Adult T-cell leukemia/lymphoma (ATL), an aggressive blood cancer caused by the human T-lymphotropic virus 1 (HTLV-1), has long frustrated clinicians with high relapse rates after chemotherapy, immunotherapy and even allogeneic stem cell transplantation, the only potentially curative intervention currently available. A new study published in EMBO Molecular Medicine now reports that infected and malignant cells carry a distinctive metabolic signature centered on a single mitochondrial enzyme, ferredoxin reductase (FDXR), and that this signature tracks disease severity, predicts patient survival and points to repurposed drugs that may exploit it.

The research team, led by senior author Iart Luca Shytaj and including collaborators in the United Kingdom, Brazil, Japan and elsewhere, combined metabolomic and proteomic profiling of lymphoid cell models to hunt for cellular markers dysregulated by HTLV-1 that could be targeted therapeutically. They compared three HTLV-1-infected or ATL-derived cell lines (MT-2, MT-4 and SP) with three uninfected lymphoid lines (Jurkat, MOLT-4 and CEM-ss), analyzing cell lysates in triplicate for metabolomics and in quadruplicate for proteomics. Metabolomic pathway enrichment analysis identified multiple mitochondrial pathways as significantly enriched in infected cells, with the electron transport chain, part of oxidative phosphorylation, emerging as the most significantly enriched pathway overall. Metabolites belonging to these pathways were among the most significantly altered and were predominantly upregulated, with the exception of acetyl-CoA.

Proteomic analysis told a consistent story. Filtering differential expression results to proteins with known mitochondrial localization using the MitoCarta repository, the researchers found FDXR to be the most significantly upregulated mitochondrial protein in infected cells. FDXR contributes to lipid homeostasis, iron-sulfur cluster biogenesis and redox regulation, and depends on flavin adenine dinucleotide (FAD) as a cofactor, a metabolite the team found significantly increased in infected cells. The finding was confirmed by qPCR, western blotting of isolated mitochondria and immunofluorescence across a panel of sixteen cell lines as well as activated and resting primary CD4 T cells and peripheral blood mononuclear cells. Two ATL models, MT-1 and ED cells, showed elevated FDXR mRNA but not protein, a discrepancy the authors note. Reanalysis of RNA-Seq data from 259 blood cancer cell lines in the Human Protein Atlas confirmed the pattern: the three HTLV-1-infected lines in that dataset (Hut-102, MJ and C8166) ranked among the top five overall for FDXR expression, a permutation test p value of 3.2e-06.

To test physiological relevance, the team reanalyzed previously published transcriptomic data from CD4 T cells and tumor samples of people living with HTLV-1, including Brazilian and Japanese patients. A meta-analysis of microarray datasets from four studies, batch-corrected and modeled with a linear mixed-effects approach, showed FDXR expression elevated across all patient groups compared with healthy controls, reaching statistical significance for acute ATL, HAM/TSP, chronic ATL and asymptomatic carriers, and mirroring disease severity. A parallel reanalysis of RNA-Seq data from a separate cohort showed the same gradient, with significant differences from healthy controls in acute and lymphoma-type ATL. In the ATL patients from that cohort, FDXR levels also showed a modest but significant correlation with proviral load (Spearman r = 0.355, p = 0.00755), a known predictor of disease progression. An independent, newly recruited Brazilian cohort confirmed higher FDXR expression in people living with HTLV-1 than in healthy donors, although differences between clinical subgroups did not reach significance there, likely reflecting the small number of individuals with advanced disease.

The prognostic signal came from 51 ATL patients with available clinical follow-up. A Cox proportional-hazards model on continuous, z-score-normalized FDXR expression linked higher levels to reduced survival, with a hazard ratio of 1.7 (p = 0.034). When the analysis was stratified by subtype, the association was significant only within indolent disease (chronic and smoldering ATL), where the hazard ratio rose to 5.5 (p = 0.014). The authors attribute this to the broader dynamic range of FDXR expression in indolent cases: most acute and lymphoma cases already show uniformly elevated FDXR, whereas indolent cases range from marked upregulation to little or no increase, so high expression in an indolent patient may signal atypically high proliferative activity. Kaplan-Meier analyses stratifying patients by FDXR levels also showed poorer outcomes among those with higher expression, reaching significance in the two-group comparison. Survival was censored at hematopoietic stem cell transplantation.

Mechanistically, the study traces the upregulation to two members of the p53 family. In silico inference of transcription factor activity from RNA-Seq data of leukemia and lymphoma cell lines, using the VIPER algorithm with regulatory networks from DoRothEA and ARACNe, highlighted p63 and p73 as significantly upregulated in cells with high FDXR expression. Isoform-specific assays showed that the transactivating isoforms, TAp63 and TAp73, were consistently expressed in HTLV-1-infected and ATL cell models, whereas uninfected cells largely lacked TAp63; the ΔNp63 isoform was undetectable across all models, and p53 itself showed no significant differences between infected and uninfected cells. Chromatin immunoprecipitation followed by qPCR demonstrated that both p63 and p73 bind a p53-responsive element within the FDXR promoter in infected cells, but not in uninfected controls. Most decisively, stable CRISPR-Cas9 knockouts of TP63 or TP73 in MT-4 and ATL-55T cells significantly reduced FDXR protein levels, with double knockouts producing the greatest reduction.

