Long-Read RNA Sequencing Reveals Hidden Radiation Signatures in Human Blood Cells

When ionizing radiation strikes the human body, it leaves molecular fingerprints in the blood long before any physical symptom appears. A research team at the University Medical Center of the Johannes Gutenberg University Mainz has now read those fingerprints with unprecedented resolution, using long-read RNA sequencing to capture how individual white blood cells rewrite their transcriptome after a radiation dose. The study, published in BMC Genomics, is the first to systematically investigate differential transcript usage in the leukocyte fraction of irradiated human whole blood, and it exposes regulatory events that conventional gene-level analyses simply cannot see.

The experiment was deliberately simple in design but technically ambitious in execution. Whole peripheral blood from three healthy donors was exposed to 4 Gy of X-rays, a dose comparable to what might be encountered in a serious radiation accident or in the upper range of clinical radiotherapy fractions, while sham-irradiated samples from the same donors served as controls. Six hours after exposure, the researchers isolated the leukocytes and profiled their RNA using Oxford Nanopore Technologies’ long-read platform. Unlike short-read sequencers that chop RNA molecules into fragments of a hundred or so bases and then computationally stitch them back together, nanopore sequencers read entire RNA transcripts from end to end, preserving the full-length structure of each isoform.

At the gene level, the analysis confirmed what decades of radiation biology have established. Compared with the sham-irradiated controls, 117 genes were significantly upregulated and 66 were downregulated, and the enriched pathways were the canonical ones: DNA damage repair and inflammatory responses. Genes involved in detecting double-strand breaks, signaling their presence, and orchestrating their repair were switched on, while the cell’s routine metabolic housekeeping was dialed back. This validation matters because it demonstrates that the long-read approach recovers the same radiation-response signatures that biodosimetry researchers have relied on for years, meaning the new method is not a departure from established science but an extension of it.

The real surprises emerged when the team descended from genes to transcripts. A single gene can produce multiple RNA isoforms through alternative splicing, and these isoforms can be regulated independently of one another. The long-read data revealed 102 transcripts that were significantly upregulated and 17 that were downregulated, a pattern of isoform-specific regulation that was completely masked at the gene level. The most striking example was IL32, an inflammatory cytokine gene that showed no significant change when measured as a whole gene, yet two of its transcript isoforms were strongly upregulated, with log2 fold-changes of 1.19 and 1.99. In other words, radiation was selectively boosting particular molecular versions of IL32 while the gene-level average diluted the signal into statistical invisibility.

IL32 is not a random bystander. It is deeply involved in inflammatory and immune signaling, and its isoform-specific induction after irradiation suggests that the immune reaction to radiation damage is more finely tuned than gene-level studies imply. The researchers also identified changes in transcript usage for WDR74, ITM2B, AK2, and RPS19, meaning that irradiation altered which isoform of these genes the cells preferentially produced rather than how much total gene product was made. WDR74 participates in DNA damage response signaling, AK2 is a kinase with roles in apoptosis and energy metabolism, and RPS19 is a ribosomal protein essential for red blood cell production, so these usage shifts point to radiation rewiring not just expression levels but the functional architecture of the transcriptome itself.

Beyond confirming known signatures and revealing hidden regulation, the long-read approach enabled genuine discovery. The team identified novel candidate isoforms that had never been annotated in reference databases, including protein-coding isoforms of PHPT1 and FBXO22 as well as several non-coding candidates. PHPT1 encodes a phosphohistidine phosphatase, and FBXO22 is part of the ubiquitin ligase machinery that tags proteins for degradation, a process with known links to DNA damage response. The newly found FBXO22 isoform showed evidence of altered protein domain architecture relative to the reference isoform, meaning that if this isoform is translated, it could produce a protein with a different functional structure than the canonical version. To assess coding potential, the researchers ran the candidates through tools including TransDecoder, the Coding Potential Calculator 2, and CPAT, and used SQANTI3 for structural quality annotation of the novel isoforms, a standard workflow for separating real transcripts from sequencing artifacts.

The clinical and biodosimetric implications of this work are substantial. Biodosimetry, the estimation of absorbed radiation dose from biological markers, is critical after radiological accidents when physical dosimeters are absent, and gene expression profiling of blood is one of the most promising approaches because the transcriptomic response scales with dose. The Mainz study shows that transcript-level markers can carry information that gene-level markers miss, potentially improving the sensitivity and specificity of dose estimates. The IL32 finding is a case in point: a biodosimetric panel built only on gene-level data would overlook a radiation-responsive signal that isoform-level analysis captures clearly.

Portability adds another dimension. Oxford Nanopore sequencers are compact enough to be carried to the field, which opens the possibility of on-site gene expression studies after radiation emergencies, in disaster zones, or in military settings where samples cannot be shipped to a central laboratory. The authors note that the diffusion of long-read RNA sequencing in radiation biology has been limited despite these advantages, and their study provides a systematic demonstration of what the technology can extract from irradiated blood. The workflow, from whole blood exposure through leukocyte isolation to full-length transcript profiling, could serve as a template for larger biodosimetric studies.

Radiotherapy is another arena where the findings could prove useful. Evaluating the transcriptomic radiation response of blood provides insight into absorbed dose, hematotoxicities, and immune reactions, and transcript-level resolution could help clinicians understand why some patients develop severe immune reactions or blood toxicity while others tolerate the same dose. Isoform-specific markers might eventually distinguish inflammatory signatures that predict toxicity from those that reflect routine tissue response, refining the personalization of radiation treatment.

The study has clear limitations that the authors acknowledge through their design choices. With only three donors, the analysis captures individual variability but cannot establish population-level biomarkers, and the single 4 Gy dose at a single 6-hour time point leaves the dose-response and time-course dimensions unexplored. The researchers addressed donor effects with leave-one-out sensitivity analyses and compared their long-read results against short-read pipelines, finding that 34 differentially expressed genes were shared between approaches while others were platform-specific. Still, the conceptual advance stands: radiation-induced transcriptomic alterations include isoform-level regulatory events and novel candidate isoforms that are masked in gene-level analyses and conventional short-read pipelines. As long-read sequencing matures, the molecular record of radiation exposure may prove far richer, and far more clinically actionable, than the gene-level view has allowed.

Subject of Research: Radiation-induced transcriptomic signatures in human leukocytes revealed by long-read RNA sequencing

Article Title: Decoding radiation-induced transcriptomic signatures of human leukocytes using long-read RNA-seq: clinical and biodosimetric implications

Article References: Salah, A., Schmidberger, H., Marini, F., & Zahnreich, S. (2026). Decoding radiation-induced transcriptomic signatures of human leukocytes using long-read RNA-seq: clinical and biodosimetric implications. BMC Genomics. https://doi.org/10.1186/s12864-026-13475-3

Image Credits: AI Generated

DOI: 10.1186/s12864-026-13475-3

Keywords: long-read RNA sequencing, biodosimetry, ionizing radiation, transcriptomics, Oxford Nanopore, human leukocytes, isoform discovery, BMC Genomics, Decoding, radiation-induced, transcriptomic, signatures