AUTHORS
Andrew Jay Portuguese, MD, Albert C Yeh, MD2, Bryce N Fukuda3, Minggang Lin4, Elizabeth Donato5, Kikkeri N Naresh, MD, MBBS, FRCPath6, Alexandre V Hirayama, MD7, Jordan Gauthier, MD8, Vikas Dharnidharka9, Cecilia CS Yeung, MD, Professor Pathology10, Christopher Blosser11
AFFILIATIONS
1 Fred Hutchinson Cancer Center, Seattle, Washington, United States
2 Fred Hutchinson Cancer Center, Seattle, Washington, United States
3 Fred Hutchinson Cancer Center, Seattle, Washington, United States
4 Fred Hutchinson Cancer Center, Seattle, Washington, United States
5 Fred Hutchinson Cancer Center, Seattle, Washington, United States
6 Fred Hutchinson Cancer Center, Seattle, Washington, United States
7 Fred Hutchinson Cancer Center, Seattle, Washington, United States
8 Fred Hutchinson Cancer Center, Seattle, Washington, United States
9 Rutgers University, New Brunswick, New Jersey, United States
10 Fred Hutchinson Cancer Center, Seattle, Washington, United States
11 University of Washington, Seattle, Washington, United States
ABSTRACT
Background: Epstein-Barr virus (EBV)-positive and EBV-negative post-transplant lymphoproliferative disorders (PTLD) are biologically distinct. In EBV-positive PTLD, EBV latent membrane protein 1 (LMP1) signaling promotes PD-L1/PD-L2 upregulation and immune evasion, and prior studies found PD-L1 expression in malignant B cells and myeloid cells, as well as monocyte enrichment. In contrast, EBV-negative PTLD more closely resembles diffuse large B-cell lymphoma (DLBCL). However, the cell-specific and spatial organization of the PD-1/PD-L1 axis across EBV-positive PTLD, EBV-negative PTLD, and conventional DLBCL remain poorly defined.
Methods: We used the 380-gene 10x Genomics Xenium Human Immuno-Oncology panel plus a 100-gene custom add-on (480 targets including EBV transcripts) on tissue microarrays from archival diagnostic specimens: 42 PTLD cores from 19 patients (pts; 11 EBV-positive, 8 EBV-negative) and 41 cores from 21 non-PTLD DLBCL samples. PTLD cases included liver (58%), heart (26%), kidney (11%), and pancreas (5%) transplant recipients; biopsies included GI (53%), nodal (26%), or extranodal sites (21%) from 2002-2024. Proseg segmentation and filtering (≥20 transcripts and 10 genes/cell) yielded 1,061,904 cells: 318,284 EBV-positive PTLD, 138,132 EBV-negative PTLD, and 605,488 control DLBCL cells. Canonical-marker rules assigned cell identities; hematopathologists reviewed concordance with matched H&E sections. High checkpoint expression was defined using lineage-specific PTLD 90th percentile thresholds. Checkpoint expression was summarized by cell type; within-core 10-nearest-neighbor graphs quantified checkpoint-cell adjacency. Gene-program scores averaged listed normalized genes; higher values indicated more lymphoma-like or proliferative transcription. Replicate cores were collapsed to the pt/sample level before statistical testing. Wilcoxon tests assessed programs and spatial metrics. All-cell and malignant B-cell pseudobulk differential expression used Limma-voom with Benjamini-Hochberg correction.
Results: Broad lineage composition and global interfaces did not differ after false-discovery correction. EBV-positive PTLD had higher B-cell PD-L1 expression than EBV-negative PTLD (median sample-level mean normalized expression, 0.0461 vs 0.0168; q=0.016) and control DLBCL (0.0461 vs 0.0230; q=0.016). Malignant B-cell PD-L2 expression was also higher in EBV-positive PTLD than in EBV-negative PTLD and control DLBCL (EBV-positive 0.0462, EBV-negative 0.0226, control 0.0268; q=0.016 and q=0.024, respectively). Control DLBCL had a higher fraction of PD-L1-high myeloid cells than EBV-positive PTLD (0.171 vs 0.089; q=0.016) and more PD-1-high T cells adjacent to PD-L1-high myeloid cells (0.267 vs 0.107; q=0.036). Among 9 EBV-positive PTLD pts evaluable for within-pt EBER-detected versus EBER-undetected analysis, EBER-detected malignant B cells had higher adjusted PD-L1 (median coefficient, 0.0072; q=0.049).
The curated B-cell lymphoma-associated program (BCL6, CD40, CD83, CXCR5, PAX5, TCL1A, TOX) was higher in EBV-negative than EBV-positive malignant B cells (median standardized pt score, 0.137 vs -0.907; q=0.024); control DLBCL scored 0.630 (vs EBV-positive, q=0.001). The malignant B-cell proliferation-associated score (MKI67, TOP2A, PCNA, MYC) trended higher in EBV-negative than EBV-positive PTLD (q=0.055), while EBV-positive PTLD scored lower than control DLBCL (q=0.010). Pseudobulk confirmed enrichment of latent and lytic-cycle EBV transcripts (EBER1/2, LMP1/2, EBNA2/3B/3C, BZLF1) in EBV-positive PTLD. EBV-positive tissue expressed higher IL12RB2, whereas EBV-negative tissue showed higher TCL1A overall (EBV-positive vs EBV-negative log2FC=-4.54; q=0.004) and higher TCL1A and SPIB within malignant B cells (q=0.002 and q=0.048). Relative to EBV-positive PTLD, control DLBCL was enriched for B-cell/lymphoma genes including CXCR5, PAX5, CD37, SPIB, MS4A1, CD40, and CD83 (all q≤0.002).
Conclusions: EBV-positive PTLD exhibited higher PD-L1/PD-L2 and latent/lytic EBV transcripts in malignant B cells, whereas non-PTLD DLBCL showed a myeloid-centered PD-L1 program with greater adjacency of PD-1-high T cells to PD-L1-high myeloid cells. EBV-negative PTLD retained a more DLBCL-like B-cell program, including higher TCL1A/SPIB and a trend toward greater proliferation. These findings define distinct EBV-associated checkpoint architectures, supporting EBV-stratified biomarker development and future immunotherapeutic investigation
DISCLOSURES
Andrew Jay Portuguese, MD: Pfizer - Research Funding; Regeneron - Research Funding; Karyopharm - Consultancy (Includes expert testimony)