Immune Cell Types and Host Gene Signatures Associated with Epstein-Barr Virus (EBV) Post-Transplant Lymphoproliferative Disorder (PTLD)
1Stanford University School of Medicine, Stanford, CA, 2Columbia University Medical Center, New York, NY, 3Mayo Clinic College of Medicine, Phoenix, AZ, 4University of Pittsburgh School of Medicine, Pittsburg, PA, 5Katholieke Universiteit Leuven, Leuven, Belgium
Meeting: 2019 American Transplant Congress
Abstract number: 52
Keywords: Gene expression, Mononuclear leukocytes, Natural killer cells
Session Information
Session Name: Concurrent Session: PTLD/Malignancies: All Topics
Session Type: Concurrent Session
Date: Sunday, June 2, 2019
Session Time: 2:30pm-4:00pm
Presentation Time: 3:30pm-3:42pm
Location: Room 311
*Purpose: PTLD is a serious post-transplant complication that encompasses a range of abnormal lymphoproliferations, including EBV+ B cell lymphomas. Although many are EBV-driven, a significant minority are not. EBV employs diverse immune evasion strategies and EBV+ PTLD lymphomas also exhibit immunomodulatory phenotypes. To better characterize the interactions between EBV and host immunity in PTLD, we used systems biology approaches on EBV+ and EBV- PTLD tumor samples to identify immune cell types and host genes associated with EBV-driven oncogenesis.
*Methods: Three gene expression datasets with EBV+ PTLD and EBV- PTLD tumor samples (n = 60) were chosen for multi-cohort analyses. We estimated various immune cell proportions in each sample via in silico deconvolution using immunoStates basis matrix with a support vector regression-based method (CIBERSORT). To identify consistently enriched/deficient immune cell types in EBV+ PTLD tumors across datasets, we calculated the Hedges’ g effect size of individual leukocyte proportions within each dataset. These were then incorporated to give an overall summary effect size for each cell type across all datasets.
To identify differentially expressed genes in EBV+ PTLD, we computed the Hedges’ g effect size of each gene’s expression in the individual datasets, and combined these into summary effect sizes using a random effects inverse variance model. We selected genes with 1) a log2(summary effect size) ≥ 0.59 (equivalent to a 1.5× gene expression fold change), 2) a false discovery rate ≤ 20%, and 3) representation in ≥ 2 of the 3 dataset microarrays.
*Results: We show that NK cells and monocytes are enriched in EBV+ compared to EBV- PTLD tumors. We also obtained a 189-gene signature distinguishing EBV+ from EBV- PTLD, which robustly discriminated between EBV+ and EBV- PTLD samples in each dataset. Amongst these, several genes associated with NK cell and monocyte/macrophage function (KIR2DL3/4, KIR3DL1/2, IL-15; and CCL3/4/8, CSF2 respectively) are upregulated in EBV+ PTLD.
*Conclusions: Our results are indicative of common interactions between EBV, the tumor, and host immunity in EBV+ PTLD, and may identify novel clinical targets for EBV+ B cell lymphomas where host immunity is compromised.
To cite this abstract in AMA style:
Toh J, Vallania F, Haynes WA, Bhagat G, Craig FE, Swerdlow SH, Tousseyn T, Krams SM, Khatri P, Martinez OM. Immune Cell Types and Host Gene Signatures Associated with Epstein-Barr Virus (EBV) Post-Transplant Lymphoproliferative Disorder (PTLD) [abstract]. Am J Transplant. 2019; 19 (suppl 3). https://atcmeetingabstracts.com/abstract/immune-cell-types-and-host-gene-signatures-associated-with-epstein-barr-virus-ebv-post-transplant-lymphoproliferative-disorder-ptld/. Accessed May 13, 2025.« Back to 2019 American Transplant Congress