Canine diffuse large B-cell lymphoma is far more than a veterinary curiosity. It accounts for roughly 50 to 60 percent of all lymphoid cancers in dogs and mirrors human diffuse large B-cell lymphoma, which makes up about 40 percent of non-Hodgkin lymphomas in people. Because pet dogs develop the disease spontaneously, live in the same environments as their owners, and possess intact immune systems, they offer a clinically relevant model that laboratory mice cannot match. Drugs such as venetoclax, ibrutinib, and acalabrutinib have already been explored through this comparative framework, and the typical time to relapse in dogs, six to eight months, is far shorter than in humans, allowing therapeutic studies to move at an accelerated pace.

In the new study, the researchers analyzed lymph node samples from 15 dogs with multicentric DLBCL and two healthy control animals treated at the Governador Laudo Natel Veterinary Hospital at UNESP in Jaboticabal. Tissue fragments were snap-frozen and stored at minus 80 degrees Celsius before total RNA was extracted and converted into sequencing libraries using the Illumina TruSeq protocol. The team sequenced the libraries on an Illumina HiScanSQ platform with paired-end chemistry, then aligned the resulting reads against the canine reference genome CanFam6 using the STAR aligner. Differential expression analysis was carried out with the EdgeR package, modeling read counts with a negative binomial distribution and controlling the false discovery rate through the Benjamini-Hochberg procedure.

The results were striking in their asymmetry. The analysis identified 410 differentially expressed genes, of which 408 were downregulated in the tumor samples compared with normal lymph nodes, while only two genes, CPLX3 and a poorly characterized locus, were upregulated. Most of the silenced genes are involved in immune function, the very machinery that healthy tissue uses to police malignant cells. The researchers then cross-referenced their findings with four previous RNA-sequencing studies of canine B-cell lymphoma, comparing the overlapping gene lists to identify candidates of highest biological relevance. From this convergence, five genes emerged as the focus of the study: STOM, TBC1D8, HMOX1, ABCB1, and CTLA4.

Each of these five genes tells a distinct mechanistic story. STOM encodes stomatin, an integral membrane protein that organizes ion channels and helps structure the immunological synapses through which immune cells communicate; its disruption has been linked to gastric MALT lymphoma, colorectal cancer, and altered glucose transport via GLUT1. TBC1D8 encodes a GTPase-activating protein that regulates Rab-mediated vesicular trafficking, a process essential for antigen presentation, and in human colorectal and ovarian cancers it has been tied to invasiveness and poor prognosis through metabolic reprogramming toward aerobic glycolysis, the Warburg effect. HMOX1, or heme oxygenase 1, plays a paradoxical role in tumor biology, acting cytoprotectively at moderate levels while promoting lipid peroxidation and ferroptotic cell death when overexpressed.

The remaining two genes connect the signature directly to therapy. ABCB1 encodes P-glycoprotein, the ATP-dependent efflux pump responsible for multidrug resistance, which exports chemotherapeutic agents from tumor cells and undermines treatment efficacy; inhibiting it has been shown to restore drug sensitivity in lymphoma models. CTLA4 is one of the most famous targets in modern immunology, an immune checkpoint receptor that dampens T-cell activation by competing with CD28 for binding to CD80 and CD86. Its dysregulation enables tumors to escape immune surveillance, and checkpoint inhibitors targeting CTLA4, often combined with anti-PD-1 therapy, have transformed human cancer treatment. The enrichment analysis and protein-protein interaction networks built with the STRING database and Enrichr tool further connected these genes to the nuclear receptor meta-pathway, anion transport regulation, and xenobiotic transport across the blood-brain barrier, hinting at how drug metabolism and tumor stress responses converge.

When the team correlated expression of the five-gene signature with clinical outcomes, the picture was suggestive but not definitive. All 15 lymphoma patients had been treated with the CHOP chemotherapy protocol, and their survival times ranged from just 2 weeks to 160 weeks, with a mean of 48.8 weeks. Kaplan-Meier analysis showed a trend toward reduced overall survival in dogs with lower expression of the individual genes and of the composite signature, but none of these associations reached statistical significance, likely reflecting the small cohort size. The authors note that a priori power analysis was performed, and that the study was designed as a vanguard investigation capable of detecting very large effect sizes, but they acknowledge that larger, multicentric validation studies will be needed before these markers enter clinical practice.

