Abstract Background: Diffuse large B-cell lymphoma (DLBCL) is a heterogenous disease, with variable response to T-cell therapies, highlighting the need for a clinically applicable patient stratifier. To dissect DLBCL complex tumor-host immune interactions, we established fully humanized DLBCL PDX models (HuMice PDX) and performed integrated multi-omics profiling of patient samples. Methods: We analyzed 132 DLBCL (112 R-CHOP, 20 CAR-T patients), 48 matched PDX, 9 HuMice PDX and CAR-19/CAR-BAFFR treated PDX, using bulk RNA-seq, single-cell RNA/TCR-seq, WES/CAPP-seq and CosMx SMI profiling (6000 RNA, 64 proteins). Large DLBCL datasets were used to build a prediction classifier. heterologous/allogeneic CD34+ Human Stem Progenitor Cell (HSPC) were implanted in NSG-SGM3xNBSGW mice, to generate HuMice. Results: We identified EPICORE, a 12-gene epigenetic gene-expression signature, able to predict whether DLBCL displays an immune-depleted (ID) versus an immune-rich (IR) tumor microenvironment. ID/EPICORE-high patients displayed significantly worse outcomes following both R-CHOP- (p0. 0001) and CAR-19 (p=0. 002) treatment, compared to IR/EPICORE-low patients. To investigate host-lymphoma interactions, we generated HuMice and implanted them with ID/EPICORE-high and IR/EPICORE-low PDX models. Mimicking patients’ outcomes, only IR models displayed reduced tumor growth in HuMice, while ID-DLBCL showed unimpeded growth. Flow cytometry, mIHC, and RNAseq/scRNAseq analyses revealed that ID-DLBCL-HuPDX were devoid of Tumor-Infiltrating Lymphocytes (TILs), whereas IR-DLBCL-HuPDX displayed abundant (TILs) (p2e-04) and clonal expansion. We then tested two different CART (CD19, BAFFR) in-vivo in nine ID- and IR-PDX-HuMice models. IR/EPICORE-low models were eradicated, while ID/EPICORE-high were refractory to both CAR. To enhance ID-DLBCL responses, we performed drug screening of clinical phase epigenetic targeting drugs shown to improve immune responses. Valemetostat and Lenalidomide treatments were associated with improved CAR-T cell killing, both in-vitro and in-vivo. Exposure to drug led to a reduction of EPICORE score, and enrichment of inflammatory signaling pathways (NFkb, IFNg/a, IFN-dependent chemokines), thereby converting ID-DLBCL into IR-DLBCL (p0. 05-0. 0005) Conclusions: EPICORE, a novel classifier, predicts DLBCL T-cell infiltration and therapy outcomes allowing patients stratification. Humanized PDX accurately reflect DLBCL physiopathology and immune-host interactions, providing a platform for validating next-generation immunotherapies. Epigenetic therapies reprogram EPICORE-high/ID- into EPICORE-low/IR-DLBCL, overcoming immune refractoriness and enhancing T-cell infiltration and tumor immunorecognition. Citation Format: Giovanni Medico, Giorgia Zanetti, Giorgio Bertolazzi, Francesco Vallania, Paul Zumbo, Maider Amiama, Doron Betel, Maria Cacciapuoti, Abigail Taylor, Clarisse Kayembe, Sanjay Patel, Arvin Ruiz, Jennifer Ziello, Jason Weirather, Gulpreet Kaur, Luca Cappelli, Luca Paruzzo, Marco Ruella, Zhenyuan Dong, Larry Kwak, Nicolás Di Siervi, Maria Revuelta, Claudio Tripodo, Leandro Cerchietti, Giorgio Inghirami. Humanized PDX models uncover drug-reversible epigenetic immune exclusion governing T-cell therapy response in DLBCL, captured in patients by EPICORE stratifier abstract. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts) ; 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86 (8Suppl): Abstract nr LB427.
Medico et al. (Fri,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: