and RSAD2, showed visually overlapping interferon-related patterns between ZIKV-FSS and interferon-beta conditions in the progenitor-focused contrasts. Both viral strains and interferon-beta stimulation produced overlapping antiviral transcriptional patterns overall, and Zika exposure was associated with detectable induction of antiviral-associated transcripts in progenitor populations even without exogenous cytokine stimulation.
One of the study’s most methodologically interesting contributions is its sensitivity analysis of ZIKV-interferon-beta transcriptional similarity. In the main descriptive analysis, based on genes present in the exploratory differential expression tables, log2 fold-change correlations between Zika-associated and interferon-beta-associated responses were high in progenitor-enriched radial glia-like and cycling neural progenitor populations. But the authors found that these correlations strongly depended on the gene universe used. Correlations calculated from the primary differential expression table genes remained high even after excluding the canonical interferon-stimulated gene panel, indicating that filtering to differentially expressed genes can overestimate apparent similarity. When correlations were recalculated across an expanded all-expressed and tested gene universe, Pearson correlations were lower and Spearman correlations were much more conservative, particularly for non-interferon-stimulated genes. Canonical interferon-stimulated genes showed strong concordance between viral exposure and cytokine stimulation, whereas broader gene sets gave a more conservative estimate of alignment. The authors conclude that Zika-associated and interferon-beta-associated responses share broad transcriptional alignment but should not be interpreted as functionally equivalent or as evidence of a single interferon-beta-driven program.
To place the single-cell findings in a wider context, the team analyzed two additional public datasets, explicitly warning that these were not validations of fetal neural lineage-specific responses. The first was a bulk RNA sequencing dataset from human cerebral organoids exposed to mock or Zika conditions, drawn from the RNA-seq SubSeries GSE97919 within the GSE104279 SuperSeries associated with Watanabe and colleagues. In that dataset, principal component analysis of variance-stabilized expression values showed partial separation between Zika-exposed and control samples, canonical interferon-stimulated genes showed coordinated upregulation after exposure, and differential expression analysis identified induction of established antiviral genes in infected samples. The second contextual dataset, GSE230571, contained Zika-infected Vero epithelial cells and primary human monocyte-derived dendritic cells. Interferon-stimulated gene module scoring showed strong activation in dendritic cells with only minimal induction in Vero cells, consistent with known differences in innate immune competence between these cell types.
The authors also acknowledge several limitations beyond the single-sample-per-condition design. The ex vivo exposure conditions may not fully represent the complexity of Zika infection in vivo. The progenitor-enriched radial glia-like cluster was not subdivided into apical radial glia, outer or basal radial glia, or intermediate progenitor subtypes, so findings for that cluster should be read as mixed progenitor-enriched patterns rather than subtype-specific radial glia responses. Differences in species, cell-type composition, experimental design, infection context, and measurement modality also limit direct comparisons across the fetal neural, cerebral organoid, Vero, and dendritic cell datasets. The analysis does not determine whether the observed strain-associated differences reflect viral genotype, infection efficiency, viral RNA burden, donor or tissue effects, cell-type composition, sample handling, or other experimental factors.
Within those boundaries, the study offers a lineage-resolved descriptive reference of antiviral transcriptional states in developing human neural tissue and a demonstration of how gene-set selection shapes conclusions about transcriptional similarity. The observation that progenitor populations respond strongly to exogenous interferon-beta as well as to viral exposure supports their characterization as interferon-responsive populations rather than demonstrated initiators of cell-intrinsic antiviral signaling. For researchers studying congenital Zika syndrome, the work underscores both the value of reanalyzing public single-cell resources with standardized cell-type comparisons and the caution required when biological replication is limited. The analysis scripts, curated sample metadata, gene lists, and documentation have been deposited in a public GitHub repository, and all datasets analyzed are available through the Gene Expression Omnibus.
Subject of Research: Antiviral transcriptional responses to Zika virus in developing human neural tissue
Article Title: Exploratory strain-associated patterns of antiviral transcriptional responses to Zika virus exposure in developing human neural tissue
Article References: Abedi, Z., Sheikh Beig Goharrizi, M. A., Abbasi, A., Soleimani Zakeri, N. S., Jangi, H., & Susanabadi Farahani, A. (2026). Exploratory strain-associated patterns of antiviral transcriptional responses to Zika virus exposure in developing human neural tissue. Genomics & Informatics, 24(1), Article 14. https://doi.org/10.1186/s44342-026-00076-5
Image Credits: AI Generated
DOI: 10.1186/s44342-026-00076-5
Keywords: Zika virus, single-cell RNA sequencing, neural progenitors, interferon-stimulated genes, fetal brain, radial glia, ZIKV-BR, ZIKV-FSS, transcriptomics, congenital Zika syndrome, gene expression, secondary analysis
News Source: Juliet Wilcox. (October 29, 2026). Progenitor cells dominate antiviral gene responses to Zika in fetal neural tissue. Scienmag.