The chain extends one step further upstream. Building on prior reports that p73 in ATL cells is induced by the IRF4/BATF3 transcriptional complex, the researchers found that IRF4 emerged as one of the strongest predicted upstream activators of both TP63 and TP73 in their regulon analysis. Mining published ChIP-seq data from ATL cells revealed binding peaks for both BATF3 and IRF4 at the TP63 locus, and sequence scanning uncovered a canonical AP1-IRF composite (AICE) motif within 50 base pairs of those peaks, overlapping an active enhancer marked by H3K27ac and BRD4 enrichment. In ATL-55T cells, CRISPR-Cas9 knockout of IRF4 significantly decreased p63 protein levels, and knockout of BATF3 approached significance. IRF4 induction is a recognized hallmark of HTLV-1 infection and can be driven by both the viral proteins Tax and HBZ, which may explain why FDXR is elevated in both ATL and the inflammatory neurological condition HAM/TSP, which typically show high expression of HBZ and Tax respectively.

Single-cell analysis added a proliferative dimension. Examining a published scRNA-seq dataset of peripheral blood cells from healthy donors and people living with HTLV-1, the team found that active FDXR expression was enriched within the proliferating cell cluster (cluster 8, marked by S-phase and G2/M cell-cycle scores) in all but one ATL sample, and that FDXR was more enriched in that cluster than the established infection markers CADM1 and HBZ. Average frequencies of FDXR-expressing cells rose from 0.47 percent in smoldering ATL to 2.29 percent in chronic and lymphoma-type ATL, against 1.20 percent in healthy donors. FDXR and its upstream regulators also correlated modestly with CADM1 expression, suggesting preferential expression in infected cells even when active viral gene expression is absent.

The therapeutic angle came from drug-susceptibility screening. Analyzing cytotoxicity and gene-expression data from the Cancer Therapeutics Response Portal, covering 481 compounds tested in up to 835 cell lines, the team found FDXR expression significantly associated with enhanced sensitivity to five agents. Restricting the analysis to hematopoietic cell lines identified fluorouracil and bleomycin A2 as the compounds with the lowest dose-response area under the curve, a pattern also seen in the dataset’s only HTLV-1-infected cell line. Neither drug has previously been evaluated in ATL. In MTS assays, fluorouracil showed significantly greater efficacy in HTLV-1-infected and ATL cell lines than in uninfected controls (p = 0.016 for AUC comparison), while bleomycin showed a similar but non-significant trend. Drug-Seq on primary CD4 T cells from five ATL patients and three healthy donors, plus two ATL cell lines, showed that both drugs activated p53-family-linked pathways, induced FDXR expression as a leading-edge gene in p53 and apoptosis pathways, and reduced proliferation; FDXR was the most significant differentially expressed gene in primary ATL cells treated with either bleomycin concentration. Immunofluorescence combined with EdU proliferation assays in primary cells from three chronic ATL patients confirmed a significant drop in proliferating cells after treatment with either drug.

The authors are careful about limits. Their analyses focused mostly on CD4 T cells, leaving other infectable cell types unexplored; the drug-seq donors were predominantly indolent cases, so efficacy in aggressive subtypes with the highest FDXR levels remains plausible but untested; and repurposing fluorouracil or bleomycin for a hematological malignancy would require dedicated in vivo studies or early-phase clinical trials, particularly since resistance mechanisms to both drugs are known in other cancers. ATL responses to fluorouracil also remained within the intermediate range reported for colorectal cancer cells, suggesting combination approaches, perhaps with BCL-2 family inhibitors, may be needed. Still, the study establishes FDXR as a biomarker linking HTLV-1 infection, mitochondrial reprogramming, proliferative activity and chemotherapeutic vulnerability, and proposes a regulatory axis, IRF4/BATF3 to TAp63 and TAp73 to FDXR, that could anchor future risk stratification and treatment selection in a disease with few good options.

Subject of Research: FDXR upregulation as a biomarker and drug-sensitivity predictor in adult T-cell leukemia/lymphoma

Article Title: FDXR upregulation by p63/p73 is a prognostic and therapeutic marker of adult T-cell leukemia/lymphoma

Article References: Gallucci, L., Josephine, M., Kogure, Y., Rosini Silva, A. A., Cantaloube-Ferrieu, V., Souza Abreu, M. E., Ramteke, T., Kibble, P., Angelo Braz, G. L. D., Santos Sebastião, R. D. A., Lambaz, L. A., Lin, A., Hoare, B., Carvalho dos Santos, F. C., Shackley, S., Quin, A., Moelas Giuzio, S. L., Lewis, P. A., Zhang, W., … Shytaj, I. L. (2026). FDXR upregulation by p63/p73 is a prognostic and therapeutic marker of adult T-cell leukemia/lymphoma. EMBO Molecular Medicine. https://doi.org/10.1038/s44321-026-00535-0

Image Credits: AI Generated

DOI: 10.1038/s44321-026-00535-0

Keywords: HTLV-1, adult T-cell leukemia/lymphoma, FDXR, p63, p73, IRF4, BATF3, fluorouracil, bleomycin, mitochondria, biomarker, drug repurposing