Beyond the gene expression data, the study offers one of its most vivid insights into the tumor microenvironment itself. Using three independent computational deconvolution methods, CIBERSORT, quanTIseq, and EPIC, the researchers estimated the proportions of immune cell types within each sample. Tumor tissue was overwhelmingly dominated by neoplastic B cells, as expected in a B-cell malignancy, but the effector arm of the immune system was conspicuously depleted: cytotoxic CD8-positive T cells and macrophages were markedly reduced compared with healthy lymph nodes, which displayed a diverse repertoire of macrophages, monocytes, CD4-positive T cells, and neutrophils. This immunosuppressive architecture suggests active evasion of immune detection by the tumor. Intriguingly, an age-stratified analysis found that younger dogs tended to show higher CD4-positive T-cell infiltration, a trend the authors interpret as possibly reflecting preserved antitumor immunity before age-related immunosenescence takes hold.

The Brazilian setting of the study adds an important dimension to the global picture. Transcriptomic data from South American canine populations had been largely absent from the literature, despite evidence that environmental factors shape cancer biology. Previous case-control studies in São Paulo and the metropolitan region of Recife found that dogs living near high-traffic roads or exposed to solvents and paints faced substantially elevated risks of developing lymphoma, with one study reporting a nearly fivefold increase in risk, positioning pet dogs as sentinels for environmental carcinogens that may also affect humans. The authors also highlight a methodological point with consequences for the entire field: because earlier studies aligned their reads against the older CanFam3.1 or even CanFam2 reference genomes, while this study used the more complete and accurately annotated CanFam6, discrepancies in the numbers and identities of differentially expressed genes across studies are partly artifacts of genome assembly rather than true biological differences.

The study is not without limitations, and the authors are candid about them. The cohort comprised only 15 tumor samples and two controls, drawn from multiple breeds, and breed predispositions were not incorporated into the power analysis. Yet the consistency of the normal-sample expression profiles across all comparable published studies supports the validity of the small control group, and the magnitude of the transcriptional differences observed lent the analysis statistical footing despite its scale. What the work ultimately delivers is a framework: five interconnected genes governing membrane dynamics, vesicular transport, oxidative stress, immune regulation, and drug resistance, set against a tumor microenvironment that has lost its cytotoxic defenders. If validated in larger cohorts, these signatures could guide individualized treatment decisions in veterinary oncology, from checkpoint inhibitor combinations to strategies for reversing multidrug resistance, while reinforcing the dog’s standing as one of the most valuable spontaneous models for advancing human lymphoma research.

Subject of Research: Transcriptomic profiling and immune microenvironment analysis of canine diffuse large B-cell lymphoma

Article Title: Transcriptomic profiling reveals prognostic gene signatures and immune landscape alterations in canine diffuse large B-cell lymphoma

Article References: Anai, L. A., Cavalca, A. M. B., Xavier, P. L. P., Semolin, L. M. S., Jark, P. C., Senhorello, I. L. S., Fukumasu, H., Demarchi Munhoz, T., Figueiredo, M. L., Fernandes, C. C., Costa, M. T., Kobayashi, P. E., Santana, Á. E., & Fonseca-Alves, C. E. (2026). Transcriptomic profiling reveals prognostic gene signatures and immune landscape alterations in canine diffuse large B-cell lymphoma. Veterinary Oncology, 3(1), Article 1. https://doi.org/10.1186/s44356-025-00051-2

Image Credits: AI Generated

DOI: 10.1186/s44356-025-00051-2

Keywords: canine lymphoma, diffuse large B-cell lymphoma, RNA sequencing, transcriptomics, tumor microenvironment, CTLA4, ABCB1, HMOX1, STOM, TBC1D8, immune evasion, comparative oncology