Prepared By
Joelle Fenger, DVM, PhD, DACVIM (Oncology)
Peripheral Blood Mononuclear Cell Gene Expression Signatures Predict Long-Term Survivorship in Canine DLBCL
Kirthana Rao, Zechuan Rao, Angelina Huang, Scott Heston, Max Wang, Ümmügülsüm Yildiz-Altay, Fatima Qutab, Danny A. Kwong, Heather L. Gardner, Jillian M. Richmond, Cheryl A. London
Background
Diffuse large B-cell lymphoma (DLBCL) is a common and clinically heterogeneous malignancy in both humans and dogs, with a subset of patients experiencing early relapse and poor outcomes despite standard chemoimmunotherapy regimens. While CHOP-based protocols improve survival, there remains a critical unmet need for predictive biomarkers to identify patients at risk of treatment failure. Pet dogs with spontaneous DLBCL serve as a valuable translational model for human disease. This study investigates whether peripheral blood mononuclear cell (PBMC) gene expression signatures can serve as minimally invasive (“liquid biopsy”) biomarkers to predict long-term survivorship and treatment response in a cohort of pet dogs with DLBCL undergoing doxorubicin-based immunomodulatory therapy.
Study Design
This is a prospective, non-randomized clinical trial undertaken in pet dogs (N=18) with spontaneous DLBCL. Enrolled dogs received treated with:
• Anti-CD20 monoclonal antibody
• Low-dose doxorubicin
• One of three small molecule inhibitors (TAK-981, KPT-9274, or RV1001)
Animals
18 pet dogs with spontaneous DLBCL were prospectively enrolled in a non-randomized clinical trial and received treatment with three different B-cell depletion (Anti-CD20 mAb) + low dose doxorubicin + immunomodulatory therapy (TAK-981, KPT-9274, or RV1001) regimens.
Methods
This study utilized samples from a previously conducted non-randomized clinical trial in dogs (N=18) with DLBCL treated with:
• Anti-CD20 monoclonal antibody
• Low-dose doxorubicin
• One of three small molecule inhibitors (TAK-981, KPT-9274, or RV1001)
PBMCs were collected longitudinally at four key time points: baseline, day 7 (post–B-cell depletion), day 21–28 (post–chemoimmunotherapy), and at relapse.
Gene expression profiling was performed using the NanoString Canine IO panel, followed by statistical analyses to identify differentially expressed genes associated with outcomes. Long-term survivorship was defined as overall survival >400 days. Key candidate biomarkers were further validated using qPCR assays.
Results
• Treatment response biology: Effective B-cell depletion (loss of CD19, MS4A1, PAX5) was observed across all treatment arms, with recovery at relapse, confirming biological consistency of therapy.
• Biomarkers of poor prognosis: Increased expression of interferon-stimulated genes (ISGs: ISG15, MX1, OAS2, PSMB9) and immune-skewing genes (CCR9, CD209, CMKLR, DDX58) were consistently associated with early relapse and shorter survival. Early timepoint markers (Day 7): THBD, NPNT, and ISG20 predicted poor outcomes and were validated via qPCR.
• Biomarkers of favorable prognosis: CD1E and CCL14 expression post-treatment correlated with long-term survivorship, suggesting a protective immune phenotype.
• Clinical utility: A simplified qPCR-based assay targeting THBD, NPNT, and ISG20 demonstrated feasibility for point-of-care prognostication.
Important Findings
This study provides proof-of-principle that PBMC-derived gene expression signatures can predict treatment response and survival in canine DLBCL using a minimally invasive approach. The findings highlight:
• The prognostic significance of systemic immune signatures (not just tumor-intrinsic features)
• A potential role for interferon signaling as a negative prognostic biomarker in this context
• The feasibility of developing rapid, clinically deployable assays to guide treatment decisions
Importantly, this work strengthens the role of comparative oncology by demonstrating translational relevance to human DLBCL, particularly in identifying biomarkers for risk stratification and precision medicine approaches.
Study Limitations
• Small sample size (N=18), particularly within treatment arms and outcome subgroups, limiting statistical power
• Non-randomized study design, introducing potential bias
• Incomplete validation of key biomarkers (e.g., CD1E and CCL14 could not be validated by qPCR)
• Heterogeneity across treatment arms, which may confound interpretation of immune signatures. Despite the variability in treatment protocols utilized in the study, clinical differences (e.g. long vs short term survivorship) were not significant between treatment groups as to allow larger cohort analyses (this has also been previously addressed in the investigative team’s prior peer-reviewed publication, Dittrich K, et al. PLoS One 2023).
• Findings require prospective validation in larger, controlled cohorts before clinical implementation

Leave a Reply
Want to join the discussion?Feel free to contribute!